UBC Faculty Research and Publications

Fluidization and Drying of Biomass Particles in a Pulsed Fluidized Bed with Vibration Jia, Dening; Cathary, Océane; Peng, Jianghong; Bi, Xiaotao; Lim, C. Jim; Sokhansanj, Shahab; Liu, Yuping; Wang, Ruixu; Tsutsumi, Atsushi, 1956- Oct 31, 2015

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

Item Metadata


52383-Jia_D_et_al_Fluidization_drying_biomass.pdf [ 4.17MB ]
JSON: 52383-1.0228320.json
JSON-LD: 52383-1.0228320-ld.json
RDF/XML (Pretty): 52383-1.0228320-rdf.xml
RDF/JSON: 52383-1.0228320-rdf.json
Turtle: 52383-1.0228320-turtle.txt
N-Triples: 52383-1.0228320-rdf-ntriples.txt
Original Record: 52383-1.0228320-source.json
Full Text

Full Text

1 FluidizationandDryingofBiomassParticlesinaPulsedFluidizedBedwithVibrationDeningJiaa,OcéaneCatharyb,JianghongPenga,XiaotaoBia*,C.JimLima,ShahabSokhansanja,c,YupingLiudandRuixuWangaĂŶĚƚƐƵƐŚŝdƐƵƚƐƵŵŝĚa. Department of Chemical and Biological Engineering, the University of British Columbia, 2360 EastMall, Vancouver, BC, V6T 1Z3, Canadab. The École Nationale Supérieure de Techniques Avancées, 828, Boulevard des Maréchaux,92120, Palaiseau Cedex, Francec. Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USAd. Collaborative Research Center for Energy Engineering, Institute of Industrial Science, TheUniversity of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, JapanAbstract Fluidization of biomass particles has been tested in a pulsed fluidized bed with vibration in the absence of inert bed materials. Results indicate that pulsation is capable of overcoming the cohesive nature of biomass introduced by strong interparticle forces. The gas-solid flow behavior in the pulsed fluidized bed up to 6.7 Hz of pulsation has been analyzed. As the pulsation frequency increases, regular bubble patterns are formed. Drying tests have been carried out to obtain the optimum operating condition for future pulsed fluidized bed dryer and torrefier in terms of good gas-solid contact and high heat and mass transfer rates. The effect of flow rate, temperature, pulsation frequency and vibration intensity has been investigated. While higher temperature and flow rate are favored in drying, there exists a sweet spot for pulsation frequency between 1 Hz and 2 Hz where gas-solid contact is enhanced. Both constant rate drying and falling rate drying are present in the drying.  2  I. IntroductionBiomass as a sustainable energy source has regained its popularity during the past decade, mainly attributable to the surging energy demand, supply shortage of fossil fuels, mitigation of greenhouse gases and the increasing environmental awareness among the public. The total energy supplied from biomass reached approximately 55.6×1018 J in 2013. The international trade of biomass has continued to grow considerably due to higher demand, particularly for biofuels and wood pellets. Global production and transport of wood pellets already exceeded 23.6 million tonnes in 2013, a 13% increase compared to 2012 [1]. Despite the global prosperity and popularity of biomass, one of the major issues that hinders its further application is high moisture content. Biomass from various sources generally contains high concentrations of water, which could be either free or chemically bounded. Sawdust directly obtained from sawmill contains over 60% of water, while in general it should be controlled under 13%  as high moisture levels normally lead to energy losses, higher risks of mold formation during storage and decreased efficiency and increased emission in combustion. Studies have also shown that for biomass pelletization the ideal moisture content ranges from 5% to 15% on dry basis [2].  There is no surprise that drying being an essential step for many biomass related operations such as pelletization. Research by Mani et al. [3] discovered that the cost of drying (capital and operation) alone could take up to 20% of the entire pelletization process while hammer mill and pellet mill combined share less than 10%. The energy consumption of biomass drying is approximately 2500 - 3000 MJ/tonne, whereas the grinding and pelletizing requires only 400 MJ/tonne. Therefore, in order to optimize pelletization in terms of process efficiency, energy consumption and product quality, not only it is obvious but also highly effective to focus on drying. Considering the current scale of operation for pelletization (900,000 tonnes of pellets per year from sustainably managed forests [1]), even a small improvement in drying efficiency would significantly alter the outlook.  Compared to traditional biomass dryers (rotary drum dryer, conveyer belt dryer, etc.) fluidized bed offers excellent heat and mass transfer rates, uniform temperature distribution. However, the unconventional nature of biomass particles such as irregular shape, low bulk density and wide particle distribution has made biomass fluidization quite problematic [4]. The surface moisture on biomass materials increases the inter-particle forces dramatically [5], at times bigger than the particle weight, which could lead to channeling, gas bypassing and eventually defluidization. Previous researchers introduced inert particles such as sand to the system to assist biomass particles in fluidizing [6-9]. However it increases ash content in biomass product due to attrition of inert materials, posing serious threats to biomass-powered boilers and turbines 3  that may as a result encounter corrosion, sintering, and slagging, which will require additional maintenance for the removal of deposits and even unscheduled shutdowns [10]. Therefore, fluidized bed dryers without the need of inert particles are greatly favored.  Previous study by Liu et al [11] showed the potential of fluidized bed drying of biomass without the aid of inert bed materials. Sawdust was dried in a 50 mm ID fluidized bed, the same one used by Li et al [12], acceptable gas-solid contact was achieved by a novel distributor with vertically inclined orifices designed to increase radial particle movement. Drying rate curves consisted of multiple peaks that were directly tied to the hydrodynamics, as the drying rate will drop when the occasional channeling and gas bypassing occur.  Besides relying on new distributors and higher flow rates, pulsation and vibration are commonly used in granulation, coating and drying of agricultural products [13-27]. Pulsed fluidized beds, in which the gas flow rate shifts periodically with time have been proven effective in transforming random and chaotic bubble behavior into regular and ordered patterns, improving the fluidization quality of coarse particles greatly [28]. Similar to many mechanical systems, the behavior of pulsation is strongly linked its frequency. Based on one-dimensional particle motions in incompressible fluid, Molerus interpreted the effect of vibration on the stabilization of particle fluidization, and confirmed that 50 Hz was very effective in reducing gas bypassing [29]. Similarly, Massimilla et al. examined three types of pulsed fluidization at different pulsation frequencies [30], namely the intermittent fluidization at low frequencies (1.2 ~ 2.7 Hz), piston-like fluidization at medium frequencies (2.7 ~ 4.8 Hz) as well as an apparently ‘normal’ fluidization at high frequencies (> 4.8 Hz). The piston-like behavior was not observed in experiments carried out by Wong and Baird [31]. A mathematical model was proposed offering approximate prediction of the natural frequency of the fluidized bed, around which the pulsation was most effective. DEM simulation by Wang and Rhodes [19, 32] confirmed the 3 ~ 10 Hz frequency register to be most successful in ameliorating gas-solid interaction. Periodically formed horizontal gas channels were spotted where they quickly gave rise to regular bubbles, and moved up to the surface of the bubbling bed. Vibrated fluidized beds bring mechanical vibration energy into conventional fluidized beds to improve the fluidization quality of sticky and cohesive particles and to reduce the consumption of drying medium. Gupta and Mujumdar [33] discussed the hydrodynamics and heat transfer properties in vibrated fluidized beds and derived a power-type correlation for bed pressure drop at minimum fluidization velocity between vibrated and conventional fluidized bed. The promotion of heat and mass transfer rates in vibrated bed was verified by Eccles and Mujumdar [34], where the bed-to-surface heat transfer coefficient was measured in a vertically vibrated fluidized bed. By modifying the diffusion term in the conservation equations for drying, Stakiü and Uroševiü successfully simulated drying of fine grains in both fixed and vibrated fluidized beds. Simulation results were then compared with corresponding experiments [14]. The numerical model 4  proposed was able to predict the drying performance in vibrated fluidized beds. It was also found that vibrated beds had a shorter residence time and more uniform temperature distribution.   To further the study of Liu and Li [11, 12], reduce the gas consumption during drying, and better understand the relationship between drying and hydrodynamics of the fluidized bed, a pulsed fluidized bed dryer with optional vibration capabilities was designed. Multiple parameters including flow rate, temperature, pulsation frequency and vibration intensity have been investigated. Drying rate was adopted as an indicator for dryer performance, the goal is to gauge said parameters to obtain good gas-solid contact and fast heat and mass transfer, which are crucial to the design and scale-up of biomass dryers.   II. ExperimentalSetupExperimental set-up of the Pulsed fluidized bed with vibration (PVFB) is depicted in Fig 1. The fluidized bed column was made of Plexiglas with a rectangular cross-section (15 cm ൈ10 cm, and a height of 1.0 m). It was installed on a vibrating base (Eriez 48A, Eriez Manufacture Co., USA). The vibrating base had a constant frequency of 60 Hz with adjustable amplitude. It was adjustable on a scale from 0 to 100%, corresponding to 0 to 0.381 mm of displacement. The drying medium was high-pressure building air. After being regulated by a pressure regulator (AR-40-N04H-Z, SMC, USA), the gas line branched out into two streams. Both streams were equipped with a rotameter (FL6212-V and FL6213-V respectively, Omega, Canada) with needle valve to monitor and control the flow rate individually. The first stream was the pulsed stream, on which a solenoid valve (8210G034-120/60, ASCO Valve, USA) was installed. The open/closed state of the valve was controlled by a computer program. In order to reduce the oscillation of the rotameter readings caused by pulsation, a surge tank was placed between the solenoid valve and the rotameter. The presence of the surge tank could also maintain a stable supply of drying air during operation. The second stream was the steady stream where the flow rate of air was maintained constant. Purpose of this stream was to provide a stable amount of gas to keep the bed mobilized. The two streams merged before entering the inline gas heater (AHP-7561, Omega, Canada) controlled by a temperature controller (CN4316, Omega, Canada) so that drying temperature could be regulated. A perforated aluminum plate with 1/8’’ holes and 40% open area was used as gas distributor. A filter bag was installed at the exit of the column to retain the fines.  Figure 1. Schematic drawing of the vibrated fluidized bed with pulsed gas flow Along the height of the column, there were six ports for temperature, relative humidity and pressure measurements. The first one was located 25 mm above the gas distributor, and the others were 102 mm 5  apart. Port 1 and 5 were used to obtain pressure drop across the bed, through a pressure transducer (PX164-010D5V, Omega, Canada). The absolute pressure of the windbox (PX142-005D5V, Omega, Canada) and the column (PX163-005BD5V, Omega, Canada) were also monitored. Temperature in the windbox, dense phase and freeboard were measured by T-type thermocouples. Temperature and relative humidity of the gas phase at both inlet and outlet of the fluidized bed column were constantly monitored and recorded by humidity indicators (HMI41 at inlet, HMT335 at outlet, Vaisala, Finland). The analog signals of the pressure, temperature and the relative humidity were captured by data acquisition devices, so that they could be processed, visualized, and saved through Labview software (Version 2014, National Instrument, USA) on a PC. Labview was also used to control the solenoid valve through a digital output channel. The specific devices used and sampling rate of the signals are listed in Table 1 below. A high-speed camera (1280ൈ720 pixel, f/2.2 aperture, 240 frames per second) was mounted in front of the Plexiglas column to capture the transient behavior of the fluidized bed. Table 1. Summary of Data Acquisition Device and Acquired Signals Biomass material used in this study was generously donated by Tolko Industries Ltd (Vernon, Canada), specifically Douglas fir and pine sawdust. Switchgrass was also used. The raw material came in the forms of particles with very high moisture content. They were passed through a 3.5 mm sieve to get rid of larger particles unsuitable for fluidization. Particle size distribution after sieving are shown in Fig. 2. The Douglas fir sawdust has more fines, while pine contains more coarse particles (1.7 mm ~ 3.5 mm). Other basic properties including Sauter diameter, bulk and true densities, as well as sphericity are listed in Table 2. True density was measured by a multi-pycnometer (Quantachrome Instruments Co. Florida, USA), and sphericity was measured by an image processing software developed in this group.  To investigate suitable operating parameters for biomass drying in terms of flow rate, pulsation frequency and duty cycle, vibration intensity, air temperature etc., batch runs of drying test were conducted in the PVFB. 200 g of samples were placed into the fluidized bed and dried for 30 min. Since the moisture content at the inlet and outlet were constantly monitored, combined with known flow rate, pressure and temperature, the drying rate Rw was obtained,    (Y )w out indwR M Ydt  ˜    (1) where M is the mass flow rate, Yout and Yin are the specific humidity, which is the ratio of mass of water vapor to mass of dry air, calculated by,  6   0.622( )sspYp p    (2) ps is the water vapor saturation pressure, obtained from the Wagner-Pruss equation [35],  1.5 363.5 4 7.5647.096ln( ) ( 7.85951783 1.84408259 11.786649722.064 1022.6807411 15.9618719 1.80122502 )spTQ Q QQ Q Q   u    (3) in which   1647.096TQ     (4) Figure 2. Particle Size Distribution of Bed Materials T is the air temperature in Kelvin. Besides the instantaneous drying rate that could be calculated directly in Labview with equations provided above, at the end of each drying test a small sample of biomass particles was also collected. Together with the wet particles, the moisture content of all samples were determined according to the ASTM D4442-07 standard by heating the samples at 103 °C for 24 h in a precision oven. The sample moisture content (dry basis, w% d.b.) is calculated by  . 100%wet dryd bdrym mXm u   (5) where mwet and mdry are the weight of moisture of the sawdust samples before and after drying. Final moisture content of biomass samples are normalized so that results from different experiments can be better compared. The normalized final moisture content Ȥ is defined as the ratio of final moisture content to initial moisture content,  100%oXXF  u   (6) with Xo  being the initial moisture content in a sample. Similarly, the drying efficiency Ș, or the efficiency of water removal could be defined as,  100%ooX XXK  u   (7) 7  III. ExperimentalResults1. Hydrodynamicsinfluidizedbeddryer[Pressure Fluctuations] In the case of pulsed fluidized bed with biomass particles, the relatively chaotic nature of the bed is structured by the pulsed gas flow as external excitation [28]. Distinct ‘Flow-On’ and ‘Flow-Off’ period could be observed at 0.33 Hz. As can be seen from Fig. 3(a) each ‘On’ period starts with a dominant peak of bed pressure drop (marked with letter A), corresponding to the onset of gas flow immediately after the solenoid valve was switched open. The major peak is followed by a decaying oscillation (B - C). The uprising gas bubbles mixing with falling particles may have contributed to the oscillatory pressure profile from point B to C. At the end of each ‘On’ period there is also a brief pressure recovery (C – D) caused by falling particles compressing the gas phase in the plenum chamber. The bed remained stationary during ‘Off’ period due to the lack of gas supply, the pressure profile flatlines (D - E).  Similar pattern could be seen at 1 Hz in Fig. 3(b), with F representing the dominant peak and G – H the pressure oscillation. It should be noted that at lower pulsation frequencies, the intensity of the pulsation is greater, which leads to longer period of decaying oscillations. Five or more peaks can be identified during the decaying oscillation at 0.33Hz, while only one is observed at the relatively low-intensity 3 Hz. As the pulsation frequency keeps increasing, particles will be carried up by gas phase from the next cycle before reaching the bottom of the fluidized bed. Consequently, the pressure goes up before it could start oscillating, which explains the disappearance of decaying oscillation altogether at 6.67 Hz, as illustrated in Fig. 3(d).  Figure 3. Pressure Fluctuations in PVFB at Various Pulsation Frequencies  [Gas-solid Behavior] Typical flow behavior of the pulsed fluidized bed at various pulsation frequencies are shown in Fig. 4. Average gas velocities in all experiments were kept constant and high enough to avoid severe channeling and gas bypassing. At the lower end of pulsation frequency investigated in this work such as 0.33 Hz, the bed was fluidized intermittently due to prolonged ‘On’ and ‘Off’ period in each pulsation cycle. The bed remained stationary during the ‘Off’ period while sufficiently high pressure has built up in the surge tank. As soon as the solenoid valve was opened, pressurized gas stored in the surge tank would be quickly discharged into the fluidized bed. Chains of slugs were formed above the distributor, as can been in the yellow rectangle of Fig. 4 (A1). The frequency of slug formation is identical to the pulsation frequency. 8  Slugs quickly rose to the top of the bed via a few channels (A2), while the remaining gas in this cycle kept mixing with biomass particles (A3). The high intensity of gas pulsation at 0.33 Hz was powerful enough to lift the entire bed, therefore the amplitude of bed expansion reached maximum with heavy elucidation. A large amount of fines was collected in the bag filter. As the pressure of the gas phase diminished and solenoid valve closed, particles eventually fell back to the bed (A4), which remained stationary until the next cycle.  The interval between two gas injections became shorter at 1 Hz. The horizontal layer of slugs (B1) quickly evolved into gas channels that zigzagged up the bed, as is shown in Fig. 4 (B2) and (B3). The bed expanded slightly less due to reduced pressure build-up and gas velocity. No regular bubble pattern could be seen at this frequency due to the cohesiveness of particles and high intensity of pulsation. However, at 3 Hz the horizontal slugs (C1) quickly gave rise to large bubbles (circled in Fig. 4 C2 and C3). The ‘Off’ period disappeared as the pulsation frequency exceeded the natural frequency of the fluidized bed, particles were lifted up again before it could reach the bottom of the bed. This is consistent with the disappearance of decaying pressure oscillation shown in Fig. 3 (c). At 6.67 Hz, the peak instantaneous gas velocity in each pulsation cycle is much smaller than that of 0.33 Hz. There will be more resistance from particles to form bubbles. The combined effect of faster gas renewal and less excessive gas available to form bubbles led to the reduction in bubble size and bed expansion. Shown in Fig. 4 (D1), slugs formed at 6.67 Hz quickly disappeared and split into regular bubbles (circled in Fig. 4 D2, D3). The pulsed fluidized bed at this frequency behaves like a conventional fluidized bed if the biomass particles were able to be fluidized under similar conditions. Figure 4. Typical Fluid Behavior of the Bed at Various Pulsation Frequencies, A1-A4 Pulsation frequency = 0.33Hz, B1-B4 1Hz, C1-C4 3Hz, D1-D4 6.67Hz Fast Fourier Transform (FFT) were also performed on the pressure signals. Similar to a conventional fluidized bed where the power spectrum of pressure drops reflects bubble frequency, including bubble formation, splitting and coalescence, the dominant frequency in pulsed fluidized bed corresponds to the pulsation frequency, as shown in Fig. 6. Peaks are more distinct for the pulsed fluidized bed, and several orders of magnitude higher than that of a conventional fluidized bed. The appearance of overtones of the dominant frequency could be attributed to the decaying pressure oscillation, and the fact the pressure fluctuations in pulsed fluidized bed are not perfect sinusoids.  [Effect of Moisture Content on Hydrodynamics] The hydrodynamics of the PVFB ties closely to the moisture content of biomass particles. Water will form bridges between particles, introducing capillary liquid bridge force into the system, which is consisted of 9  two major components, the surface tension force and the force caused by the pressure difference outside and inside the bridge. As the particles are being dried in the system, moisture is consistently being removed. The liquid bridge force will change as a result, which affects the hydrodynamics of the fluidized bed dramatically. The relationship between moisture content and pressure fluctuations is especially important. Fig. 6 presents the time-averaged standard deviation of pressure fluctuations at various pulsation frequencies during drying process. Similar to conventional fluidized beds, the standard deviation of pressure fluctuations is mainly associated with bubble size. It is obvious that 0.33 Hz exhibits the highest amplitude of pressure fluctuations, and it decreases with increasing pulsation frequency. This goes along with visual observations obtained during the drying tests. At lower pulsation frequencies, larger voids and bed expansion were observed. The standard deviation also decreases during drying, indicating the bubble size reduces as the particles are being dried. This may have been caused by the breaking down of liquid bridges between particles as the thin film of water being evaporated, and the inhibition of bubbles splitting from such forces are weakened.  Figure 5. Spectrum Analysis of Pressure Fluctuations at Various Pulsation Frequencies Figure 6. Standard Deviation of Pressure Fluctuation at Various Pulsation Frequencies  2. DryingintheVibratedFluidizedBedwithPulsedGasFlow[Effect of Flow Rate] Flow rate has a direct impact on drying. Unlike conventional fluidized beds, pulsed fluidized beds could be operated at an average gas velocity lower than the minimum fluidization velocity. Successful fluidization behaver was observed at / 0.9mfU U  combined with a low pulsation frequency such as 0.33 Hz or 0.5 Hz. As the pulsation frequency increases, higher flow rate is required to maintain good gas-solid contact. 1.2Umf was found to be sufficient for pulsations of 3 ~ 6.67 Hz. Fig. 7 depicts the efficiency of fluidized bed dryer at different gas flow rates. Since 200 g of Douglas fir samples with the same initial moisture content were placed in to the fluidized bed before each run, a low final moisture content, or X/Xo would suggest faster drying and subsequently better gas-solid contact. Certain experiments were repeated three times, and good repeatability was obtained. Higher average gas velocity seems to lead to higher drying efficiency. For 0.5 Hz and 3 Hz pulsation, as the average superficial gas velocity increases from 1.5Umf to 2Umf , the normalized final moisture content Ȥ decreased by 8.4% and 13.8% respectively. In addition, it appears that for dryers that is being operated at a moderate flow rate (1.3Umf  Ud  < 1.9Umf ), higher pulsation frequency is favored. 10  In the case of very low ( < 1.2Umf ) or very high ( t 1.9Umf ) flow rates, a lower pulsation frequency is preferred. Figure 7. Effect of Flow Rate on the Final Moisture Content of Douglas Fir Sawdust after 30 min of Drying  [Effect of Pulsation Frequency] As a crucial parameter in pulsed fluidized beds, the impact of pulsation frequency on drying has been investigated and is given in Fig. 8. Due to the limitations of the solenoid valve, the maximum pulsation frequency investigated in this study was 6.67 Hz. Since the velocity profile at various pulsation frequencies are different, it is important that the average gas velocities be kept the same to ensure the comparability of the results. Two average gas velocities were studied, 0.8 mfU U  and 1.3Umf .  Figure 8. Effect of Pulsation Frequency on the Final Moisture Content of Douglas Fir Sawdust after 30 min of Drying With the relatively high gas velocity of 1.3Umf , the bed could be successfully fluidized at a wider range of pulsation frequencies compared to 0.8Umf, from 0.33 Hz to 6.67 Hz, without severe channeling or gas bypassing that could hinder the drying process. An increase in drying efficiency is observed with increasing pulsation frequency, as the final moisture content in biomass sample drops from 19.3% at 0.33 Hz to 14.5% at 6.67 Hz. For 0.8 mfU U  channeling was observed at approximately 3.5 Hz and above. Gas phase bypasses the particles and travels through the channels with minimum gas-solid contact, which explains the dramatic decrease in drying efficiency at 3.3 Hz and 5 Hz. No channeling was observed below this frequency, as the pulsation was strong enough to overcome the cohesiveness of biomass particles. However, the prolonged ‘Off’ period of 0.33 Hz and below hindered drying, causing the final moisture content in biomass samples to increase slightly, from 22.4% at 0.75 Hz to 25.3% at 0.33 Hz. This makes 1 ~ 3 Hz among the most effective frequencies for drying, a sweet spot for the operation of pulsed fluidized bed dryer under such conditions.   [Effect of Vibration Intensity] Similar drying tests were carried out to verify the contribution of vibration to the fluidization of biomass particles. Two gas velocities were used, with the amplitude of vibration set to 0%, 25%, 75%, 50% and 11  100%. Similar to previous tests, final moisture content of biomass samples after 30 min of drying was also chosen as an indicator of fluidization quality. Figure 9. Effect of Vibration Intensity on Drying, Douglas Fir Sawdust under 1 Hz Pulsation Frequency It is obvious that biomass dries faster at the higher gas velocity of 1.3 mfU U  compared to 0.8Umf, as Shown in Fig. 9. A minimum 17.8% of difference in normalized final moisture content could be seen between the two gas flow rates. At 0.8Umf, a reduction in channeling and gas bypassing was observed in the fluidized bed immediately after the vibration was turned on. Even a low vibration intensity of 25% is sufficient to break up the cohesive forces between biomass particles, consequently the final moisture content decreases by 7% compared to the case where no vibration was applied. Higher amplitudes yielded only marginal improvement. In the case of 1.3Umf where pulsation is capable of overcoming the inter-particles forces, the contribution of vibration is less obvious. Less than 2% improvement in drying efficiency suggests that pulsation alone should be enough to handle biomass particles, provided that suitable operating conditions are implemented.    [Effect of Temperature] Biomass materials, in which the initial moisture content is high and the final moisture content requirement is notably strict, drying times are normally long, and the temperature of drying medium becomes important. Series of drying test were conducted at different air temperatures. Limited by the Plexiglas column, temperatures up to 50 Ԩwere tested. The fluidized bed column was pre-heated to designated temperatures for 30 min prior to any drying tests. Despite the pulsating nature of the gas flow, temperature of the inlet air could be controlled within 0.5 Ԩ. Fig. 10 demonstrates the effect of temperature on drying of Douglas fir sawdust at various pulsation frequencies. All three curves possesses similar trends, as higher temperature does promote dying rate. One third of the experiments were repeated three times to ensure credibility of the results, the standard errors were less than 1 %. For 0.33 Hz, the final moisture content of Douglas fir samples after 30 min of drying decreased from 44.2 % to 10.4% as the temperature rises from 20 Ԩ–‘ͷͲԨǡconsequently the drying efficiency improves by 33.8%. At 1 Hz and 3 Hz, the normalized final moisture content decreased by 24% and 21% respectively. In the case of drying porous particles, temperature affects both external and internal conditions of drying. Higher temperature will speed up the evaporation of free water on the surfaces of biomass particles, which leads to faster drying in constant rate period. In the falling rate period where heat conduction from the surface to the core of a particle is important, higher gas temperature will increase the temperature gradient between particle surface and core, leading to faster 12  transfer of bound water from core to surface and evaporation at the surface. Little difference in final moisture content can be found for 1 Hz and 3 Hzpulsation at 50 Ԩ, indicating that gas velocity/external mass transfer may have become the rate-controlling step here.   Figure 10. Effect of Temperature on the Drying of Douglas Fir Sawdust at Various Pulsation Frequencies, U/Umf=1.25  3. DryingMechanismsTypical drying curves for all three biomass species including Douglas fir, pine and switchgrass are shown in Fig. 11(a). Since fresh wet particles are placed into the fluidized bed dryer prior to each run, it is inevitable that the surface moisture on the wet particles evaporate and saturate the freeboard. Therefore, the decline of drying rate during the first 5 min of drying is most likely caused by incoming dry air diluting the humid air in the freeboard. The constant rate drying is finally revealed after the initial system response, with the relative humidity of the exhaust gas remaining constant. It is represented by the horizontal section on the drying curve. At this stage, a thin film of water covers the surface of particles, on which the evaporation mainly takes place. The rate of drying is limited by the evaporation of water, which could be characterized as a heat-transfer limited process, only to be influenced by external conditions such as temperature, humidity and flow rate of the drying air.  Figure 11. Typical Drying Curves of Three Biomass Species (a) Drying Rate vs. Time (b) Bed Temperature vs. Time Bed temperature remains constant in during constant rate drying. However, as drying proceeds, dry spots appear on the surface of particles once critical moisture content is reached. Although drying rate per unit wet area remains the same, a reduction in wet areas causes the drying rate to decline. As the surface of the particles completely dries out, surface temperature increases dramatically. Drying becomes a mass-transfer limited process that is mainly driven by the concentration gradient between the core and surface of the particles. Temperature profiles during drying are illustrated in Fig. 11(b), in which three distinct areas could be seen, representing two drying mechanisms. The initial temperature drop is most likely caused by the higher temperature of the sample (room temperature, 25 Ԩ) than the drying air (20 Ԩ), and heat taken away by evaporation.     13  IV. DiscussionAs proven in the previous section, one of the advantages of pulsed fluidized bed is that it could be operated below minimum fluidization velocity with reasonable heat and mass transfer rates to carry out biomass drying, which could benefit the process economically by reducing the consumption of drying medium. Demonstrated in Fig. 8 that operating the fluidized bed at 0.8 mfU U  with 0.75 Hz pulsation, the remaining moisture content in the biomass sample after 30 min of drying is only 5% higher than of 1.3Umf  with the same frequency. Further increase or decrease of pulsation frequency actually reduces the drying efficiency at 0.8 mfU U , while for 1.3 mfU U the drying efficiency monotonically increases. A possible explanation is provided in Fig. 12, which demonstrates two velocity profiles in the pulsed fluidized bed, mfU U  and mfU U! , corresponding to the velocities used in the drying experiments. Though theoretically the velocity profile should be square waves, in reality due to the actuation time of the solenoid valve (less than 1 second) they are closer to sinusoidal waves. The gas phase is only capable of lifting the biomass particles when it possesses enough momentum, i.e. when instantaneous gas velocity is higher than minimum fluidization velocity ( mfU U! ). Fortunately, in each pulsation cycle there should be a portion of time that satisfies this condition. These portions are marked as ‘fluidized zones’ in Fig. 12, the duration of which depend on the pulsation frequency and the average flow rate. Fig. 12 (a) represents a typical scenario wheremfU U! . Of the two frequencies portrayed in this figure, 1.5 Hz has a larger portion of gas velocity under Umf in each cycle compared to 4.5 Hz, which suggests that increasing the pulsation frequency from 1.5 Hz to 4.5 Hz will yield a larger fluidized zone, i.e. a larger portion of time in each pulsation cycle where the gas phase is strong enough to fluidize the particles. This explains the increasing water removal rate with increasing pulsation frequency at 1.3 mfU U . Situation is revered for mfU U  when most of the gas flow in a pulsation cycle is too weak to overcome the cohesive forces in the fluidized bed, leaving most of the drying and gas-solid mixing done in the limited fluidized zones. A higher pulsation frequency simply becomes weaker, with less portions of gas velocity exceeding Umf in each cycle, which is consistently with the fact that only at lower pulsation frequencies could the bed be effectively fluidized at this flow rate. On the other side of the spectrum however, an extremely low pulsation frequency bring no benefits either. For instance, 0.33 Hz will leave the bed stationary for 3 seconds, providing enough time for the liquid bridges 14  to be formed again. This indicates that frequencies too high or too low are equally detrimental to the drying of biomass when mfU U . Based on the experimental results it seems that gas pulsation between 1 Hz and 3 Hz is most effective in improving the fluidization quality during biomass drying as the water removal efficiency is the highest among the frequencies investigated at a decent flow rate shown in Fig. 8. Interestingly this frequency register is close to the natural frequency of the system. Moreover, it has been adequately proven gas-solid fluidized beds could be characterized as a second-order mechanical vibration system, in which the pressure waves represent the output response of such a dynamic system to an external excitation [36], and the undamped natural frequency fn could be expressed as,  12n mfgfHS   (8) where Hmf is the steady bed height of the fluidized bed. However due to particle wall friction and interparticle forces, the fluidized bed is mostly a damped system, specifically an underdamped system in which the damping factor ȗ  is less than 1. The natural frequency could then be rewritten as,  2' 12n mfgfH]S    (9) Simple calculation revealed that the natural frequency of the system studied in this work being  f = 1.1 Hz ~ 2.6 Hz with damping factor reported in the literature of being 0.3. It should also be noted that the dominant frequency of the fluidized bed that is normally associated with the input excitation source such as bubble eruption, may not be the same as the natural frequency of the system. In the case of fluidized bed with pulsed gas flow, as shown in the spectrum analysis in Fig. 5, dominant frequency is the same as the pulsation frequency. As the dominant frequency of the system approaches its natural frequency, resonance will occur in the fluidized bed just like in a mechanical system. Higher amplitude of pressure fluctuation as well as improved gas-solid have been observed, which is one of the causes for the increase in drying efficiency.    15  V. ConclusionPulsed fluidized bed with optional assistance of vibration has proven to be successful in the fluidization of various biomass materials such as Douglas fir, pine and switchgrass. The biomass particles fluidized and dried in the system remained ‘as received’ except for a simple sieving process to eliminate extremely large branches and barks, proven pulsed fluidized bed a versatile and highly potential alternative to conventional biomass dryers and torrifiers. The gas-solid flow behavior at different pulsation frequencies were analyzed. By increasing the pulsation frequency, the intensity of the pulsation becomes weaker with less entrainment and bed expansion. The fluidized bed is only fluidized intermittently below 3 Hz with evident ‘On’ and ‘Off’ periods. Each ‘On’ period commences with a horizontal layer of slugs formed above the gas distributor that quickly give rise to gas channels when pulsation frequency is below 3 Hz. Large bubbles are formed at 3 Hz instead of channels, above which frequency regular bubble patterns could be observed. At even higher pulsation frequencies, the pulsed fluidized bed behaves similar to a conventional one. In order to find the optimum operating condition for future biomass drying and torrefaction reactors with the aim of reaching good gas-solid contact, high heat and mass transfer rates, various drying tests were performed. Results indicate that temperature and gas velocity played a significant role in biomass drying, as they both are able to boost the mass transfer rate from surface of the particles, to the drying medium. Higher temperature and flow rates usually lead to faster drying. Pulsation frequency on the other hand, being an external excitation to the fluidized bed that structures the chaotic motion of the fluidized bed into an ordered system, its influence on the hydrodynamics ties closely with the natural frequency, as well as the flow regime. Since within each pulsation cycle the useful part of the pulsed gas flow that could carry biomass particles is the one in which instantaneous gas velocity is higher than the minimum fluidization velocity, or the ‘fluidized zone’. This fluidized zone in each cycle becomes smaller with increasing pulsation frequency and decreasing average flow rate, which generated the interesting results in Fig. 8 and elucidated in Fig. 12. However, pulsation frequency between 1 Hz and 3 Hz with the average gas velocity right below minimum fluidization velocity seems to be a good spot, offering comparable drying performance with drying test conducted at much higher gas flow rates. Vibration was found to be a useful in cohesive particle fluidization, especially when the extra acceleration brought by pulsation is not adequate to break up the interparticle forces. However, as pulsation is almost always sufficient in tackling the cohesive nature of biomass particles investigated, vibration should only be considered as an optional add-on. Drying characteristics during the 30 minutes drying process was also studied. Two major mechanisms were observed during the drying of all three species of biomass, namely the constant rate drying and falling rate 16  drying. However due to the placement of the humidity indicator, a system delay contributed to the decline in the drying curve prior to the constant rate drying section. Further investigation of the relationships between certain parameters during drying, for example drying rate vs. bed pressure fluctuation, or bed temperature vs. average cycle frequency is crucial, as it furthers the knowledge of the intricate relationships between hydrodynamics and drying.   17  Nomenclature nf   Natural frequency of the fluidized bed, Hz 'nf   Modified natural frequency of the fluidized bed, Hz g Gravitational acceleration, m/s2 Hmf Static bed height, m mwet Water content in wet biomass samples, g  mdry Water content in dried biomass samples, g  M Mass flow rate of drying air, g/s p Water vapor pressure, Pa ps Saturated water vapor pressure, Pa Rw Drying rate, g/s t Time, s  T Temperature, K U Instantaneous gas velocity, m/s U  Average gas velocity, m/s mfU   Minimum fluidization velocity, m/s w Weight of the sample, g X Moisture content of biomass samples, dry basis Xo Initial moisture content of biomass samples, dry basis Y Specific humidity, g-water/kg-air Greek Letters Ș Drying efficiency Ȟ Simplified term, K-1 ȟ  Damping factor  Ȥ Normalized final moisture content  18  References  [1] R. Secretariat, Renewables Global Status Report, REN21, Paris, 2014. [2] R. Renström, J. Berghel, Drying of Sawdust in an Atmospheric Pressure Spouted Bed Steam Dryer, Drying Technology, 20 (2002) 449. [3] S. Mani, S. Sokhansanj, X. Bi, A. Turhollow, Economics of Producing Fuel Pellets from Biomass, Applied Engineering in Agriculture, 22 (2006). [4] R. Legros, C. Alan Millington, R. Clift, Drying of tobacco particles in a mobilised bed, Drying Technology, 12 (1994) 517-543. [5] J.P.K. Seville, C.D. Willett, P.C. Knight, Interparticle forces in fluidisation: a review, Powder Technology, 113 (2000) 261-268. [6] S.M. Tasirin, I. Puspasari, A.W. Lun, P.V. Chai, W.T. Lee, Drying of kaffir lime leaves in a fluidized bed dryer with inert particles: Kinetics and quality determination, Industrial Crops and Products, 61 (2014) 193-201. [7] M.J.C. van der Stelt, H. Gerhauser, J.H.A. Kiel, K.J. Ptasinski, Biomass upgrading by torrefaction for the production of biofuels: A review, Biomass and Bioenergy, 35 (2011) 3748-3762. [8] H.J. Ciro-Velásquez, R.L. Cunha, F.C. Menegalli, Drying of Xanthan Gum Using a Two-Dimensional Spouted Fluidized Bed (2DSFB) with Inert Particles: Performance and Rheological Considerations, Drying Technology, 28 (2010) 389-401. [9] R. Moreno, G. Antolín, A. Reyes, Quality of Fluidisation for the Drying of Forestry Biomass Particles in a Fluidised Bed, Biosystems Engineering, 94 (2006) 47-56. [10] J.H. Turnbull, Use of biomass in electric power generation: the california experience, Biomass and Bioenergy, 4 (1993) 75-84. [11] Y. Liu, J. Peng, Y. Kansha, M. Ishizuka, A. Tsutsumi, D. Jia, X.T. Bi, C.J. Lim, S. Sokhansanj, Novel fluidized bed dryer for biomass drying, Fuel Processing Technology, 122 (2014) 170-175. [12] H. Li, X. Liu, R. Legros, X.T. Bi, C.J. Lim, S. Sokhansanj, Torrefaction of sawdust in a fluidized bed reactor, Bioresource technology, 103 (2012) 453-458. [13] C. Liu, L. Wang, P. Wu, F. Xiang, Size distribution in gas vibration bed and its application on grain drying, Powder Technology, 221 (2012) 192-198. [14] M. Stakiü, T. Uroševiü, Experimental study and simulation of vibrated fluidized bed drying, Chemical Engineering and Processing: Process Intensification, 50 (2011) 428-437. [15] F.C. Godoi, N.R. Pereira, S.C.S. Rocha, Analysis of the drying process of a biopolymer (poly-hydroxybutyrate) in rotating-pulsed fluidized bed, Chemical Engineering and Processing: Process Intensification, 50 (2011) 623-629. [16] L. Meili, R.V. Daleffe, M.C. Ferreira, J.T. Freire, Analysis of the Influence of Dimensionless Vibration Number on the Drying of Pastes in Vibrofluidized Beds, Drying Technology, 28 (2010) 402-411. [17] L.F.G. de Souza, M. Nitz, P.A. Lima, O.P. Taranto, Drying of Sodium Acetate in a Pulsed Fluid Bed Dryer, Chemical Engineering & Technology, 33 (2010) 2015-2020. [18] V. Bubnovich, C. Villarreal, A. Reyes, Computer Simulation of the Drying of Seeds and Vegetables in a Discontinuous Fluidized Bed, Numerical Heat Transfer: Part A -- Applications, 54 (2008) 255-278. [19] A. Akhavan, J.R. van Ommen, J. Nijenhuis, X.S. Wang, M.-O. Coppens, M.J. Rhodes, Improved Drying in a Pulsation-Assisted Fluidized Bed, Industrial & Engineering Chemistry Research, 48 (2008) 302-309. [20] A. Reyes, P. Moyano, J. Paz, Drying of Potato Slices in a Pulsed Fluidized Bed, Drying Technology, 25 (2007) 581-590. [21] M. Nitz, O.P. Taranto, Drying of beans in a pulsed fluid bed dryer: Drying kinetics, fluid-dynamic study and comparisons with conventional fluidization, Journal of Food Engineering, 80 (2007) 249-256. [22] A. Reyes, C. Campos, R. Vega, Drying of Turnip Seeds with Microwaves in Fixed and Pulsed Fluidized Beds, Drying Technology, 24 (2006) 1469-1480. [23] R.V. Daleffe, M.C. Ferreira, J.T. Freire, Drying of Pastes in Vibro-Fluidized Beds: Effects of the Amplitude and Frequency of Vibration, Drying Technology, 23 (2005) 1765-1781. 19  [24] Z. Li, N. Kobayashi, S. Deguchi, M. Hasatani, Investigation on the Drying Kinetics in a Pulsed Fluidized Bed, JOURNAL OF CHEMICAL ENGINEERING OF JAPAN, 37 (2004) 1179-1182. [25] T. Kudra, Z. Gawrzynski, R. Glaser, J. Stanislawski, M. Poirier, DRYING OF PULP AND PAPER SLUDGE IN A PULSED FLUID BED DRYER, Drying Technology, 20 (2002) 917-933. [26] Z. Gawrzynski, R. Glaser, Drying in a Pulsed-Fluid Bed with Relocated Gas Stream, Drying Technology, 14 (1996) 1121-1172. [27] M. Hasatani, Y. Itaya, K. Miura, DRYING OF GRANULAR MATERIALS IN AN INCLINED VIBRATED FLUIDIZED BED BY COMBINED RADIATIVE AND ONVECTIVE HEATING, Drying Technology, 9 (1991) 349-366. [28] M.-O. Coppens, J.R. van Ommen, Structuring chaotic fluidized beds, Chemical Engineering Journal, 96 (2003) 117-124. [29] A.A.H. Drinkenburg, Proceedings of the International Symposium on Fluidization, June 6-9, 1967, Eindhoven, Netherlands University Press, 1967. [30] G.V. L. Massimilla, G. Raso, A study on pulsating gas fluidization of beds of particles, in:  Chemical Engineering Progress Symposium Series, American Institute of Chemical Engineers, 1966, pp. 63. [31] H.W. Wong, M.H.I. Baird, Fluidisation in a pulsed gas flow, The Chemical Engineering Journal, 2 (1971) 104-113. [32] X.S. Wang, M.J. Rhodes, Pulsed fluidization—a DEM study of a fascinating phenomenon, Powder Technology, 159 (2005) 142-149. [33] R. Gupta, A.S. Mujumdar, Aerodynamics of a vibrated fluid bed, The Canadian Journal of Chemical Engineering, 58 (1980) 332-338. [34] E.R.A. Eccles, A.S. Mujumdar, CYLINDER-TO-BED HEAT TRANSFER IN AERATED VIBRATED BEDS OF SMALL PARTICLES, Drying Technology, 10 (1992) 139-164. [35] W. Wagner, A. Pruss, International Equations for the Saturation Properties of Ordinary Water Substance. Revised According to the International Temperature Scale of 1990. Addendum to J. Phys. Chem. Ref. Data 16, 893 (1987), Journal of Physical and Chemical Reference Data, 22 (1993) 783-787. [36] B. Hao, H.T. Bi, Forced bed mass oscillations in gas–solid fluidized beds, Powder Technology, 149 (2005) 51-60. `V2 V3V4FL1 FL2VS1HighFlowBuildingAirTCR2HT1VibratorFilteredGasOutletV5V1 BallvalveV2,V3 NeedlevalveV4 SolenoidvalveV5 ThreeͲwayvalvePR1 PressureregulatorFL1FL2 RotameterVS1 SurgetankTC TemperaturecontrollerP1ͲP3 PressuretransducerT1ͲT3 TͲtypethermocoupleHT1 InlinegasheaterR1,R2 HumidityandtemperatureprobeV1PR1P1DAQR1T3T2T1P3P2123456CameraFigure 1. Schematic drawing of the vibrated fluidized bed with pulsed gas flow   Figure 2. Particle Size Distribution of Bed Materials   Figure 3. Pressure Fluctuations in PVFB at Various Pulsation Frequencies   Figure 4. Typical Fluid Behavior of the Bed at Various Pulsation Frequencies, A1-A4 Pulsation frequency = 0.33Hz, B1-B4 1Hz, C1-C4 3Hz, D1-D4 6.67Hz   Figure 5. Spectrum Analysis of Pressure Fluctuations at Various Pulsation Frequencies   Figure 6. Standard Deviation of Pressure Fluctuation at Various Pulsation Frequencies    Figure 7. Effect of Flow Rate on the Final Moisture Content of Douglas Fir Sawdust after 30 min of Drying    Figure 8. Effect of Pulsation Frequency on the Final Moisture Content of Douglas Fir Sawdust after 30 min of Drying   Figure 9. Effect of Vibration Intensity on Drying, Douglas Fir Sawdust under 1 Hz Pulsation Frequency    Figure 10. Effect of Temperature on the Drying of Douglas Fir Sawdust at Various Pulsation Frequencies, U/Umf=1.25    Figure 11. Typical Drying Curves of Three Biomass Species (a) Drying Rate vs. Time (b) Bed Temperature vs. Time    Figure 12. Schematic Drawing of the Pulsations at Different Frequencies and Flow Rates    Table 1. Summary of Data Acquisition Device and Acquired Signals SignalName Sensor/Actuator DAQModel SamplingRatePressure PressureTransducer MCDASͲ08PCI 400HzTemperature TͲtypethermocouple NI9214incDAQͲ9171 20HzRelativehumiditywithtemperature Vaisalahumidityprobes NIUSBͲ6008 10HzDigitalControl SolenoidValve NIUSBͲ6009 100Hz    Table 2. Basic Properties of Biomass Particles BiomassSpecies BulkDensity(xx%,d.b),kg/m3 TrueDensity,kg/m3 SauterDiameter,mm SphericityDouglasͲfir 164 1375 1.449 0.42Pine 139 1242 1.469 0.43Switchgrass 184 1446 0.755 0.65 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items