UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evolution and stratification of off-gasses in stored wood pellets Yazdanpanah, Fahimeh 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_fall_yazdanpanah_fahimeh.pdf [ 6MB ]
Metadata
JSON: 24-1.0165614.json
JSON-LD: 24-1.0165614-ld.json
RDF/XML (Pretty): 24-1.0165614-rdf.xml
RDF/JSON: 24-1.0165614-rdf.json
Turtle: 24-1.0165614-turtle.txt
N-Triples: 24-1.0165614-rdf-ntriples.txt
Original Record: 24-1.0165614-source.json
Full Text
24-1.0165614-fulltext.txt
Citation
24-1.0165614.ris

Full Text

EVOLUTION AND STRATIFICATION OF OFF-GASSES IN STORED WOOD PELLETS by Fahimeh Yazdanpanah M.A.Sc, University of British Columbia, 2009 B.Sc., Amirkabir University of Technology, 2006  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) October 2013 ? FahimehYazdanpanah, 2013  iiAbstract Storage of wood pellets has resulted in several deathly accidents in connection with off-gassing and self-heating. The goal of the present study was to quantify off-gassing characteristics of white wood pellets when stored in an experimental silo. Wood pellets properties were characterized with respect to gas adsorption-desorption and spatial and temporal concentrations of off-gases and thermal conditions within the pilot storage were quantified. In the last part, the effectiveness of purging the silo in reducing off-gas concentration was evaluated. To assess the adsorption of off-gases by wood pellets in storage, Temperature Programmed Desorption was used. Highest CO2 adsorption was seen by torrefied wood pellets while lowest uptake was showed to be for steam exploded pellets. Quantifying the uptake of CO was challenging due to chemical reaction and therefore strong bonds between the material and carbon monoxide. Studies on emission and stratification of off-gases showed higher emission factor compared to work done with white wood pellets in small scale. Some stratifications were observed for CO2 and CH4 over the first days of storage. However for CO the stratification was much clear and related to high uptake of CO by wood pellets over time. During the entire period of storage, maximum temperature in the silo was recorded on day 15 of storage (storage time was 63 days) at the elevation of 2.5 m (silo dimension was 1.2m diameter and 4.6m height). Measured temperature in the silo during 5.5 hour purging experiments with air at 18-18.5 oC, helped the temperature decrease in the lower parts and slightly middle parts of the silo after 200 minutes of purging. To evaluate the effectiveness of a purging system to sweep the off-gases from the experimental silo, multiple purging tests were done. Mixing experiments  iiishowed large deviations from plug flow and thus better mixing for all superficial velocities used. Predicted results showed the concentration model fitted best to the measured off-gas concentration at the bottom and in the middle of the silo while the model overestimated the exponential decay of the off-gases in the head-space of the silo.       ivPreface This PhD thesis is divided to seven chapters. The author, Fahimeh Yazdanpanah, has done the experimental design, performing experiments, material characterization, literature review, data processing and analysis, manuscript and thesis preparation under the supervision of Professor Shahab Sokhansanj, Professor Jim Lim and Professor Anthony K. Lau in the Chemical and Biological Engineering Department at the University of British Columbia.   Chapter 5 of the thesis has been accepted for publication in Biomass and Bioenergy Journal. A part of  Chapter 4 is prepared for publication. Material presented in  Chapter 6 is prepared in the manuscript format for publication.  The list of publication is presented below: 1- Yazdanpanah, Fahimeh; Sokhansanj, Shahab; Lim, C. Jim; Lau, Anthony; Bi, Xiaotao and Melin, Staffan, ?Stratification of Off-gases in Stored Wood pellets?, Biomass and Bioenergy, In Press.  2- Yazdanpanah, Fahimeh; Sokhansanj, Shahab; Lim, C. Jim; Lau, Anthony; Bi, Xiaotao; Lam, Pak Yiu and Melin, Staffan, ?Potential for Flammability of Gases Emitted from  Stored Wood Pellets?, Canadian Chemical Engineering Journal, In Press. DOI: 10.1002/cjce.21909 3- Yazdanpanah, Fahimeh; Sokhansanj, Shahab; Lim, C. Jim; Lau, Anthony; Bi, Xiaotao. ?Wood Pellet Gas Adsorption Capacity?, Internal Review.   v4- Yazdanpanah, Fahimeh; Sokhansanj, Shahab; Lim, C. Jim; Lau, Anthony; Bi, Xiaotao, ?Temperature and Relative Humidity Pattern Inside a Pilot Scale Wood Pellet Storage?, Internal Review.  5- Yazdanpanah, Fahimeh; Sokhansanj, Shahab; Lim, C. Jim; Lau, Anthony; Bi, Xiaotao, ?Effectiveness of Purging on Gas Emission in Stored Wood Pellets?, Internal Review.    viTable of Contents Abstract .................................................................................................................................... ii?Preface ..................................................................................................................................... iv?Table of Contents ................................................................................................................... vi?List of Tables ......................................................................................................................... xii?List of Figures ....................................................................................................................... xiv?Nomenclature ..................................................................................................................... xxvi?Acknowledgements ............................................................................................................ xxxi?Dedication ......................................................................................................................... xxxiii?Chapter 1 Introduction........................................................................................................... 1?1.1? Background ............................................................................................... 1?1.2? Objectives of the Thesis ............................................................................ 4?1.3? Approach of the Thesis ............................................................................. 4?Chapter 2 Literature Review ................................................................................................. 6?2.1? Biomass Storage........................................................................................ 6?2.2? Off-gassing .............................................................................................. 10?2.2.1?Condensable and Non-condensable Gases ....................................................... 10?2.2.2?Oxygen Depletion ............................................................................................. 18?2.3? Self-heating ............................................................................................. 20? vii2.4? Dust Explosion ........................................................................................ 22?2.5? Kinetics of Off-gassing ........................................................................... 23?2.6? Safety Measures for Storage of Pellets ................................................... 27?2.7? Summary of Literature Review ............................................................... 31?Chapter 3 Experimental Apparatus and Set-up ................................................................ 32?3.1? Pilot Scale Silo Configuration ................................................................ 32?3.1.1?Off-gas Sampling Ports .................................................................................... 34?3.1.2?Temperature and Relative Humidity Monitor .................................................. 35?3.1.3?Pressure Monitor .............................................................................................. 36?3.1.4?Pellet Sampling Ports ....................................................................................... 36?3.1.5?Apparatus for Residence Time Measurement .................................................. 37?3.1.6?Purging System ................................................................................................. 41?3.2? Instrumentation for Off-gas Analysis ..................................................... 45?Chapter 4 Material Characterization ................................................................................. 46?4.1? Materials ................................................................................................. 46?4.2? Density, Porosity, Moisture Content and Calorific Value ...................... 49?4.3? Elemental (CHN) Analysis ..................................................................... 50?4.4? Microbial Analysis .................................................................................. 51?4.5? Material Surface Scanning Electron Microscopy (SEM) ....................... 53? viii4.6? Adsorption Analysis by Micromeritics AutoChem 2920 II .................... 56?4.6.1?Carbon Dioxide TPD Procedure ....................................................................... 56?4.6.2?Carbon Dioxide TPD Results ........................................................................... 60?4.6.3?Carbon Monoxide TPD Procedure ................................................................... 63?4.6.4?Carbon Monoxide TPD Results ....................................................................... 64?4.6.5?Oxygen TPD ..................................................................................................... 67?4.6.6?Mass Spectrometry ........................................................................................... 68?4.7? Concluding Remarks ............................................................................... 71?Chapter 5 Stratification and Effect of Temperature and Relative Humidity on Off-gases in Stored Wood Pellets .......................................................................................................... 72?5.1? Materials and Methods ............................................................................ 72?5.1.1?Gas Measurement ............................................................................................. 72?5.1.2?Temperature and Relative Humidity Measurement ......................................... 73?5.2? Results and Discussion ........................................................................... 75?5.2.1?Off-gas Concentration in Silo Head-space ....................................................... 75?5.2.2?3D Analysis of Longitudinal Distribution of Emitted Gases ........................... 80?5.2.3?Oxygen Depletion ............................................................................................. 88?5.2.4?3D Analysis of Radial Distribution of Emitted Gases ..................................... 91? ix5.2.5?Effect of Temperature and Relative Humidity on Off-gas Stratification ......... 92?5.2.6?Longitudinal and Radial Distribution of Temperature ..................................... 97?5.2.7?Changes in Relative Humidity During Storage of Wood Pellets ................... 102?5.3? Concluding Remarks ............................................................................. 106?Chapter 6 Effectiveness of Purging on Gas Emission Buildup ....................................... 108?6.1? Materials and Methods .......................................................................... 109?6.1.1?Purging Experiments ...................................................................................... 109?6.1.1.1? Gas Measurements ............................................................................ 110?6.1.1.2? Temperature and Relative Humidity Measurement .......................... 110?6.1.2?Gas Mixing Experiments ................................................................................ 111?6.1.2.1? Calibration of Thermal Conductivity Detectors ............................... 111?6.1.2.2? Residence Time Distribution (RTD) Experiments ........................... 114?6.2? Results and Discussions ........................................................................ 115?6.2.1?RTD Experiments ........................................................................................... 115?6.2.2?Purging Efficiency .......................................................................................... 124?6.2.2.1? Gas Analysis ..................................................................................... 124?6.2.2.2? Temperature and Relative Humidity Changes During Purging Experiments  .......................................................................................................... 129? x6.3? Concluding Remarks ............................................................................. 134?Chapter 7 Conclusions and Recommendations ................................................................ 135?7.1? Conclusions ........................................................................................... 135?7.2? Limitations and Contributions of This Research .................................. 138?7.3? Recommendations ................................................................................. 139?References ............................................................................................................................ 141?Appendices ........................................................................................................................... 152?Appendix A? Pilot Silo and All Attached Monitoring and Measurement Instruments     .......................................................................................................... 153?Appendix B? Adsorption Tests .............................................................................. 160?B.1?Equipment and Sample Tube Pictures .............................................................. 160?B.2?TCD Calibration................................................................................................ 161?B.3?Typical TPD Curves Obtained for CO2, CO and O2 ......................................... 162?Appendix C? SEM Images of Treated and Untreated Wood Pellets ..................... 164?Appendix D? GC Scan for Gas Sample Collected from Pilot Silo ........................ 166?Appendix E? Emission of Off-gases in Oxygen-free and Oxygen-rich Environments     .......................................................................................................... 167?Appendix F? Peak Emission Factor for CO, CO2, CH4 and H2 as Effected by Head-space Percentage ...................................................................................................... 172?Appendix G? Sensor and Gas Sampling Ports Coordinates ................................... 176? xiAppendix H? Contour Plot of Temperature and Relative Humidity ...................... 179?Appendix I?CO and CO2 Concentration Fitted to 1st Order Kinetic Reaction (Data obtained from [32] ) ................................................................................................ 184?Appendix J?OPI Temperature and Rh Recording System ........................................ 186?Appendix K? Measured F(t) Curves at Different Velocities .................................. 189?Appendix L? TCD Reading During RTD Experiment .......................................... 191?Appendix M? Measured and Predicted Concentration of off-gases as a Function of Time [Using the fmix concept] .................................................................................. 192?    xiiList of Tables Table  2.1 Fatalities due to off-gassing in biomass storage enclosures. .................................... 8?Table  2.2 The threshold limit value of carbon monoxide, carbon dioxide, methane and oxygen [70, 71] ....................................................................................................... 17?Table  3.1 Location of the gas sampling ports on the experimental silo ................................. 34?Table  4.1 Initial characteristics of wood pellets used in the pilot silo tests ............................ 47?Table  4.2 Physical properties of wood pellet samples used in lab-scale gas adsorption tests 48?Table  4.3 Properties of wood pellets used in pilot silo experiments before and during storage ................................................................................................................................ 52?Table  4.4 Description of events in gas adsorption process ..................................................... 59?Table  4.5 Supplemantry data needed for calculating energy of desorption for CO2 and CO tests ......................................................................................................................... 67?Table  5.1 Compositions (volumetric) of standard gases for GC calibration .......................... 73?Table  5.2 Comparison of peak emission factor for CO, CO2 and CH4with previous research ................................................................................................................................ 79?Table  6.1 Residence time, variance and system dispersion number at different purging velocities ............................................................................................................... 121? xiiiTable  6.2 Off-gas concentration inside the silo before first, second and third purging experiment ............................................................................................................ 124 Table E. 1 Reactor and test conditions for 12 reactors used in oxygen-free and oxygen-rich experiments ........................................................................................................... 168?Table E. 2 Off-gassing results from oxygen-free and oxygen-rich experiments .................. 171 Table F. 1 Peak emission factor for gases emitted from wood pellets reactors with different head-space percentage .......................................................................................... 173 Table G. 1 Coordinate of all temperature sensors located on 9 vertical thermocouple cables .............................................................................................................................. 177?Table G. 2 Coordinate of all relative humidity sensors located on 5 vertical cables ............ 178?    xivList of Figures Figure  2.1 Temperature development and processes responsible for self-heating in stored biological material [73] ........................................................................................... 20?Figure  2.2 CO2 concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets .......................................................... 24?Figure  2.3 CO concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets .......................................................... 24?Figure  2.4 CH4 concentration in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets .......................................................... 25?Figure  3.1 Location of gas and pellet sampling ports (left) and location of horizontal thermocouples in the silo (right) ............................................................................. 33?Figure  3.2.  Position of the vertical cables along the silo (top view) ...................................... 36?Figure  3.3 Schematic of calibration set-up for TCD ............................................................... 39?Figure  3.4 Schematic of set-up for RTD experiment .............................................................. 40?Figure  3.5 Schematic of gas distributors for purging ............................................................. 42?Figure  3.6 Schematic of nitrogen/air purging system for the experimental silo ..................... 43?Figure  3.7 Schematic of set-up for purging experiment ......................................................... 44?Figure  4.1 Two SEM micrographs of regular pellet samples before adsorption tests ............ 54? xvFigure  4.2 Two SEM micrographs of regular pellet samples after CO adsorption tests ........ 55?Figure  4.3 Temperature program and gas flow events for gas adsorption experiments using Micromeritics device .............................................................................................. 59?Figure  4.4 CO2 uptake for three different wood pellet samples ............................................. 61?Figure  4.5 Typical CO2 TPD curves for a) Regular pellet b) Torrified pellet and c) Steam-treated pellet at desorption temperature of 110 oC ................................................. 63?Figure  4.6 Typical CO TPD curve for regular wood pellets at desorption temperature of 150 oC ............................................................................................................................ 65?Figure  4.7 Typical CO TPD curve for regular wood pellets at desorption temperature of 180 oC ............................................................................................................................ 65?Figure  4.8 CO uptake rate as a function of desorption temperature ....................................... 66?Figure  4.9 Typical O2 TPD curve for regular wood pellets at desorption temperature of 150 oC ............................................................................................................................ 68?Figure  4.10 Profile of effluent gases during thermal treatment of pellet samples as measured by mass spectrometer using 0.8 g of sample .......................................................... 70?Figure  5.1 Top view of the pilot silo with the coordinate [r,y] of all sensors located on 9 vertical cables ......................................................................................................... 74? xviFigure  5.2 CO2  concentration in head-space of the silo as a function of storage time for wood pellets. (G0: head-space gas sampling port) [Concentrations are volume percentage] .............................................................................................................. 76?Figure  5.3 CH4  concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage] ................................................................................................................................ 77?Figure  5.4 CO concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage] ................................................................................................................................ 77?Figure  5.5 CO/CO2 ratio in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage] ................................................................................................................................ 78?Figure  5.6 O2 concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port)  [Concentrations are volume percentage] .............................................................................................................. 78?Figure  5.7 3D-map of CO2 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ....................................... 81?Figure  5.8 (a) 3D- map of CO2 concentration in the silo at G0 to G12 (Even sampling ports at one side) during 63 days of storage (b) Contour of CO2 concentration over time [Concentrations are volume percentage] ................................................................ 83? xviiFigure  5.9 (a) 3D- map of CO2 concentration in the silo at G0 to G13 (Odd sampling ports at one side) during 63 days of storage (b) Contour of CO2 concentration over time [Concentrations are volume percentage] ................................................................ 84?Figure  5.10 3D- map of CO concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ....................................... 86?Figure  5.11 3D- map of CH4 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ....................................... 87?Figure  5.12 3D- map of O2 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ....................................... 90?Figure  5.13 CO2 concentration in the silo head-space and the bottom of the silo at different radial positions ........................................................................................................ 92?Figure  5.14 Temperature in the silo recorded by cable 3 (sensors 1 and 29) and cable 1 (sensors 1 and 29) ................................................................................................... 93?Figure  5.15 Contour plot of temperature for all sensors on cable C1M during 63 days of storage ..................................................................................................................... 94?Figure  5.16 (Left) Contour plot of CO2  and (Right) CO concentration at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ............. 95?Figure  5.17 (Left) Contour of CH4 and (Right) O2 concentration at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] ....................... 95? xviiiFigure  5.18 Contour plot of relative humidity for all sensors on cable C1M during 63 days of storage ..................................................................................................................... 96?Figure  5.19 Contour plot of temperature for all sensors on the cables C1M, C2M, C3M, C5M, C1 and C3 during 63 days of storage ......................................................... 101?Figure  5.20 Contour plot of relative humidity for all sensors on cables C1M to C5M during 63 days of storage ................................................................................................. 104?Figure  5.21 Contour plot of temperature inside the pilot silo on day 15 of storage (r=0 refers to the centre of the silo) ........................................................................................ 105?Figure  5.22 Contour plot of relative humidity inside the pilot silo on day 15 of storage (r=0 refers to the centre of the silo) .............................................................................. 105?Figure  6.1 Plot of detected signals as a function of volume percent He injected for A) TCD#1, and B)TCD#2 [Signal amplification ratio=100; Current=70mA; TCD sample flow rate = 8.4?10-7 m3.s-1] ..................................................................... 113?Figure  6.2 Measured F(t) for the bed of pellets, for air velocity U = 8.22E-4 to 4.11E-3 m.s-1 . Tracer gas was injected from the lower diffuser. Sampling was done at the silo outlet. .................................................................................................................... 118?Figure  6.3 Measured F(t) for the bed of pellets, for air velocity U = 5.70E-3 m.s-1 . Tracer gas was injected from the lower diffuser. Sampling was done at the silo outlet. ....... 119?Figure  6.4 Plot of Peclet number versus Reynolds number .................................................. 122?Figure  6.5 Mean residence time of the tracer gas as a function of superficial velocity ........ 123? xixFigure  6.6 Predicted CO2 concentrations over time during silo 1st purging using numerical solution (DL=6.2E-4 m2.s-1 and initial CO2 concentration was 2.38%) ................ 126?Figure  6.7  Predicted and experimental CO2 concentrations over time during silo purging using numerical solution (CO2 initial concentration was 2.38%) ......................... 127?Figure  6.8 CO2 concentration measured and predicted as a function of time during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts] .............................................................................................................................. 127?Figure  6.9 CO concentration measured and predicted as a function of time during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts] .............................................................................................................................. 128?Figure  6.10 CO2 concentration predicted as a function of time and different dispersion coefficients during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts] .................................................................................... 128?Figure  6.11 Number of turnover as a function of velocity to reduce CO2 concentration from 2 to 0.5% in the silo head-space .............................................................................. 129?Figure  6.12 Contour plot of temperature for all sensors on the cables C1M-C5M during first purge experiment done on April 26th (Data shown represent the minute-to-minute temperature data from 00:00 AM to 23:59 PM on April 26th , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min) ........ 132? xxFigure  6.13 Contour plot of relative humidity for all sensors on the cables C1M to C5M during first purge experiment done on April 26 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on April 26 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min) ........ 133 Figure A.1 Pilot scale silo installed at Clean Energy Research Centre, Department of Chemical and Biological Engineering, University of British Columbia .............. 153?Figure A.2 Gas sampling ports located at different levels of the silo ................................... 154?Figure A.3 View from the top of the pilot silo showing the installed rods inside the gas sampling ports ....................................................................................................... 154?Figure A.4 Solid sampling ports located at 6 different elevations of the silo ....................... 154?Figure A.5 Pressure transducers located at different elevations of the silo .......................... 155?Figure A.6 Side view of the silo showing wall thermocouples as well as horizontal thermocouples located at 5 different levels .......................................................... 155?Figure A.7 View of the silo from the top. The top end of the vertical thermocouples are shown with the black wires attached to them. 4 of the thermocouple cables are located in inner ring and 4 in the outer ring. ......................................................... 155?Figure A.8 Purging system attached to the silo(connected to both nitrogen and air). .......... 156?Figure A.9 Manometer attached to the silo outlet to avoid pressure build up inside the silo 156?Figure A.10 TCD #1 and #2, amplifier, pump, flowmeters and data logging system .......... 157? xxiFigure A.11 Flowmeters attached to TCDs to measure sampling rate ................................. 157?Figure A.12 TCD#1, TCD #2 and attached amplifier .......................................................... 158?Figure A.13 Reference (air) gas cylinder, He gas cylinder and building air line ................. 158?Figure A.14 Nitrogen/Air distributor 1 and 2 ....................................................................... 159?Figure A.15 Air/N2 and He flowmeters, 3-way and needle valves used in purging and RTD experiments ........................................................................................................... 159 Figure B. 1 Micromeritics Autochem 2920 II ...................................................................... 160?Figure B. 2 Sample tube loaded with wood pellet sample for adsorption test...................... 160?Figure B. 3 Micromeritics TCD calibration curve for O2 ..................................................... 161?Figure B. 4 TCD signal and temperature reading over time for carbon dioxide .................. 162?Figure B. 5TCD signal and temperature reading over time for carbon monoxide ............... 163 Figure C. 1 SEM micrographs of white wood pellet surface before CO adsorption ............ 164?Figure C. 2 SEM micrographs of white wood pellet surface after CO adsorption ............... 165 Figure D. 1 (a) FID reading for CO2, CO and CH4 and (b) TCD reading for Helium, Oxygen and Nitrogen ......................................................................................................... 166 Figure E. 1 Empty containers equipped with 2 valves, fittings and stainless steel rod for purging and gas sampling ..................................................................................... 170? xxiiFigure E. 2 Reactors filled to 75% with wood pellets for experiments ................................ 170 Figure F. 1 Peak emission factor for carbon monoxide for different head-space percentage after 60 days .......................................................................................................... 174?Figure F. 2 Peak emission factor for carbon dioxide for different head-space percentage after 60 days .................................................................................................................. 174?Figure F. 3 Peak emission factor for methane for different head-space percentage after 60 days ....................................................................................................................... 175?Figure F. 4 Peak emission factor for hydrogen for different head-space percentage after 60 days ....................................................................................................................... 175 Figure G. 1 Coordinate of gas sampling ports and some sensors located on 2 cables.......... 176 Figure H. 1 Contour plot of temperature for all sensors on the cables C4M, C2 and C4 during 63 days of storage ................................................................................................. 179?Figure H. 2 Contour plot of temperature for all sensors on the cables C1 to C4 during first purge experiment done on April 26 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on April 26 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min) ............................ 180?Figure H. 3 Contour plot of temperature for all sensors on the cables C1 to C4 during 3rd purge experiment done on October 19 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on October 19 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=590 min) ........ 181? xxiiiFigure H. 4 Contour plot of temperature for all sensors on the cables C1M to C5M during 3rd purge experiment done on October 19th (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on October 19 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=590 min) ........ 182?Figure H. 5 Temperature recorded in the lab from January 14 to April 24 2011 (The thermometer was located next to the lower part of the silo which was close to the lab door and exposed to outside air) ..................................................................... 183 Figure I. 1 CO concentration for pellets with 4% MC at 25 oC (stored for 60 days) ........... 184?Figure I. 2 CO concentration for pellets with 4% MC at 40 oC (stored for 60 days) ........... 184?Figure I. 3 CO2 concentration for pellets with 4% MC at 25 oC (stored for 60 days) .......... 185?Figure I. 4 CO2 concentration for pellets with 4% MC at 40 oC (stored for 60 days) .......... 185 Figure J. 1 Temperature reading for each cable at any moment ........................................... 186?Figure J. 2 Minimum, Maximum and average temperature recording ................................. 186?Figure J. 3 Relative humidity data reading for cables C1M to C5M at any moment ........... 187?Figure J. 4 EMC (Equilibrium Moisture Content) at any point [EMC is calculated and not measured] .............................................................................................................. 187?Figure J. 5 Pressure reading from 6 transducers and temperature reading from horizontal and wall thermocouples ............................................................................................... 188?  xxivFigure K. 1 F(t) measured for the bed of pellets, for U = 8.22E-4 to 4.11E-3 m.s-1. [TCD 1 Replicate 2(left) and Replicate 3 (right)] .............................................................. 189?Figure K. 2 F(t) measured for the bed of pellets, for U = 8.22E-4 to 4.11E-3 m.s-1. Tracer gas was injected with from the lower diffuser. Sampling was done at the reactor outlet. (TCD 2 Replicate 1, 2 and 3) ................................................................................ 190 Figure L. 1 TCD#1 detected signals as a function of time [Signal amplification ratio=100; current=70mA; TCD sample flow rate = 8.4?10-7 m3.s-1]. .................................. 191 Figure M. 1 Predicted off-gas concentrations over time during silo purging using equation 6.15 ....................................................................................................................... 193?Figure M. 2 CO2 concentration measured and predicted at different elevation in the silo as a function time during 1st purging experiment [U=1.23E-3m.s-1] ........................... 195?Figure M. 3 CO concentration measured and predicted at different elevation in the silo as a function time during 1st purging experiment [U=1.23E-3 m.s-1] .......................... 195?Figure M. 4 CO2 concentration measured and predicted at different elevation in the silo as a function time during 2nd purging experiment [U=1.64E-3m.s-1] .......................... 196?Figure M. 5 CO2 concentration measured and predicted at different elevation in the silo as a function time during 3rd purging experiment [U=1.64E-3 m.s-1] .......................... 196?Figure M. 6 CO concentration measured and predicted at different elevation in the silo as a function of time during 2nd purging experiment [U=1.64E-3 m.s-1] ..................... 197? xxvFigure M. 7 CO concentration measured and predicted at different elevation in the silo as a function of time during 3rdpurging experiment [U=1.64E-3 m.s-1] ...................... 197?    xxviNomenclature ACGIH American Conference of Governmental Industrial Hygienists A0  Quantity Adsorbed As  Mean surface area of particles C  Constant (related to desorption rate) Ci  Concentration of gas species i CSTR  Continues Stirred-Tank Reactor CN0/CNt Correction factor D  Diameter Dp  Equivalent particle diameter DL  Axial dispersion coefficient DT  Transverse dispersion coefficient Ed  Energy of desorption EMC  Equilibrium Moisture Content E(t)  Residence time distribution function f  Fine percentage  xxviifmix  Mixing factor F(t)  Cumulative distribution FID  Flame Ionization Detector h  Hour HD  Head-Space K  Kelvin kV  Kilovolts MC  Moisture content Mp  Pellets? weight mm  Millimeter Mwt  Molecular weight of gas species i Nturnover  Number of turnover L  Characteristic length OSHA  Occupational Safety and Health Administration P  Pressure Pe  Peclet number Per  System peclet number  xxviiiPEL  Permissible Exposure Limit ppm  Part per million R  Universal gas constant Rh  Relative humidity RTD  Residence Time Distribution SEM  Scanning Electron Microscopy SPF  Spruce, Pine and Fir t  Time T  Temperature TCD  Thermal Conductivity Detector TLV  Threshold Limit Value TP  Desorption peak temperature oC  Degree celsius U  Superficial velocity u  Interstitial velocity U?   Volumetric flow rate  xxixV  Experimental silo volume VC  Cell volume Vg  Gas volume VOC  Volatile Organic Compounds VP  Particle volume VR  Reference volume %wb  Moisture content (wet basis) Wt %  Weight percent Greek Symbols ?  Heating rate  ?  Fluid density ?b  Bulk density ?p  Pellet density ?? ? Porosity ?2  Variance ?  Residence time  xxx?  Fluid viscosity     xxxiAcknowledgements Foremost, my sincere gratitude and appreciation goes to my PhD Supervisors, Dr. Shahab Sokhansanj, Dr. Jim Lim and Dr. Anthony Lau for their excellent guidance, time dedication, support and patience over the course of this PhD study. Without their insight both into my research and personal development, I don?t believe I could have ever achieved all that I have.  Also I extend my sincere thanks to my committee members Dr. Xiaotao Bi and Dr. Victor Lo for their time and being a part of my examination committee and for their invaluable advice on my research. I am also grateful to Mr. Staffan Melin for his continuous help during my work in facilitating our relationship in receiving material and technical instruments for my experiments.  I owe much gratitude to all staff members at Chemical and Biological Engineering Department at the University of British Columbia specially Richard Ryoo, Gordon Cheng, Helsa Leong, David Roberts, Alex Thng, Doug Yuen, Joanne Dean and Amber Lee for their administrative and technical assistance.  The financial and support from Natural Science and Engineering Research Council of Canada, Wood Pellet Association of Canada and the Oak Ridge National Laboratory is also acknowledged. Thanks are also extended to Premium Pellet Ltd. and Fibreco Export Inc. for donating wood pellets for this study. I would like to thank OPI Integris for donating temperature and relative humidity cables and their technical support throughout my experimental work.  xxxiiI would like to thanks all my colleagues in Biomass and Bioenergy Research Group especially Zahra Tooyserkani and Ladan Jafari Naimiwho helped me in the course of PhD study. Also I would like to thank Marius W?hler for providing the opportunity for me to meet his colleagues in Bioenergy 2020+ for insightful discussions and suggestions. I am also very much indebted to all my friends for their continuous support; Special thanks to Carolina Casas-Cordero, Ayat Shariati, Shahrzad Jooya, Sona Kazemi, Shahrzad Talachian and Siduo Zhang to name a few.  Special and deepest thanks are expressed to my mom, Nahid and my brothers, Fardad, Farzad and Fahim for their continuous support and encouragement to work hard and to continue pursuing a PhD abroad.  Without them I could not have made it here.  Last but not least, I am greatly indebted to my devoted husband and my love, Hooman for his kindness, understanding, patience and continuous support and assistance during the course of this work.         xxxiii  Dedication   To My Loving Husband, Hooman My Extraordinary Mom, Nahid & My Wonderful Brothers, Fardad, Farzad and Fahim   1  Chapter 1 Introduction 1.1 Background Bio-fuels derived from renewable resources have been receiving increased attention due to a steady rise in energy prices and price volatility. Increases in greenhouse gas emissions have been attributed to fossil fuels. Wood pellets are renewable non-fossil source of energy. They are made from excess forest residues. The majority of wood pellets are produced by milling wood chips, bark, planer shavings or sawdust into a fine powder, which after drying is compressed into pellets [1]. Unlike a moist bulky raw biomass, wood pellets are adequately dry and dense to be suitable for long term storage and transport. Combustion of wood pellets generates less particulate emission than other forms of solid biomass.  The pelletization process involves pressuring saw dust through a circulate die. This produces a dense cylindrical shaped pellet. The compression coupled with heating causes the lignin to plasticize and act like glue that provides for cohesion of the pellets [2, 3]. Generally the bulk density of wood pellets are around 650 kg.m-3 based on wood pellets with less than 10% moisture content compared to 200 kg.m-3 for wood chips at 25% moisture content. Wood chips require at least three times the storage space as wood pellets for the same energy content. The world consumption of wood pellets in 1995 was less than 200,000 metric tons. It increased to 14.5 million tonnes in 2012 and expected to exceed to 50 million tonnes in 2020.The pellet industry is a well-established industry in British Columbia. Out of 1.9 million tonnes of wood pellets produced in Canada in 2011, 90% is exported to Europe and 2  are transported in large ocean vessels in 7,000 to 20,000 tonne lots. Pellets may also be transported in bulk containers, railcars or tank trucks. Pellets from BC have a typical moisture content of 5% and an energy density of 18 GJ/tonne. A number of wood pellet mills are in development with most companies focusing on the pellet production while the finished product storage is usually an afterthought. One major issue is associated with safe handling and storage of wood pellets. Before shipping, wood pellets are stored in silos for a few days up to a few months. Moreover, since the major consumption of wood pellets is not where most of the pellets are manufactured, they are exported in increasing volumes as a bulk commodity by ocean vessels. The temperature inside the wood pellet bulk increases over time due to self-heating. If the storage time is long, the temperature can reach 60?C or higher. The initial rise in temperature is usually due to microbial activity and it is followed by the exothermic oxidation of fatty acids when the temperature becomes sufficiently high. At the last stage the endothermic cracking or pyrolysis at high temperature will occur [4]. Particularly for wood pellets the second stage - oxidation of fatty acids- is suspected to contribute most to self-heating [5]. Fires and explosions in wood pellet handling operations are frequent [6-13]. These accidental fires are some due to dust explosion and some due to spontaneous combustion. These incidences cause major loss of capital investment or revenue.  One major fire that happened in pellet silo as reported by Melin [14] was at the Port of Harnosand, Sweden in September 2004. Another fire occurred in pellet storage facilities in September 2007 in Kristinehamn, Sweden. In July, 2008, a fire broke out in one of the newly 3  constructed silos at the Fibreco wood chip and handling facility in North Vancouver, BC, Canada. The local newspaper (North Shore News, July 6, 2008) reported that firefighters arrived to find black smoke and flames pouring from the structure's roof vents. "It was just a very tedious, very difficult, exhausting fire to put out." Wood pellets can also cause significant health issues because of the dust from mechanical degradation and off-gases from chemical degradation. Wood pellets decompose over time and emit gases such as CO, CO2 and CH4 during storage especially in high temperature conditions due to microbial growth on pellets, chemical oxidation of lipids in the pellet as well as hydro-thermal migration of moisture between pellets in bulk. These biochemical processes consume oxygen causing oxygen depletion in the space. Off-gassing is especially dangerous during ocean transportation as it may rapidly produce lethal levels of CO and an oxygen-deficient environment that could affect adjacent access space and make those spaces dangerous for people to work [15]. The surrounding temperature and humidity and the state of pellets are known to control the degree of off-gassing from wood pellets storage.  Developing methods of mitigating the evolution and dispersion of the toxic gasses within a pellet storage space will reduce the danger of working around this product. For instance, frequent purging or ventilation is needed to slow down and prevent self-heating in bulk pellets. Industry practice is to ventilate silos when the temperature exceeds a set point of about 40oC. Pellets can also be transferred from the heated silo to another silo in an attempt to cool down the pellets. This transfer reduces the pellet temperature by about 10oC degrees. At present, there is not yet a well-designed purging and ventilation system for these silos to bring down the gas concentration to a safe level. Ventilation is needed to control the silo 4  temperature and prevent spontaneous combustion.  Also purging of the storage space with an inert gas or even air is necessary to bring down the concentration of the emitted gases to a safe level. There is no information available on stratification of off gases in silo enclosures and how a purging and ventilation systems may reduce the gas concentrations in a bulk storage of pellets. 1.2 Objectives of the Thesis The main goal of this research is to quantify off gassing characteristics of white wood pellets when stored in an experimental silo.  Specific objectives are as follows: 1) To characterize wood pellet properties with respect to gas adsorption-desorption processes.  2) To quantify spatial and temporal concentrations of the off-gasses and thermal conditions in the pilot scale silo. 3) To investigate the efficacy of ventilating silo to reduce the concentration of off gases. This research addresses storages of pellets in sealed containers similar to ship holds in ocean going vessels. Purging studies are done to investigate a method of reducing high concentrations of gases or lowering the temperature of pellets within the sealed containers. 1.3 Approach of the Thesis The approach to achieve the objectives of the thesis combined experimental studies to quantify the emission of gases during storage of wood pellets in a pilot scale silo, study the stratification of emitted gases over time, study the changes in temperature and relative 5  humidity in the storage and finally to look into the purging and ventilation requirements of such silos. Chapter 1 introduces the problems and explains the needs for studying the purging systems for wood pellet silos. Chapter 2 includes a comprehensive literature review on storage of biomass, off-gassing, self-heating and decomposition of biomass material during storage, kinetics of off-gassing as well as approaches for inerting and ventilating storage silos.  In chapter 3 details of the experimental set-ups and the design of experiments are explained. Chapter 4 outlines the material characterization procedures and techniques using different instruments. It also includes data on the material properties before and after storage under different conditions. Chapter 5 is focused on the emission of gases over time in a pilot scale silo filled with wood pellets. The emission of individual non-condensable gases are measured and studied during a long storage time period. The changes in concentration are further related to the spatial temperature and degree of gas reaction with the stored material. Changes of temperature and relative humidity in the pilot scale silo during storage of wood pellets are presented in this chapter. In particular the stratification of temperature during short term and long term storage was analyzed and discussed.  Chapter 6 outlines the experiments for determining the gas residence time in the pilot scale silo. Discussions are made on the degree of mixing in the bed, the results from the purging cycles as well as purging effectiveness.  Finally,  Chapter 7 includes a summary of the main results and contributions to this field of study, as well as recommendations for future works. 6  Chapter 2 Literature Review 2.1 Biomass Storage With the emphasis on using biomass recently it is getting clear that there is a need to understand biomass storing requirements. When wood pellets are produced, they will be stored on the plant before being shipped. Also when shipped, they are stored in very large ocean going dry-bulk tankers. They are shipped from the ports of Vancouver and Prince Rupert in the West and Halifax in the East to Europe. A typical amount of transportation is reported to be 7,000 to 30,000 tons [16]. The storage amount of 40,000 tons and 47,000 tons in ships is also recorded for transports between Canada and Europe [17]. When pellets are shipped in large volumes, they are classified as hazardous material due to the potential for off-gassing and self-heating.  The first fatal accident happened onboard MV Weaver Arrow on May 9, 2002 in the Port of Rotterdam during discharge of pellets from British Columbia. Before this incident there was not any knowledge of off-gassing from wood pellets. The safety regulations that issued by International Maritime Organization (IMO) earlier had warnings of oxygen depletion caused by wood such as lumber, chips and logs which all have relatively high moisture content. The second recorded incident happened on November 19, 2006 in the port of Helsingborg, Sweden when discharging wood pellets from BC [14, 15, 18].  After this, studies on off-gassing were considered urgent and necessary. In this incident, one seaman was killed, a stevedore was seriously brain injured and a large number 7  of stevedores and rescue workers sustained less serious injuries after entering the unventilated stairway. The report investigating the accident indicates that as the workers entered the stairway adjacent to the cargo hatch, they were collapsed by the high concentration of CO [14].  A number of fatal accidents have occurred due to personnel entering spaces where large amount of wood pellets were stored. Table  2.1 lists all recorded fatalities and available data on injuries due to biomass handling. Major concerns with handling and storage of wood pellets are connected to off-gassing, self-heating, and formation of dust.  8  Table  2.1 Fatalities due to off-gassing in biomass storage enclosures. Year Place Number of people killed Number of people severely injured circumstances Source of information Material 2002 Rotterdam, The Netherlands  1 2 A stevedore entered a stairway between two cargo holds onboard MV Arrow Weaver  World Pellet Conference 2002, Stockholm  Pellets 2005 Gruv?n, Sweden  1  Sailor entered a stairway next to a cargo hold onboard MV Eken with pulp logs  Transportstyrelsen, Swedish Maritime Administration Report B2006-2  Pulp logs 2005 US East Coast port 1  Stevedore was working in an open cargo hold onboard a Saga Forest Carrier vessel with green lumber and was overcome by oxygen depletion  Saga Forest Carrier Green Lumber 2006 Port of Helsingborg, Sweden  1 1 Sailor entered a stairway between two cargo hold onboard MV Saga Spray with pellets  Transportstyrelsen, Swedish Maritime Administration Report B2007-1  Pellet 2006 Port of Skelleftehamn, Sweden  1  Sailor entered a stairway next to a cargo hold onboard MV Noren with wood chips  Swedish Maritime Magazine, Issue 1-2007  Wood chips 2007 Port of Timr?, Sweden  2  One sailor and the captain of the vessel MV Fembriaentered the stairway leading down to the cargo hold and was overcome by oxygen depletion and CO  Secotidningen June 2007  Saw logs 9  Year Place Number of people killed Number of people severely injured circumstances Source of information Material 2007 Finland  1  One person entering a pellet storage with 10 tonnes Vasa Arbetsskyddsdistrikt, Finland  Pellet 2008 Finland 1  One person entering a pellet storage with 10 tonnes Vasa Arbetsskyddsdistrikt, Finland  Pellet 2009 Bornholm, Denmark  2  Two crew members onboard MV Amirante loaded with pellets entered a stairway leading down to the cargo hold  Police Authority, R?nne, Bornholm  Pellet 2010 Germany 1  One person entering a pellet storage with 150 tonnes Propellets, Austria  Pellet 2010 Ireland 1  One person entering a pellet storage with 10 tonnes Health and Safety Authority, Dublin, Ireland  Pellet 2011 Switzerland 1  One pregnant woman entered a storage with 100 tonnes pellets  NeueLuzernerZeitung February 9, 2011  Pellet Total 14 3    10  2.2 Off-gassing 2.2.1 Condensable and Non-condensable Gases Stored wood pellets are known to emit non-condensable gases like CO, CO2, CH4 and small amounts of aldehydes (alcohol dehydrogenated) and ketones including hexanal and pentanal in addition to acetone and methanol [19-22].More volatile organic compounds emit from stored sawdust and wood chips compared to stored wood pellets [23-25]. This is because of the fact that most of the VOCs easily evaporate during heating and drying processes [26, 27] and as wood pellets pass through high temperature during drying and pelletization, decomposition of extractives would result in less VOCs emission. During pelletization, sawdust is dried at temperatures from 100 to 400 ?C so the moisture content goes below 10%. Roffael [28], Rupar and Sanati [29] also reported the emission of VOCs during biomass storage. Large amount of monoterpenes and other volatile organic compounds are emitted during drying [26, 27, 30, 31]. Arshadi and Gref [20] have mentioned VOCs being emitted from stored pellets and the circumstances under which the concentration of these gases reached high levels. However, a large portion (>30%) of the emission from stored wood pellet is CO as opposed to emission of VOCs [20, 21, 32]. High levels of CO were first measured in compartments of a ship in 2002 when the ocean vessel was discharging pellets in Rotterdam, the Netherlands. [14]. In that incident one person was killed and several people were severely injured. The causes of fatal accidents and injuries have been traced to high concentrations of CO and low concentrations of O2 [33]. 11  Emissions of CO2, CH4 and N2O are reported in stored wood chips by [24]. Svedberg et al. [21] found that there is a high level of hexanal and carbon monoxide associated with storage of wood pellets and can lead to an occupational and domestic health hazard. It is not just limited to wood pellets as it was due to general degradation process of wood and got facilitated by drying at elevated temperature. Emission of carbon monoxide, carbon dioxide, methane and aldehydes were also detected during ocean transportation of wood pellets. It could produce lethal levels of carbon monoxide and create an oxygen-deficient atmosphere that may affect adjacent access space and make these spaces dangerous for people to enter [21, 34]. CO binds to hemoglobin about 200 times stronger than oxygen and inhibits the ability of oxygen to bind with hemoglobin. The reduced partial pressure of oxygen increases the available sites for CO-bindings.  When organic matters are stored especially in the presence of air and light, they emit small amounts of carbon monoxide and the emission rate increases with temperature [35]. Stored biomass, decomposes over time and emits toxic gases and creates an oxygen deficient environment [20, 36, 37]. Microorganisms break down the sugar; produce heat, CO2 and water as well as creating a more favorable environment for further microbial activities [38]. The water content in wood plays an important role on microbial activities and according to a study [38], the temperature in a biomass pile depend on material initial moisture content. Even at low temperatures, depending on the composition of the biomass and its moisture content fungi and bacteria can colonize the material and lead to heat generation. Microbial activity is reduced by drying biomass but fungi and bacteria can survive in a dormant stage and will revive by rewetting [39]. 12  Some studies show that an increase in humidity could also accelerate biomass decomposition by releasing heat and gas [25, 40, 41]. Depending on the access of the microorganisms to nutrition in the material, the moisture content and the size of the material pile, the degree and rate of gas generation due to microbial reaction would be different [38]. At higher temperature the chemical process is dominant over biological process. The gas and heat production from chemical processes in biomass piles are caused by several types of reactions such as hydrolysis, pyrolysis and oxidation [42]. Pyrolysis and oxidation progress exponentially as temperature rises. According to [19, 23], oxidation of unsaturated fatty acids and other extractives under certain conditions would cause the release of volatile aldehydes such as hexanal and pentanal compounds. The aldehydes formed during storage depend on the storage conditions, the biomass species and characteristics and its pre-treatment and pellet manufacturing parameters. The aldehydes that are usually reported as emission from biomass are methanol, ethanol, acrolein, 2-propanone, propanal, butanal, benzaldehyde, pentanal, toloulaldehyde, hexanal, octanal and nonanal [19]. Further auto-oxidation of aldehydes and ketones can lead to formation of carboxylic acids. For example, further oxidation of hexanal can result in formation of hexanoic acid [19]. Linoleic acid which is a polyunsaturated acid contains most of the free acids and triglycerides in the wood [43, 44]. The oxidation of this acid and its esters lead to the emission of hexanal [45]. Break down of wood hemicellulose could also cause emission of some organic acids such as acetic [46]. 13  Measurement and analysis of the gas in a sealed warehouse of 7000 tons of rapeseed showed CO concentration of 300-400 ppm and emission rate of 200 mg/ton/day [36]. CO emission rate in a wheat grain warehouse was 9 mg/ton/day [47]. Measurement done on the emission of total hydrocarbon during drying of pine lumber showed emission factors of 2.21 and 1.49 g/kg of oven dry wood for commercial and laboratory kiln respectively. The total monoterpene emissions were 63% and 48% of total hydrocarbon emissions for the two replicated tests [48]. Researchers have mentioned CO as the main component being emitted from storage of wood pellets and emission of other compounds such as methanol, formic acid and aldehydes are much lower than CO [49]. A recent study to determine the formation rate of CO from oxidation of pure linoleic acid and using it as a prediction parameter for CO formation concluded that linoleic acid content is not the sole determining factor for CO prediction rates. However they have proposed it to be a good predictive parameter in estimating the critical maximum rate of CO emission for spruce wood [50].  A study [51] was done by Austrian wood pellets industry to determine the typical CO levels in small scale residential storages. Their study showed that air-tight storage systems are highly prone to CO build up much higher than the threshold values. It was also revealed that emission of CO from wood pellets made out of pine sawdust was three to five times of the amount emitted from pellets made out of spruce sawdust. Parameters that influence the formation of CO and VOCs are listed as age, moisture content, mechanical stress, wood species, extractives, temperature, pelletizing process, oxygen availability and relative 14  humidity [32, 52-57]. Extractives in the wood are reduced during storage and the composition of wood is modified by microbiological and auto-oxidative chemical reactions [58, 59]. The results from a study on off-gassing from pellets made of Norway spruce, European larch and loblolly pine showed that a direct correlation between the organic extractive content and the release of CO and VOCs could not be established. Moisture content plays an important role in initiating chemical and microbial reactions. However the emissions from pine samples were higher than those of spruce and larch. The organic extractive content decreased in that order [60]. Moisture content together with high storage temperature could cause severe off-gassing from biomass [61]. Pellets that are made from pine are known to emit more condensable gases [20-22, 62, 63]. Arshadi et al. [19, 20] studied the fatty/resin acid compositions for fresh produced pellets as well as 2 and 4 weeks stored pellets made out of Pine and Spruce. There was a strong relationship between the amount of pine in the pellets and the fatty/resin acid content but the influence decreased over time. As a result 3-week-old pellets showed 28 times more pentanal and 8 times more hexanal than the reference pellets. A high drying gas temperature used in sawdust drying also results in higher emissions of aldehydes/ketones from pellets. Finell et al. [64] exposed sawdust to electron beams with different strength in an effort to remove fatty and resin acids from sawdust.  A number of researches are done by Biomass and Bioenergy Research Group at UBC on off-gassing of wood pellets during storage [32, 56, 57, 65, 66]. Kuang et al. [56, 57, 65] measured the emission of the gases (CO, CO2 and CH4) using 45 liter reactors. They looked into the impact of temperature, oxygen level and relative humidity on the rate of emission as 15  well as the peak concentration for each gas. They showed that temperature is a key parameter that affects the peak emission factor of off-gases for stored wood pellets.  In another study [55] to investigate the effect of low and high storage temperature on head-space gas concentrations, in storage temperature of <15 oC oxygen and carbon dioxide concentration was maintained within the safe threshold level for 77 and 257 days respectively but only for 1 day for CO. Faster emission rates and higher peak emissions are associated with higher temperatures. Increased head-space and humidity were also associated with higher emission factors. Most of the studies have focused on off-gassing of wood pellets with moisture contents below 10% but Yazdanpanah et al. [32] examined the concentration of emitted gases in an enclosed space at high moisture (4-50%) and high temperature (20-60 oC). As expected the gas emission profiles followed an exponential profile. At higher temperatures, the maximum CO2 concentration increased from 2.8% (40 oC, 4% MC) to 5.6% (40 oC, 50% MC) and from 9.6% (60 oC, 4% MC) to 10.9% (60 oC, 50% MC). For pellets with 4 and 15% MC, CO concentration increased as the temperature increased from 25 oC to 60 oC. However at a constant temperature of 25 oC, the peak concentration of CO was the same for wood pellets with 4% and 9% moisture contents and lower for pellets with 15%, 35% and 50% MC. Least depletion of oxygen was observed for samples at 25 oC and 50% MC and accelerated as the temperature increased. High temperature conditions can aggravate gas emissions due to microbial growth on pellets, chemical oxidation of lipids in the pellet and hydrothermal migration of moisture 16  between pellets. Emissions of CO2, CH4 and N2O from stored wood pellets and wood chips have been reported [21, 24]. CO and CO2 emissions have been attributed to the auto-oxidative degradation of fatty acids while for non-thermally processed forest products such as logs and chips, emissions are attributed to microbiological activities [21, 67, 68]. Study done by Boddy [69] on the effect of temperature and moisture content on off-gassing from wood in aerobic condition also confirmed the increase in CO emission when temperature and moisture is increased. CO ? ??O? ? CO?           2.1 CH?	 ? 2O? ? CO? ? 2H?O          2.2 According to (2.1) and (2.2), a high temperature will favor a high CO/CO2 and CH4/CO2 ratio [65]. Table  2.2 includes the threshold levels of CO, CO2, CH4 and O2 according to US department of labor occupational safety and health administration.    17  Table  2.2 The threshold limit value of carbon monoxide, carbon dioxide, methane and oxygen [70, 71] Chemical Substance Threshold Level  CO2 5,000 ppm for 8 hours Maximum exposure allowed by OSHA in the workplace over an eight hour period  30,000 ppm and above (short exposure) headache, loss of judgment, dizziness, drowsiness, and rapid breathing  CO 25 ppm for 8 hours Maximum exposure allowed by OSHA in the workplace over an eight hour period  200 ppm for  2-3 hours Mild headache, fatigue, nausea and dizziness  400 ppm for 1-2 hours Serious headache- other symptoms intensify. Life threatening after 3 hours  800 ppm for 45 minutes  Dizziness, nausea and convulsions. Unconscious within 2 hours. Death within 2-3 hours  1600 ppm for 20 minutes Headache, dizziness and nausea. Death within 1 hour  3200 ppm for 5-10 minutes Headache, dizziness and nausea. Death within 1 hour  6400 ppm for 1-2 minutes Headache, dizziness and nausea. Death within 25-30 minutes  12800 ppm for 1 minutes Death CH4 500,000 ppm- 8hours Could asphyxiate by displacing oxygen this concentration. The main danger with CH4 is explosions. CH4 is one of the main constitutes of natural gas. Being lighter than air, it tends to be removed through ventilation as the gas is being produced.  O2 17% Breathing is faster and deeper; impaired judgment may result  16% The first signs of anoxia appear  < 6% Convulsive movements and gasping respiration occurs; respiration stops and soon after the heart also stops  18  2.2.2 Oxygen Depletion The biological and chemical processes consume oxygen causing its depletion in the storage environment. Low oxygen concentrations can lead to suffocation of the handling personal when entering closed biomass storage without proper ventilation. Depletion of oxygen is partially due to the formation of carbon monoxide. A larger amount of oxygen losses could be due to the radical-induced oxidative degradation of natural lipids, particularly the polyunsaturated linoleic acid [14]. In this process, much oxygen is believed to become chemically bound within the wood structure. Hexanal and other alkanals are hypothesized to be formed by this radical-induced oxidative process of lipids. These oxidation processes and hence gas emission are temperature dependent.  The mechanism through which shorter compounds are formed is not clear; again the oxidation of fatty acids and other compounds seems likely. These oxidation processes would occur at room temperature but will be enhanced at higher temperatures. Svedberg et al. [34] conducted research on off-gassing of carbon dioxide and oxygen depletion during ocean transportation. They monitored five different vessels carrying wood pellets from BC to Sweden. Plotting the oxygen level from each vessel versus the CO level, a linear relationship was found with a high degree of correlation.  O? ? ?0.0016	CO ? 20.9																						?R? ? 0.92?       2.3 A very low concentration of oxygen (0.8%) was found in one vessel hold. The results from this study emphasized the hypothesis of oxidative degradation of natural lipids and other organic materials naturally present in wood pellets.  19  Svedberg et al. [67] indicated oxygen depletion and toxic gas emissions in the marine vessel stairway adjacent to the cargo hatches that contained wood chips and logs. Melin et al. [14] measured emissions from wood pellets during ocean transportation. In this study an analysis was done on the gas condition in the cargo hatch. To do so the gas conditions were measured to get a better understanding of the dynamics between oxygen, carbon monoxide and carbon dioxide as a function of temperature. Monitoring oxygen condition in the head-space and stairway clearly showed a rapid decrease in oxygen concentration while a steady-state 10% oxygen level is reached after 3 weeks. L?nnermark et al. [72] has studied different types of biomass pellets off-gassing with the aim of providing some methods for estimating risks for self-heating and off-gassing. To investigate the oxygen depletion of biomass pellets, they used OxipresTM which is an oxygen bomb instrument. Measurement is based on the consumption of oxygen at elevated temperature and pressure. They recorded different amount of oxygen consumption for different pellet samples. They concluded that Oxipres generates reproducible results and can be used as a screening tool for pellets though the mechanism of the reactions causing the pellets to use oxygen is not yet understood. As stated by Tumuluru et al. [61] off-gassing and self-heating issues are insignificant in torrefied biomass as most of the chemically and microbiologically active solid, liquid and gaseous products are removed during the process. Research done on torrefied wood chips showed that CO ad CO2 emissions are one third of the emissions from regular wood chips at room temperature. 20  2.3 Self-heating When storage of biomass in long term is needed, one factor that needs to be considered is heat development as there is a high tendency in the material to decompose [20, 25, 38, 48]. Biological and biochemical degradation as well as chemical oxidation processes result in heat development which can cause self-ignition in some cases. A schematic diagram of temperature development due to self-heating in stored biological material is given in the following graph [73].    Figure  2.1 Temperature development and processes responsible for self-heating in stored biological material [73] Results from the research that studied the self-heating phenomenon on different biofuels showed a peak in energy production after 10 days of storing material at 50 ?C [74]. Several researches reports are available from Sweden and Finland on self-ignition of 21  biomass. Investigations done in this area [4, 40, 75, 76] show that at about 300?C woody material start to pyrolyze and charcoal is formed. If a pile of stored material is exposed to oxygen during pyrolysis, it would turn into a fire. The results from these researches show the rapid rise in temperature in a pile of fuel chips as the material starts to decompose and self-ignition has even cause fires in fuel storages. It is also reported by [77] that the temperature in a pile of chipped fresh or naturally dried forest residue rises quickly after about 1 week.  As mentioned earlier, spontaneous heating for wood chips piles is known to be due to oxidation of unsaturated fatty acids and some other extractives [78-80]. They are oxidized to volatile Aldehydes (hexanal and pentanal) compounds [23]. The temperature behavior is very dependent on the moisture level of the material. For example some storages practices show that biomass piles should not be compacted as this can lead to concentration of moisture on local spots in the pile, leading to an increased tendency for self-heating. Also biomass piles should contain homogeneous material. In piles of different materials or different batches of the same material (e.g. with different moisture content) the process of self-heating can start in a niche and can spread over the rest of the storage pile [73].  Another factor affecting the temperature behavior seems to be the size of the chips [24]. For example when storing fresh woodchips or bark, the temperature in the core of the pile normally rises up to 60?C within the first days. No increase of temperature occurs when material with a particle size of more than about 20cm is stored [73]. Larsson et al. [81] monitored temperature in six large scale silos for wood pellets over 7 months. During their 22  study the silos were charged and discharged a few times. Temperature inside the silo showed an additive pattern where temperature increased over time and by increasing height.  Data gathered from industrial wood pellet silos by Guo [82] for model verification also shows high temperatures of around 60 oC recorded in the silos. In a series of fire tests done [83] in SP Swedish National Testing and Research Institute on a pilot scale silo, they simulated the auto-ignition by a coiled heating wire which was placed in the centre of pellet bulk. Measurements showed a self-sustaining pyrolysis zone was created very quickly and preferred to spread vertically whereas the spread in the horizontal direction was marginally.  2.4 Dust Explosion There are wood pellets fire accidents happened [6-13] so far that some are caused due to dust explosion [6, 9-11]. Wood dust fires and explosions that are caused by sparks and electrostatic discharge are common [84]. Dust explosion happens during transporting and operating and caused by fast combustion of powders in enclosed space. Dust surface area compared to its mass is quite large and ignition of biomass occurs at the interface between biomass. Thus dust is much more flammable than bulk material. Depending on biomass type, size and shape of the particles, explosive suspensions can be formed at different mass to oxygen ratios [85].  This could happen in many combustible powders when they are mixed with an oxidant such as oxygen. Wood pellet material safety data sheet (MSDS) has the information on safe limits of dust during operation [86]. To understand gas emission and dust explosions that are happening in large commercial wood pellet operations, estimation of fines in the pellets are 23  necessary [1]. Presence of dust in the material increases the resistance to airflow and cause disparate temperature distribution [87]. Explosion properties of dust are characterized by some parameters such as chemical composition and freshness of the dust, particle size of, volumetric concentration of dust particles in a cloud, thickness of the dust layer access of oxygen, temperature of the ambient air and temperature of the dust. If the dust particles are smaller than 100?m, could cause health issues [86, 88].  The MSDS of wood pellets includes information on the limit of dust for safe operation. Results obtained from explosibility tests on wood pellets show the auto-ignition temperature for a dust layer is much lower (225?C) compared to dust cloud (450 ?C). Explosibility results from untreated sawdust (softwood and hardwood) show the auto-ignition temperature of 220 ?C [89, 90]. 2.5 Kinetics of Off-gassing Kuang et al. [56, 57, 65] showed that production rate of CO, CO2 and CH4 is a function of temperature. On the other hand the temperature rises are due to pellets? tendency to self-heat. Figure  2.2, Figure  2.3 and Figure  2.4 show the typical results on carbon dioxide, carbon monoxide and methane emissions from their experiments from 45 liter containers filled with BC wood pellets. 24   Figure  2.2 CO2 concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets  Figure  2.3 CO concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets 25   Figure  2.4 CH4 concentration in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets Kuang et al. [65] developed the simple kinetic model for off-gas emissions from stored wood pellet and showed that a first-order reaction kinetic equation fits the data well. C??t? ? C?,??1 ? exp??k?t??          2.4 C?,?, represents the maximum concentration at the plateau, and ki is the kinetic rate constant of first order kinetic equation. For first order reactions, the reaction rate constant k follows the Arrhenius relationship: k ? A? exp?? ????           2.5 where A? is the pre-exponential factor, E is the activation energy and R is the gas constant. Comparing the activation energies for CO2, CO and CH4 shows that CO2 is emitted 26  with the least energy of formation followed by CH4 and CO. Results obtained from this research showed a linear relationship between the peak emission factor and temperature for all three gases (CO2, CO and CH4). According to Kuang et al. [65] one of the major factors which will affect the emission rate is temperature. Temperature was shown to affect both the peak emission factor and the emission rate. The other two factors mentioned are humidity and head-space. Concentrations of all three gases even at room temperature exceeded the safety threshold of the accepted TLV-TWA standards in most jurisdictions. Kuang et al. [56, 57] explored the sensitivity of rate and peak concentrations of off-gasses to temperature, relative humidity, oxygen level and head-space volume. The investigation showed significant impacts of head-space volume and temperature on gas emission factors especially for CO2. Increased head-space and increased Rh level in the head-space would both lead to higher emissions of carbon monoxide, carbon dioxide and methane. In terms of the impact of oxygen level of the rate of emission, their studies showed a direct relationship between the emission rate and oxygen availability for CO and CO2 while seems to be insensitive for emission rate of CH4. Fan and Bi [91] developed a lumped three reaction kinetic model for off-gassing from wood pellets in sealed containers. As a small amount of hydrocarbon is found from storage of wood pellets, in the developed kinetic model they have neglected the reactions leading to formation of hydrocarbons. They have considered the following 3 reactions in their modelling: Pellets ? ?? O? ? CO																																			r?? ? k????O??      2.6 Pellets ? O? ? CO?																																					r????k????O??      2.7 27  Pellets ?	O? ? ?O???Pellets? ? Hydrocarbon	intermediates				r?? ? k???O??   2.8 They used the experimental data from Kuang et al. [65] in their model. They have suggested that based on all experimental data analysed, the simple lumped kinetic model developed can predict the off-gassing process. Equations 2.9 to 2.11 along with fitted kinetic parameters and initial conditions can be used to predict the evolutions of CO, CO2, and O2 concentrations over time within sealed containers. ??? ?CO? ? mk????O??          2.9 ??? ?CO?? ? mkCO??O??                        2.10 ? ??? ?O?? ? mk????O?? ? m?k??? ? k????O??                 2.11 where m is a parameter introduced into the model and is a function of head-space, bulk density of wood pellet bed and porosity of the bed. Equations 2.9 to 2.11 were solved simultaneously to match the experimental data presented by Kuang et al. [57]. From the fitted kinetic model for data recorded for different head-space and temperature it is seen that compared with temperature, head-space ratio have only a secondary influence on the reaction rate constants. By the same fitting process for Rh and temperature data, it was confirmed that Rh and oxygen to pellet mass ratio had a secondary effect on the reaction rate compared to temperature.  2.6 Safety Measures for Storage of Pellets As mentioned in  2.2.2 and  2.3 when wood pellets are stored, they may start to 28  decompose with emission of toxic gases and heat generation. Different guidelines are published by the Association of German Engineers, German pellets institute and Nordic Innovation Centre on safe handling and storage of solid biofuels [92-94]. The off-gassing in enclosed environment could rapidly produce lethal levels of carbon dioxide and create an oxygen deficient environment. Moreover, significant increase in temperature could occur. Using the present technologies, it will be technically difficult to have temperature monitoring, CO and CH4 monitoring as well as an emergency inerting (N2 or CO2) system in place and as of now suppliers of protective systems can offer partial solutions at best [95]. In Europe where wood pellets are commonly used CO poisoning is a problem in residential wood pellet storages too. Researches are done by Emhofer [52, 53] did some monitoring experiments by inserting some data logging devices in 31 wood pellet storages in Austria with the capacity of 5 to 10 tons of wood pellets for a year. Highest CO concentration in residential wood pellets storages was recorded right after fresh pellets were unloaded. The concept of purging and inerting storages vessels are used to prevent fires and explosions in the vapor spaces of equipment. In such systems correct selection of the oxygen concentration and the inert gas-flow rate are critical to ensuring safety. Purging and inerting (or blanketing) process vessels and equipment are two common, yet distinctive, practices to control the concentration of oxygen to reduce fire and explosion hazards. Purging usually refers to the short-term addition of an inert gas to a piece of process equipment that contains flammable vapors or gases to render the space non-ignitable for a specific time period. In contrast, inerting (or blanketing) is the long-term maintenance of an inert atmosphere in the vapor space of a container or vessel during operation. Siphon, vacuum, pressure and sweep-through 29  are the most common batch purging methods while the continuous inerting methods are fixed and variable-rate. Sweep-through purging method introduces a purge gas into a vessel at one opening and withdraws the mixed gas at another opening and vents it to the atmosphere, thus sweeping out residual flammable vapour. The concentration of flammable vapour is dependent upon the volumetric feed of inert gas and flow patterns within the vessel. Studies have been done on the effect of different parameters on gas dispersion. Several researchers have worked on the effect of the ratio of column diameter to particle diameter and ratio of column length to particle diameter [96-102]. Particle shape and particle size distribution would also affect the gas dispersion to a noticeable extent [102-111]. Viscosity and temperature of the gas have also been investigated to some extent. Modern storages of wood pellets require forced ventilation systems due to thermal conditions in silos. Yazdanpanah et al. [112-114] measured the permeability of bulk wood pellets in respect to airflow.  They developed equations to measure the airflow resistance of stored wood pellets with different sizes and moisture contents and in beds with the presence of fine materials. If the ambient temperature is high as compared to the temperature in the pellets, the thermal content of the injected water vapor will contribute to the heating of the pellets. In this condition, storage ventilation with air containing high relative humidity could result in temperature increase in the storage. To prevent this, a dehumidifier should be included in the ventilation system [2]. Without proper ventilation in bulk wood pellets temperature can reach 60 ?C or higher. All pellet storage types should have the facilities to discharge pellets in case the temperature approaches the runaway temperature. It could be done by transferring the pellets into another storage or an outdoor storage. In Canada ventilation is done when the 30  ambient temperature is lower than the temperature inside the pellets. Guo [82] developed a two-dimensional axi-symmetric self-heating model and used the model to predict the self-heating process as well as critical conditions for thermal runaway in large wood pellet silo using properties that she measured and kinetics data. She studied the influencing parameters such as cooling airflow rate, wall insulation, and dimension of the storage container, ambient temperature and ambient wind condition. Results from her studies showed ventilation in silo is a very effective method to reduce self-heating and preventing thermal runaway when ambient temperature is lower than 330 K while ventilation effect will be very limited if its higher than 330 K. If no ventilation is done, the critical ambient temperature for a 21 m diameter silo can be as low as 36 oC. Some of the general recommendations [2] given to avoid self-heating and self-ignition of biomass are:  o To avoid storage and transport of large volumes if the fuel?s tendency for self-heating is unknown.  o Avoid mixing of different types of biomass fuels in one storage o Avoid mixing of biomass fuels with different moisture content o Avoid large parts of fines in the fuel bulk o Measure and monitor the distribution temperature and gas composition within the stored material Technical solutions tested in Austria to provide necessary volumes of ventilation in order to remove carbon dioxide within the residential pellet storages were mainly ventilated coupling covers, some mechanical ventilation and chimneys. The results from 2 identical storages filled with the same type of wood pellets, showed CO maximum concentration 31  dropped from 510 to 220 PPM for the storage naturally ventilated continuously with ventilated coupling covers [52, 53]. The Austrian standard [115], requires continuous natural ventilation for storages capacity smaller than 30 tonnes. Additional CO level measurement as well as forced ventilation is required for storages with capacity greater than 30 tonnes. In case of self-heating, if fire suspected, inertation is used to prevent air reaching the smoldering zone. In Sweden mobile emergency equipment are used and for practical reasons Nitrogen is used. Nitrogen should be inserted from the bottom to ensure that off-gases are replaced by an inert gas. There is yet no study done to determine the minimum required air exchange rates to confirm risk free access to the storage silos of wood pellets. 2.7 Summary of Literature Review The literature review in this chapter highlight the challenges associated with storage and handling of wood pellets with refers to incidents and causes. It is concluded that decomposition of biomass during storage causes emission of lethal levels of toxic gases as well as heat development. A number of parameters such as age, moisture content, mechanical stress, wood species, extractives, temperature, pelletizing process, oxygen availability and relative humidity influence on off-gassing. Although there are studies available on the emission of these toxic gases and kinetics development, no study is known to author on the stratification of emitted gases and determination of the minimum required air exchange rates to confirm risk free access to the storage silos of wood pellets. This study focuses on gas adsorption properties of pellets and stratification of gases in an experimental silo located in a relatively constant environment condition. 32  Chapter 3 Experimental Apparatus and Set-up 3.1 Pilot Scale Silo Configuration All measurements of the gases (CO, CO2 and CH4) were made in an experimental silo designed and installed at the Clean Energy Research Centre (CERC), the Department of Chemical & Biological Engineering, University of British Columbia. The experimental silo has a volume of 5.65m3 (1.2m diameter, 4.6m height) and is made of mild steel. The silo is designed to represent the storage of substantial quantities of wood pellets. A two level structure was designed and built to support the pilot scale silo. The set-up was further equipped with sensors for temperature, gas pressure and relative humidity measurements. Thermocouples and pressure transducers were installed at various levels along the height of the silo. Two EXP-32 cards were used to log the temperature and pressure data during the experiments. The experimental data of pressure and temperature were logged into a computer running on the LABTECH software using a data acquisition board (PCI-DAS08, Techmatron Instruments Inc, Canada) consisting of pressure transducers and thermocouples. However the pilot size of the set-up made it challenging to have it equipped with different monitoring systems and make it sealed.  Figure  3.1 shows the silo configuration, locations of the gas and pellet sampling ports, pressure transducers and thermocouples.  Appendix A shows pictures of the silo, gas and solid sampling ports, thermocouples, thermal conductivity detectors and purging system.  33    Figure  3.1 Location of gas and pellet sampling ports (left) and location of horizontal thermocouples in the silo (right)  34  3.1.1 Off-gas Sampling Ports Fourteen 12mm diameter gas sampling ports with 3mm ball valves are provided at 2 sides of the silo along the wall at different levels for gas sampling. The rod (Figure A.3 in  Appendix A) inside each sampling valves is about 0.66m which make it possible to measure the gas concentration in the silo center as well as close to the silo wall. The positions of the gas sampling ports in the large silo are indicated in Table  3.1. A close up look of the sampling port is shown in Figure A.4. Table  3.1 shows the details of the location of 14 access points for gas sampling on two sides of the silo. Table  3.1 Location of the gas sampling ports on the experimental silo Gas sampling ports on side A Gas sampling ports on side B Gas Port Name Elevation (m) Gas Port Name Elevation (m) G0 4.16 G0 4.16 G2 3.55 G3 3.35 G4 2.94 G5 2.74 G6 2.33 G7 2.13 G8 1.65 G9 1.52 G10 1.21 G11 0.91 G12 0.50 G13 0.30 Six of the gas sampling ports are located on one side of the silo and another 7 on the other side. Gas samples were collected using an airtight syringe (SG-009770-100mL SGE Gas-Tight Syringe, Luer-Lock, Mandel Scientific, Canada). The syringe had a Luer lock device to help in collecting a known quantity of gas sample. The Shimadzu GC needle used 35  with the 100mL SGE syringe was Togas Loop Fill Interface N711 Needle (Model number: 220-90615-00, Mandel Scientific Inc.).  3.1.2 Temperature and Relative Humidity Monitor In order to provide a suitable storage management system for wood pellets, changes in temperature and relative humidity in the silo need to be monitored. This is done through vertical and horizontal thermocouples installed in the silo. The silo wall temperature was also recorded by K-type thermocouples (Precision fine wire thermocouple, Omega Engineering Inc.) regularly. Figure A.6 shows a picture of these thermocouples. K-type thermocouples were installed at 5 levels along the height of the silo to record temperature during the storage. Five horizontal cables at a distance of 0.3m (1ft) with thirty thermocouple points on each level was used in this study. Figure  3.1 shows the positions of thermocouples in the silo. Temperature was recorded using a data logging system with frequency of 1 minute. Nine vertical cables were also installed (OPI Systems Inc., Calgary, Alberta).  4 of these cables plus the centre cable carry both temperature and relative humidity sensors. Due to the curvature of the silo and the modified bottom fixture, the centre cable was 0.22m shorter (14 ft in total) than the other 8 cables (14 feet and 9 inches in total). Sensors on all these 5 cables were 304mm (12 inches) apart. For the centre cable, the same spacing was used for sensors with 9 inches out at the top. Each of the 4 temperature thermocouples had 29 temperature sensors along the cable from the top to the bottom of the silo which were spaced every 152 mm (6 inches) apart. Figure  3.2 shows the cross sectional view of the cable configuration for 36  cables with and without Rh sensors.  Appendix A show the details of the temperature monitoring systems.   Figure  3.2.  Position of the vertical cables along the silo (top view) 3.1.3 Pressure Monitor Seven pressure transducers (Model No: PX142-005D5V) with 0-35 kPa (0-5 psi) range were installed at different levels to record pressures that might develop in the pilot silo during storage. Figure A.5 in  Appendix A show a picture of the pressure transducers installed on the experimental silo. For safety reasons, especially during purging tests, a U-tube manometer is built and attached to the outlet of the silo (Figure A.9). The overflow of water inside the manometer will indicate substantial pressure rise in the silo for any reason. 3.1.4 Pellet Sampling Ports Six 50 mm diameter ports provided with 50mm ball valve were used for the pellet 37  collection during storage studies. Pellet sampling ports were arranged at 6 different levels of the silo to collect pellets during storage studies and measure the physical and chemical properties such as moisture content, size distribution, durability, bulk density, particle density, ash content, calorific value and elemental analysis and investigate how these properties change during storage. Pellet sampling ports were opened for sample analysis only when the purging experiments were done. Figure A.4 shows the wood pellets sampling ports. 3.1.5 Apparatus for Residence Time Measurement Gas mixing experiments were carried out in pilot silo. Helium (He) was used as the tracer using step injection for residence time distribution (RTD) measurement. The tracer concentration at the silo outlet was monitored using two thermal conductivity detectors (TCD). A schematic diagram of the calibration set-up is shown in Figure  3.3. Red lines show the specific tubing used for the TCD calibration tests. Figure A.10 to Figure A.13 in  Appendix A show pictures of the RTD system. Figure  3.4 shows the schematic diagram of the RTD set up attached to the silo. The green lines are specifically used for RTD tests. As shown in Figure  3.3 for calibration tests, air from air cylinder flows to the TCD as the reference gas. Air from compressed air facilities in the building or nitrogen from gas cylinders flows to the mixing point where it mixes with helium coming from helium cylinder and the mixture then passes through the septum (where a sample is taken for analysis and He concentration analysis), the 3-way valve on the 2nd floor of the set-up enters TCD as the sample. Different amount of helium and air was mixed to build the calibration curve for each TCD. The response voltage was then amplified and sent to the data acquisition system. 38  For gas mixing tests, the set-up shown in Figure  3.4 was used where air from compressed air facilities flowed toward the lower diffuser in the silo. Helium was injected to this flow at time zero using the helium cylinder. The green line on top of the silo shows the tubing on the outlet of the silo where the effluent was pumped into the TCD and analyzed for helium concentration continuously. Air from air cylinder flows to the TCD as the reference gas. The response voltage was then amplified and sent to the data acquisition system. The valve on the top of the silo remained closed during TCD calibration and RTD experiments.   39   Figure  3.3 Schematic of calibration set-up for TCD 40   Figure  3.4 Schematic of set-up for RTD experiment 41  3.1.6 Purging System Two gas distributers were designed and installed in the lower section of the pilot silo for use with nitrogen and air during purging. A detailed drawing of the distributors as well as the configuration is presented in Figure  3.5 and Figure  3.6. Two float-type in-line flowmeters (Model FL-7313 Omega Engineering Inc.) were installed right before the distributors (one flowmeter for each distributor).One flowmeter (Model FR- 2A04, Ki Key Instruments) was also placed on helium line before mixing point with N2/air.  The distributors can get connected either to compressed air or nitrogen for purging purposes. The 3 flowmeters are mounted on a panel along with 3 way vales and needle valves (Figure A.15). The 3 way valves were installed to switch between nitrogen and air flow and the needle valves used to adjust the flow rate of the purging gas. A picture of the distributors is shown in Figure A.14. A schematic diagram of the purging system is shown in Figure  3.7. Blue lines show the specific tubing used for the purging tests. Air from building facilities compressed or nitrogen from cylinders pass through the two in-line flowmeters and enter the upper and lower diffusers. Flow into both diffusers could be done at the same time or one at a time. The blue line on top of the silo shows the flow of effluent to vent while the valve shown in blue remained open. Flow rates of 1.23E-3 to 1.64E-3 m.s-1 were used in purging experiments.   42      Figure  3.5 Schematic of gas distributors for purging  43   Figure  3.6 Schematic of nitrogen/air purging system for the experimental silo44   Figure  3.7 Schematic of set-up for purging experiment He Flowmeter3-W ValveFlowmeter forlower diffuserFlowmeter forUpper diffuserVentilation3-W Valve pumpTCDTo VentilationAir from compressed air facilityN2N2N2 N2HeAirReference gasSample gasValve A AmplifierSeptum for He sampling45  3.2 Instrumentation for Off-gas Analysis The composition of the sampled gas was analyzed by a Shimadzu GC-14A (Shimadzu Corporation, Japan), equipped with TCD (Thermal Conductivity Detector) and FID (Flame Ionization Detector). The GC has three packed columns in series: Porapak-N (80/100 mesh, 3 m), Porapak-Q (80/100 mesh, 3 m) and a MS-5A (60/80, 2.25 m). For CO, CO2 and CH4 the FID detector was used; and TCD was used for N2, O2 and H2. Argon gas provided the carrier and reference gas for the TCD. Compressed air and Argon were used as the reference and carrier gases for the FID.    46  Chapter 4 Material Characterization The physical and chemical properties of the wood pellets were studied in this chapter. These include the following: bulk density, pellet density, solid density, moisture content, calorific value, elemental analysis, microbial analysis and surface imaging analysis. The adsorption capacity of the pellets for non-condensable gases and oxygen was also investigated. Information on how these gases interact with the material would help in understanding the distribution of off-gases in confined space. The methodology and results are presented in this chapter. 4.1 Materials White wood pellets produced in British Columbia were used in the experiments. Three tonnes of wood pellets (Premium Pellets Ltd., Princeton, BC) were loaded into the pilot silo. Physical properties were measured initially and during storage. Table  4.1 lists the initial characteristics of the wood pellets before storage.  For lab-scale experiments on gas adsorption by the materials, three types of pellets were used: white wood pellets, steam exploded (treated) pellets and torrefied pellets. A summary of their initial properties is shown in Table  4.2. The white wood pellets were made mostly from sawdust and shavings with no bark content. Spruce, Pine, and Fir (SPF) were the main sources of sawdust. The torrefied wood pellets were made from sawdust that was heated at 270oC for 30 minutes. The steam exploded pellets were made from saw dust that was treated with steam at 220oC for 5 minutes.  The samples? size (weight) was chosen based 47  on the analysis gas and procedure, as explained later in this chapter. Sample preparation of steam exploded materials is described in [RW.ERROR - Unable to find reference:63]. Details on sample preparation of torrefied materials are also available in [117]. Table  4.1 Initial characteristics of wood pellets used in the pilot silo tests Physical property Mean (Range*) Number of samples Diameter (mm) 6.38 (6.30-6.85)1 50 Length (mm) 15 (7-28) 50 Moisture content (%wb) 7.5 (7.3-7.7) 5 Pellet density (g cm-3), ?p 1.15 (1.01- 1.32) 50 Bulk density2 (kg m-3), ?b 769 in-situ Bed porosity3 0.38  Ash content (%) 0.2631 (0.2591-0.2667) 3 Higher Heat Value (MJ/kg) 19.30 (19.05-19.54) 3 1 The values in parentheses are the ranges. 2 Bulk density=mass of pellet/volume occupied 3 Bed porosity=1 - ?b/?p    48  Table  4.2 Physical properties of wood pellet samples used in lab-scale gas adsorption tests White wood pellet  Physical property Mean (Range) Number of samples Diameter (mm) 6.61(6.51-6.79) 50 Length (mm) 14.21(6.50-24.60) 50 Moisture content (%wb) 4.5 (4.4-4.6) 3 Particle density (g cm-3) 1.14 (1.08-1.35) 50 Solid density (g cm-3) 1.17 (1.16-1.23) 50 Torrefied wood pellet Diameter (mm) 6.75 (6.73-6.76) 3 Length (mm) 22.77(22.58-22.99) 3 Moisture content (%wb) 4.6 (4.6-4.7) 3 Particle density (g cm-3) 0.93 (0.92-0.95) 3 Solid density (g cm-3) 1.32 (1.28-1.36) 3 Steam exploded wood pellet Diameter (mm) 6.60 (6.60-6.61) 3 Length (mm) 19.08 (18.99-19.15) 3 Moisture content (%wb) 5.1 (5.1-5.2) 3 Particle density (g cm-3) 1.16(1.16-1.17) 3 Solid density (g cm-3) 1.28 (1.22-1.33) 3   49  4.2 Density, Porosity, Moisture Content and Calorific Value Bulk density was measured in situ for tests in the pilot silo. It was calculated from the mass of material (3000kg) stored in the silo divided by the volume occupied. Pellet density (i.e. density of a single wood pellet) was determined for all experiments from the measurements of its length, diameter and mass. The length and diameter of each pellet were measured by a digital caliper with an accuracy of 0.01mm (Mastercraft?, Miami, FL, USA). An electronic balance (GR200, A&D Inc. Tokyo, Japan) with 0.1 mg precision was utilized to weigh the sample. With known mass and dimensions, and assuming cylindrical shape of pellet, the average pellet and solid densities were computed from 50 pellets. Solid density was determined for all experiments using the Quantachrome Multipycnometer (Model No.: MVP-D160-E, Quantachrome, Boyton Beach, FL, USA). The bed porosity (or bulk porosity of the bed) was determined using the following expression [118]: ? ? 1 ? ????            4.1 where ?b is the bulk density and ?? is the particle density of a single wood pellet, which was determined from the measurement of the weight and volume of each pellet. Porosity of each pellet is also determined from the measured particle density and solid density of the pellet. After 6 and 8 months of storage, samples of pellets were collected from all 6 solid sampling ports and tested for particle density. Results showed a slight decrease in solid density (Table  4.3).  50  Moisture content of the pellets was analyzed according to ASABE S358.2 [119]. After 6 and 8 months of storage, samples of pellets were collected from all 6 solid sampling ports and tested for moisture content. Results show that the moisture content of pellets dropped to 5.8 and 5.3 respectively (Table  4.3).  The higher heat value for wood pellets was measured using bomb calorimeter (Model 6600, Parr Instrument Company, Moline, IL). The measurements were repeated three times. The calorific value of pellets as received was estimated from: q? ? q?? ??????????? ? ? 0.02443mc?        4.2 where qp is the calorific value (MJ/kg) at constant pressure at the moisture content of mcw (% wet basis), qpd is the calorific value for dry matter measured with bomb calorimeter [120]. Results obtained from calorific value measurement of pellets showed slight changes of higher heat value over time (Table  4.3), however it was not a consistent drop or increase. The same changes in higher heat value of wood pellets were observed when they were stored at 25, 40 and 60 oC [121].  4.3 Elemental (CHN) Analysis Samples of initial material as well as samples from pilot silo after 6 and 8 months of storage were taken and sent to the Chemistry Department at UBC for elemental analysis. Ultimate analysis of the samples shows that the samples contain about 48% C, 7% H and 0.12 % N. The elemental analysis did not show large variations over time. C, H, N and O assumed to be the only elements in wood pellets and thus oxygen content is calculated by 51  subtracting the percentage of C, H and N. The ultimate analysis of wood pellets done in a study [122] showed <0.2 sulphur in the samples. Elemental analysis of 132 different types of wood pellet samples done by Chandrasekaran et al. [123] showed that pellets contained Cl, S, Hg, Na, Mg, Al, K, Ca, V, CR, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Rb, Sr, cd, Sb, Ba,Pb,Tl and SO4.  4.4 Microbial Analysis Samples are taken from pilot silo and sent to a microbiology laboratory in Vancouver, BC for microbial analysis. The methodology used for total bacterial counts followed the MFHPB-18 Standard [124]. The results obtained showed that the microbial counts were less than 5 cfu/g of wood pellets. Microbial analysis was also done by Svedberg et al. [67] on dry and wet pellets respectively. The results showed microbial count of 5 and 13 cfu/g for dry and wet pellets respectively. 52  Table  4.3 Properties of wood pellets used in pilot silo experiments before and during storage Date Solid density (g cm-3) SD Date Solid density* (g cm-3) SD Date Solid density (g cm-3) SD February 10 2011 1.15 0.15 August 02 2011 1.12 0.02 October19 2011 1.09 0.03 Date Moisture content (%) SD Date Moisture content* (%) SD Date Moisture content (%) SD February 10 2011 7.57 0.09 August 02 2011 5.77 0.12 October19 2011 5.31 0.27 Date Higher heat value (MJ/kg) SD Date Higher heat value* (MJ/kg) SD Date Higher heat value (MJ/kg) SD February 10 2011 19.30 0.24 August 02 2011 19.21 0.21 October 19 2011 19.27 0.13 Date Elemental analysis (%) Date Elemental analysis (%) Date Elemental analysis (%) February 10 2011 C 48.05 48.21 August 02 2011 C 47.96 48.22 October 19 2011 C 48.40 48.15 H 6.32 6.30 H 6.36 6.36 H 6.33 6.30 N 0.12 0.12 N 0.13 0.13 N 0.15 0.11 O (by difference) 45.51 45.37 O (by difference) 45.55 45.29 O (by difference) 45.12 45.44 * Recorded properties are the average taken from 6 solid sampling ports. The lowest solid sampling port is named port 1 and the solid sampling port on top is called port 6. 53  4.5 Material Surface Scanning Electron Microscopy (SEM) SEM analysis of surface was carried out at UBC Bio-image Facility. The pellet samples were mounted on specimen stubs and coated with gold under vacuum. Images were taken at 10 to 20 kV accelerating voltage by using a field emission scanning electron microscope, (Model S4700, Hitachi, Japan). Figure  4.1 and Figure  4.2 show two of the SEM micrographs from the surface of wood pellets before and after CO adsorption. More pictures are available in  Appendix C. As shown in the images the structure of pellet is seen as different layers of flakes on top of each other. Images obtained from the cross sectional of the material after gas adsorption shows clearly thicker fibers of wood standing out. Comparing the surface of the samples before and after adsorption shows differences occurred in the pellet surface after gas adsorption. Some agglomeration is seen on the surface of pellets after adsorption. More SEM micrographs of pellets after adsorption presented in  Appendix C show the same type of surface.    54    Figure  4.1 Two SEM micrographs of regular pellet samples before adsorption tests  A B 55    Figure  4.2 Two SEM micrographs of regular pellet samples after CO adsorption tests 56  4.6 Adsorption Analysis by Micromeritics AutoChem 2920 II1 The temperature Programmed Desorption (TPD) analysis technique used to measure the adsorption of gases (CO, CO2 and O2) by the pellets. TPD was first proposed in 1967 by Amenomiya and Cvetanovic [125]. As temperature increases and the thermal energy exceeds the adsorption energy of molecules previously adsorbed during pre-treatment, these molecules desorb from the surface of the material. These desorbed molecules are then swept to a detector (such as thermal conductivity detector TCD, mass spectrometry MSD and so on) by carrier gas and quantified by the detector. If it is a TCD, the differential thermal conductivity measured by the detector at any moment is proportional to the instantaneous concentration of desorbed molecules.  Researches have been done on TPD of carbon?based samples such as activated carbon to better understand the surface properties of the material [126-130]. In this study, the Micromeritics AutoChem 2920 II device (Micromeritics, Norcross, GA, USA) was used to quantify the amount of CO2, CO and O2 adsorbed by the pellet samples. 4.6.1 Carbon Dioxide TPD Procedure Gas uptake of wood pellet samples was measured by a Micromeritics AutoChem 2920 II Analyzer unit equipped with a thermal conductivity detector. TPD analysis was conducted                                                  1 A version of this part of chapter 4 is prepared for Publication: F Yazdanpanah, S. Sokhansanj, C.J. Lim, A. Lau, X. Bi, S. Melin. Wood Pellet Gas Uptake Capacity.Internal Review. 57  to determine the amount of gas desorbed from the surface of the material. Around 0.7g of wood pellet sample was loaded over quartz wool in the U-tube sample holder.Pictures of the sample tube and Micrometritics AutoChem 2920 II are shown in  Appendix B. TCD was calibrated for CO2, CO and O2 before conducting TPD tests. More information is included in  Appendix B. Prior to the desorption tests, the sample was degassed in a stream of helium at flow rate of 50 cm3/min and 120oC temperature. After degassing, the sample was cooled down to 40oC. A continuous flow of 10% CO2 in helium passed through the sample for 210 min. Then the flow of CO2 stopped and the sample was left for another 180 min to maximize the possible adsorption of carbon dioxide [optimized procedure and thus gas residence time was decided after running the experiments with different amount of samples, residence time and gas flow time]. Programmed desorption began by raising the temperature linearly to 110oC at 0.5oC/min while a steady stream of inert carrier gas flowed through the sample. Again, helium was used as the carrier gas since it has a very low thermal conductivity. The single species analysis gas (carbon dioxide) having a much higher thermal conductivity was blended in a fixed proportion with helium. The TPD test for CO2 adsorption by wood pellet samples started with a small portion of the wood pellet sample (0.1741 g out of wood pellet mass about 1g). CO2 was passed through the sample for 60 minutes followed by another 120 minutes with no CO2 flow to maximize adsorption. The recorded results showed that either this amount of sample or duration of analysis gas passing was not enough for the desorbed amount to be detectable. By increasing the amount of sample gradually and changing the gas residence time, the optimized results 58  were found with the operating conditions depicted in Figure  4.3 (sample size ~ 0.74 g). As seen in Figure  4.3, the process starts with increasing the temperature to 120oC (10oC/min) while passing helium through the sample to remove any moisture in the sample (A-B). Temperature is maintained at 120oC (B-C) while helium continues flowing for another 30 minutes. The sample was then cooled down to 40oC (C-D). When it reached 40oC, a stream of 10% CO2 in helium was passed through the sample. After 210 minutes (D-E), the inflow of gas mixture was stopped and the sample was left for another 180 minutes while the sample container temperature was maintained at 40oC (E-F).  For the next 60 minutes a baseline was established before desorption process starts (F-G). Then the temperature was increased to 110 oC at a heating rate of 0.5oC/min in helium stream (flow rate 50 cm3/min) (G-H). The sample would gradually desorb the amount of analysis gas it has previously adsorbed during this process. The evolution of carbon dioxide was monitored in the outlet stream by the TCD. The maximum desorption temperature of 110 oC (point H in Figure  4.3) was chosen based on the results obtained from trials at different temperatures. For instance, desorption results at 60 oC showed that the maximum desorption was not achieved. Increasing the temperature gradually to a maximum of 110oC was found to be optimal for maximum CO2 desorption from the pellets. Temperature was maintained at 110 oC before stopping the test (H-I). 59   Figure  4.3 Temperature program and gas flow events for gas adsorption experiments using Micromeritics device Table  4.4 Description of events in gas adsorption process Segment Thermal conditioning Gas treatment Measurement A-B 23oC to 120oC, 10oC/min Prep Gas: He Carrier Gas: He Moisture removal B-C 120 ? C; 30 min Prep Gas: He Carrier Gas: He ----- C-D 120oC to 40oC; 10oC/min Prep Gas: He Carrier Gas: He Sample cool down D-E 40oC; 210 min Prep Gas: 10% CO2 in He Carrier Gas: He Gas treatment E-F 40 oC ,180 min Prep Gas: None Carrier Gas: He Maximize adsorption F-G Baseline at 40 oC; 60 min Prep Gas: None Carrier Gas: He Desorption baseline G-H 40 oC to 110 oC, 0.5oC/min Prep Gas: None Carrier Gas: He Desorption processH-I 110 oC, 30 min Prep Gas: None Carrier Gas: He ------ 0 100 200 300 400 500 600 700020406080100120140Temperature ( o C)Time (min)AB CD E F GH I120 oC, 30 minutesPrep Gas: He, Carrier Gas: He120 oC, 10 oC/minPrep Gas: He , Carrier Gas: He40 oC , 10 oC/min (~8 minutes)110 oC, 0.5 oC/min CO2 desorption being recordedas temperature increasesPrep Gas: None Carrier Gas: HePrep Gas: None Carrier Gas: He40 oC, 210 minutesPrep Gas: CO2-HeCarrier Gas: He180 minutesPrep Gas: None Carrier Gas: HeGas Desorption Recording StartsBaseline at 40 oC for 60 minutesPrep Gas: None Carrier Gas: He60  Reading from the equipment was TCD signal versus time and temperature. TCD was calibrated for CO, CO2 and O2 and the calibration was associated with the signal reading (more information is available in  B.2 in  Appendix B). To correlate the signal readings collected in the analysis with the volume of gas uptake at any given point in the analysis, the equipment default calibration test was run with a series of known gas concentrations. The calibration file was then associated with the signal reading to calculate the gas concentration. During the calibration test, the analyzer decreases the proportion of the analysis gas in 10% increments, beginning with 100% and ending with 0%.Using the calibration for TCD, readings were converted to cm3/min versus time. Area under the curve was calculated and divided by the mass of sample for gas uptake (cm3/g). For samples that multiple peaks were detected, peaks were analyzed and area under each was calculated using Originlab.  4.6.2 Carbon Dioxide TPD Results Figure  4.4 and Figure  4.5 show the CO2 uptake and the typical curves, respectively, for the three types of wood pellet sample used. Steam exploded pellets have the lowest CO2 uptake whereas the torrefied pellets have the highest CO2 uptake.  Torrefied pellets adsorbed about 20% more CO2 compared to regular pellets. As all experiments were done at the same temperature and pressure, this may be explained by the rate of reaction between gas and solid which is generally proportional to the accessible surface area of the solid.  Torrefaction process causes dehydration and thus initiates and propagates cracks in the lignocellulosic structure of material. Moreover, mass loss in the form of solid, liquid and gas 61  can occur, thus, inducing changes in density and porosity. At higher temperature and residence time, the emission of volatiles becomes more intensive, resulting in increased porosity and material surface area [61, 131]. In steam explosion process hemicellulose is released from the wood cell wall and becomes accessible to chemical and biochemical degradation. Degradation of hemicellulose makes wood more rigid. Cellulose and lignin are also affected and deconstructed by steam explosion conditions. The structure deconstruction could change the surface properties. Graphs obtained from CO2 adsorption shows 2 detected peaks for all steam treated pellets which could be attributed to different surface functional groups.  Figure  4.4 CO2 uptake for three different wood pellet samples 1.0 1.5 2.0 2.5 3.00.10.20.30.40.50.60.70.80.91.0  CO2 Uptake (Torrefied Pellets) CO2 Uptake (Steam Treated Pellets) CO2 Uptake (Regular Pellets)CO2 Uptake (cm3  CO 2 min-1 )Test number62    0 50 100 150 200 250 300-0.006-0.004-0.0020.0000.0020.0040.0060.008CO2 Uptake (cm3 CO2.min-1 )Time (min)a0 50 100 150 200 250 300-0.0020.0000.0020.0040.0060.0080.0100.0120.014bCO2 Uptake (cm3 CO2.min-1 )Time (min)63   Figure  4.5 Typical CO2 TPD curves for a) Regular pellet b) Torrified pellet and c) Steam-treated pellet at desorption temperature of 110 oC 4.6.3 Carbon Monoxide TPD Procedure The test on CO-TPD was similar to CO2 measurements. Initially, moisture was removed following similar steps for CO2. A stream of 10% CO in helium passed through the sample for 120 min. Then the passage of gas mixture stopped and the sample was left for 90 min to maximize adsorption. TPD started with the same procedure with maximum desorption temperatures of 150 oC and 180 oC respectively.  An optimum amount of sample (~0.7 g) was again used in the test. The analysis gas residence time and waiting time started from 60 min and 120 min respectively and increased to 210 and 180 min. At a desorption temperature of 100 oC, even after 210 min of passing CO through the material, no desorption was detected. Figure  4.6 shows the typical CO-TPD 0 50 100 150 200 250 300-0.004-0.0020.0000.0020.0040.006cCO2 Uptake (cm3 CO2.min-1 )Time (min)64  curve for regular pellet sample when the desorption temperature, was increased to 150 oC. It can be seen that the adsorbed carbon monoxide did not fully desorb. It shall be noted that the temperature could not be raised much higher than 150oC as there is a possibility of sample breakdown. Although an optimum operating condition could not be found for carbon monoxide due to much stronger bonds between CO and the sample as compared to CO2, the maximum desorption temperature was set at 180oC. 4.6.4 Carbon Monoxide TPD Results Figure  4.6 and Figure  4.7 show typical CO-TPD curves for regular wood pellets at 150 oC and 180 oC respectively. The maximum uptake rate of 17 cm3/g was two orders of magnitude higher when compared to CO2 uptake rate (Figure  4.5). The results from desorption at 150 oC for carbon monoxide indicates that quite a large amount of CO (7-17cm3/g) is still attached to the sample and the temperature was not sufficiently high to overcome the bonds energy. Figure  4.8 shows the relative differences in CO adsorption by pellets (or total amount of CO uptake in cm3/g) as a function of desorption temperature for regular wood pellet samples. Total amount of gas desorbed from material is considered as gas uptake. The values were (4.33, 7.66 and 17.20 cm3/g) for (120,150 and 180oC). Figure B. 3 in  Appendix B includes typical TPD and temperature profile in adsorption tests.   65   Figure  4.6 Typical CO TPD curve for regular wood pellets at desorption temperature of 150 oC  Figure  4.7 Typical CO TPD curve for regular wood pellets at desorption temperature of 180 oC -10 0 10 20 30 40 50 60 70 800.00.20.40.60.81.0  CO Uptake Rate Fit Peak 1 Fit Peak 2 TemperatureTime (min)CO Uptake Rate (cm3. min-1 )20406080100120140160 Temperature (C)-20 0 20 40 60 80 100 120 140-0.20.00.20.40.60.81.01.21.41.61.8  CO Uptake Rate Fit Peak 1 Fit Peak 2 Cumulative Fit Peak TemperatureTime (min)CO Uptake Rate (cm3. min-1 )020406080100120140160180200 Temperature (C)66   Figure  4.8 CO uptake rate as a function of desorption temperature Amenomia et al. [132] suggested the following relation between heating rate and active gas molecules assuming there is homogeneous adsorption on the surface and no re-adsorption or diffusion occurs: log? ???? ? ????.???? ?? ? log??????? ?           4.3 Where: Tp?desorption peak temperature (K)  ??heating rate (K/min)  Ed?energy of desorption (kJ/mol)  A0?quantity adsorbed  C?constant (related to desorption rate) 120 130 140 150 160 170 1804681012141618  CO Uptake RateCO Uptake (cm3  CO g-1)Desorption Temperature (oC)[CO]=-0.081T+0.0034T2+52.867  R: universal gas constant (J/(mol.K))  Change in heating rate shifts the desorption peak temperature. Table  4.5 lists the experimental data needed for calculating the energy of desorption for gases. Eqn 4.4 is in the form of a straight line. The energy of desorption (Ed) can be calculated from the slope of this line, as ?(log Tp2/?) divided by  ?(1/Tp).  Table  4.5 Supplemantry data needed for calculating energy of desorption for CO2 and CO tests Gas ? (K?min?1) Tp(K) log(TP2/?) 1/TP CO2 0.5 383 5.4674 0.0026 CO2 10 453 5.3121 0.0022 CO 5 423 4.55 0.00236 CO 0.5 393 5.48 0.00254  The calculated value was found to be 97.8 kJ/mol for CO and 7.24 kJ/mol for CO2.  The large difference in magnitude of the desorption energy demonstrates that CO adsorption is a chemical sorption phenomenon, while CO2 adsorption is more indicative of a physical sorption mechanism. Much higher temperature is needed to overcome the adsorption energy of CO and material surface. 4.6.5 Oxygen TPD When wood pellets are stored in confined spaces, the oxygen content in the void space can be used up by fatty acids rapidly. To study adsorption of oxygen by wood pellets, TPD was done following the same procedure as CO and CO2 TPD. Initial tests started with a maximum desorption temperature of 100 oC, and no desorbed oxygen was detected in the 68  effluent. Desorption temperature was increased to 120, 150 and 180 oC in various trials. Based on the temperature and TCD signal profiles, the adsorbed oxygen was observed to start desorption at about 140 oC (Figure  4.9). More oxygen was desorbed from the samples at 180 oC. However due to limitation of material breakdown at high temperature, the oxygen desorbed by the pellets cannot be quantified. Experimental results showed very high activation energy was needed to overcome the sorption energy between oxygen and the samples. This is a clear indication of chemical adsorption of oxygen, and intermediate products could be formed as a result of material oxidation.  Figure  4.9 Typical O2 TPD curve for regular wood pellets at desorption temperature of 150 oC 4.6.6 Mass Spectrometry A series of tests was subsequently done using mass spectrometry to verify that no gases were emitted from the sample due to material breakdown at the temperatures used in the TPD 0 40 80 120 160 200 240 280 3204.904.955.005.055.105.155.20 O2 Uptake Rate TemperatureTime (min)O 2 Uptake Rate (cm3. min-1 )020406080100120140 Temperature (C)69  experiments. A residual gas analyzer (RGA) (Model RGA 200, SRC Stanford Research Systems) was used for the measurements. RGA is a mass spectrometer with small physical dimensions, and its function is to analyze the gases inside the vacuum chamber. During operation, a small fraction of the gas molecules are ionized, which are then separated, detected and measured according to their molecular masses.  Tests were conducted using a quartz u-tube reactor placed in a temperature programmable muffle furnace with a mass spectrometer connected to the reactor effluent in order to continuously monitor the products generated during the temperature-programmed desorption. About 0.8 g of samples was put in the sample holder in which a thermocouple is located and put inside the furnace. It was then purged with helium (50 cm3/min) for 10 min. The set point of the furnace was adjusted to 120 and 150 oC. Temperature was increased at the same rate (0.5 oC/min) as the TPD tests done by Micromeritics Autochem II 2920. The mass spectrometer analyzes the effluent by ionizing some of the molecules (positive ions) and then separating them with respect to their mass. Partial pressure measurements were determined with the help of previously calculated sensitivity (calibration) factors by reference to the abundance of the individual mass numbers attributed to each gas type. As the temperature was increasing, the gas outlet was monitored for any possible gas emitting from the sample. Tests are done for maximum desorption temperatures of 120, 150 and 180 oC in three replicates for each. The SRC Residual Gas Analyzer results were recorded every minute as the partial pressure versus the mass of the emitted gas. Evidently, no significant amount of gas was 70  emitted from the pellet sample during the test. Moreover, based on the TPD profiles, it was confirmed that the composition of the sample remains intact up to 180 oC. Figure  4.10 shows the profile of effluent gases during thermal treatment of samples in helium as measured by the mass spectrometer. High partial pressure was only observed for a mass of 4 g, which represents the flow of helium through the sample. The subsequent small peak observed for a mass of 18 g represents vaporization of the moisture inside the material; another peak was seen for a mass of about 28 g; although the identity is not known, the partial pressure was too low to represent any substantial breakdown.  Figure  4.10 Profile of effluent gases during thermal treatment of pellet samples as measured by mass spectrometer using 0.8 g of sample 0 20 40 60 80 1000.000000.000020.000040.000060.000080.000100.00012H2OPartial Pressure (Torr)Mass (g)He71  4.7 Concluding Remarks Ultimate and proximate analysis was done for wood pellets before and after storage. Measurement of solid density and moisture content over time showed a continuous drop. However changes in higher heat value did not follow a consistent increase or decrease pattern. Elemental analysis of material before and after storage shows about 48% of carbon, 6% hydrogen and about 0.1% nitrogen with minimal changes over time. Microbial count analysis for wood pellets before storage clearly showed low amount (<5 cfu/g). Experiments carried out in respect to adsorption/desorption of gases by wood pellets showed the highest adsorption of carbon dioxide by torrefied wood pellets possibly due to higher porosity and thus surface area. Steam exploded pellets showed minimum adsorption of carbon dioxide due to changes resulted from steam explosion process. The results obtained from carbon monoxide experiments showed a very high energy of desorption (~98 KJ/mol) which was an indication of chemical adsorption. For carbon dioxide a very low energy of desorption (~7 KJ/mol) indicated the presence of physical adsorption. Moreover the high desorption temperature that was needed to overcome the bonds energy for oxygen adsorption verified the existence of chemical adsorption of oxygen and possible formation of intermediate material.   72  Chapter 5 Stratification and Effect of Temperature and Relative Humidity on Off-gases in Stored Wood Pellets2 5.1 Materials and Methods 5.1.1 Gas Measurement The pilot silo as described in in  Chapter 3 was used for experiments. The silo has the volume of 5.2 m3 (1.2m diameter, 4.6m height). Three quarters of the silo was filled with 3000 kg of fresh wood pellets on February 10, 2011 and the surface of the pellets was leveled. Gas sampling started 2 days after loading the silo and continued for the first 9 weeks (63 days) of storage when most gases reached a plateau. Gas samples were taken from the sampling ports for composition analysis daily during the first 2 weeks and sampling became less frequent for the next 7 weeks. The sampling ports G0 to G13 at different elevations are for studying the axial (longitudinal) gas distribution/stratification, whereas sampling from 3 locations at the same level are for studying the radial gas distribution/stratification.                                                   2 A version of this chapter has been accepted for Publication: F Yazdanpanah, S. Sokhansanj, C.J. Lim, A. Lau, X. Bi, S. Melin. Stratification of Off-gases in Stored Wood Pellets, Biomass and BioEnergy.In Press. 73  For each gas sampling event, 100mL of gas sample was collected using an airtight syringe. The samples were analyzed using a GC-14A gas chromatograph as described in  Chapter 3. The GC got calibrated regularly with 3 standard calibration gases (Table  5.1). The samples were analyzed for CO, CO2 and CH4 concentrations as this study is only focused on the emission of non-condensable gases. All gas concentrations presented are in volume percentage. Table  5.1 Compositions (volumetric) of standard gases for GC calibration Standard gas 1 Standard gas 2 Standard gas 3 Components Concentration (%) Components Concentration (%) Components Concentration (%) CO 2.5 CO 0.05 CO 0.50 CO2 6.0 CO2 0.10 CO2 0.50 CH4 1.5 CH4 0.50 CH4 0.10 N2 3.0 H2 0.50 H2 1.00 O2 10.0 Air balance Air balance He 2.0     Ar balance     5.1.2 Temperature and Relative Humidity Measurement For these experiments, the pilot silo explained in  Chapter 3 was used. Temperature and relative humidity were recorded during 12 month of storage [January - December 2011]. Temperature was recorded by the 5 horizontal and 9 vertical thermocouple cables explained in  3.1.2. Four of the vertical cables plus the centre cable carry 14 temperature and relative humidity sensors each. Each of the other 4 temperature thermocouples had 29 temperature sensors along the cable.  74  Table G. 1 in  Appendix G shows the coordinate for each sensor. The point where the centre cable (C5M) is attached to the bottom of the silo is chosen as reference point T(r,y)=T(0,0). Figure  5.1 has the top view of silo illustrating the location of all cables and all sensors coordinate. Table G. 2 in  Appendix G shows the coordinate of some of the sampling ports and sensors on thermocouples installed inside the pilot silo. Equilibrium moisture content (EMC) is also theoretically calculated inside the silo and shown over time ( Appendix J). Temperature and relative humidity in the silo were recorded by each sensor every minute, from January 2011 to December 2011. About 88,387,000 records in total for temperature and 33,264,000 records for relative humidity were made.   Figure  5.1 Top view of the pilot silo with the coordinate [r,y] of all sensors located on 9 vertical cables C1MC4C1C4MC5MC2MC3M C3C2r=0.3my=0.5696 to 4.532mr=0.5my=0.5696 to 4.532mr=0my=0.5696 to 4.2272mr=-0.3my=0.5696 to 4.532mr=-0.5my=0.5696 to 4.532mr=0.3my=0.2648 to 4.532mr=0.5my=0.2648 to 4.532mr=-0.3my=0.2648 to 4.532mr=-0.5my=0.2648 to 4.532m75  5.2 Results and Discussion 5.2.1 Off-gas Concentration in Silo Head-space Gas composition analysis was performed for the first 9 weeks after loading the pilot silo.  Results (Figure  5.2 to Figure  5.4) showed a rapid increase in the concentration of CO2, CO and CH4 in the head-space of the silo. Off-gas concentrations in the first 7 days were 1.0-1.6% CO2, 0.8-1.0% CO and 0.73- 0.85% CH4. All gas concentrations increased with storage time.  Figure  5.5 shows the increase in the CO/CO2 ratio with time, which reached a constant value after 50 days of storage. Some scattering of data is seen between days 10 to 15 with a peak CO/CO2 ratio of about 0.59 on day 10. Minimum oxygen concentration in the bed was also seen on day 10 (Figure  5.6). It could be explained by slightly (~2%) lower relative humidity recorded by cable 5M (Figure  5.20) and higher H2 and thus the shift of water-gas shift reaction towards reactants. Oxygen concentrations were also measured according to the same gas sampling schedule. As shown in Figure  5.6 the O2 level dropped dramatically from 21% to 7-8% after 3 days and reached close to 0% after 1 week and again after 3 weeks of storage.  The concentration of carbon dioxide in the head-space increased to about 2.7% (27,000 ppm), and this is higher compared to the CO2 concentrations in the pellet bed which could be attributed to the exposure of pellet surface to head-space oxygen and thus higher local emission. Lab-scale experiments (as described in  Appendix E and  Appendix F) had been previously performed to determine the effect of oxygen availability on the emission rates of 76  off-gasses. Highest emission of CO2 was seen in containers with higher percentage of head-space or where oxygen was pumped in regularly. When wood pellets were stored in helium, minimum emission of CO2 was detected. For carbon monoxide, its concentration increased from the very first days of storage and reached a maximum value of about 1.7% after 9 weeks of storage. This accumulated concentration was well above the threshold [71] limit value (TLV) for human health, and it can cause injuries and immediate death.   Figure  5.2 CO2  concentration in head-space of the silo as a function of storage time for wood pellets. (G0: head-space gas sampling port) [Concentrations are volume percentage] 0 10 20 30 40 50 600.51.01.52.02.53.0T=25?C Head-space = 25%CO2 Concentration (%)Time (Day)77   Figure  5.3 CH4  concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage]  Figure  5.4 CO concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage] 0 10 20 30 40 50 600.030.060.090.120.15CH4 Concentration (%)Time (Day)T=25oCHead-space =25%0 10 20 30 40 50 600.40.60.81.01.21.41.61.8CO Concentration (%)Time (Day)T=25?C Head-space =25%78   Figure  5.5 CO/CO2 ratio in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port) [Concentrations are volume percentage]  Figure  5.6 O2 concentration in head-space of the silo as a function of storage time for wood pellets. (G0 head-space gas sampling port)  [Concentrations are volume percentage] 0 10 20 30 40 50 600.480.520.560.600.64T=25?C Head-space = 25%CO/CO2Time (Day)0 10 20 30 40 50 60048121620 T=25?C Head-space =25%O 2 Concentration (%)Storage days79  The gas emissions are converted to emission factors using Eqn (5.1) in order to compare the results with other researchers? work on the same basis: f? ? ????????????????????            5.1 where fi is emission factor of species ?i?, Ci is the gas concentration, Mwt is the molecular weight (g),Vg is the gas volume (m3), R is universal gas constant (8.314 J/mol K), T is  temperature (K),P is the pressure (Pa), MP is the mass of pellets (kg). Assuming that the amount of N2 remains constant during off-gassing process, CN0/CNt is introduced as a correction factor to balance the change in gas volume where CNt is the concentration of nitrogen measured at any time and CN0 is the concentration of nitrogen at the beginning of test. The gas emission results for 25% head-space are compared to those obtained by Kuang et al. [65] for the storage of wood pellets. As shown in Table  5.2, the peak emission factors of all three gases derived from this study (using a large pilot silo) are higher than their findings (using a small lab-scale reactor); though both sets of values are in the same order of magnitude.  Table  5.2 Comparison of peak emission factor for CO, CO2 and CH4with previous research Peak emission factor (g/kg) This study (Storage volume=5200 L) Kuang et al.[65] (Storage volume=45 L) CO  0.0191 0.0124 CO2 0.0485 0.0200 CH4 0.0009 0.0002 Methane emission is due to the activities of anaerobic microorganisms. For instance, in a typical anaerobic digestion process for biogas production from readily biodegradable 80  feedstock, the methane content can range from 50-70%. In this study, the CH4 concentration was observed to increase from 0.03% to 0.14% after 9 weeks, while an oxygen deficient environment was created in the silo just a few days after the start of the test.  The relatively low CH4 contents are in line with the results of total bacteria counts, which were less than 5 cfu/g of pellets as compared to orders of magnitude higher microbial counts in a typical anaerobic digester.  5.2.2 3D Analysis of Longitudinal Distribution of Emitted Gases The measured concentrations of the gases derived from all 13 gas sampling ports over time are plotted in 3D graphs and contours. The contour plots would provide a supplemental view of how the various gases are stratified in the bed. Figure  5.7 show the 3D and contour plots of the concentration of carbon dioxide in all locations. For the first several days, the gas concentration was quite the same for all locations. After 10 days, CO2 concentrations were different; yet, after about 40 days, the concentrations became uniform again everywhere in the silo.  The contour plot clearly illustrates that the CO2 concentration was always higher in the head-space compared to the other areas within the pellet bulk. CO2 emission is more sensitive to temperature versus CO; this could explain why higher CO2 concentration was observed in the silo head-space. While the peak CO2 concentration of 2.7% was reached just after 2 weeks in the head-space of silo, such a high concentration was not seen in the pellet bulk section until after 45 days. Carbon dioxide concentration was above the threshold limit value for worker safety and this limit was reached sooner in the silo head-space. In terms of worker safety, both the Occupational Safety and Health Administration (OSHA) and the 81  American Conference of Governmental Industrial Hygienists (ACGIH) have set the permissible exposure limit (PEL) and the threshold limit value (TLV), respectively for CO2 of 5,000 ppm (0.5%) by volume over an 8-hour average exposure ( Chapter 2, Table  2.2).  Figure  5.7 3D-map of CO2 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] Figure  5.8 (a,b) and Figure  5.9 (a,b) show the concentration of CO2 at even-number (G0-G12) and odd-number (G0-G13) sampling ports respectively. The sampling port in the head-space (G0) is included in both graphs. During the planning stage for the experiment, the gas emission was assumed to be symmetrical on both sides of the silo and this was treated as a replicate of the experiment. As seen from the data the CO2 concentration profiles are 1020304050601.01.52.02.53.00.51.01.52.02.53.03.54.0Concentration (%)Height (m)Time (day)0.0000.40000.80001.2001.6002.0002.4002.8003.20082  practically the same on both sides with the maximum concentration reached in the head-space at all times.   83    Figure  5.8 (a) 3D- map of CO2 concentration in the silo at G0 to G12 (Even sampling ports at one side) during 63 days of storage (b) Contour of CO2 concentration over time [Concentrations are volume percentage] 1020304050601.01.52.02.53.01.01.52.02.53.03.54.0Concentration (%)Height (m)Time (days)0.0000.40000.80001.2001.6002.0002.4002.8003.2001.62.02.41.22.02.42.82.810 20 30 40 50 601.01.52.02.53.03.54.0Time (days)Height (m)0.00.400.801.21.62.02.42.83.284    Figure  5.9 (a) 3D- map of CO2 concentration in the silo at G0 to G13 (Odd sampling ports at one side) during 63 days of storage (b) Contour of CO2 concentration over time [Concentrations are volume percentage] 1020304050601.01.52.02.53.00.51.01.52.02.53.03.54.0Concentration (%)Height (m)Time (days)0.0000.40000.80001.2001.6002.0002.4002.8003.2001.62.0 2.42.81.22.01.62.010 20 30 40 50 600.51.01.52.02.53.03.54.0Time (days)Height (m)0.00.400.801.21.62.02.42.83.285  Figure  5.10 show the overview of carbon monoxide emission, accumulation and dispersion over 63 days of storage. As distinguished from the concentration profile for carbon dioxide, a higher CO concentration was observed in the head-space of the silo at the beginning of the test, and the CO concentrations were similar in most other locations within the pellet bulk. This difference may be due to the gas dispersion phenomenon. Nevertheless, after 18 days of storage, stratification was observed with the highest concentration of CO at the G5, G6 and G7 locations (2.03-2.54 m from the bottom of the silo). Due to close value of density for air and carbon monoxide, not much gravitational stratification of CO was expected. Although carbon monoxide emissions are attributed to oxidation of unsaturated acids in wood pellets, the exact pathway through which CO is emitted hasn?t been identified yet. More distinct stratification of gas was seen in this study for carbon monoxide compared to other off-gases, which could be attributed to the high uptake of CO by pellets during storage as explained in  Chapter 4.  86   Figure  5.10 3D- map of CO concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] Figure  5.11 illustrate the distribution of emitted methane (CH4) during 63 days of sealed storage of wood pellets. After 4 days, the concentration of methane was the highest close to the bottom of the silo. Over time the emitted gas started to stratify; but after 48 days almost the same concentration was observed in all locations, with slightly higher CH4 concentration at the upper sections of the silo near the interface of wood pellets and head-space. The same stratification was noted until the end of the test. Methane emission has a similar behavior as carbon dioxide; both gases reached higher concentrations more quickly in the head-space than in the other areas within the pellet bed.  1020304050600.20.40.60.81.01.21.41.61.80.51.01.52.02.53.03.54.0Concentration (%)Height (m)Time (days)0.0000.20000.40000.60000.80001.0001.2001.4001.6001.8002.00087   Figure  5.11 3D- map of CH4 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] Diffusion of gases in air and its effect on oxygen deficiency could also partly account for the created environment. Studies [133, 134] have been conducted to understand how easily the released gases (carbon dioxide, helium and sulfurhexafluoride) would mix with air and whether they would remain fully mixed. The gases were found to diffuse in air more readily than expected and modest gas velocities due to natural convection would fully mix the released gases with fresh air.  In the wood pellet storage, gases were emitted over time from the first day till the gas concentrations reached a plateau. Some gas stratification was detected in the early weeks of storage especially for carbon monoxide which is due to chemical and physical adsorption 1020304050600.020.040.060.080.100.120.140.160.51.01.52.02.53.03.54.0Concentration (%)Height (m)Time (day)0.0000.020000.040000.060000.080000.10000.12000.140088  between the material and the gases as well as differences in temperature and relative humidity at various levels in the silo.  However after a few weeks, gases started to mix and remained mixed. In this regard, natural convection within the silo can bring about gas mixing. 5.2.3 Oxygen Depletion Figure  5.6 demonstrates the oxygen depletion in the silo head-space during storage of wood pellet, which is much faster than the oxygen within the bed of wood pellets. Oxygen concentration dropped very quickly in less than 10 days of storage and this can be partially explained by the emission of carbon monoxide. Depletion of oxygen is partly related to CO formation but a greater amount could be due to the radical-induced oxidative degradation of natural lipids, particularly the polyunsaturated linoleic acid [14]. As explained in section  4.6.5, the high energy of desorption required to overcome the bond indicates chemical adsorption of oxygen to the pellets. The auto-oxidation of fatty acids starts with formation of free radicals. In the presence of oxygen, hydroxyperoxide radicals are formed and in interaction with an unsaturated fatty acid produce two hydroxyperoxides and a new free radical. When pellets stored in air, hydroxyperoxides are formed from oxidation of fatty acids. Depending on whether oxygen bond or carbon bond break in them alcoxi radical, aldehydes, acids, hydrocarbons or ketones will be formed. In order to examine the role of oxygen availability in the storage space on the rate of off-gassing, a set of experiments was conducted under controlled environment conditions. The same reactors described in [32]were used in these experiments (Details are presented in in  Appendix E). Pellets were placed in 6 89  sealed reactors (2L by volume) that were purged with different gases. In two of the reactors pellets were stored in oxygen-free environments. When pellets were stored in an environment dominated by N2 after purging, CO emission was as high as when pellets were stored under regular conditions. However when the reactors were purged with He (helium), the peak CO emission decreased to 25% of the values when pellets were stored under regular conditions. A hypothesis is that, although oxygen availability can accelerate the emission of non?condensable gases, the O2 content of the pellet material could be high enough to induce high amount of CO emission compared to pellets stored in air. Emission of carbon monoxide for pellets stored in CO2 showed 50% less emission compared to pellets stored in air. Results obtained from stored pellets exposed to different head-space (HD) percentage ( Appendix F) showed that an increase in head-space is related to the peak emission factor for carbon monoxide and carbon dioxide with the following linear relationships (average values of two replicates each): f?? ? 3 ? 10???HD%? ? 0.0065         5.2 f??? ? 1.9 ? 10???HD%? ? 0.0041         5.3 Oxygen concentration was also measured in all 13 locations including the head-space. Ever since the first gas sample was taken on day 2 of storage, the O2 concentration decreased to 15% in the reactor. As seen in Figure  5.12, O2 concentration was depleted rapidly to less than 5% within the first 7 days thus generating an oxygen-deficient atmosphere in a confined space. More oxygen was available within the bed of pellets  compared to the head-space as a 90  result of oxygen being trapped within the pellets until  40 days of storage when oxygen was consumed everywhere in the reactor.  When oxygen levels fall below 19.5% by volume, air cannot support metabolism for an unlimited period of time. At 17% oxygen, the symptoms might simply be worse as reflected by hyper-ventilation. The oxygen-deficient atmosphere becomes more dangerous when oxygen content is further lowered to 15%; people can quickly progress to dizziness and rapid heartbeat. Finally, oxygen levels below 13% can lead to unconsciousness and eventually to death at around 6% oxygen [49].   Figure  5.12 3D- map of O2 concentration in the silo at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage] 010 20 30 40 50 60-15-10-5051015200.51.01.52.02.53.03.54.0Concentration (%)Height (m)Time (days)0.0002.0004.0006.0008.00010.0012.0014.0016.0018.0020.0091  5.2.4 3D Analysis of Radial Distribution of Emitted Gases Figure  5.13 shows CO2 concentrations in the head-space and the bottom of the silo respectively. The gas concentration was measured at the same elevation but at 3 different radial positions (0.0, 0.6 and 1.2 m from the center) in order to investigate any possible radial concentration gradient. The results showed no significant variations in the CO2 concentrations in different radial positions, due to the uniform radial temperature profile.  The same measurements were made for all three gases (CO2, CO and CH4); again no significant differences in any of the gas concentrations were observed at different radial positions. However, Figure  5.13 clearly shows the difference in CO2 concentrations along the axial locations, with the head-space concentration higher than that at the silo bottom which could be due to lower rate of mixing in the silo compared to the rate of reaction. The same observations were obtained for radial concentrations of carbon monoxide and methane over time. 92   Figure  5.13 CO2 concentration in the silo head-space and the bottom of the silo at different radial positions 5.2.5 Effect of Temperature and Relative Humidity on Off-gas Stratification Figure  5.14 shows temperatures recorded by the two sensors (S1 and S29) on vertical cables C1 and C3. The cable C1 is located in the outer ring and cable C3 is located in the inner ring.  S29 is located at the top of the cable while S1 represents the sensor located at the bottom of the cable. The temperatures ranged from 18 to 24.5oC. There were little radial temperature variations in the silo but a longitudinal temperature gradient of 6-7 oC was recorded in the first 60 days of storage. This temperature gradient could lead to differential off-gassing across the pellet bed as well as vertical natural convection within the silo, influencing the distribution of gas concentration in the bed.  0 10 20 30 40 50 600.51.01.52.02.53.0  Concentration at head-space radius=0m          Concentration at the bottom radius=0m Concentration at head-space radius=0.60m     Concentration at the bottom radius=0.6m Concentration at head-space radius=1.2m       Concentration at the bottom radius=1.2mCO2 Concentration (%)Time (day)93   Figure  5.14 Temperature in the silo recorded by cable 3 (sensors 1 and 29) and cable 1 (sensors 1 and 29) Slightly elevated temperatures are evident at several locations within the bed. As temperature is one of the main factors affecting the rate of off-gassing, local higher temperature would cause higher emission rate of gases over time and consequently higher gas concentrations in a poorly mixed vessel. Figure  5.15 shows the temperature varying from 18-24 oC, as recorded by the cable located in the outer ring close to the gas sampling locations (C1M) during 63 days of storage.  Higher temperature was found in the upper parts of the pilot silo. By comparing the local temperatures in all parts of the silo with the contour plots of off-gases and oxygen (Figure  5.16 and Figure  5.17), it may be deduced  that higher concentrations of off-gases and lower concentration of oxygen were  attained  faster in the silo head-space where the temperature was always higher compared to the bed of pellets. The -10 0 10 20 30 40 501819202122232425 C3/S29          C3/S1 C1/S29          C1/S1Temperature (o C)Time (Day)94  higher concentration of off-gases and lower concentration of oxygen in head-space could be attributed to exposure of wood pellets in the bed surface to head-space. Recorded relative humidity in the head-space showed that the humidity was higher in the bed surface and that resulted in higher emission rate. This was especially noticeable for the first 4 weeks of storage when the temperature was highest in the head-space and there was a 6 oC difference between the top and bottom of the silo.   Figure  5.15 Contour plot of temperature for all sensors on cable C1M during 63 days of storage   20.521.221.922.621.221.919.821.221.923.319.1 19.8 19.822.618.421.910 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.795   Figure  5.16 (Left) Contour plot of CO2  and (Right) CO concentration at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage]   Figure  5.17 (Left) Contour of CH4 and (Right) O2 concentration at all locations (G0 to G13) during 63 days of storage [Concentrations are volume percentage]   1.62.02.41.22.81.62.82.82.02.41.62.010 20 30 40 50 600.51.01.52.02.53.03.54.0Time (day)Height (m)0.00.400.801.21.62.02.42.83.20.801.01.21.41.61.6.600.80 1.21.61.410 20 30 40 50 600.51.01.52.02.53.03.54.0Time (days)Height (m)0.00.200.400.600.801.01.21.41.61.82.0.0600.0800.100.120.14.0400.140.1010 20 30 40 50 600.51.01.52.02.53.03.54.0Time (day)Height (m)0.00.0200.0400.0600.0800.100.120.1418161412108.06.04.02.02.00.04.00.00.04.00.04.00.00.020 0.0202.04.02.04.04.00.04.00 10 20 30 40 50 600.51.01.52.02.53.03.54.0Time (days)Height (m)0.02.04.06.08.010121416182096  Another factor that plays an important role on emissions from wood pellet storage is relative humidity. Figure  5.18 shows the contour plot of relative humidity recorded every minute by the cable located in the outer ring close to the gas sampling locations (C1M) in the pilot silo from day 1 to day 60. Relative humidity varied from 25 to 31%, and they did not exhibit significant changes over time; however, vertical stratification was clearly seen, with the highest relative humidity occurring in the head-space. Higher humidity could contribute to the accelerated biological decomposition process during storage and thus increase the emissions of CO2, CO, and CH4.   Figure  5.18 Contour plot of relative humidity for all sensors on cable C1M during 63 days of storage 2830302826263224243026 302222320 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C1M97  5.2.6 Longitudinal and Radial Distribution of Temperature Temperature and relative humidity patterns over time are plotted in 3D graphs and contours. The 3D-graphs contain data derived from measurements taken from all 9 vertical thermocouple cables. Experimental silo wall temperature was also recorded over time. Short term and long term plot series and analysis of temperature and relative humidity are presented. The contour plots provide a supplemental view of how temperature and relative humidity were stratified in the bed. More than 260,000 temperature and relative humidity data were recorded every day. Daily analysis of temperature showed that minimum and maximum recorded temperature in the silo before loading the material was 16 oC and 20 oC; however this minimum and maximum were increased to 20 oC and 25 oC just after 3 days of storage. Readings from 2 cables (one located in the outer ring and the other one located in the pilot silo centre) showed almost constant reading for each sensor in one day with increase over time. Minute-to-minute observations were converted into a daily observation by taking the average and used for long term analysis of data such as in Figure  5.19. However, minute-to-minute data were used in short term (e.g. purging experiments) analysis of data as shown in Figure  6.12.  For wood pellets, the tendency for self-heating is different depending on the material quantities and is observed most, shortly after production. When wood pellets were loaded in the silo they were 1 week old. In Figure  5.19, day 0 is referred to as the day that pilot silo was loaded. Right after loading the material, temperature increased over time in all locations. Highest temperature was observed during the first 4 weeks of storage as expected.  98  Ambient temperature in the lab was about 20 oC. However, due to natural ventilation in the room, temperature in the lower parts of the pilot silo was lower compared to the upper parts. Temperature below 20 oC was seen in the lower parts of the silo mostly due to close location of the silo to the door where fresh air came in most of time (Outside temperature in February was between 0-15 oC). This resulted in lower wall temperature in lower parts of the silo too. As shown in Figure  5.19, clear segregation of temperature was seen in the silo. Before material was loaded, temperature recorded by all cables especially by the sensors in lower parts was between 17-19 oC. A daily record of lab temperature is shown in  Appendix H.  Temperature inside the wood pellet bulk was between 18-19 oC when it was loaded. Temperature in all locations increased over time and reached the wall temperature in less than 70 hours of storage. It then increased further and reached the maximum temperature of about 3 oC higher ? a clear indication of self-heating- than wall temperature close to 25 oC. Depending on the location in the silo, temperature started to decrease to 21 oC after 600 ? 700 h of storage which was still 1-2 oC higher than the wall temperature. Self-heating is initiated by physical processes as well as chemical and biological reactions which are all exothermic. However the very low moisture content of wood pellets (7.5% in this study) make the effect of biological activities minimum. Adsorption (gas physical and chemical adsorption; explained in  4.6.5), moisture adsorption and oxidation are exothermic reactions. The oxidation reactions take place on the surface of wood pellets [49]. Guo [82] measured the heat release rate from stored wood pellets experimentally and concluded that 99  it?s very sensitive to pellet age. This could explain why the highest temperature during the experiments was observed in the first 600-700 h of experiment. As shown in Figure  5.2 to Figure  5.4 in  Chapter 5, the first 600-700 h after storage is when the highest off-gassing rate was recorded. Decline of temperature after 4weeks of storage and decreased reaction rate in oxidation products emission in the same time period, could be due to the decrease in fatty acid contents over time as also mentioned by Arshadi et al. [63].  The contour plots (Figure  5.19) display very similar patterns of radial temperature distribution over time. However, it is clear that temperature within the pellets bulk was higher in the centre of the silo where the heat is more difficult to be dissipated. At any given time, temperature is highest on C5M (centre of silo) followed by C2M (halfway between the centre and silo wall) and then C1M (close to the silo wall). In the study done by Larsson et al. [81] on temperature pattern inside a wood pellet silo, temperature showed an additive pattern where pellet temperatures increased over time and with increasing height in the silo.  As mentioned in  2.5, according to [56, 57, 65] the decomposition reaction of wood pellets can be assumed as a first order reaction where CO2, CO and CH4 and heat (q) are the main products. As explained by Guo [82], decomposition reactions can be separated into oxygen dependent reactions and oxygen independent reactions where q is expressed as the sum of ??? and ????. ??? is the heat released from oxygen dependent reactions, in J and  ???? is the heat released from oxygen independent reactions. Data on heat release rate [82] indicated that temperature has a significant impact on the heat release rate and it was higher at the elevated temperatures. A maximum self-heating release rate of 0.33 mW/g was 100  observed at 50oC. In the first days of storage, where the heat production from the pellets is highest, the heat release should have been dominated by oxygen dependent reactions. Over time, as the heat production decreases and the oxygen is consumed, the heat should mainly be released from oxygen independent reactions.  As mentioned above, ambient temperature is one of the critical factors affecting self-heating in wood pellet storage. In this study, the pilot silo was located in the laboratory with minimum increase in ambient temperature and thus small degree of self-heating inside the pellet bulk. Temperature fluctuations in silo head-space were recorded by sensors on all cables but no pattern for maximum temperature was found. Contour plots of temperature for other cables (C4M, C2 and C4 during 63 days of storage are presented in  Appendix H.    101     Figure  5.19 Contour plot of temperature for all sensors on the cables C1M, C2M, C3M, C5M, C1 and C3 during 63 days of storage 20.521.221.921.922.621.221.919.821.221.923.319.1 19.8 19.822.618.40 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C1M19.819.820.521.221.922.620.21.923.323.319.819.822.621.921.921.223.3 21.20 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C2M19.819.819.821.221.922.623.320.521.922.619.823.321.9 23.324.00 10 20 30 40 50 601.01.52.02.53.03.54.0Cable C3MTime (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7 20.521.221.922.621.221.923.322.23.323.320.21.9 23.320.524.021.20 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C5M21.221.921.922.620.521.922.621.923.319 822.619 123.3 24.02 60 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C121.221.921.922.620.521.923.323.321.922.619 821.923.3 24.00 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C3102  5.2.7 Changes in Relative Humidity During Storage of Wood Pellets Relative humidity (Rh) was recorded during storage by cables C1M to C5M. Figure  5.20 shows contour plots of relative humidity during 63 days of storage. For long term analysis of Rh minute-to-minute data were converted to daily average and used as the representative of the day relative humidity.  Appendix G includes the coordinate for all temperature and Rh sensors located on 5 cables. Before the wood pellets were loaded and on the very first days of experiment, Rh was between 20-24%. Rh increased to 26 and 28% in some locations over time and over 30% in the head-space. If we divide the silo to 2 sections of wood pellets and head-space, relative humidity pattern inside the pellet bulk was very similar for all 5 cables located at different radial locations. Close to the silo wall and in the centre, Rh was mostly between 26-30%. Rh at the surface of pellets was clearly higher. In general, according to the principles of psychometrics, relative humidity of air decreased with increasing dry bulb temperature.  However Rh in the silo head-space was higher at all times and above 28%. This could be explained by the temperature gradient inside the silo which causes moisture-carrying convection current inside the silo. Contour plots of relative humidity in Figure  5.20 show that the hydrothermal moisture migration was towards upper parts (head-space) of the silo. With moisture moving in the bed, there is also the possibility of promoting microbial growth and as a result more off-gassing (CH4) over time. Figure  5.17 clearly shows higher concentration of methane in the head-space of the silo at all times and it increased with time.   103  Figure  5.21 and Figure  5.22 show the contour plots of temperature and relative humidity at all heights and radius on a particular day of storage (day 15) when the maximum temperature was recorded. This measurement was before any purging took place. Again, recorded temperature clearly shows the highest spot located at the centre of the material (close to upper elevations) as was also predicted by the model developed by Guo [82].    Higher temperature will lead to higher local emission factors for CO, CO2 and CH4 as explained in chapter 5. Comparing off-gassing on day 15 (Figure  5.16 and Figure  5.17) at the elevation of 2.5 m where the highest temperature was recorded in the pilot silo, it is noticeable that concentration of gases and specifically CO and depletion of O2 was higher. Effect of temperature compared to relative humidity on gas emission is more significant in this study as changes in Rh within the pellet bulk were not high. Comparing the results of Yazdanpanah et al. [32] shows clearly that peak emission factor and reaction rate constant for off-gasses were much sensitive to temperature than relative humidity.  For pellets with 4% moisture content kept at 25 oC and 40 oC, reaction rate constant (k) increased from 0.20 s-1 to 1.73 s-1 for CO and from 0.05 s-1 to 0.99 s-1 for CO2. Details are presented in  Appendix I. A hypothetical situation can arise if temperature inside the pellets becomes high enough, then a pyrolysis zone would be created, and the concentrations of CO and CO2 would increase significantly.    104       Figure  5.20 Contour plot of relative humidity for all sensors on cables C1M to C5M during 63 days of storage 2830302826263224243026 302222320 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C1M2828282830302630280 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C2M26262828303028 280 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C4M283028262632320 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C3M28282628 280 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)2022242628303234Cable C5M105   Figure  5.21 Contour plot of temperature inside the pilot silo on day 15 of storage (r=0 refers to the centre of the silo)  Figure  5.22 Contour plot of relative humidity inside the pilot silo on day 15 of storage (r=0 refers to the centre of the silo) 21.2021.9022.6023.3020.5022.60 24.0019.8020.50-0.4 -0.2 0.0 0.2 0.41.01.52.02.53.03.54.04.5Radius (m)Height (m)17.0017.7018.4019.1019.8020.5021.2021.9022.6023.3024.0024.7027.0029.0029.0027.0027.0029.00 31.0031.0031.0033.0-0.4 -0.2 0.0 0.2 0.41.01.52.02.53.03.54.04.5Radius (m)Height (m)23.0025.0027.0029.0031.0033.0035.00106  5.3 Concluding Remarks Experiments were carried out in pilot silo to study the stratification of emitted gases during bulk storage of wood pellets. Gas composition is analyzed over time at different axial and radial positions in the silo. Local temperature, relative humidity and oxygen concentration in the silo are also measured. The results obtained clearly show high concentration (compared to threshold limits stated in Table  2.2) of gases as well as depletion of oxygen. Non-condensable gases emitted from the storage of 3tonnes of wood pellets were mainly carbon monoxide, carbon dioxide, methane and hydrogen. Maximum emission concentration was of 2.7 % CO2, 0.14% CH4 and 1.7% CO in the silo head-space.  Some stratification for carbon dioxide and methane were observed after a few days of storage, though almost uniform concentration was achieved over time. The stratification observed for carbon monoxide is attributed to high CO uptake of wood pellets during storage. Higher temperature and relative humidity play an important role in the higher emission factors, particularly carbon dioxide, associated with gas concentrations in the silo head-space compared to the bed of material. A very rapid drop of oxygen concentration to less than 5% was recorded within a week of storage; it was depleted first in the head-space. Although oxygen availability is important for accelerating CO emission, material storage in oxygen-free environment (dominated by N2) showed the same lethal CO concentrations, possibly due to the consumption of oxygen within the pellets. Experiments were carried out in pilot silo to study the temperature and relative humidity during storage of wood pellets. Temperature and Rh data recorded for 12 months of 107  storage. Temperature increased over time during storage at all levels reaching the wall temperature within 70 hours of experiment.  Maximum temperature increase was in the first 600-700 h of storage where temperature increased to 25 oC. A decrease in temperature was seen after that which could be due to less fatty acid contents of material over time. Highest temperature was recorded on day 15 of storage in the centre of the silo at the elevation of 2.5 m. Temperature changes in radial direction showed that temperature at all times was highest in this order: temperature in the centre, temperature half way between centre and wall, and temperature close to wall.  Study of relative humidity during the first 60 days of storage, shows that Rh increased in the head-space of the silo right after material was loaded. Changes within the pellet bulk were not significant. Larger gradients were recorded by cables located close to the silo wall and between wall and centre while changes in the silo centre was minimum. Higher Rh was also seen at the surface of the material throughout the experiment.     108  Chapter 6 Effectiveness of Purging on Gas Emission Buildup3 As explained in previous chapter, low temperature oxidation of pellets result in formation of aldehydes and low molecular weight carboxylic acids, CO2, CO and CH4 which constitute health and safety risks. Under certain circumstances, especially in enclosed areas with low ventilation, this results in acutely toxic environment with very high concentrations of these gases. As the non-condensable gases are poisonous and result in oxygen depletion, sufficient ventilation of storage rooms is necessary. Cargo holds are also sealed during ocean transportation of wood pellets and thus leads to a very fast oxygen depletion and generation of CO, CO2, CH4 and H2. Entry into cargo holds and communicating spaces are prohibited unless the spaces have been thoroughly ventilated and the gas concentration is verified by measuring CO and O2.The same situation exists when wood pellets are stored in silos before being shipped.  A forced ventilation system should be in place for controlling the thermal conditions in the stored pellets as explained in section  2.3. The fans? capacity can be calculated using the permeability of bulk wood pellets. This capacity must be high enough to provide effective cooling as too low capacity could make the situation worse. The effectiveness of such                                                  3A version of this chapter prepared for publication: F Yazdanpanah, S. Sokhansanj, C.J. Lim, A. Lau, X. Bi. Effectiveness of Purging on Gas Emission Buildup in Wood Pellet Storage. 109  ventilation system in reducing the concentration of off-gases in a wood pellet enclosure was studied in this chapter. Information on the degree of mixing in the bed would help in better evaluating the effectiveness of the purging system. Measurements were done on the gas concentration over time in a pilot scale silo and compared with the predicted values. The methodology and results are presented in this chapter.  6.1 Materials and Methods All experiments presented in this chapter were done in the pilot silo described in  Chapter 3. Gas mixing experiments were carried out by means of the tracer gas (He) and thermal conductivity detectors using the set-up illustrated in Figure  3.4. Three purging experiments were conducted at different airflow rates in the experimental silo. 6.1.1 Purging Experiments Purging experiments were carried out in the pilot silo in order to dilute and sweep the off-gasses inside the silo. Purging system was connected to both building air and nitrogen cylinders. Schematic of the set-up is shown in Figure  3.7 ( Chapter 3). Valve A that is located on top of the pilot silo (shown in Figure  3.7 and Figure A.7) remained open during all purging experiments to allow the effluent gas to flow to the ventilation system. Gas concentration, relative humidity and temperature were measured and monitored continuously. 110  6.1.1.1 Gas Measurements The composition of the gas in all locations (G1 to G13 gas sampling ports) within the pilot silo was analyzed by the GC before running the purging tests. Gas samples were drawn from 3 gas sampling ports (G0 in the head-space, G7 in the middle and G13 close to the bottom of the silo) and analyzed for the composition over time during purging tests. Three purging experiments were performed in this study. The first one was done on April 26, 2011 (1:30pm to 7:00pm), 11 weeks after the silo was loaded. The second and third experiments were done on August 2, 2011 (12:50pm to 5:40pm) and October 19 2011 (9:50am to 1:50pm) which were 25 weeks and 36 weeks after loading the pilot silo respectively. An experiment started by introducing air at a rate of 1.23E-3m.s-1 or 1.64E-3 m.s-1(superficial velocity with respect to the cross section of the silo) to the pilot silo through the gas diffuser. The purging experiment stopped once the oxygen concentration in the reactor reached 20% and off-gas concentrations were 0% in all locations. In order to determine the time required to reduce the off-gassing in the silo, the degree of mixing was evaluated first. Thus a series of experiments were done (section  6.1.2) to measure the gas residence time inside the silo.  6.1.1.2 Temperature and Relative Humidity Measurement Temperature and relative humidity were recorded every minute before and during the three purging experiments. Data collected by the nine vertical thermocouple cables (explained in section  3.1.3) are presented in this chapter. 111  6.1.2 Gas Mixing Experiments The experimental silo was filled to 75% with wood pellets as explained in  Chapter 3. Air was the gas flowing inside the bed. Tests were operated at ambient temperature and pressure. Helium was used as the tracer gas due to its properties (inert, non-adsorbing, mix intimately with flowing gas). The non-adsorbing property of the tracer allows studying that portion of the mixing which is solely related to the hydrodynamics. The tracer concentration was monitored and recorded every 5 sec at the outlet of the silo using two identical thermal conductivity detectors (TCD) assembled in our department. 6.1.2.1 Calibration of Thermal Conductivity Detectors The two thermal conductivity detectors (TCDs) were calibrated for different known concentrations of helium tracer gas in a stream of air (~ 0.8-12% He by volume) to confirm the linearity of the TCD output as well as obtaining calibration equations. A schematic diagram of the calibration set-up is shown in Figure  3.3. Red lines show the specific tubing configurations used for the calibration tests.  The calibration procedure is described below. A controlled amount of He and air were mixed before entering the TCDs. The tubes had sufficient length (tube diameter=1.27cm, tube length=4m) to ensure proper mixing of air and helium before entering the silo. A sample of He and air mixture after mixing was drawn by an air-tight GC syringe (25mL SGE Gas-Tight Syringe, Luer-Lock and TOGAS Luer Lock Adapter, Mandel Scientific Company) from a septum located after the mixing point (shown in red in Figure  3.3) and analyzed by GC/FID (Flame Ionization Detector) and GC/TCD (Thermal Conductivity Detector) for the 112  composition of the gas samples. A needle with a side-hole was used with the GC syringe to draw samples through the septum (Model number NLL-5/23H Luer Needle, Mandel Scientific Inc.). This is done for all concentrations of He used in calibration experiments. Technical specification on the GC is mentioned in section  3.2.  Mixture of He and air was used as the sample gas and air from cylinder as the reference gas for TCDs. A sample of known concentration was then drawn from the He-Air mixture into the TCD using a vacuum pump (EW-07531-40, Air Cadet Vacuum/Pressure Pump, Cole-Parmer, Canada) connected to the TCDs. Two identical flowmeters (4112K93, Panel-Mount glass flowmeter, McMaster-Carr, USA) were installed before the TCDs to ensure consistent and equal flow rates of sample and reference gas on both sides of the TCDs. The response voltage was amplified and recorded automatically. Results from different signal amplification ratio, sampling rate and current showed that the response signal is dependent on these parameters as also mentioned in [135, 136] using the same type of TCDs.  Optimum conditions in the experiments (with minimum fluctuations and maximum detected signal) were current of 70 mA, signal amplification of 100 and sampling flow rate of 50 mL/min (8.4?10-7 m3/s). Final calibration curves for both TCDs are presented in Figure  6.1. The voltage values presented in Figure  6.1 are the average values of 3 replicates. Calibration for each TCD was done using 5 replicates. The calibration curves for TCDs confirm that the calibration outputs were linear for both TCDs. Also the TCD signal measurements were reproducible within the examined range. 113    Figure  6.1 Plot of detected signals as a function of volume percent He injected for A) TCD#1, and B)TCD#2 [Signal amplification ratio=100; Current=70mA; TCD sample flow rate = 8.4?10-7 m3.s-1] 0 2 4 6 8 10 120.00.20.40.60.81.01.2Decetcted Output Signal (V)He Concentration (%)Ay=0.094x+0.0425R2=0.9970 2 4 6 8 10 12 140.00.20.40.60.81.01.21.41.61.82.0By=0.151x+0.0608R2=0.985Decetcted Output Signal (V)He Concentration (%)114  6.1.2.2 Residence Time Distribution (RTD) Experiments RTD experiments were conducted by injecting a step change of Helium to the silo at time t=0 and measuring the tracer concentration C in the exhaust stream as a function of time. A number of gases such as CO2 and He have been used as tracers for such studies with transient (step or pulse) injection or continuous injection. The tracer gas is a non-reactive species that is easily detectable. Moreover it should not adsorb on the walls or other surfaces in the silo [137]. Pulse input and step input are the most common methods. Advantages and disadvantages of different types of injections are listed by Arena [138].  For a pulse input, a fixed volume of tracer is injected in one shot in the gas stream flowing into the silo. The outlet concentration of the tracer is then measured as a function of time to give a curve of concentration versus time. The main difficulty with applying the pulse technique lies with producing a reasonable pulse input large enough to allow detection over an extended time period, yet small enough not to upset the flow patterns. The injection must happen over a very short time period compared to residence times in the silo, and there must be negligible dispersion between the point of injection and the silo entrance [137]. In this study, a positive step-change injection was employed. For step-change method, a constant (C=C0) rate of tracer is injected into the feed at t=0. Before this time (t<0), no tracer is added to the feed (C=0) [137]. In the experiment, helium was used as the tracer gas. The gas from the outlet of the silo was drawn to the TCDs using the pump shown in Figure  3.4. Helium concentration was measured in the outlet every 5 sec until the concentration in the effluent was indistinguishable from that of the feed. The step change is 115  easier to conduct experimentally compared to pulse injection. The shape of the residence time distribution (RTD) curve identifies the mixing pattern and deviation from plug flow and continuous stirred-tank reactor. 6.2 Results and Discussions 6.2.1 RTD Experiments In these experiments, a positive step tracer input with air superficial velocity ranging from 8.22E-4 to 5.76E-3 m.s-1 was used. Samples from the silo outlet and air reference were drawn continuously into the TCDs by the vacuum pump. For each velocity 3 sets of data were logged for 30-100min sampling period by each TCD, depending on the operating conditions. Both TCDs were located at the exit of the silo. The main factors that affect the accuracy of the RTD measurement are the sampling rate of the gas and possible differences in the characteristics of the two thermal conductivity cells. Corrections were made by controlling the sampling rates during the experiments in a way that both detectors will respond to a change in tracer concentration at precisely the same time. Besides, the possible differences in the characteristics of the two thermal conductivity cells were checked by placing the sampling tubes at the same location and recording their responses to the change in tracer concentration.  The concentration of the tracer gas was zero at t<0 and constant (C0) for t? 0.	Tracer gas concentration was kept at this level; once the same concentration was reached in the effluent the experiment discontinued. The cumulative distribution, F(t),represents the fraction 116  of molecules leaving the system that experiences a residence time less that t. It is determined directly from a step input as follows: F?t? ? ??????? ?????           6.1 The residence time distribution function, E(t) may be obtained by differentiating Eqn 6.1: E?t? ? ??? ??????? ?????           6.2 The mean residence time is: ? ? ? ?????????? ????????? ? t	E?t?dt??           6.3 Figure  6.2 shows the measured F(t) for the bed of pellets at different velocities. The F(t) values were obtained by converting the TCD recorded signal (voltage) to helium concentration using the calibration equations (Figure  6.1). E(t) was calculated based on the F(t) curves for all velocities. Finally,? was calculated using equation 6.3. Other replicates of the experiments done with TCD1 and TCD2 are presented in  Appendix K.  It can be seen from Figure  6.2 that as the velocity increased from 8.22E-4 to 4.11E-3 m.s-1, the mean residence time decreased from 36.0 to 14.3 min. When the velocity was increased further to 5.7E-3 m.s-1, the F(t) curve indicates channeling of airflow inside the silo (Figure  6.3).The first peak reflects channelling of the tracer flow directly from the input to the output, while the second peak represents the main flow of the fluid inside the system. 117  The first peak occurred at a time smaller than the space-time (t= V/U) indicating early exit of fluid, and some fluid exits at a time greater than the space-time which could be representative of the RTD for a system with channeling and dead zones. To characterize the axial spread of the tracer, axial dispersion in the experimental silo was studied. The magnitude of axial dispersion in the bed can affect the effectiveness of ventilation to a large extent. However, transverse dispersion may be neglected compared to axial dispersion if the D/L ratio is small or the fluid velocity is large. In this study one dimensional formulation and analysis is used.   118       Figure  6.2 Measured F(t) for the bed of pellets, for air velocity U = 8.22E-4 to 4.11E-3 m.s-1 . Tracer gas was injected from the lower diffuser. Sampling was done at the silo outlet.  0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=8.22x10-4 m/s?=36.00 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (Time)?=29.01 minU=1.23x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=1.64x10-3 m/s?=26.51 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)?=22.46 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)?=14.29 minU=4.11x10-3 m/s119   Figure  6.3 Measured F(t) for the bed of pellets, for air velocity U = 5.70E-3 m.s-1 . Tracer gas was injected from the lower diffuser. Sampling was done at the silo outlet. The longitudinal dispersion coefficient is generally larger than the radial dispersion coefficient by a factor of 5 for Reynolds number (Re) greater than 10 [139]. In this study, Re becomes closer to 10 at the higher velocities. As an example, Re at U=5.75E-3 m.s-1 is calculated as: Re ? ?????	????? ???.??????.??????.????????.?????	?????.??? ? 5        6.4 where ? is the fluid density (kg.m-3), U is the superficial velocity (m.s-1), Dp is equivalent particle diameter (m), ? is the fluid viscosity (kg.m-1.s-1) and ? is the bed porosity. For non-spherical particles, the equivalent particle diameter, Dp, can be written as [140]: 0 5 10 15 20 250.00.20.40.60.81.0 U=5.70x10-3 m/sC(t)/C0Time (min)120  Dp=6/Sv            6.5 Sv=As/Vp            6.6 where As is the mean surface area of the particles and Vp is the volume of particles. For a small control volume, the mass balance for the tracer gas ?considering no chemical reaction and for a two dimensional case - leads to: D? ?????? ?????? ?D?r????? ? u???? ?????          6.7 If we assume constant velocity, for axial direction only, equation 6.7 in 1 dimension is reduced to: D? ?????? ? u???? ?????            6.8 where DL is the longitudinal dispersion coefficient and u is the interstitial velocity (u=U/?). Putting equation 6.8 in dimensionless form and considering the closed-closed boundary condition (assuming no dispersion in concentration occurs upstream or downstream of the reaction section): ???? ????? ?????? 	?1 ? e?????          6.9 Where Per is the Peclet number (uL/DL) for the silo. u is the velocity in m.s-1, DL is the dispersion coefficient in m2.s-1 and L is the silo length in m.  ???? ? 2?????? ? 2???????	?1 ? e???/???                   6.10 121  The variance in residence time, which is an indication of the spread of distribution, is also calculated for all conditions using [141]: ?? ? ? ?t ? ???E?t?dt??                     6.11 ? and ? were calculated from t, F(t) and E(t) curves. ? was obtained from the area under the curve using equation 6.3. ? was calculated using equation 6.11. The system dispersion number (1/Per=DL/uL) was calculated using equation 6.10. For a velocity of 8.22E-4 m.s-1, DL/uL was found to be 0.11. Dispersion numbers of the system at different velocities are listed in Table  6.1: Table  6.1 Residence time, variance and system dispersion number at different purging velocities U (m.s-1) u (m.s-1) ? (min) ?(min) DL/uL Re Per 8.22E-4 2.16E-03 36.00 16.04 0.1100 0.77 9.1 1.23E-3 3.24E-03 29.01 8.24 0.0420 1.16 23.8 1.64E-3 4.32E-03 26.51 6.71 0.0345 1.55 29.0 2.47E-3 6.50E-03 22.46 4.79 0.0230 2.33 43.5 4.11E-3 1.08E-02 14.29 2.98 0.0220 3.89 45.5  As seen from Table  6.1, DL/uL is greater than 0.01 for all velocities and thus a large deviation from plug flow exists. A number of factors such as length of the bed, particle diameters, velocity, porosity, temperature, density and viscosity of the fluid would affect the dispersion coefficient. The system dispersion number characterizes the spreading rate of flow in the silo caused by different affecting factors. As the velocity of the fluid is increased, the 122  contribution of advection becomes dominant over that of dispersion. Figure  6.4 shows an increase in Peclet number as the velocity increases which indicates higher rate of transport due to advection rather than dispersion.   Figure  6.4 Plot of Peclet number versus Reynolds number The mean residence time distribution as a function of superficial velocity is plotted in Figure  6.5. As expected the mean residence time was found to decrease with increasing velocity. 1 2 3 45101520253035404550Peclet Number (u.L/D L)Re (?Udp/?(1-?))123   Figure  6.5 Mean residence time of the tracer gas as a function of superficial velocity The structure of the bed is one of the main factors that would affect dispersion in the bed, and it changes depending on the particle size distribution, surface friction and filling method. As a result the degree of packing and porosity in the bed are different. Studies [104, 106] show that axial dispersion coefficient increases over a wide range of particle size distribution. However, in a bed of wood pellets the anisotropy and non-uniformity of the voids between the pellets might cause deviation in the sole effect of particle size. Some of the properties of the fluid used as the purging gas, including viscosity and density, velocity and temperature in a bed of wood pellets would also affect the dispersion and should be considered.  The density and viscosity of air and nitrogen over the range of temperature from 0 to 100 oC are very close. However if carbon dioxide is to be used for ventilation purposes in the 0.0008 0.0016 0.0024 0.0032 0.0040 0.0048010203040???min?U (m/s)124  bed, the difference in properties should be taken into account. For the temperature range of 0 to 100 oC, CO2 density is 1.50-1.55 times that of air or nitrogen, and its viscosity is 7.95-8.45 times that of air and 8.20-8.80 times that of nitrogen depending on temperature.  6.2.2 Purging Efficiency 6.2.2.1 Gas Analysis Before introducing the purging gas into the silo, off-gas concentrations at all locations were measured. Results are shown in Table  6.2 for all three purging tests. The concentration of off-gases decreased after each purge. The third purging test was done after more than 8 months of storage where the gas concentration was still above threshold (Table  2.2). Table  6.2 Off-gas concentration inside the silo before first, second and third purging experiment  Purge 1 Purge 2 Purge 3 Location CO2 (%) CO (%) CH4 (%) CO2 (%) CO (%) CH4 (%) CO2 (%) CO (%) CH4 (%) G0 2.38 1.70 0.140 1.83 1.42 0.057 1.71 1.12 0.021 G2 2.35 1.70 0.135 1.80 1.47 0.054 1.70 1.10 0.023 G3 2.38 1.71 0.125 1.92 1.40 0.048 1.68 1.14 0.020 G4 2.35 1.68 0.132 1.91 1.41 0.054 1.73 1.16 0.023 G5 2.35 1.64 0.138 1.87 1.44 0.050 1.74 1.16 0.022 G6 2.38 1.60 0.129 1.83 1.47 0.056 1.70 1.14 0.022 G7 2.58 1.62 0.140 1.94 1.47 0.055 1.72 1.14 0.022 G8 2.32 1.66 0.145 1.87 1.48 0.052 1.75 1.15 0.023 G9 2.38 1.68 0.141 1.92 1.42 0.058 1.74 1.17 0.021 G10 2.39 1.62 0.129 1.87 1.48 0.049 1.72 1.16 0.022 G11 2.34 1.67 0.130 1.91 1.47 0.051 1.69 1.17 0.023 G12 2.30 1.61 0.135 1.80 1.42 0.056 1.74 1.11 0.021 G13 2.37 1.68 0.132 1.91 1.49 0.052 1.78 1.16 0.022 The number of turnovers is defined as the volume of purging gas divided by the volume of the silo: 125  N????????? ? U? t/V                     6.12 According to the results obtained from the gas mixing experiments, deviation from plug flow is large for low-flow velocities, and as velocity increased the deviations became smaller.  Equation 6.8 was solved by MATLAB for the concentration of off-gas in the bed of wood pellets at different locations over time. The measured concentrations of the off-gas in  Chapter 5 showed that gas concentrations did not vary much in all locations inside the silo except for some stratification that occurred during the early days of storage. Thus the average concentration of the off-gas component (CO, CO2 or CH4) in the bed was used as the initial concentration in numerical simulation and one curve is presented for each purging velocity.  Figure  6.6 shows the predicted off-gas concentration in the bed for the velocity of U=1.23E-3 m.s-1 after t=1,2,3 and 4h. As time proceeds, the speed of frontal movement slows down. Results of gas concentrations (CO2) during first purging experiment using numerical solution and experimental data at different elevations are shown in Figure  6.7. The results obtained from the numerical solution predict best the off-gas concentration in the bottom and middle of the silo whiles it over-predicts the gas concentration in the silo head-space. Calculated concentrations of CO2 at locations G7 and G13 at different times using numerical solution show good agreement with the experimental data obtained in the first purging test (Figure  6.8) while it always over-predict the gas concentration in the head-space (G0) (Figure  6.10). Calculated concentrations of CO at locations G7 and G13 at different times using numerical solution is shown in Figure  6.9. The predicted values for CO show 126  good agreement with the experimental data obtained in the purging test at location G13 and G7). Depending on the degree of mixing in the bed, an empirical approach is also evaluated and compared with the experimental data ( Appendix M).                             Figure  6.6 Predicted CO2 concentrations over time during silo 1st purging using numerical solution (DL=6.2E-4 m2.s-1 and initial CO2 concentration was 2.38%)  0 2 40.00.51.01.52.02.5Concentration (%)Z (m)U=1.23E-3 m.s-1t=1h0 2 40.00.51.01.52.02.5 U=1.23E-3 m.s-1t=2hConcentration (%)Z (m)0 2 40.00.51.01.52.02.5 U=1.23E-3 m.s-1t=3hConcentration (%)Z (m)0 2 40.00.51.01.52.02.5 U=1.23E-3 m.s-1t=4hConcentration (%)Z (m)127   Figure  6.7  Predicted and experimental CO2 concentrations over time during silo purging using numerical solution (CO2 initial concentration was 2.38%)  Figure  6.8 CO2 concentration measured and predicted as a function of time during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts]  0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50.00.51.01.52.02.5t=4hG0G0G0G13G13G7G7t=1ht=2hConcentration (%)Z (m)t=3hG7-40 0 40 80 120 160 200 240 280 3200.00.51.01.52.02.5 Experimental (G13)   Experimental (G7) Experimental (G0) Numerical (G7) Numerical (G13)CO2 Concentration (%)Time (min)128   Figure  6.9 CO concentration measured and predicted as a function of time during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts]  Figure  6.10 CO2 concentration predicted as a function of time and different dispersion coefficients during 1st purging experiment [U=1.23E-3m.s-1, Negative time refers to time before purging starts] -40 0 40 80 120 160 200 240 280 3200.00.51.01.52.0 Experimental (G13) Experimental (G7) Experimental (G0) Numerical (G13) Numerical (G7)CO Concentration (%)Time (min)-40 0 40 80 120 160 200 240 280 320 360 4000.00.51.01.52.02.5DL=2.7E-04 m2.s-1 Experimental (G0) Numerical (G0) DL=6.2E-4 m2.s-1 Numerical (G0) DL=2.7E-4m2.s-1CO2 Concentration (%)Time (min)DL=6.2E-04 m2.s-1129  Figure  6.11 shows the number of turnovers for the vessel increased by more than 2.5 times as the superficial velocity increased from 8.22E-4 to 4.11E-3 m.s-1 in order to reduce the CO2 concentration from 2 to 0.5% using air for purging.   Figure  6.11 Number of turnover as a function of velocity to reduce CO2 concentration from 2 to 0.5% in the silo head-space 6.2.2.2 Temperature and Relative Humidity Changes During Purging Experiments Temperature and relative humidity were also recorded during the purging tests. Figure  6.12 is a contour plot of temperature during the first purging experiment on April 26, 2011. Purging experiment was conducted from 1:30 to 7:00pm. The graph shows data recorded every minute from 00:00 to 23:59 on that day. Time is converted to min whereby t=0 refers to time 00:00 and t=1439 refers to time 23:59. Purging was done using the purging system explained in section  3.1.6 with compressed air at 18.5 oC. More than 20,000 0.001 0.002 0.003 0.0041.52.02.53.03.54.04.55.0N TurnoverU (m.s-1)130  temperature and relative humidity data were recorded for each day that experiment was done. Temperature inside the pilot silo during the first 63 days of storage (before purging) clearly shows a decline in temperature after 720 hours of storage (Figure  5.19). Figure  6.12 shows similar temperature profile after 720 hours of storage. Purging started at 810 min with a flow rate of 1.23E-3 m.s-1 and continued for 5.5 h. The purging air started to bring down the temperature at 1000 min in the lower parts of the silo. Purging also moved the higher temperature slightly upwards; yet the duration of purging was not long enough to lower the temperature as the cooling front moved too slowly at ~1cm/min. Resistance of bulk pellets to flow could have also increased due to the formation of fines inside the bulk as a result of the extended storage time period. More plots for temperature on other cables during the 2ndand 3rd purging experiments are presented in  Appendix H. As shown in the graphs, slightly cooler air that was injected into the silo brought down the temperature in the lower parts of the silo. However it should be noted that the air used for purging was quite dry. In humid areas, ventilation using ambient air, can lead to temperature escalation in the storage if its relative humidity is too high. Eventually, it can also moisten the pellets and contribute to greater microbial activities thus causing higher gas emissions.  In this study, air was used as the purging gas as temperature was not high. However if temperature rises to critical values, the flow of air should be blocked and replaced with nitrogen or carbon dioxide. Lower air flow velocities showed larger deviations from plug flow and thus better mixing in the system. On the other hand, the output of the blower should be able to overcome the system pressure in order to deliver certain amount of air. Yazdapanah [113] developed the relationship between the airflow velocity and pressure drop 131  based on experimental data. Using the pressure drop versus flow rate data and blower curve, the operating point is determined for the system. Taking into account the efficiency of the blower, the required flow rate for use in the event that temperature should rise to critical value can be estimated. Figure  6.13 shows the contour plot of relative humidity during the first purging experiment on April 26, 2011 on all cables C1M to C5M. Before purging, the relative humidity patterns inside the bulk of pellets were seen to be very similar for all 5 cables located at different radial locations, with clearly higher relative humidity close to the surface of the pellets. In the contrast, the Rh in the silo head-space was higher at all times and above 28%. The relative humidity pattern continued to be the same, with the maximum changes in humidity occurring in the head-space and the minimum changes in humidity in the silo centre. Purging started at 810 min with a flow rate of 1.23E-3 m.s-1 and continued for 5.5 h. Around 1100 min, slightly lower (24-15%) Rh values were recorded by most of the cables close to the bottom of the silo. However, as the relative humidity of the injected air was about 24-25%, minimum changes are observed during the purging experiment.   132     Figure  6.12 Contour plot of temperature for all sensors on the cables C1M-C5M during first purge experiment done on April 26th (Data shown represent the minute-to-minute temperature data from 00:00 AM to 23:59 PM on April 26th , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min) 20.5 20.520.520.5 20.520.5 20.520.521.221.221.221.219.821.920.521.219.80 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0Time (min)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C1M20.521.219.819.121.221.218.421.921.217.721.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0Time (min)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C2M20.521.219.819.121.221.2 21.918.421.221.921.921.9 21.921.921.91. 21.921.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0Time (min)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C3M20.521.221.219.821.2 21.919.121.921.9 21.921.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0Time (min)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C4M21.221.221.2 21.919.821.220.519.121.920.520.520.520.5 20.520.520.520.520.520.520.521.921.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0Time (min)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C5M133     Figure  6.13 Contour plot of relative humidity for all sensors on the cables C1M to C5M during first purge experiment done on April 26 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on April 26 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min) 28.028.028.030.032.032.032.030.034.030.00 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)20.022.024.026.028.030.032.034.0Cable C1M28.030.030.030.030.030.030.026.00 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)20.022.024.026.028.030.032.034.0Cable C2M28.028.028.030.026.032.032.032.032.0 32.032.00 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)20.022.024.026.028.030.032.034.0Cable C3M28.028.028.030.030.0 30.026.00 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)20.022.024.026.028.030.032.034.0Cable C4M28.028.028.026.030.00 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)20.022.024.026.028.030.032.034.0Cable C5M134  6.3 Concluding Remarks Experiments were done in the experimental silo to study the effectiveness of a purging system in diluting off-gas concentrations over time. Study of the degree of mixing was needed to calculate the time and volume of purging gas to reduce the off-gas concentrations in the bed from C1 (the off-gas concentration at t=0) to C2 (the off-gas concentration at time t). Superficial air velocity ranging from 8.22E-4 to 4.11E-3 m.s-1 was used. The results obtained from RTD experiments showed large deviations from plug flow at velocity of 8.22E-4 and the deviations decreased as the velocity increased. As velocity increased some channeling was observed inside the bed. Using numerical solution for the gas axial dispersion, off-gas concentration over time was calculated and compared to the off-gas concentration during purging the experiments inside the silo. The calculated results showed to fit best to the concentration measured in the bottom and middle of the silo while the results for the head-space over-predicted the off-gas concentration. The number of turnovers also increased as velocity increased. Temperature and relative humidity changes were recorded during purging of the experimental silo. During a 5.5 h purging experiment with air at 18.5 oC, decrease in temperature was seen in the lower parts and slightly in the middle part after 200 min of purging. Right before the purging experiment, maximum and minimum variations of relative humidity were recorded in the head-space and centre of the silo respectively. The effect of purging on Rh variation was minimal due to low Rh of the purging gas.  135  Chapter 7 Conclusions and Recommendations 7.1 Conclusions Emission and stratification of off-gasses from storage of wood pellets have been studied. The focus of the study was to investigate the spatial and temporal concentration of off-gases and purging effectiveness. A summary of the major findings is as follows: 1- In order to analyse and measure the adsorption of off-gases and oxygen by wood pellets during storage, Temperature Programmed Desorption (TPD) analysis was used. Steam exploded pellets showed the lowest CO2 uptake compared to regular and torrefied pellets. Torrefied pellets showed highest CO2 adsorption possibly due to their porous structure and thus higher available surface area. Due to chemical reaction and therefore strong bond between the material and carbon monoxide, quantifying the uptake of CO by pellets was challenging.  The very high calculated energy of desorption for carbon monoxide (97.8 kJ/mol) and relatively low energy of desorption for carbon dioxide (7.24 kJ/mol) demonstrated the chemical and physical adsorption for CO and CO2 respectively. For oxygen, similar to carbon monoxide a very high temperature was needed to overcome the energy of adsorption. As the desorption temperature increased to 140 and 180 oC, more oxygen came off from the material. Strong bonds formed between the material and oxygen verified the existence of chemical adsorption and formation of intermediate material.  136  2- Emission and stratification of off-gases was studied in pilot scale storage for over one year. Non-condensable gases that emitted from storage of material were carbon monoxide, carbon dioxide, methane and hydrogen. To study the stratification of gases, analysis of gas composition was done over time for different axial and radial positions. Moreover, temperature, relative humidity and oxygen depletion was studied over time. The emitted gases showed to have higher emission factor compared to work done with white wood pellets in small scale. It could be explained by the fact that off-gassing is a surface phenomenon and thus much active surface is available for reaction when larger amount of wood pellets are stored.  Concentration of gases at plateau after 9 weeks of storage was2.7 % CO2, 0.14% CH4 and 1.7% CO in the rector head-space. After a few days of storage some stratification were observed for carbon monoxide and methane. The clear stratification of carbon monoxide could be due to high uptake of CO by wood pellets over time. The unstable environment in the storage of wood pellets in the first few weeks of storage was due to very rapid consumption of oxygen in the space, reaction of emitted CO and CO2 with pellets, development of high temperature spots within the bulk due to higher activity of pellets and thus higher local temperatures, and moisture migration within the bed. Oxygen plays an important role in accelerating the emission of carbon monoxide. Results from experiment where material was kept in oxygen-free environment (N2-rich) confirmed the same lethal concentration of CO possibly through consuming the oxygen within the material.   137  3- Studies done in the same pilot silo on temperature and relative humidity changes during long term storage of 3 tonnes of material. Recorded data for 12 months of storage, showed the increase in temperature inside the silo from the beginning of tests where it reached the wall temperature within 70 hours of storage. Highest temperature increase (25 oC) was observed in the first 600-700 hours of storage. During the whole period of storage, maximum temperature in the silo was seen on day 15 of storage at the elevation of 2.5 m. At any time, highest temperature was seen in the silo centre and lowest temperature was close to the silo wall.  An increase in relative humidity in the silo head-space was clear once the silo was loaded. Higher Rh was also seen at the interface of material while changes in Rh within the wood pellet bed were minimum. Measured temperature in the silo during 5.5 hour purging experiments with air at 18-18.5 oC, helped the decrease in the lower parts and slightly middle parts of the silo after 200 minutes of purging.  4- In order to evaluate the effectiveness of a purging system and quantify the time and volume of the gas needed to sweep the off-gases from the experimental silo, multiple purging tests were done. To identify the degree of mixing in the silo, experiments were done on residence time distribution of the gas. Large deviations from plug flow and thus better mixing was seen for all superficial velocities used. However as the velocity increased, system dispersion number became smaller which indicated less mixing and more volume of purging gas.  One dimensional numerical simulation of the off-gas concentration in the bed showed to predict the concentration best at the bottom and middle of the silo while it over-predicted 138  the off-gas concentration in the head-space. In an empirical approach, the results from equation M.4 considering the mixing factor fitted to measured off-gas concentration during purging experiment. The results showed to fit best to the measured off-gas concentration in the silo head-space while it overestimated the exponential decay of the off-gases in the bottom and middle of the silo.  7.2 Limitations and Contributions of This Research Although the research reached its objectives, there were some unavoidable limitations. First due to the silo scale, installation and preparation of the reactor was more challenging than expected. Moreover, as the study aimed to investigate the distribution of off-gases for fresh materials, it was impractical to repeat the tests with the same batch of wood pellets. Ventilation studies are limited to constant conditions in temperature. Temperature and relative humidity vary and thus more precise ventilation schedule must be investigated and developed. The results here refer to a condition in which pellets would have a similar bulk density across and along the height. Pellets in bulk commercial full size silos probably will not have uniform bulk density and that would affect stratification in flow of gases. The present work has yielded the following contributions to knowledge: It adds new contribution to rare available data on quantifying spatial and temporal distribution of off-gases in large scale experimental silo. Findings from material characterization in respect to physical and chemical reaction with the material supported the distribution of gases and thermal condition existed in the silo. The capability of a purging system effectiveness in reducing off-gas concentration was examined experimentally and compared with predicted 139  values. The data obtained from this study could also fill the gap between studies that are done in smaller scale experiments and industrial data taken from full scale pellet transport in ocean vessels. 7.3 Recommendations This study has experimentally measured the off-gassing and its stratification in pilot scale of wood pellets during long term storage.  o Effects of local temperature and relative humidity were investigated. However, one of the main factors that affects the rate of off-gassing and helps self-heating is environment temperature. In order to stimulate the solar radiation, it is essential to heat up one side of the silo and measure off-gas concentration and stratification, temperature and relative humidity and oxygen depletion. Also this study could be done for torrefied wood pellets and wood chips in the experimental silo to study the rate of off-gassing and possibility of using such purging cycles. o To further study the depletion of oxygen in storages of wood pellets and investigate the possible formation of intermediate material, a detailed analysis of fatty acids content of the wood pellet sample before and after storage is suggested. Moreover, to better quantify the adsorption capacity of gases by different types of pellets, specific surface area and pore volume of regular, torrefied and steam-exploded pellets can be determined. 140  o In this study 1-D analysis is done for the gas dispersion and it is not far from reality as the silo diameter to length ratio was not large. However in real case scenario where silo diameter to length ratio could be close to one, transverse dispersion should be considered. It will be useful to identify system transverse dispersion number.  o Effects of factors that play the highest role on rate of off-gassing such as temperature, material moisture content, relative humidity and oxygen availability are known. For temperature and moisture content, these effects are known at extreme limits (e.g. 50% w.b. MC and 60 oC). Data on the changes in emission factor of each gas as a function of amount of stored material is also available. A numerical model can be developed that could incorporate the effect of main factors to simulate the off-gassing emission and stratification as the critical parameters change over time.  o The model could be further developed to simulate the decrease in off-gas concentration at all locations in the bed during purging practices considering the real scenarios with the assumptions made in the last chapter of this study in calculating purging gas volume and turn over. For example, transverse dispersion should be taken into account. Temperature and gas velocity in bed are not constant where they were assumed to be constant in this study. The desorption of gases and specifically carbon dioxide (at lower temperature) during purging could also attribute to development of higher concentration in bed and thus more consumed purging gas to bring down the concentration to a safe level. Higher temperatures would lead to higher rate of off-gas desorption from pellets which should be taken into account in numerical simulations.  141  References [1] Lehtikangas P. Quality properties of pelletised sawdust, logging residues and bark. Biomass and Bioenergy 2001; 20: 351-360. [2] Obernberger I, Thek G. The pellet handbook: The production and thermal utilization of biomass pellets. London, UK ; Washington, DC: Routledge; 2010. [3] BC Ministry of Energy, Mines and Petroleum Resources, and BC Ministry of Forests and Range. An information guide on pursuing biomass energy opportunities and technologies in British Columbia. Biomass & Bioenergy Conference 2008. [4] Liodakisa S, Bakirtzisa D, Loisb E, Gakisa D. The effect of (NH4)(2)HPO4 and (NH4)(2)SO4 on the spontaneous ignition properties of Pinus halepensis pine needles. Fire Safety Journal 2002; 37: 481?494. [5] Wadso L. Measuring chemical heat production rates of biofuels by isothermal calorimetry for hazardous evaluation modeling. Fire and Materials 2007; 31(4): 241-255. [6] Explosion at Armstrong plant. www.globaltvbc.com; 2011. [7] Fire destroys wood stove business in Norwich. [Accessed March  11 2013]. http://www.norwichbulletin.com/news/x780391163/Blaze-breaks-out-at-Norwich-business#axzz2RttagHg9. 2011. [8] Firefighters battle huge biomass fire at Port of Tyne. [Accessed March 11 2013]. http://www forestbusinessnetwork com/10442/firefighters-battle-huge-biomassfire-at-port-of-tyne/ 2011. [9] Explosion damages Waycross plant, no injuries reported. [Accessed March 11 2013]. http://www forestbusinessnetwork com/4314/explosion-damages-waycross-plantno-injuries-reported/ 2011. [10] Firefighters leave power station. in BBC 2012. [11] Power station engulfed in blaze. in BBC 2012. [12] Firefighters battle blaze at Scotland plant. [Accessed March 11 2013]. http://www laurinburgexchange com/view/full_story/17804602/articleFirefighters-battle-blaze-at-Scotland-plant?instance=popular 2012. [13] Units fight another blaze at NC wood pellet plant. . in WMBF News 2012. 142  [14] Melin S, Svedberg U, Samuelsson J. Emission from wood  pellets during ocean transportation 2008; WPAC Research Report. [15] Melin S. Review of Off-gassing from Wood Pellets: A Canadian Perspective 2010; WPAC Research Report. [16] Selkimaki M, Mola-Yudego B, Roser D, Prinz R, Sikanen L. Present and future trends in pellet markets, raw materials, and supply logistics in Sweden and Finland. Renew Sust Energ Rev 2010; 14: 3068-3075. [17] Emhofer W. Market research, national incidents, problems and lacks within national guidelines 2013. [18] Melin S.  Research on off-gassing and self-heating in wood pellets during bulk storage 2011; WPAC Research Report. [19] Arshadi M, Geladi P, Gref R, Fj?llstr?m P. Emission of volatile aldehydes and ketones from wood pellets under controlled conditions. In Ann  Occup  Hyg 2009; 53(8): 797-805. [20] Arshadi M, Gref R. Emission of volatile organic compounds from softwood pellets during storage. For Prod J 2005; 55: 132-135. [21] Svedberg URA, Hogberg HE, Hogberg J, Galle B. Emission of hexanal and carbon monoxide from storage of wood pellets, a potential occupational and domestic health hazard. Ann Occup Hyg 2004; 48: 339-349. [22] Hagstr?m K. Occupational exposure during production of wood pellets in Sweden. Doctoral Dissertation 2008. [23] Risholm-Sundman M, Lundgren M, Vestin E, Herder P. Emissions of acetic acid and other volatile organic compounds from different species of solid wood. Holz Als Roh-Und Werkstoff 1998; 56: 125-129. [24] Wihersaari M. Greenhouse gas emissions from final harvest fuel chip production in Finland. Biomass & Bioenergy 2005; 28: 435-43. [25] Wihersaari M. Evaluation of greenhouse gas emission risks from storage of wood residue. Biomass & Bioenergy 2005; 28(5): 444-453. [26] Manninen A, Pasanen P, Holopainen JK. Comparing the VOC emissions between air-dried and heat-treated Scots pine wood. Atmos Environ 2002; 36: 1763-1768. 143  [27] Banerjee S. Mechanisms of terpene release during sawdust and flake drying 2001; 55: 413-416. [28] Roffael E. Volatile organic compounds and formaldehyde in nature, wood and wood based panels. Holz Als Roh-und Werkst 2006; 64: 144-149. [29] Rupar K, Sanati M. The release of terpenes during storage of biomass. Biomass & Bioenergy 2005; 28: 29-34. [30] McGraw GW, Hemingway RW, Ingram LL, Canady CS, McGraw WB. Thermal degradation of terpene: camphene, delta-3-carenic, limonene and alfa-terpinene. Environmental Science & Technology 1999; 33: 4029-4033. [31] Granstrom KM. Emissions of Hexanal and Terpenes during Storage of Solid Wood Fuels. For Prod J 2010; 60: 27-32. [32] Yazdanpanah F, Sokhansanj S, Lim CJ, Lau A, Bi X, Lam PY, et al. Potential for flammability of gases emitted from stored wood pellets. The Canadian Journal of Chemical Engineering In Press. [33] Meyer S. Fatalities in enclosed spaces 2008. [34] Svedberg U, Samuelsson J, Melin S. Hazardous off-gassing of carbon monoxide and oxygen depletion during ocean transportation of wood pellets. Ann Occup Hyg 2008; 52: 259-266. [35] Levitt MD, Ellis C, Springfield J, Engel RR. Carbon monoxide generation from hydrocarbons at ambient and physiological temperature: a sensitive indicator of oxidant damage?. Journal of Chromatography A 1995; 695: 324-328. [36] Reuss R, Pratt S. Accumulation of carbon monoxide and carbon dioxide in stored canola. J Stored Prod Res 2001; 37: 23-24. [37] Johansson LS, Leckner B, Gustavsson L, Cooper D, Tullina C, Potterc A. Emission characteristics of modern and old-type residential boilers fired with wood logs and wood pellets. Atmos Environ 2004; 38 (25): 4183?4195. [38] Meijer R, Gast CH. Spontaneous combustion of biomass: experimental study into guidelines to avoid and control this phenomenon. Proceedings of the 2nd world conference on Biomass for BioEnergy,Industry and Climate Protection 2004; II: 1231-1233. [39] Richardson J, Bj?rheden R, Hakkila P, Lowe AT, Smith CT. Bioenergy from sustainable Forestry 2002. 144  [40] Yang L, Guo Z, Zhou Y, Fan W. The influence of different external heating ways on pyrolysis and spontaneous ignition of some woods. Journal of Analytical and Applied Pyrolysis 2007; 78(1): 40-45. [41] Fasina OO, Sokhansanj S. Bulk thermal properties of alfalfa pellets. Canadian Agricultural Engineering 1995; 37(2): 91-95. [42] Kubler H. Heat generating processes as cause of spontaneous ignition in forest products. Forest Products Abstracts 1987; 10: 299-327. [43] Hoell W, Piezconka K. Lipids in sap and heartwood of Picea abies (L.) Karst. Z Pflantzenphysiol 1978; 87: 191-198. [44] Piispanen R, Saranpaa P. Neutral lipids and phospholipids in Scots pine (Pinus sylvestris) sapwood and heartwood. Tree Physiology 2002; 22(9): 661. [45] Back E, Allen L. Pitch control, wood resin and deresination. Atlanta, GA: Tappi Press; 2000. [46] Johansson A, Rasmuson A. The release of monoterpenes during convective drying of wood chips. Drying Technology 1998; 16(7): 1395 -1428. [47] Whittle CP, Waterford. CJ, Annis PC, Banks HJ. The production and accumulation of carbon monoxide in stored dry grain. Journal of Stored Products Research 1994; 30: 23-26. [48] Lavery MR, Milota MR. Measurement of VOC emissions from ponderosa pine lumber using commercial and laboratory kilns. Drying Technology 2001; 19(9): 2151?2173. [49] Koppejan J, L?nnermark A, Persson H, Larsson I, Blomqvist P, Arshadi M, et al. Health and safety aspects of solid biomass storage, transportation and feeding. IEA Bioenergy 2013. [50] Emhofer W, P?llinger-Zierler B, Siegmund B, Haslinger W, Leitner E. Correlation between CO off-gassing and linoleic fatty acid content of wood chips and pellets 2013. [51] Emhofer W, Christian P. Report Lagertechnik und Sicherheit bei der Pelletslagerung. Bioenergy2020+ GmbH 2009. [52] Emhofer W. Safety Measures for the Storage of Wood Pellets at the End Customer ? Status of the discussion   in VDI and ?NORM. 11th Pellets Industry Forum. [53] Emhofer W. BTEC NYSERDA pellet storage project webinar. 145  [54] Tumuluru J, Kuang X, Sokhansanj S, Lim CJ, Bi X, Melin S. Development of  laboratory studies on the off-gassing of wood pellets. Canadian Biosystems Engineering 2010; 52: 8.1-8.9. [55] Tumuluru JS, Sokhansanj S, Lim CJ, Bi T, Kuang XY, Melin S. Effect of low and high storage temperatures on headspace gas concentrations and physicalproperties of wood pellets. International Wood Products Journal 2012. [56] Kuang X, Shankar TJ, Bi XT, Lim CJ, Sokhansanj S, Melin S. Rate and peak concentrations of off-gas emissions in stored wood pellets-sensitivities to temperature, relative humidity and headspace volume. Ann Occup Hyg 2009; 53: 789-796. [57] Kuang X, Shankar TJ, Sokhansanj S, Lim CJ, Bi XT, Melin S. Effects of headspace and oxygen level on off-gas emissions from wood pellets in storage. Ann Occup Hyg 2009; 53: 807-813. [58] Hemingway RW, Nelson PJ, Hillis WE. Rapid oxidation of the fats and resins in Pinus radiata chips for pitch control. TAPPI Journal 1971; 54: 95-98. [59] Martinez-Inigo MJ, Immerzeel P, Gutierrez A, Del Rio JC, Sierra-Alvarez R. Biodegradability of extractives in sapwood and heartwood from scots pine by sapstain and white rot fungi. Holzforschung 1999; 53: 247-252. [60] Schmutzer-Roseneder I, Emhofer W, Haslinger W. Emissions from wood pellets during storage reffering to the extractive content. Proceedings of WSED next 2013. [61] Tumuluru JS, Sokhansanj S, Wright CT, Hess JR, Boardman RD. A Review on Biomass Torrefaction Process and Product Properties. Symposium on Thermochemical Conversion August 2011. [62] Svedberg U, Calle B. Evaluation of the Time Correlated Tracer (TCT) Method for Assessment of Diffuse Terpene Emissions from Wood Pellets Production. Stockholm, Sweden: V?rmeforsk; 2001. [63] Arshadi M, Nilsson D, Geladi P. Monitoring chemical changes for stored sawdust from pine and spruce using gas chromatography-mass spectrometry and visible-near infrared spectroscopy. Near Infrared Spectroscopy 2007; 15: 379-386. [64] Finell M, Arshadi M, Gref R, Scherzer T, Knolle W, Lestander T. Laboratory-scale production of biofuel pellets from electron beam treated scots pine (Pinus silvestris L.) sawdust. Radiation Physics and Chemistry 2009; 78(4): 281-287. 146  [65] Kuang X, Shankar TJ, Bi XT, Sokhansanj S, Lim CJ, Melin S. Characterization and kinetics study of off-Gas emissions from stored wood pellets. Ann Occup Hyg 2008; 52: 675-683. [66] Yazdanpanah F, Sokhansanj S, Lim CJ, Lau A, Bi X, Melin S. Stratification of off-gases in stored wood pellets. Biomass and Bioenergy In Press. [67] Svedberg U, Petrini C, Johanson G. Oxygen depletion and  formation of toxic gases following sea transportation of logs and wood chips. Ann Occup Hyg 2009; 53: 779?787. [68] Madsen A, Martensson L, Schneider T, Larsson L. Microbial dustiness and particle release of different biofuels. Ann Occup Hyg 2004; 48: 327-338. [69] Boddy L. Carbon dioxide release from decomposing wood: effect of water content and temperature. Soil Biology and Biochemistry 1983; 15(5): 501-510. [70] ACGIH. TLVs and BEIs: Based on the Documentation of the Threshold Limit Values for Chemical  Substances and Physical Agents & Biological Exposure Indices. American Conference of Governmental Industrial Hygienists 2004. [71] US. Department of Labor. Occupational Safety and Health Administration (www.osha.gov). [72] L?nnermark A, Persson H, Blomqvist P, Larsson I, Rahm M. Self-heating and off-gassing from biomass pellets during storage. World Bioenergy Conference 2012 proceedings 2012: 9-15. [73] Loo SV, Koppejan J. The handbook of biomass combustion and co-firing. London: Sterling, VA : Earthscan; 2008. [74] Rupar-Gadd K. Biomass pre-treatment for the production of sustainable energy - emissions and self-ignition. Vaxjo University 2006. [75] Miao Y, Yoshizaki S. Mechanism of spontaneous heating of Hay. 1. Necessary conditions and heat generation from chemical reactions. Transactions of the ASAE 1994; 37(5): 1561-1566. [76] Bilbao R, Mastral JF, Lana JA, Ceamanos J, Aldea ME, Betr?n M. A model for the prediction of the thermal degradation and ignition of wood under constant and variable heat flux. Journal of Analytical and Applied Pyrolysis 2002; 62 (1): 63-82. [77] Nurmi J. The storage of logging residue for fuel. Biomass and Bioenergy 1999; 17(1): 41-47. 147  [78] Springer EL, Hajny GJ. Spontaneous heating in piled wood chips. TAPPI Journal 1970; 53: 85-86. [79] Kubler H. Air convection in self heating piles of wood chips. TAPPI Journal 1982; 65: 79-83. [80] Kubler H, Wang YR, Barkalow D. Generation of heat on wood between 80 C and 130 C. Holzforschung 1985; 39: 85-89. [81] Larsson SH, Lestander TA, Crompton D, Melin S, Sokhansanj S. Temperature patterns in large scale wood pellet silo storage. Appl Energy 2012; 92: 322-7. [82] Guo W. Self heating and  spontaneous combustion of wood pellets during storage. PhD Dissertation , University of British Columbia 2013. [83] Persson H, Blomqvist P, Yan-LTH Z. Fire and fire suppression in silos, an experimental study. Interflam ?07 2007. [84] US Chemical Safety and Hazards Investigation Board, Investigation Report. Combustible Dust Hazard Study 2006; Report No 2006-H-1. [85] Stelte W. Guideline: Storage and Handling of Wood Pellets. Danish Technological Institute 2012. [86] Wood pellet association of canada. Material safety data sheet of wood pellets in bulk 2009. [87] Yazdanpanah F. Permeability of bulk wood pellets with respect to airflow. University of British Columbia, Chemical and Biological Engineering 2009. [88] Melin S. Testing of explosibility and fammability of airborne dust from wood pellets. Wood Pellet Association of Canada 2008. [89] S.A. Sawmilling Pty Ltd. Material safety data sheet of sawdust from non-treated softwood 2000. [90] Hyne & Son PTY Limited. Material safety data sheet of sawdust from non treated hardwood. Biomass & Bioenergy 2007. [91] Fan C, Bi X. Development of Off-Gas Emission Kinetics for Stored Wood Pellets. Ann  Occup  Hyg 2013; 57(1): 115?124. [92] Nordic Innovation Centre. Guidelines for storing and handling of solid biofuels 2008; NT ENVIR 010. 148  [93] Deutscher Energieholz- und Pellets-Verband e.V. Empfehlung zur Lagerung von Holzpellets 2011. [94] Department of Environmental Protection. Emissionsminderung - Lagerung von Holzpellets beim Verbraucher - Anforderungen an das Lager unter Sicherheitsaspekten. Commission on Air Pollution Prevention of VDI and DIN - Standards Committee KRdL 2012. [95] European Biomass Association. First international workshop on pellet safety: Results and main lines of action. [96] Schwartz CE, Smith JM. Flow Distribution in Packed Beds. Industrial and Engineering Chemistry 1953; 45: 1209-1218. [97] Cairns EJ, Prausnitz JM. Longitudinal Mixing in Packed Beds. Chemical Engineering Science 1960; 12: 20-34. [98] Coelho D, Thovert JF, Adler PM. Geometrical and transport properties of random packings of spheres and aspherical particles. Physical Review E 1997; 55: 1959-1978. [99] Stephenson JL, Stewart WE. Optical Measurements of Porosity and Fluid Motion in Packed-Beds. Chemical Engineering Science 1986; 41: 2161-2170. [100] Gunn DJ. Axial and Radial Dispersion in Fixed-Beds. Chemical Engineering Science 1987; 42: 363-373. [101] Akehata T, Sato K. Flow distribution in packed beds. Chem Eng Japan 1958; 22: 430-436. [102] Hiby JW. Longitudinal dispersion in single-phase liquid flow through ordered and random packings. Interact between Fluid &Particles 1962: 312-325. [103] Raimondi P, Gardner GHF, Petrick CB. Effect of pore structure and molecular diffusion on the mixing of miscible liquids flowing in porous media 1959. [104] Niemann EH, Greenkorn RA, Eckert RE. Dispersion during flow nonuniform heterogeneous porous media 1971. [105] Eidsath A, Carbonell RG, Whitaker S, Herrmann LR. Dispersion in Pulsed Systems .3. Comparison between Theory and Experiments for Packed-Beds. Chemical Engineering Science 1983; 38: 1803-1816. [106] Wronski S, Molga E. Axial dispersion in packed beds: the effect of particle size non-uniformities. Chem Eng Process 1987; 22: 123-135. 149  [107] Bernard RA, Wilhelm RH. Turbulent Diffusion in Fixed Beds of Packed Solids. Chem Eng Prog 1950; 46: 233-244. [108] Ebach EA, White RR. Mixing of Fluids Flowing through Beds of Packed Solids. AICHE J 1958; 4: 161-169. [109] Carberry JJ, Bretton RH. Axial Dispersion of Mass in Flow through Fixed Beds. AICHE J 1958; 4: 367-375. [110] Strang DA, Geankoplis CJ. Longitudinal Diffusivity of Liquids in Packed Beds. Industrial and Engineering Chemistry 1958; 50: 1305-1308. [111] Guedes de Carvalho JRF, Delgado JMPQ. Radial dispersion in liquid flow through packed beds for 50 < Sc < 750 and 103 < Pem < 105. Proceedings of 5th World Conference on Experimental Heat Transfer, Fluid Mechanics Thermodynamics 2001. [112] Yazdanpanah F, Sokhansanj S, Lau AK, Lim CJ, Bi X, Melin S, et al. Permeability of wood pellets in the presence of fines. Bioresour Technol 2010; 101: 5565-5570. [113] Yazdanpanah F, Sokhansanj S, Lim CJ, Lau A, Bi X, Melin S. Airflow versus pressure drop for bulk wood pellets. Biomass & Bioenergy 2011; 35(5): 1960-1966. [114] Yazdanpanah F, Sokhansanj S, Lim CJ, Lau A, Bi X, Melin S. Resistance of wood pellets to low airflow. Canadian Journal of Chemical Engineering 2012; 90(6): 1479-1483. [115] ?NORM M 7137, Austrian Standards Institute. Compressed wood in natural state - Woodpellets - Requirements for storage of pellets at the ultimate consumer 2012. [116] Lam PS, Sokhansanj S, Bi X, Lim CJ. Effect  of steam explosion on wood pellet quality. Annual Meeting of AIChE 2010. [117] Peng JH, Bi HT, Sokhansanj S, Lim JC. A study of particle size effect on biomass torrefaction and densification. Energy Fuels 2012; 26: 3826-3839. [118] Mohsenin NN. Physical properties of  plant and animal materials. New York: Gordon and Breach Science Publishers; 1986. [119] ASABE. Moisture measurement  - forages. Standard S358.2. ASABE 2006: 608. [120] Solid biofuels - Fuel specifications and classes. . EN 14961-1 2010. 150  [121] Yazdanpanah F, Sokhansanj S, Lim CJ, Lau A, Bi X, Lam PY, et al. Study of wood pellet quality change during storage at different moisture and temperature condition. internal review. [122] Teixeira G, Van de Steene L, Martin E, Gelix F, Salvador S. Gasification of char from wood pellets and from wood chips: Textural properties and thermochemical conversion along a continuous fixed bed. Fuel 2012; 102: 514-524. [123] Chandrasekaran SR, Hopke PK, Rector L, Allen G, Lin L. Chemical composition of wood chips and wood pellets 2012; 26 (8): 4932 - 4937. [124] Health C. HPB methods of microbiological analysis of foods (MFHPB 18) 2001; 2. [125] Cvetanovic RJ, Amenomiya Y. Application of a temperature-programmed desorption  technique to catalyst studies. Advan Catal 1967; 17: 103-149. [126] Haydar S, Moreno-Castilla C, Ferro-Garcia M, Carrasco-Marin F, Rivera-Utrilla J, Perrard A, et al. Regularities in the temperature-programmed desorption spectra of CO2 and CO from activated carbons. Carbon 2000; 38: 1297-1308. [127] Wang Z, Chen Y, Zhou C, Whiddon R, Zhang Y, Zhou J, et al. Decomposition of hydrogen iodide via wood-based activated carbon catalysts for hydrogen production. Int J Hydrogen Energy 2011; 36: 216-223. [128] Haydar S, Joly J. Study of the evolution of carbon dioxide from active carbon by a threshold temperature-programmed desorption method. J Therm Anal 1998; 52: 345-353. [129] Petkovic LM, Ginosar DM, Rollins HW, Burch KC, Deiana C, Silva HS, et al. Activated carbon catalysts for the production of hydrogen via the sulfur?iodine thermochemical water splitting cycle. Int J Hydrogen Energy 2009; 34: 4057-4064. [130] Khezami L, Chetouani A, Taouk B, Capart R. Production and characterisation of activated carbon from wood components in powder: Cellulose, lignin, xylan. Powder Technol 2005; 157: 48-56. [131] Acharya B. Torrefaction and Pelletization of Different Forms of Biomass of Ontario March 2013. [132] Amenomiya Y, Cvetanovic RJ. Application of Flash-Desorption Method to Catalyst Studies .1. Ethylene-Alumina System. J Phys Chem 1963; 67: 144. [133] Carpenter TM, Fox EL. Absence of stratification and rapidity of mixing of carbon dioxide in air samples. J  Biol  Chem 1927; 73:379-381. 151  [134] Theilacker JC, White MJ. Diffusion of gases in air and its affect on oxygen deficiency hazard abatement. Transactions of the cryogenic Engineering conference - CEC 2006; 51. [135] Abba I. A Generalized Fluidized Bed Reactor Model across the Flow  Regimes. PhD Thesis 2001. [136] Al-Sherehy FA. Distributed addition of gaseous reactants in fluidized beds 2002. [137] Fogler HS. Elements of Chemical Reaction Engineering. 2nd ed. ed. London: Prentice Hall Inc.; 1992. [138] Arena U. Gas Mixing, Chapter3. In: Grace JR, Knowlton TM, Avidan AA, editors. Circulating Fluidized Beds, London,: Chapman and Hall; 1197. [139] Delgado J. A critical review of dispersion in packed beds. Heat Mass Transf 2006; 42: 279-310. [140] Comiti J, Renaud M. A new model for determining mean structure parameters of fixed beds from pressure drop measurements: application to beds packed with parallelepipedal particles. Chem  Eng  Sci 1989; 44(7): 1539?1545. [141] Levenspiel O. Chemical reaction engineering: Wiley; 1972.         152     Appendices     153  Appendix A Pilot Silo and All Attached Monitoring and Measurement Instruments  Figure A. 1 Pilot scale silo installed at Clean Energy Research Centre, Department of Chemical and Biological Engineering, University of British Columbia  154   Figure A.2 Gas sampling ports located at different levels of the silo  Figure A.3 View from the top of the pilot silo showing the installed rods inside the gas sampling ports  Figure A.4 Solid sampling ports located at 6 different elevations of the silo 155   Figure A.5 Pressure transducers located at different elevations of the silo  Figure A.6 Side view of the silo showing wall thermocouples as well as horizontal thermocouples located at 5 different levels  Figure A.7 View of the silo from the top. The top end of the vertical thermocouples are shown with the black wires attached to them. 4 of the thermocouple cables are located in inner ring and 4 in the outer ring. Wall thermocouplesHorizontal thermocouples156   Figure A.8 Purging system attached to the silo(connected to both nitrogen and air).  Figure A.9 Manometer attached to the silo outlet to avoid pressure build up inside the silo 157   Figure A.10 TCD #1 and #2, amplifier, pump, flowmeters and data logging system  Figure A.11 Flowmeters attached to TCDs to measure sampling rate 158   Figure A.12 TCD#1, TCD #2 and attached amplifier  Figure A.13 Reference (air) gas cylinder, He gas cylinder and building air line Helium Air Building Air Line TCD #1 TCD #2 Amplifier 159   Figure A.14 Nitrogen/Air distributor 1 and 2  Figure A.15 Air/N2 and He flowmeters, 3-way and needle valves used in purging and RTD experiments Distributor 1 Distributor 2160  Appendix B Adsorption Tests B.1 Equipment and Sample Tube Pictures  Figure B. 1 Micromeritics Autochem 2920 II  Figure B. 2 Sample tube loaded with wood pellet sample for adsorption test  161  B.2 TCD Calibration TCD was calibrated for CO2, CO and O2 before running TPD tests. Analysis yield data on signal reading, peak area, temperature and time. To correlate the signal readings collected in the analysis with the volume of gas uptake at any given point in the analysis, the equipment default calibration test was run with a series of known gas concentrations. The calibration file was then associated with the signal reading to calculate the gas concentration. During the calibration test, the analyzer decreases the proportion of the analysis gas in 10% increments, beginning with 100% and ending with 0%. The resultant data appears as a series of ten stepwise changes in the TCD signal. Figure B. 3 shows the graph of TCD O2 calibration carried out prior to O2 desorption tests.   Figure B. 3 Micromeritics TCD calibration curve for O2162  B.3 Typical TPD Curves Obtained for CO2, CO and O2  Figure B. 4 TCD signal and temperature reading over time for carbon dioxide Minutescm?/min0.0000.0050.0100 20 40 60 80 100 120 140 160 180 200 220 240 260Temperature (?C)50100163   Figure B. 5TCD signal and temperature reading over time for carbon monoxideMinutescm?/min0.00.50 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75Temperature (?C)50100150164  Appendix C SEM Images of Treated and Untreated Wood Pellets             Figure C. 1 SEM micrographs of white wood pellet surface before CO adsorption    165             Figure C. 2 SEM micrographs of white wood pellet surface after CO adsorption   166  Appendix D GC Scan for Gas Sample Collected from Pilot Silo (a) FID Reading  (b) TCD Reading  Figure D. 1 (a) FID reading for CO2, CO and CH4 and (b) TCD reading for Helium, Oxygen and NitrogenH2 O2 N2 CO2 CH4 CO 167  Appendix E Emission of Off-gases in Oxygen-free and Oxygen-rich Environments 12 reactors (2 liter capacity) were used in these experiments. Each reactor was filled to 75% with wood pellets (977 g of wood pellets). 6reactorswere equipped with proper fittings to allow purging the reactor with different gases. Figure E. 1 show the configuration of purging and sampling instruments for each reactor. To purge the reactors, one valve of the 2 valves connected to the top of the reactor was connected to the desired purging gas and the other valve left open and was connected to ventilation. Flow rate used in purging was approximately 20 mL.min-1. Each reactor was purged for 200 minutes. A 20 cm 1/8? stainless steel rod was connected to one valve and located in the centre of the reactor to ensure through purge of reactor. After 200 minutes, the purging gas source was shut down and the 2 valves were closed. For 2 reactors where oxygen (air) was injected every 4 days, a septum was installed for injecting air through syringe needle.  One sample was taken from the reactor by an air-tight GC syringe (25mL SGE Gas-Tight Syringe, Luer-Lock and TOGAS Luer Lock Adapter, Mandel Scientific Company) and analyzed by GC to confirm the gas is 100% the purging gas and no oxygen is available. Experiments were carried out in 2 replicates for 60 days of storage. Table E. 1 has the information on each reactor condition for tests. Gas sample was taken from all reactors on days 14, 24, 31, 49 and 60 except for reactors A-O3-PEAK-REP1 and A-O3-PEAK-REP2 where only the peak concentration was measured on day 60. Results obtained from these experiments are presented in Table E. 2.   168  Table E. 1 Reactor and test conditions for 12 reactors used in oxygen-free and oxygen-rich experiments Reactor Name Conditions Sampling Days A-O1-REG-REP1 Reactor was filled to 75% with wood pellets and sealed for regular off-gassing experiments. (replicate 1) 14,24,31,49,60 A-O2-REG-REP2 Reactor was filled to 75% with wood pellets and sealed for regular off-gassing experiments. (replicate 2) 14,24,31,49,60 A-O3-PEAK-REP1 Reactor was filled to 75% with wood pellets and sealed for regular off-gassing experiments (replicate 1). However one sample of gas was taken at the end of experiment (after 60 days when plateau was reached.) 60 A-O4-PEAK-REP2 Reactor was filled to 75% with wood pellets and sealed for regular off-gassing experiments (replicate 2). However one sample of gas was taken at the end of experiment (after 60 days when plateau was reached.) 60 A-O5-N2 PURGED-REP1 Reactor was filled to 75% with wood pellets, purged with nitrogen and sealed for off-gassing experiments (replicate 1).  14,24,31,49,60 A-O6-N2 PURGED-REP2 Reactor was filled to 75% with wood pellets, purged with nitrogen and sealed for off-gassing experiments (replicate 2).  14,24,31,49,60 169  Reactor Name Conditions Sampling Days A-O7-CO2 PURGED-REP1 Reactor was filled to 75% with wood pellets, purged with carbon dioxide and sealed for off-gassing experiments (replicate 1).  14,24,31,49,60 A-O8-CO2 PURGED-REP2 Reactor was filled to 75% with wood pellets, purged with carbon dioxide and sealed for off-gassing experiments (replicate 2).  14,24,31,49,60 A-O9-He PURGED-REP1 Reactor was filled to 75% with wood pellets, purged with helium and sealed for off-gassing experiments (replicate 1).  14,24,31,49,60 A-O10-He PURGED-REP2 Reactor was filled to 75% with wood pellets, purged with helium and sealed for off-gassing experiments (replicate 2). 14,24,31,49,60 A-O11-O2 PUMPED-REP1 Reactor was filled to 75% with wood pellets sealed for off-gassing experiments (replicate 1). However oxygen was injected to the reactor every 4 days. 14,24,31,49,60 A-O12-O2 PUMPED-REP2 Reactor was filled to 75% with wood pellets sealed for off-gassing experiments (replicate 2). However oxygen was injected to the reactor every 4 days. 14,24,31,49,60  170   Figure E. 1 Empty containers equipped with 2 valves, fittings and stainless steel rod for purging and gas sampling  Figure E. 2 Reactors filled to 75% with wood pellets for experiments171  Table E. 2 Off-gassing results from oxygen-free and oxygen-rich experiments Regular Test N2 Purged Day CO2 CO CH4 O2 N2 H2 Day CO2 CO CH4 O2 N2 H2 14 0.690 0.870 0.033 6.088 91.450 0.479 14 0.658 0.705 0.017 0.340 97.760 0.100 24 1.687 1.204 0.048 5.077 90.870 0.628 24 1.440 1.040 0.022 0.303 96.550 0.118 31 2.319 1.409 0.055 4.471 90.554 0.752 31 1.700 1.201 0.024 0.332 96.054 0.124 49 2.301 1.393 0.067 3.142 91.740 0.868 49 1.671 1.184 0.027 0.270 96.012 0.130 60 2.311 1.405 0.071 2.110 92.450 0.971 60 1.705 1.281 0.030 0.291 96.220 0.143 CO2 Purged He Purged Day CO2 CO CH4 O2 N2 H2 Day CO2 CO CH4 O2 N2 H2 14 97.200 0.181 0.024 0.255 1.670 0.299 14 0.235 0.155 0.017 0.600 3.220 0.187 24 85.740 0.512 0.030 0.341 12.360 0.362 24 0.456 0.252 0.027 0.452 5.210 0.264 31 73.010 0.684 0.035 0.340 24.904 0.406 31 0.589 0.330 0.034 0.448 7.397 0.303 49 69.310 0.701 0.042 0.354 28.477 0.510 49 0.574 0.414 0.035 0.509 10.269 0.433 60 68.960 0.713 0.049 0.035 28.870 0.581 60 0.582 0.443 0.035 0.451 11.700 0.458 O2 Pumped Peak Only Day CO2 CO CH4 O2 N2 H2 Day CO2 CO CH4 O2 N2 H2 14 0.952 0.859 0.039 10.080 87.230 0.511 60 2.3347 1.3920 0.0705 0.5100 94.1400 0.9580 24 2.030 1.320 0.054 7.086 87.201 0.668 60 2.3315 1.4011 0.0712 0.5074 94.2800 0.9589 31 2.871 1.472 0.062 6.400 87.760 0.812          49 3.790 1.500 0.074 5.550 87.840 0.950          60 3.910 1.671 0.078 6.078 86.710 0.974                172  Appendix F Peak Emission Factor for CO, CO2, CH4 and H2 as Effected by Head-space Percentage 10 reactors (2 liter capacity) were used in these experiments (same reactor as the ones in  Appendix E). 5 Reactors were filled with white wood pellets with 10, 25, 50, 60 and 75% head-space (2 replicates for each). All reactors were equipped with proper fittings to draw gas samples and sealed. One gas sample was taken from the reactors by an air-tight GC syringe (25mL SGE Gas-Tight Syringe, Luer-Lock and TOGAS Luer Lock Adapter, Mandel Scientific Company) after 60 days of experiment and analyzed by GC for composition. Table F. 1 has the results on peak emission factor for all reactors. Experiment was done in 2 replicates. Figure F. 1, Figure F. 2, Figure F. 3 and Figure F. 4 present the peak emission factor for carbon monoxide, carbon dioxide, methane and helium versus the reactor head-space respectively. Higher head-space percentage and as a result higher oxygen in the reactors showed to have an important effect on peak emission of carbon monoxide, carbon dioxide and hydrogen. However the peak concentration of methane was insensitive to oxygen concentration.   173  Table F. 1 Peak emission factor for gases emitted from wood pellets reactors with different head-space percentage 10% Head-space Day CO2 Peak Emission Factor COPeak Emission Factor CH4 Peak Emission Factor H2 Peak Emission Factor 60 0.035 0.010 0.0004 0.0004 60 0.031 0.010 0.0004 0.0004 25% Head-space 60 0.048 0.013 0.0004 0.0004 60 0.050 0.013 0.0005 0.0005 50% Head-space 60 0.088 0.019 0.0003 0.0005 60 0.091 0.021 0.0004 0.0006 60% Head-space 60 0.096 0.021 0.0004 0.0005 60 0.102 0.024 0.0004 0.0006 75% Head-space 60 0.149 0.024 0.0003 0.0005 60 0.149 0.025 0.0003 0.0006  174   Figure F. 1 Peak emission factor for carbon monoxide for different head-space percentage after 60 days  Figure F. 2 Peak emission factor for carbon dioxide for different head-space percentage after 60 days 0 10 20 30 40 50 60 70 800.0100.0150.0200.0250.030 CO emission factor (replicate 1)   CO emission factor (replicate 2)CO emission factor (g/kg)HD percentage (%)0 10 20 30 40 50 60 70 800.020.040.060.080.100.120.140.160.18  CO2 emission factor (replicate 1 )  CO2 emission factor (replicate 2)CO2 emission factor (g/kg)HD percentage (%)175   Figure F. 3 Peak emission factor for methane for different head-space percentage after 60 days  Figure F. 4 Peak emission factor for hydrogen for different head-space percentage after 60 days 0 10 20 30 40 50 60 70 800.000280.000300.000320.000340.000360.000380.000400.000420.000440.00046  CH4 emission factor (replicate 1)  CH4 emission factor (replicate 2)CH4 emission factor (g/kg)HD percentage (%)0 10 20 30 40 50 60 70 800.000400.000450.000500.000550.000600.00065 CH4 emission factor (replicate 1)   CH4 emission factor (replicate 2)H 2 emission factor (g/kg)HD percentage (%)176  Appendix G Sensor and Gas Sampling Ports Coordinates  Figure G. 1 Coordinate of gas sampling ports and some sensors located on 2 cables177  Table G. 1 Coordinate of all temperature sensors located on 9 vertical thermocouple cables Elevation and Radius (r,y) Cable Number C1 C2 C3 C4 C5M C1M C2M C3M C4M S1 (-0.5,0.2648) (-0.3,0.2648) (0.3,0.2648) (0.5,0.2648) (0,0.5696) (0.5,0.5696) (0.3,0.5696) (-0.3,0.5696) (-0.5,0.5696) S2 (-0.5,0.4172) (-0.3,0.4172) (0.3,0.4172) (0.5,0.4172) (0,0.8744) (0.5,0.8744) (0.3,0.8744) (-0.3,0.8744) (-0.5,0.8744) S3 (-0.5,0.5696) (-0.3,0.5696) (0.3,0.5696) (0.5,0.5696) (0,1.1792) (0.5,1.1792) (0.3,1.1792) (-0.3,1.1792) (-0.5,1.1792) S4 (-0.5,0.7220) (-0.3,0.7220) (0.3,0.7220) (0.5,0.7220) (0,1.484) (0.5,1.484) (0.3,1.484) (-0.3,1.484) (-0.5,1.484) S5 (-0.5,0.8744) (-0.3,0.8744) (0.3,0.8744) (0.5,0.8744) (0,1.7888) (0.5,1.7888) (0.3,1.7888) (-0.3,1.7888) (-0.5,1.7888) S6 (-0.5,1.0268) (-0.3,1.0268) (0.3,1.0268) (0.5,1.0268) (0,2.0936) (0.5,2.0936) (0.3,2.0936) (-0.3,2.0936) (-0.5,2.0936) S7 (-0.5,1.1792) (-0.3,1.1792) (0.3,1.1792) (0.5,1.1792) (0,2.3984) (0.5,2.3984) (0.3,2.3984) (-0.3,2.3984) (-0.5,2.3984) S8 (-0.5,1.3316) (-0.3,1.3316) (0.3,1.3316) (0.5,1.3316) (0,2.7032) (0.5,2.7032) (0.3,2.7032) (-0.3,2.7032) (-0.5,2.7032) S9 (-0.5,1.4840) (-0.3,1.4840) (0.3,1.4840) (0.5,1.4840) (0,3.008) (0.5,3.008) (0.3,3.008) (-0.3,3.008) (-0.5,3.008) S10 (-0.5,1.6364) (-0.3,1.6364) (0.3,1.6364) (0.5,1.6364) (0,3.3128) (0.5,3.3128) (0.3,3.3128) (-0.3,3.3128) (-0.5,3.3128) S11 (-0.5,1.7888) (-0.3,1.7888) (0.3,1.7888) (0.5,1.7888) (0,3.6176) (0.5,3.6176) (0.3,3.6176) (-0.3,3.6176) (-0.5,3.6176) S12 (-0.5,1.9412) (-0.3,1.9412) (0.3,1.9412) (0.5,1.9412) (0,3.9224) (0.5,3.9224) (0.3,3.9224) (-0.3,3.9224) (-0.5,3.9224) S13 (-0.5,2.0936) (-0.3,2.0936) (0.3,2.0936) (0.5,2.0936) (0,4.2272) (0.5,4.2272) (0.3,4.2272) (-0.3,4.2272) (-0.5,4.2272) S14  (-0.5,2.2460) (-0.3,2.2460) (0.3,2.2460) (0.5,2.2460) --- (0.5,4.532) (0.3,4.532) (-0.3,4.532) (-0.5,4.532) S15 (-0.5,2.3984) (-0.3,2.3984) (0.3,2.3984) (0.5,2.3984) --- --- --- --- --- S16 (-0.5,2.5508) (-0.3,2.5508) (0.3,2.5508) (0.5,2.5508) --- --- --- --- --- S17 (-0.5,2.7032) (-0.3,2.7032) (0.3,2.7032) (0.5,2.7032) --- --- --- --- --- S18 (-0.5,2.8556) (-0.3,2.8556) (0.3,2.8556) (0.5,2.8556) --- --- --- --- --- S19 (-0.5,3.0080) (-0.3,3.0080) (0.3,3.0080) (0.5,3.0080) --- --- --- --- --- S20 (-0.5,3.1604) (-0.3,3.1604) (0.3,3.1604) (0.5,3.1604) --- --- --- --- --- S21 (-0.5,3.3128) (-0.3,3.3128) (0.3,3.3128) (0.5,3.3128) --- --- --- --- --- S22 (-0.5,3.4652) (-0.3,3.4652) (0.3,3.4652) (0.5,3.4652) --- --- --- --- --- S23 (-0.5,3.6176) (-0.3,3.6176) (0.3,3.6176) (0.5,3.6176) --- --- --- --- --- S24 (-0.5,3.7700) (-0.3,3.7700) (0.3,3.7700) (0.5,3.7700) --- --- --- --- --- S25 (-0.5,3.9224) (-0.3,3.9224) (0.3,3.9224) (0.5,3.9224) --- --- --- --- --- S26 (-0.5,4.0748) (-0.3,4.0748) (0.3,4.0748) (0.5,4.0748) --- --- --- --- --- S27 (-.05,4.2272) (-0.3,4.2272) (0.3,4.2272) (0.5,4.2272) --- --- --- --- --- S28 (-0.5, 4.3796) (-0.3, 4.3796) (0.3, 4.3796) (0.5,4.3796) --- --- --- --- --- S29 (-0.5, 4.532) (-0.3, 4.532) (0.3, 4.532) (0.5, 4.5320) --- --- --- --- --- 178  Table G. 2 Coordinate of all relative humidity sensors located on 5 vertical cables Elevation and Radius (r,y) Cable Number C5M C1M C2M C3M C4M S1 (0,0.5696) (0.5,0.5696) (0.3,0.5696) (-0.3,0.5696) (-0.5,0.5696) S2 (0,0.8744) (0.5,0.8744) (0.3,0.8744) (-0.3,0.8744) (-0.5,0.8744) S3 (0,1.1792) (0.5,1.1792) (0.3,1.1792) (-0.3,1.1792) (-0.5,1.1792) S4 (0,1.484) (0.5,1.484) (0.3,1.484) (-0.3,1.484) (-0.5,1.484) S5 (0,1.7888) (0.5,1.7888) (0.3,1.7888) (-0.3,1.7888) (-0.5,1.7888) S6 (0,2.0936) (0.5,2.0936) (0.3,2.0936) (-0.3,2.0936) (-0.5,2.0936) S7 (0,2.3984) (0.5,2.3984) (0.3,2.3984) (-0.3,2.3984) (-0.5,2.3984) S8 (0,2.7032) (0.5,2.7032) (0.3,2.7032) (-0.3,2.7032) (-0.5,2.7032) S9 (0,3.008) (0.5,3.008) (0.3,3.008) (-0.3,3.008) (-0.5,3.008) S10 (0,3.3128) (0.5,3.3128) (0.3,3.3128) (-0.3,3.3128) (-0.5,3.3128) S11 (0,3.6176) (0.5,3.6176) (0.3,3.6176) (-0.3,3.6176) (-0.5,3.6176) S12 (0,3.9224) (0.5,3.9224) (0.3,3.9224) (-0.3,3.9224) (-0.5,3.9224) S13 (0,4.2272) (0.5,4.2272) (0.3,4.2272) (-0.3,4.2272) (-0.5,4.2272) S14 --- (0.5,4.532) (0.3,4.532) (-0.3,4.532) (-0.5,4.532)    179  Appendix H Contour Plot of Temperature and Relative Humidity   Figure H. 1 Contour plot of temperature for all sensors on the cables C4M, C2 and C4 during 63 days of storage   19.819.819.820.521.221.922.620.521.923.319.119.821.918.422.19.82.621.919.123.323.323.3 21.918.40 10 20 30 40 50 601.01.52.02.53.03.54.0Cable C4MTime (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.720.521.922.620.521.921.923.323.39.122.68.4 19 822.69.121.223.318.419.117.717.722.60 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C120.521.221.922.620.521.923.323.319.121.922.68.4 19.822.69.117.718.419.118.423.30 10 20 30 40 50 601.01.52.02.53.03.54.0Time (day)Height (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C4180    Figure H. 2 Contour plot of temperature for all sensors on the cables C1 to C4 during first purge experiment done on April 26 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on April 26 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=810 min)    20.521.221.221.221.919.821.921.221.221.2 21.921.221.221.221.221.221.221.2 20.521.921.221.2 21.221.221.921.921.221.221.219.821.921.221.221.9.21.222.61.221.221.222.62.622.622.621.921.921.9 22.6 22.60 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C120.521.221.219.821.919.121.218.4 17.721.917.721.921.221.221.221.221.221.221.221.221.221.221.221.221.221.221.920.521.2 21.221.220.520.521.220.520.521.220.521.220.520.521.221.2 21.921.922.622.620.521.9 21.921.921.221. 21.221.921.221.221.221.222.621.20 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C220.521.221.919.821.920.521.221.221.921.221.221.2.21.221.221.221.221.221.221.221.221.221.221.921.921.920.520.521.2.19.820.521.9.21.221.220.50 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C320.521.221.221.219.821.919.1 18.421.221.221.221.921.221.221.9 21.2 21.921.220.520.520.520.519.821.220.520.5 20.521.221.921.921.220.521.20 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C4181    Figure H. 3 Contour plot of temperature for all sensors on the cables C1 to C4 during 3rd purge experiment done on October 19 (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on October 19 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=590 min)      19.820.521.221.922.619.122.622.622.62 .62 .6 22.6 22.619.10 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C119.820.521.221.922.619.122.622.622.621.922.619.1.19.822.622.621.9 21.922.619.8.19.819.821.9 21.921.9 21.921.922.621.921.919.822.619.822.6 22.622.622.622.622.6 22.622.622.620.50 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C219.820.521.221.922.623.323.3 23.323.323.319.122.622.619.119.10 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C319.820.521.221.922.619.123.323.3 23.323.3 23.32 .3 23.319.123.323.319.1.19.119.1 19.119.1 19.119.9.123.3 23.3 23.323.320.523.321.9 20.521.9 20.520.5 20.520.5 20.520.5 20.50 200 400 600 800 1000 1200 14000.51.01.52.02.53.03.54.04.5TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C4182     Figure H. 4 Contour plot of temperature for all sensors on the cables C1M to C5M during 3rd purge experiment done on October 19th (Data shown represent the minute-to-minute temperature data from 00:00 to 23:59 on October 19 , with time 00:00:00 represented as time 0 and 23:59:00 as 1439. Purging started at t=590 min) 19.820.521.221.922.6 22.619.121.922.622.621.921.9 21.921.921.9 21.921.921.921.921.9 21.921.921.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C1M19.820.521.221.922.619.123.323.321.921.921.90 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C2M19.820.521.221.922.619.123.323.30 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C3M19.820.521.221.922.60 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C4M20.520.520.520.521.221.922.623.319.823.30 200 400 600 800 1000 1200 14001.01.52.02.53.03.54.0TimeHeight (m)17.017.718.419.119.820.521.221.922.623.324.024.7Cable C5M183   Figure H. 5 Temperature recorded in the lab from January 14 to April 24 2011 (The thermometer was located next to the lower part of the silo which was close to the lab door and exposed to outside air)    1/14/2011 2/3/2011 2/23/2011 3/15/2011 4/4/2011 4/24/201115161718192021Temp (o C)Date184  Appendix I CO and CO2 Concentration Fitted to 1st Order Kinetic Reaction (Data obtained from [32] )  Figure I. 1 CO concentration for pellets with 4% MC at 25 oC (stored for 60 days)  Figure I. 2 CO concentration for pellets with 4% MC at 40 oC (stored for 60 days) 0 20 40 600.20.40.60.81.01.2CO Concentration (%)Day4% MC and 25 oC0 20 40 601.31.41.51.64% MC and 40 oCCO Concentration (%)Day185   Figure I. 3 CO2 concentration for pellets with 4% MC at 25 oC (stored for 60 days)  Figure I. 4 CO2 concentration for pellets with 4% MC at 40 oC (stored for 60 days)  0 20 40 600.00.20.40.60.81.04% MC and 25 oCCO2 Concentration (%)Day0 20 40 602.02.53.04% MC and 40 oCCO2 Concentration (%)Day186  Appendix J OPI Temperature and Rh Recording System  Figure J. 1 Temperature reading for each cable at any moment  Figure J. 2 Minimum, Maximum and average temperature recording 187   Figure J. 3 Relative humidity data reading for cables C1M to C5M at any moment  Figure J. 4 EMC (Equilibrium Moisture Content) at any point [EMC is calculated and not measured] 188   Figure J. 5 Pressure reading from 6 transducers and temperature reading from horizontal and wall thermocouples    189  Appendix K Measured F(t) Curves at Different Velocities   Figure K. 1 F(t) measured for the bed of pellets, for U = 8.22E-4 to 4.11E-3 m.s-1. [TCD 1 Replicate 2(left) and Replicate 3 (right)]0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=8.22x10-4 m/s?=35.89 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=1.64x10-3 m/s?=25.68 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)?=21.78 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)?=14.02 minU=4.11x10-3 m/s0 20 40 60 80 1000.00.40.81.2C(t)/C0Time (min)U=8.22x10-4 m/s?=36.98 min0 20 40 60 80 1000.00.40.81.2C(t)/C0Time (min)U=1.64x10-3 m/s?=25.60 min0 20 40 60 80 1000.00.40.81.2C(t)/C0Time (min)?=22.56 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.40.81.2C(t)/C0Time (min)?=14.32 minU=4.11x10-3 m/s190     Figure K. 2 F(t) measured for the bed of pellets, for U = 8.22E-4 to 4.11E-3 m.s-1. Tracer gas was injected with from the lower diffuser. Sampling was done at the reactor outlet. (TCD 2 Replicate 1, 2 and 3)0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=8.22x10-4 m/s?=36.93 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=1.64x10-3 m/s?=26.50 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=22.40 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=13.17 minU=4.11x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=8.22x10-4 m/s?=36.82 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=1.64x10-3 m/s?=27.16 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=22.57 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=14.26 minU=4.11x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=8.22x10-4 m/s?=36.70 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time (min)U=1.64x10-3 m/s?=26.51 min0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=22.67 minU=2.47x10-3 m/s0 20 40 60 80 1000.00.20.40.60.81.0C(t)/C0Time?=14.15 minU=4.11x10-3 m/s191  Appendix L TCD Reading During RTD Experiment    Figure L. 1 TCD#1 detected signals as a function of time [Signal amplification ratio=100; current=70mA; TCD sample flow rate = 8.4?10-7 m3.s-1].  192  Appendix M Measured and Predicted Concentration of off-gases as a Function of Time [Using the fmix concept] In order to predict the time required to reduce the concentration of the off-gas inside the silo to a specific concentration, the mixing factor (fmix) is needed. It is known that fmix=0 for plug flow situation and fmix=1 for perfect mixing situation. Depending on the velocity used in the purging experiments, an appropriate mixing factor between these two limits was used. Figure M. 1 to Figure M. 7 show the measured gas concentrations at three locations over time during the 3 purging experiments. The three locations were the head-space (G0), middle of the bulk materials (G7) and close to the silo bottom (G13). Time zero refers to the moment when purging gas was introduced to the silo, and negative time refers to the time before the test. Assuming perfect mixing within the vessel, constant temperature, and constant pressure, the mass or volumetric flow rate for the exit stream is equal to the inlet stream. Thus: V ???? ? C?U? ? CU?                     M.1 where V is the silo volume, C is the concentration of the off-gas to be diluted at time t, C0 is the concentration of off-gas in the inlet flow (which is zero in most cases), U?  is the volumetric flow rate and t is the time. Assuming a simple mixing model: n? ????	???,? ? x????	???,?n? ????	??? ? f???x??t?n? ????	?? ? ?1 ? f????x????	??,?n? ????	??              M.2 where ??  denotes the molar flow of species to be diluted and i (CO, CO2 or CH4) is the species of off-gas to be diluted,x is the fraction of the species and fmix is the mixing factor. 193  Integrating from the initial time: U? ? dt?? ? V? ????????????                     M.3         ????U? t ? V ln????????????                               M.4 Where C1 is the concentration of the off-gas inside the silo at t=0 and C2 is the concentration after time t. According to the results obtained from the gas mixing experiments, deviation from plug flow is large for low-flow velocities, and as velocity increased the deviations became smaller. The decrease in off-gas concentration in the silo over time is shown for different velocities in Figure M. 1. By increasing the gas velocity from 8.22E-4 to 4.11E-3 m.s-1, purging time to reach 0.05% CO2 decreased from 550 to 240 min.  Figure M. 1 Predicted off-gas concentrations over time during silo purging using equation 6.15 0 50 100 150 200 250 300 350 400 450 5000.00.51.01.52.0fmix =0.3fmix =0.45 fmix =0.65fmix =0.7 u=8.22E-4 ms-1 u=1.23E-3 ms-1 u=1.64E-3 ms-1 u=2.47E-3 ms-1 u=4.11E-3 ms-1CO2 Concentration (%)Time (min)fmix =0.75194  The measured concentrations of the off-gas in  Chapter 5 showed that gas concentrations did not vary much in all locations inside the silo except for some stratification that occurred during the early days of storage. Thus one line is presented for each velocity assuming equal concentrations inside the silo. Results of gas concentrations during purging are shown in Figure M. 2 to Figure M. 7. In the first purging test, gas velocity of U=1.23E-3 m.s-1 was used, and in the second and third purging tests U=1.64E-3 m.s-1 was used. The predicted curves are closest to the measured gas concentrations in the head-space of the silo. Figure M. 4 shows the measured and predicted CO2 concentrations at different elevations in the silo as a function of time during the 2nd purging experiment with U=1.64E-3m.s-1. The predicted values assuming perfect mixing underestimated the off-gas concentrations over time for the silo head-space and the middle section; however, when the degree of mixing is taken into account, the predicted and measured values were close in the head-space.    195   Figure M. 2 CO2 concentration measured and predicted at different elevation in the silo as a function time during 1st purging experiment [U=1.23E-3m.s-1]  Figure M. 3 CO concentration measured and predicted at different elevation in the silo as a function time during 1st purging experiment [U=1.23E-3 m.s-1] -40 0 40 80 120 160 200 240 280 3200.00.51.01.52.02.5     Predicted Experimental (G13) Experimental (G7) Experimental (G0)CO2 Concentration (%)Time (min)fmix=0.7-40 0 40 80 120 160 200 240 280 3200.00.51.01.52.0     Predicted    Experimental (G13) Experimental (G7) Experimental (G0)CO Concentration (%)Time (min)fmix=0.7196   Figure M. 4 CO2 concentration measured and predicted at different elevation in the silo as a function time during 2nd purging experiment [U=1.64E-3m.s-1]  Figure M. 5 CO2 concentration measured and predicted at different elevation in the silo as a function time during 3rd purging experiment [U=1.64E-3 m.s-1] -40 0 40 80 120 160 200 240 280 3200.00.51.01.52.0     Predicted (non-ideal mixing) Experimental (G13) Experimental (G7) Experimental (G0)    Predicted (Perfect mixing)CO2 Concentration (%)Time (min)fmix=0.65-40 0 40 80 120 160 200 240 280 3200.00.51.01.52.0    Predicted Experimental (G13) Experimental (G7) Experimental (G0)CO2 Concentration (%)Time (min)fmix=0.65197   Figure M. 6 CO concentration measured and predicted at different elevation in the silo as a function of time during 2nd purging experiment [U=1.64E-3 m.s-1]  Figure M. 7 CO concentration measured and predicted at different elevation in the silo as a function of time during 3rd purging experiment [U=1.64E-3 m.s-1] -40 0 40 80 120 160 200 240 280 3200.00.51.01.5    Predicted Experimental (G13) Experimental (G7) Experimental (G0)CO Concentration (%)Time (min)fmix=0.65-40 0 40 80 120 160 200 240 280 3200.00.51.0    Predicted Experimental (G13) Experimental (G7) Experimental (G0)CO Concentration (%)Time (min)fmix=0.65

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0165614/manifest

Comment

Related Items