UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An experimental investigation of thermal effects on the axial resistance to relative ground movement… Huber, Michael 2014

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

Item Metadata


24-ubc_2014_september_huber_michael.pdf [ 24.4MB ]
JSON: 24-1.0167554.json
JSON-LD: 24-1.0167554-ld.json
RDF/XML (Pretty): 24-1.0167554-rdf.xml
RDF/JSON: 24-1.0167554-rdf.json
Turtle: 24-1.0167554-turtle.txt
N-Triples: 24-1.0167554-rdf-ntriples.txt
Original Record: 24-1.0167554-source.json
Full Text

Full Text

An experimental investigation of thermal effects on theaxial resistance to relative ground movement of burieddistrict heating pipesbyMichael HuberBSc, University of Innsbruck, 2012A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMaster of Applied ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Civil Engineering)The University of British Columbia(Vancouver)July 2014© Michael Huber, 2014AbstractDistrict heating (DH) systems are commonly used in urban areas to distributethermal energy from central heat sources. Buried pipes, with a composite cross-sectional construction, are used to transport a heated medium, usually water. Thesepipes expand and contract radially and axially due to changing water temperatures,invoking soil-pipe interaction situations during operation, and potentially leadingto significant pipeline material strains. Measures to account for these soil-pipeinteractions are an important consideration and a significant cost factor when de-signing and installing robust and cost-effective DH pipe systems.A series of full-scale tests were undertaken to provide experimental data on theaxial and lateral soil resistance of DH pipes. An existing soil chamber that is partof the Advanced Soil Pipe Interaction Research™ (ASPIRe™) facility at the TheUniversity of British Columbia (UBC) was adapted to test full-size water-filledpipes. As a part of this project, a heating system was developed specifically toapply different heating histories to the water mass before the pipe is pulled. Straingauges were mounted on the pipe at the soil interface to contribute to understandingthe mechanisms involved in soil-pipe interaction.It was shown that changes in the temperature of the water mass have a signifi-iicant influence on axial pullout resistance of the DH pipe. After heating the watermass by ∆T = 50 °C, large-strain resistance increased by roughly 15 % comparedto the control tests. Three full cooling and heating cycles reduced the axial soilresistance of the pipe, potentially due to an arching mechanism in the soil.Considerable strain was measured at the soil-pipe interface both in axial andradial direction during heating of the water mass. Based on the development ofstrain with the heating history, it was inferred that the expansions at the pipe sur-face result from a combination of strains from both the steel pipe at the core andthe high-density polyethylene (HDPE) cover. Consequently, DH pipes have to betreated as a complete system in combination with the surrounding soil mass inorder to accurately model their mechanical behaviour under thermal load.iiiPrefaceA version of parts of Introduction and Chapters 3, 4 and 5 were accepted for pub-lication. [Huber, M. and Wijewickreme D.: Response of Buried District HeatingPipelines Under Relative Axial Movements. 10th International Pipeline Confer-ence 2014 in Calgary, Alberta] [Huber, M. and Wijewickreme D.: Thermal influ-ence on axial pullout resistance of buried district heating pipes. 14th InternationalSymposium on District Heating and Cooling in Stockholm, Sweden] I was thelead investigator for this project and I am responsible for all major areas of con-cept formation, data collection and analysis, as well as manuscript composition.Dr. Dharma Wijewickreme was the supervisory author on this project and wasinvolved throughout the project in concept formation and manuscript edits to theaforementioned publications.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixGlossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Background to district heating (DH) . . . . . . . . . . . . . . . . 11.1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Purpose of the research project . . . . . . . . . . . . . . . . . . . 41.3 Organization of the thesis . . . . . . . . . . . . . . . . . . . . . . 42 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6v2.1 General pipe behaviour . . . . . . . . . . . . . . . . . . . . . . . 72.1.1 Soil force development during axial pipe movement . . . 72.1.2 Soil force development during lateral pipe movement . . . 112.2 HDPE-soil interface friction . . . . . . . . . . . . . . . . . . . . 122.3 Soil-pipe interaction of DH pipes . . . . . . . . . . . . . . . . . . 142.3.1 Early analytical work . . . . . . . . . . . . . . . . . . . . 142.3.2 Experimental and numerical investigation of cyclic axialdisplacement . . . . . . . . . . . . . . . . . . . . . . . . 162.4 Thermal conductivity of sand . . . . . . . . . . . . . . . . . . . . 192.5 Research needs identified from the literature . . . . . . . . . . . . 203 Experimental Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . 223.1 Full-scale tests in soil chamber . . . . . . . . . . . . . . . . . . . 223.1.1 Equipment . . . . . . . . . . . . . . . . . . . . . . . . . 233.1.2 Properties of materials . . . . . . . . . . . . . . . . . . . 343.1.3 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 383.2 Direct shear testing for interface friction . . . . . . . . . . . . . . 453.2.1 Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . 453.2.2 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 453.2.3 Testing program . . . . . . . . . . . . . . . . . . . . . . 463.2.4 Test results . . . . . . . . . . . . . . . . . . . . . . . . . 473.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 504.1 Soil-pipe interaction testing program . . . . . . . . . . . . . . . . 504.2 Axial pullout tests . . . . . . . . . . . . . . . . . . . . . . . . . . 52vi4.2.1 Baseline and repeatability tests . . . . . . . . . . . . . . . 534.2.2 Tests with increased water temperature inside the pipe . . 554.2.3 Tests with water mass inside the pipe subjected to full ther-mal cycles . . . . . . . . . . . . . . . . . . . . . . . . . . 594.2.4 Ageing . . . . . . . . . . . . . . . . . . . . . . . . . . . 624.2.5 Strain measurements . . . . . . . . . . . . . . . . . . . . 644.2.6 Comparison with previous work . . . . . . . . . . . . . . 714.2.7 Comparison with interface friction from direct shear test . 754.3 Lateral pullout tests . . . . . . . . . . . . . . . . . . . . . . . . . 774.3.1 Test results . . . . . . . . . . . . . . . . . . . . . . . . . 774.3.2 Comparison with previous work . . . . . . . . . . . . . . 794.3.3 Visual observations . . . . . . . . . . . . . . . . . . . . . 805 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . 865.1 Contributions to experimental equipment . . . . . . . . . . . . . . 875.2 Axial pullout tests . . . . . . . . . . . . . . . . . . . . . . . . . . 885.3 Lateral pullout tests . . . . . . . . . . . . . . . . . . . . . . . . . 905.4 Direct shear tests for interface friction . . . . . . . . . . . . . . . 905.5 Recommendations for future research . . . . . . . . . . . . . . . 90Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92A Temperature Histories of the Water Mass . . . . . . . . . . . . . . . 96viiList of TablesTable 3.1 Layers of pre-insulated 8” pipes made by LOGSTOR A/S usedin this project . . . . . . . . . . . . . . . . . . . . . . . . . . 36Table 3.2 Layers of pre-insulated 16” pipes made by LOGSTOR A/S usedin this project . . . . . . . . . . . . . . . . . . . . . . . . . . 36Table 3.3 Moisture content measurements from axial tests . . . . . . . . 39Table 3.4 Density measurements from axial tests . . . . . . . . . . . . . 40Table 3.5 Moisture content measurements from lateral tests . . . . . . . 40Table 3.6 Density measurements from lateral tests . . . . . . . . . . . . 40Table 3.7 Results of direct shear tests between HDPE pipe and FraserRiver sand . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48Table 4.1 Testing schedule for full-scale tests in soil chamber . . . . . . 53viiiList of FiguresFigure 2.1 Results of pressure sensor measurements for cyclic axial pull-out tests on pipes with a diameter of 140 mm, a soil cover ofthree diameters and a medium-dense sand; from Weidlich andAchmus (2006). . . . . . . . . . . . . . . . . . . . . . . . . . 18Figure 2.2 Thermal conductivity as a function of volumetric water contentfor a sand with an average grain size of 0.297 to 0.420 mmunder dense conditions, after Smits et al. (2010). . . . . . . . 20Figure 3.1 Overview of the Advanced Soil Pipe Interaction Research™(ASPIRe™) soil chamber. (1) A DH pipe with a diameter of520 mm is buried in the chamber, (2) the red conveyor beltused for removing soil is located on the right, (3) an emptybulk bag beside it in white. (4) another DH pipe with an outerdiameter of 315 mm is stored on the left side. . . . . . . . . . 24Figure 3.2 Layouts of the flexible divider wall. Divider wall in red, pipein yellow, areas filled with soil in brown, pulling direction in-dicated with blue arrows. . . . . . . . . . . . . . . . . . . . . 25ixFigure 3.3 Schematic layout of the loading mechanism used in axial pull-out tests including hydraulic actuators, load cells and pullingcables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Figure 3.4 Schematic layout of the loading mechanism used in lateralpullout tests including hydraulic actuators, load cells and pullingcables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Figure 3.5 Schematic layout of the heating system. (1) Welded steel endcap (2) Removable aluminium end cap (3) Hose connection foroutflow from pipe (4) Water tank (5) Electric water heater (6)Pump (7) Propane water heater (8) Hose loops. . . . . . . . . 30Figure 3.6 Different heaters used in this project. . . . . . . . . . . . . . 31Figure 3.7 Schematic layout in longitudinal section (distances in mm) ofthermal sensors buried in the soil chamber for axial tests withheating history. . . . . . . . . . . . . . . . . . . . . . . . . . 31Figure 3.8 Schematic layout in longitudinal section with embedded crosssections (distances in mm) of strain gauges attached to theHDPE layer of the pipe for one test. . . . . . . . . . . . . . . 33Figure 3.9 Strain gauges installed at the HDPE surface of the pipe. (1)Metallic foil pattern, sensing strain (2) Stress relief loop (3)Soldering pad connecting strain gauge to cable (4) Cable todata acquisition system (5) Rubber tape protecting gauge setup(6) Cable mounted on pipe with zip tie looped through holesin HDPE layer. . . . . . . . . . . . . . . . . . . . . . . . . . 34Figure 3.10 Grain size distribution of Fraser River sand used in this study. 35xFigure 3.11 DH pipes manufactured by Logstor (Denmark) (1) ProtectiveHDPE layer (2) Polyurethane foam as insulation (3) Transportpipe made of steel (4) Steel end cap welded in (5) Restoredinsulation rings (6) Continuous insulation in original condition. 37Figure 3.12 Hole in plywood wall of soil chamber with sand level raised tobottom of pipe, before and after pipe is introduced. . . . . . . 41Figure 3.13 Sand preparation . . . . . . . . . . . . . . . . . . . . . . . . 41Figure 3.14 Material handling . . . . . . . . . . . . . . . . . . . . . . . . 42Figure 3.15 Installations at the north end of the pipe. . . . . . . . . . . . . 43Figure 3.16 Bare pipe end covered in fibreglass insulation to minimize heatloss. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Figure 3.17 Cable and shackle setup to transfer the pulling force from theactuators to the pipe. . . . . . . . . . . . . . . . . . . . . . . 44Figure 3.18 Overview picture of the direct shear device used in this project. 46Figure 3.19 Preparations for direct shear tests on soil-HDPE interface fric-tion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Figure 4.1 Results of two axial pullout tests on pipes with an outer diam-eter of 520 mm without heating, Tests No. A-520-H0-1 andA-520-H0-2. . . . . . . . . . . . . . . . . . . . . . . . . . . 54Figure 4.2 Visual comparison of the soil surface before and after the pull-out test A-520-H0-1. . . . . . . . . . . . . . . . . . . . . . . 55Figure 4.3 Soil loss at the gasket in pulling direction during test A-520-H0-1. Note: photo taken of the visible part of the pipe as itcomes out of the soil chamber. . . . . . . . . . . . . . . . . . 56xiFigure 4.4 Results of two axial pullout tests on pipes with an outer diam-eter of 520 mm with heating and pullout with hot water mass,Tests No. A-520-H1-1 and A-520-H1-2. . . . . . . . . . . . . 57Figure 4.5 Temperature history during A-520-H1-1, at soil-pipe interface,150 mm away from the pipe in the soil mass and ambient tem-perature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Figure 4.6 Temperature history during A-520-H1-2. . . . . . . . . . . . 59Figure 4.7 Comparison of average results from pullout tests without heat-ing and with pulling with hot water mass, Tests No. A-520-H0-1 and -2 as well as A-520-H1-1 and -2. . . . . . . . . . . 60Figure 4.8 Temperature history during Test No. A-520-H1C1. . . . . . . 61Figure 4.9 Temperature history during Test No. A-520-H3C3. . . . . . . 62Figure 4.10 Comparison of results from two axial tests with one (Test No.A-520-H1C1) and three (Test No. A-520-H3C3) full thermalcycles as well as results from baseline tests. . . . . . . . . . . 63Figure 4.11 Comparison of results from a test with 45 days of soil agingbefore pullout (Test No. A-520-H0-A) with results from base-line tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64Figure 4.12 Temperature history during Test No. A-520-H1C1-S. . . . . . 65Figure 4.13 Temperature history of the water mass during Test No. A-520-H1C1-S. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Figure 4.14 Axial strain history from strain gauges mounted on the DHpipe at a diameter buried in the soil chamber during Test No.A-520-H1C1-S, on top of the pipe and on the sides. . . . . . . 66xiiFigure 4.15 Radial strain history from strain gauges mounted on the DHpipe at a diameter buried in the soil chamber during Test No.A-520-H1C1-S. . . . . . . . . . . . . . . . . . . . . . . . . . 67Figure 4.16 Axial strain history from strain gauges mounted on the DHpipe at a diameter outside the soil chamber during Test No.A-520-H1C1-S. . . . . . . . . . . . . . . . . . . . . . . . . . 67Figure 4.17 Radial strain history from strain gauges mounted on the DHpipe at a diameter outside the soil chamber during Test No.A-520-H1C1-S. . . . . . . . . . . . . . . . . . . . . . . . . . 68Figure 4.18 Comparison of results from the test with strain gauges attachedafter one full thermal cycle (Test No. A-520-H1C1-S) with thetest with one cycle without gauges (Test No. A-520-H1C1) aswell as results from baseline tests. . . . . . . . . . . . . . . . 72Figure 4.19 Comparison of normalized axial pullout resistance from thisstudy with previous work on steel pipes. . . . . . . . . . . . . 74Figure 4.20 Comparison of baseline Tests No. A-520-H0-1 and A-520-H0-2 with estimations based on Equation (2.1) with δ from directshear test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76Figure 4.21 Results of two lateral pullout tests on pipes with an outer di-ameter of 315 mm, Tests No. L-315-1 and L-315-2. . . . . . . 78Figure 4.22 Results of two lateral pullout tests on pipes with an outer di-ameter of 520 mm, Tests No. L-520-1 and L-520-2. . . . . . . 79Figure 4.23 Comparison of normalized lateral resistance of pipes with anouter diameter of 315 and 520 mm, Tests No. L-315-2 andL-520-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80xiiiFigure 4.24 Comparison of normalized lateral pullout resistance from thisstudy with previous work on steel pipes . . . . . . . . . . . . 81Figure 4.25 Side view of the soil specimen during test L-315-2. . . . . . . 83Figure 4.26 Side view of the soil specimen during test L-520-2. . . . . . . 84Figure 4.27 Typical view of the soil surface behind the pipe after pullout inL-520-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Figure A.1 Temperature history of the water mass during Test No. A-520-H1-1st. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Figure A.2 Temperature history of the water mass during Test No. A-520-H1-2nd. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Figure A.3 Temperature history of the water mass during Test No. A-520-H1C1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Figure A.4 Temperature history of the water mass during Test No. A-520-H3C3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Figure A.5 Temperature history of the water mass during Test No. A-520-H1C1-S. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99xivGlossaryASPIRe™ Advanced Soil Pipe Interaction Research™. ii, ix, 4, 8, 21, 22, 24, 28,71, 87DH district heating. ii, iii, v, vi, ix, xi–xiii, 1–4, 6–8, 12–14, 16, 17, 19–22, 24,28, 34, 37, 45, 46, 48, 50, 52, 64, 66–68, 70–75, 79–81, 86–91EN European Standard. 36HDPE high-density polyethylene. iii, viii, x, xi, 4, 11–15, 17, 32–35, 37, 45–48,66, 68–70, 73, 88, 89ISO International Organization for Standardization. 37LVDT linear variable differential transformer. 45MDPE medium-density polyethylene. 4, 7, 14PE polyethylene. 13UBC The University of British Columbia. ii, xvii, 2, 4, 8, 10, 12, 21, 22, 27, 34,71, 86, 87xvAcknowledgmentsFirst of all, I would like to thank my supervisor Dr. Dharma Wijewickreme forhis manifold support that made this undertaking possible. I greatly appreciate theopportunity to carry out this project with the advice and trust you offered wheneverneeded throughout the various stages.I am very thankful for the technical support from my fellow students, most no-tably Ruslan Amarasinghe, Jeremy Groves and Santiago Quinteros. Special thanksare due to those who helped to implement the technical aspects: Doug Hudniuk,Scott Jackson, Bill Leung, Mark Rigolo, Harald Schrempp and Ken Taggart fromthe workshop of the Department of Civil Engineering as well as undergraduate stu-dents Michael Ang, Lynn Machacek and Norman Richardson. This project wouldnot have been possible without your hard work.I would like to thank The University of British Columbia (UBC) Project Ser-vices with Mike Champion as well as Andre´ Harrmann and Sam Orr for providingus with pipe material and valuable discussions.Last but certainly not least I want to thank my family for encouraging andsupporting me to follow my passion with this stay in Vancouver. I thank you andall my friends at home for being sympathetic to my adventures abroad.xviChapter 1Introduction1.1 Background to district heating (DH)1.1.1 GeneralDH systems transport thermal energy from central sources to multiple locationswithin cities or urban regions. In most cases, heat is distributed through a buriedpipe network, using two parallel pipes. One pipe transports hot water mass fromheating plants, while the other brings water with a lower temperature back from thecustomers. The heat energy can be used for space heating, hot water and industrialprocess.DH systems can be very efficient, as waste energy from different sources canbe used instead of being emitted to the atmosphere as waste. Examples includeindustrial plants, like paper mills, that generate heat in some of their processes thatcannot be used on site. Wherever electricity is generated from burning fuel, wasteheat is generated that can be fed into a DH system to increase the overall yield at1the same output of greenhouse gases. DH is a flexible system that allows changingthe source of energy without modifying hardware in every customers´ house. Newtechnologies can be included by simply adding new generators or replacing exist-ing plants. The ability to add multiple sources to a system also gives redundancyand the capability to react to different market prices. DH systems, especially lowtemperature systems, are compatible with sustainable energy sources like geother-mal or solar energy.Finally, DH systems are very simple suppliers of heat for customers. Thereis no need to maintain stoves and chimneys, or change heaters with improvingtechnologies or changing market environments. Customers only have to pay a feefor the connection to the system and the amount of energy actually consumed fromthe system. All that is required in each house is a heat meter and a heat exchangerto bring energy into the domestic heating system.After decades of widespread use in Europe, an increasing number of modernhot water based DH piping networks are currently planned and being built acrossNorth America. Local examples include the Neighbourhood Energy Utility in theSoutheast False Creek area (Vancouver) and the conversion to a new water basedDH system on the UBC Vancouver campus. This increasing demand will translateinto designs involving larger diameter pipes than what is currently used.Geotechnical engineering challengesDH pipeline systems are typically buried in the soil, with the pipelines having acomposite cross-section. An inner steel or plastic pipe transports the hot medium,with a foam insulation around the pipe to reduce heat loss. An external plasticcoating protects the foam material from mechanical damage and water.2Pipes are buried in the ground at ambient temperatures. The temperature ofthe medium can change with seasons to accommodate different optimization goals.When there is a high energy demand from the customers in a system, increasing themedium temperature allows for the transporting of more energy at the same pipediameter. Burying larger pipes might not be desirable if the peak demand is onlyneeded for a limited portion of the year, as increased pipe diameters result in highermaterial and installation costs. When energy demands in a system decrease, a lowermedium temperature lowers the temperature gradient in the insulation including thesoil mass around the pipe and, in turn, decreases heat loss. There is however a limitto how low medium temperatures can be chosen. A lower limit is in place to ensurequality of service to customers, for example to guarantee sufficient heating of waterto avoid formation of bacteria.These temperature changes pose several engineering challenges for the designof DH pipelines. Expansions and contractions of the buried pipes in both axial andradial directions are a result of these changes. This invariably leads to differentialmovements of pipelines with respect to the surrounding soil mass. As a result ofthis, significant soil-pipe interactions could occur under operating conditions, andin certain instances, potentially causing significant changes in the frictional forcesat the soil-pipe interface. Accounting for this soil-pipe interaction is an importantconsideration and a significant cost factor when designing and installing robust andcost-effective DH pipe systems. Current understanding of the soil-pipe interactionaspects with respect to DH pipe systems is extremely limited (Weidlich and Wijew-ickreme, 2012). This, combined with the increasing demand for designs calling forthe installation of larger diameter pipes than what is currently used, shows that abetter understanding of soil-pipe interaction mechanisms in DH systems is of criti-3cal interest. Full-scale experiments conducted under controlled conditions play animportant role in this regard.1.2 Purpose of the research projectConsidering the lack of present understanding, basic data on the behaviour of DHpipes under lateral and axial movement relative to the ground should be generatedto understand the soil-pipe interaction mechanisms related to this problem. In thiseffort, there is need to compare with and learn from the wide knowledge that hasso far been developed with respect to soil-pipe interaction in buried pipes madeof different materials (e.g. ASCE, 1984 and PRCI, 2009), including extensiveprevious work undertaken at the UBC ASPIRe™ facility.Specific to DH pipes is an interest in the change of axial pullout resistanceof pipes that are subjected to different heating histories, thus forming the core ofthis project. Finally, data to support the understanding of radial pipe expansionunder thermal influence and thus the aforementioned change in axial resistance isdesirable.1.3 Organization of the thesisChapter 1 introduces the topic of DH and presents an outline of the purpose of theresearch project.Chapter 2 summarizes relevant literature. Previous studies on soil-pipe inter-action in pipes made of various materials (e.g. steel, medium-density polyethylene(MDPE)), interface behaviour of HDPE and soil as well as specific work on DHpipes are presented. This chapter is meant to demonstrate the value of generatingnew experimental data on pullout behaviour of DH pipes under thermal influence.4Chapter 3 describes the experimental aspects (equipment and materials) usedin this project. Properties of pipe backfill and pipes are listed, along with infor-mation on the soil chamber, heating and data acquisition systems. Limitations arediscussed.Results of the tests are presented in Chapter 4, along with the motivation forthe chosen schedule explained. Results of the tests conducted are presented.Chapter 5 presents the conclusions with future research needs identified.5Chapter 2Literature ReviewWith the need to develop sustainable energy solutions for the heating of buildings,the option of using district heating pipeline systems is becoming increasingly pop-ular, particularly in regions having relatively cold climates.District heating (DH) pipes are composite products, made of different materialsfor each layer in accordance with its purpose.Published work on DH pipe includes diameter optimisation to reduce energyloss (e.g. Kalinci et al., 2008), optimisation of the network layout (e.g. Li et al.,2010) or adaption for low-energy systems including passive houses (e.g. Dalla Rosaet al., 2011).The variable temperature changes in buried DH pipelines under operating con-ditions invariably lead to significant differential movements with respect to thesurrounding soil mass, and in turn, cause soil-pipe interaction problems that needsto be well accounted in the engineering design of DH systems.Previous research on soil-pipe interaction for these pipes is scarce comparedwith single-material pipes used in other industries, such as the oil and gas sector.6So far, most of the research has been focused on steel pipes, and it is reasonable toconsider that the findings on steel pipes can provide a good starting point for theunderstanding of complex interactions in DH pipes. In addition, some related workon the behaviour of other pipe materials, such as medium-density polyethylene(MDPE), will also be considered in this chapter in an effort to provide a contextfor findings in this project. Research on interface friction between medium-densitypolyethylene (MDPE) and soil as well as pullout tests with steel and MDPE pipeswill be reviewed. Finally this chapter will describe the limited existing work onsoil-pipe interface behaviour of DH pipes.2.1 General pipe behaviourKnowledge of expected loads on buried pipe systems is of interest to design en-gineers to plan durable and safe pipeline networks. Formalized equations andguidelines have been developed to estimate soil forces on pipelines mainly throughknowledge developed from experimental and numerical research. Observationsmade on steel and MDPE pipes particularly with respect to the development offorces under relative axial and lateral movements are reviewed herein.2.1.1 Soil force development during axial pipe movementA traditionally used formalized approach to model axial resistance of a straightpipe buried in cohesionless soil can be found for example in ASCE (1984) andALA (2001). Using the following formula, FA, the axial resistance per unit lengthdue to normal stress on the pipe as a result of earth pressure, is modelled:FA = pi ·D ·σ ′n,av · tan δ = pi ·D ·1+K02· γ ·H · tan δ (2.1)7with D = diameter of the pipe, σ ′n,av = average normal soil loads on the pipe in atrest conditions, δ = interface friction angle between pipe and soil, K0 = coefficientof lateral earth pressure at rest, γ = effective unit weight of the soil and H = burialdepth from ground surface to spring line of the pipe. The soil pressure distributionaround the pipe is idealized for this equation. Usually, the interface friction angleis defined as:δ = f ·φ (2.2)with f = friction angle reduction coefficient, and φ = friction angle of the soil.Axial pullout resistance is also often normalized as:F ′A =FAγ ·H ·D (2.3)with FA’ = normalized axial resistance. Dimensionless displacement is accordinglydefined as:Y ′ =YD(2.4)with Y’ = normalized displacement and Y = displacement.Notable research on the soil-pipe interaction of steel and polyethylene pipes inpullout tests was carried out at Cornell University and The University of BritishColumbia (UBC). As DH pipes used in this project have a steel core, research atUBC on full-scale tests of steel pipes is also included in this subsection.A study on the pullout behaviour of steel pipes was presented in Karimian(2006); Wijewickreme et al. (2009). Using testing carried out in the newly devel-oped soil chamber as a part of the ASPIRe™ initiative (the soil chamber is de-scribed in Chapter 3), it was shown that significant differences between predicted8and measured loads using the above formula can occur due to a number of reasonsprimarily arising due to the limitations in the accuracy of the δ and K0 values used.This can be due to measurements of friction angle at different vertical stresses orvariations in soil density.Wijewickreme et al. (2009), using soil pressures measured on physical full-scale pipe specimen buried in dense sand, using pressure transducers flush-mountedon the surface of the pipe, noted that earth pressures using axial pullout can bemuch higher than those estimated using the coefficient of earth pressure at rest inEquation (2.1). The thickness of the circumferential shear zone around the pipeduring axial pullout was observed to be roughly 2 mm by careful post-shear mea-surements made on strips of coloured sand zones placed around the pipe.Two-dimensional finite-difference numerical modelling (using (alias?)) wasundertaken to investigate these observations further. The case of a buried pipe ex-periencing an increase in its diameter, by an amount equal to the expected increasein the size of the annular soil shear zone around the real pipe due to soil dilation,was specifically simulated. The magnitude of this diametrical increase was basedon independent normal dilative displacements experimentally observed in inter-face direct shear tests between sand-steel interfaces. It was found that the amountof stress increase computed from the numerical model was comparable to that ob-served from full-scale experiments, thus supporting the notion that soil dilation inthe shear zone is likely the reason for observed high axial loads in pullout testingof pipes buried in dense sand. In consideration of this support from independentnumerical simulation, a variable lateral earth pressure coefficient K was suggestedfor use in Equation (2.1) (instead of K0) to account for dilation-induced normalsoil stress changes on buried pipes during relative axial ground movements.9As the extent of the dilation influence was found to be larger in smaller pipes,K depends on the pipe diameter. Other parameters are the elastic modulus andinternal friction angle of the soil, while interface friction angle and soil dilationwere deemed to be of insignificant influence.Straight pipes (diameters 60 mm and 114 mm) as well as branched layouts weretested, with the former being of interest herein. It was shown that the peak pulloutresistance occurred after around 20 mm, much more than the 2 to 3 mm observedfor steel pipes, due to elastic/plastic elongation of the pipe. Peak normalized axialresistance in dense conditions was 2.3 for pipes with a diameter of 114 mm, and1.9 for a diameter of 60 mm. It was observed in Anderson et al. (2005) that thiswas higher than predicted from ASCE (1984), and resistance for loose conditionswas lower than expected. However, understanding of the earth pressure on pipeswas not enhanced yet with findings from Karimian (2006) mentioned above.At UBC, axial pullout tests on polyethylene pipes used in natural gas distribu-tion networks were conducted by Anderson (2004). Straight pipes (diameters 60mm and 114 mm) as well as branched layouts were tested, with the former being ofinterest herein. It was shown that the peak pullout resistance occurred after around20 mm, much more than the 2 to 3 mm observed for steel pipes, due to elastic/plas-tic elongation of the pipe. Peak normalized axial resistance in dense conditions was2.3 for pipes with a diameter of 114 mm, and 1.9 for a diameter of 60 mm. It wasobserved in Anderson et al. (2005) that this was higher than predicted from ASCE(1984), and resistance for loose conditions was lower than expected. However, un-derstanding of the earth pressure on pipes was not enhanced yet with findings fromKarimian (2006) mentioned above.Weerasekara (2007) provided data on the axial resistance of polyethylene pipes.10K was found to be smaller in polyethylene pipes than in steel pipes, potentially dueto diameter reduction. Overall behaviour was found to depend on pipe propertieslike flexibility and nonlinear stress-strain behaviour.At Cornell, Stewart et al. (1999) tested buried HDPE pipes with a diameterof 150 mm in a temperature controlled environment. Axial pullout resistancedecreased at lower temperatures due to arching of the soil around the pipe, andincreased at higher temperatures due to changes in diameter. According to theresults, temperature change has a larger effect on pipe diameters with a larger di-ameter. Cyclic axial tests (pulling with direction reversal) were also performed andshowed a decrease of axial resistance with the number of cycles.2.1.2 Soil force development during lateral pipe movementLateral soil restraints on buried pipes was initially modelled considering the be-haviour of rigid piles (Hansen and Christensen, 1961) or vertical anchor plates(Ovesen, 1964). Audibert and Nyman (1977) provided some of the first lateralpullout tests on steel pipes with diameters of 25, 60 and 114 mm. It was shownthat vertical anchor plates would overpredict loads on pipes, and that the stress dis-tribution is significantly different in both problems. Results were found to correlatebetter with those from (Hansen and Christensen, 1961). Trautmann and O’Rourke(1985) at Cornell found their results matching those from Ovesen (1964), leadingto an unclear overall picture.Findings from aforementioned studies served as basis for the guidelines pre-sented in ASCE (1984) and ALA (2001). Ultimate lateral soil resistance Pu per11unit length can be calculated therein as:Pu = Nh · γ ·H ·D (2.5)with Nh = dimensionless horizontal bearing capacity factor - values for differentH/D ratios and soil friction angles are included in the guidelines. Lateral pulloutresistance per unit length is also often normalized as:Nqh =Nhγ ·H ·D (2.6)with Nqh = normalized lateral resistance.Guo and Stolle (2005) worked with numerical modelling to investigate scaleeffect, soil properties, burial depth, and burial depth ratio on Nqh. The resultsindicated a dependence of horizontal bearing capacity factor on the scale of thepipe, currently not accounted for in guidelines.At UBC, Karimian (2006) conducted lateral pullout tests as well as numericalmodelling on steel pipes with diameters of 324 and 457 mm. His results weregenerally in line with estimations by Guo and Stolle (2005).Weerasekara (2007) presented limited results from lateral soil restraint tests onburied polyethylene pipes. Strain gauges were attached to the pipe to calculatebending moments and tensile forces in the pipe. Results showed mainly bendingfor the lengths of 1.5 and 2.5 m used in the tests.2.2 HDPE-soil interface frictionAs noted at the outset of this thesis, typical DH pipes used in the industry have anouter jacket made of high-density polyethylene (HDPE). Because of this, the soil-12pipe interaction of buried DH pipes essentially involved interface shear betweenHDPE and the soil backfill. In consideration of this, it is of value to examine thecurrent understanding of the mechanical characteristics at HDPE-soil interfaces.Useful information on this subject has been generated since polyethylene (PE) ma-terials have been used in earth-structures as well as piping materials in natural gasdistribution networks. It is known that polyethylene (PE) makes characterizationof the stress-strain properties complicated due to its nonlinearity and dependenceon strain rate and temperature. In particular, significant effort has been placed tocharacterize the soil-PE interface behaviour.O’Rourke et al. (1990) used a direct shear apparatus to investigate shear strengthof different soil polymer interfaces, including HDPE that is of interest herein. Thispaper is based on an extensive study with a total of 450 direct shear tests. An upperframe made of wood allowed them to perform tests at a low normal stress range of3.5 to 35 kPa, which covers the stress levels encountered in typical buried pipes aswell as in this study. They noted an increasing peak shear stress with increasingdensity of the soil under the same normal stress. A linear relationship betweennormal stress level and shear stress was found within the range of 3.5 and 35 kPa,in tests where the soil unit weight was kept constant.Although the geosynthetic surfaces displayed increasing scratches (depth ofup to 0.03 mm) due to abrasive action during interface shear with soils under re-peated loading O’Rourke et al. (1990) did not find any significant changes in shearresistance due to surface roughness changes in their tests.Based on this work, a dependence of the soil-PE interface friction on the ShoreD hardness of the polymer surface was suggested. Softer surfaces could favourrolling of grains, as their corners can cut into the polymer. On harder surfaces,13grains would rather slide along the surface. Thus, the friction would decrease withincreasing surface hardness.Bilgin et al. (2007) investigated the influence of temperature on the mechani-cal behaviour of high-density polyethylene (HDPE) and medium-density polyethy-lene (MDPE) pipes. Temperature ramp tests should simulate seasonal temperaturechanges in the field. It was found that the pipes behave according to linear vis-coelastic theory for the temperature range used in the study. The stress-strain re-sponse was found not to depend on axial loads in the pipe. A strong dependence ofpolyethylene modulus on temperature was found for both materials.2.3 Soil-pipe interaction of DH pipesNotable research in the field of soil-pipe interaction of district heating (DH) pipeshas been conducted at the University of Hannover in Germany. The followingsubsections give an overview of the two Ph.D. theses have been completed thereon this topic. A summary of potential influence factors was published in Englishby Weidlich and Wijewickreme (2012).2.3.1 Early analytical workAchmus (1995) prepared a Ph.D. thesis on calculations of loads and displacementsof buried DH pipes. These calculations are intended to serve as a basis for furtherinvestigation of soil-pipe interaction behaviour.On the topic of axial resistance, analytical or numerical models were devel-oped for temperature distribution in the system, thermal expansion of the pipe andchange in radial stress around the pipe. To achieve this, the DH pipe is first dis-sociated from the surrounding soil. Linear elastic behaviour for the components14of the pipes - steel, foam and HDPE - are assumed to calculate expansion. Thetemperature gradient throughout the layers is estimated analytically and validatedusing referenced publications. Finally, the increase in radial stress due to radialexpansion is determined using finite element modelling. A constitutive model as-suming hyperbolic stress-strain relation developed by Duncan and Chang (1970) isused in the numerical analysis.For example, the increase in radial stress due to an increase in medium temper-ature of 100 K was described in the following equation:κl = 1.18−0.1 ·h[m]+1.22 ·Dr (2.7)with κl = increase factor, h = soil cover and Dr = relative soil density. Based onreferenced, but mostly unpublished material, it was noted that the axial resistanceafter a full heating-cooling cycle is lower than in a non-heated system. This effectwas claimed to increase with larger pipe diameters.To calculate the lateral bedding resistance, the pipe was assumed to be rigidwhile the soil around was described by a finite element model. Results of simu-lated even displacement along the length in a homogeneous soil showed a decreasein normalized bedding resistance with increased pipe diameter. Different distancesfrom pipe to surrounding original soil - i.e. trench width - and embedment depthswere investigated. Small horizontal distance from original soil and large embed-ment depth increased the significance of the stiffness of the surrounding soil.Additional aspects of lateral resistance such as areas around joints, curves andfoam padding in corners were also investigated with numerical or analytical mod-els. Finally a global model for straight and curved sections was developed. Pipe15sections were represented as structural members, with soil reactions as springs.Prestresses from burial at ambient temperature were considered. For straight sec-tions, it was found that the reduction of maximum displacements of thermally pre-stressed pipelines compared to pipelines installed at ambient temperature was notas big as estimated with previously used approaches. It was also shown that theloads on pipe networks with 90 degree-bends could be underestimated with guide-lines from AGFW (1983), which are commonly used in Germany.This thesis serves as an overview of most aspects of soil-pipe interaction ofDH pipes. Analytical formulations could be a useful tool for further research. Anumber of other investigations from other unpublished sources (companies andorganizations) were included to provide experimental data with which to compare.2.3.2 Experimental and numerical investigation of cyclic axialdisplacementWeidlich (2008) presented an experimental and numerical study on reduction offriction on buried DH pipes subjected to cyclic axial displacement. Multiple paperson this research were published in English - for example Weidlich and Achmus(2006). The change in axial resistance of DH pipes under cyclic thermal loadingwas separated into influence arising from cyclic axial displacement and thermalradial expansion. This Ph.D. thesis by Weidlich mainly focused on an investigationof the former aspect.A soil chamber was built by Weidlich (2008) to embed DH pipes in sand anddisplace them axially, while measuring loads. A finite element model was used tosimulate different dimensions of the chamber. The soil behaviour was describedusing a hypoplastic model, which can account for the influence of cyclic loading.16The void ratio is a state variable in this model, which allows visualization of thechanges in packing due to cyclic displacements of the pipe throughout the soilmass. Details about the hypoplastic model used can be found in von Wolffersdorff(1996). Cyclic shear box tests under constant normal stiffness were conductedusing sand and small HDPE specimen to investigate the interaction behaviour be-tween soil and pipe in the shear zone.The dimensions of the physical test chamber were 1.2 m long, 0.9 m wide and0.9 m tall. Axial pullout tests were conducted on DH pipes with outer diametersof 110, 140 and 160 mm. For each cycle, the pipes were pulled out 50 mm beforereversing the direction to bring them back to the initial position. Different densitiesand amounts of soil cover were used.Weidlich (2008) noted that the radial stress at the soil-pipe interface and thedensity could be identified as influential parameters through the variation of pa-rameters in the preparation of physical experiments. Empirical relationships weredeveloped for the aforementioned pipe diameters to allow an evaluation of the re-duction of friction as a function of soil cover for dense sand. The following equa-tion for the reduction in axial resistance after ten cycles was suggested based on 60soil chamber tests:DF =−0.0388 ·H/D+0.73 (2.8)with D = pipe diameter, H = soil cover and DF = reduction factor.A pressure film was attached to circumference of the pipe in 10 tests to inves-tigate changes in soil pressures experienced by the pipe surface. As pressure cannot be measured during shearing without considerable calibration efforts, only the17Figure 2.1: Results of pressure sensor measurements for cyclic axial pullouttests on pipes with a diameter of 140 mm, a soil cover of three diametersand a medium-dense sand; from Weidlich and Achmus (2006).initial stress state and the state after 10 cycles were presented. Figure 2.1 showsone set of data from a test with a pipe diameter of 140 mm, a soil cover of threepipe diameters and medium-dense sand. At the initial state, a typical pattern ofrelatively higher stress on top and bottom of the pipe is visible. After the shearing,generally lower stresses can be observed, with the highest reductions at the top andbottom of the pipe. It is of interest to note that the observed initial higher stressesare in general accord with those observed by Wijewickreme et al. (2009).Finally, it was postulated that within 5 to 10 cycles a zone of soil around thepipe that has undergone large plastic strains is formed. Since the available stiffnessand strength in this zone is reduced, the soil outside this zone would experienceincreased stresses due to stress redistribution. It was claimed that, consequently,this would lead to the reduction of radial stresses on the pipe, and in turn, reductionof soil frictional force along the axial direction. This hypothesis was verified byWeidlich (2008) using experimental data from axial tests and pressure sensors as18described above.2.4 Thermal conductivity of sandA significant temperature gradient between the DH pipe and the environment,which is at ambient temperature, is inevitably present when the pipe is filled withheated water. Therefore, the temperature at various locations throughout the sys-tem, most importantly at the soil-pipe interface, would depend on the thermal con-ductivities of the respective materials (i.e. soil and materials in the composite pipesection). These thermal parameters are typically specified by the manufacturers forthe pipe material. However, they are generally unknown for the soil surroundingthe pipe.An overview of previous work on the different factors influencing thermal con-ductivity of soils has been presented by Coˆte´ and Konrad (2005). A generalizedmodel was developed to account for variations in parameters such as porosity andgrain shape present in different soil states. A new device to measure thermal con-ductivity during wetting and drying cycles in sand was developed by Smits et al.(2010). It has to be noted that the results in Figure 2.2 are presented in terms ofvolumetric water content. In soil mechanics, water content is commonly defined asthe mass ratio of water to solids. Sands with four different grain size distributionswere tested, only results for the sand resembling Fraser River sand the closest areincluded in Figure 2.2. All of them show a similar pattern, where thermal con-ductivity was observed to increase steeply from 0.25 W/m/K at 0% to more than 2W/m/K at 1% of water content. Thermal conductivity rises more gradually there-after, to reach a plateau value of approximately 2.7 W/m/K at around 20%.1900.511.522.530 0.05 0.1 0.15 0.2 0.25 0.3 0.35Thermalconductivityλ(Wm-1K-1)Water content θ (cm3cm-3)Figure 2.2: Thermal conductivity as a function of volumetric water contentfor a sand with an average grain size of 0.297 to 0.420 mm under denseconditions, after Smits et al. (2010).2.5 Research needs identified from the literatureAs may be noted from the above, so far, only very limited published informationis available to understand the soil-pipe interaction prevalent in DH pipes. It isnecessary to undertake further research on this subject, and in particular to conductphysical modelling of DH soil-pipe interaction problems to provide a database ofexperimental data to validate existing observations and analytical predictions.Specific needs that have been identified for the present research are summarizedin the following list:• Provide information on basic pullout behaviour of DH pipes in both axial andlateral direction under controlled conditions. This will help understand if andhow their behaviour is similar to other pipes made of a single material. Thelarger database of tests on these pipes could then be used for some aspects20in DH pipe research.• Provide data on the change in axial resistance with cyclic change in mediumtemperature. Together with data on the influence of cyclic axial displace-ment of the pipe prepared by Weidlich (2008), this will yield a completeframework of the most significant influences on axial resistance due to cyclicthermal changes.• Provide data on the expansion behaviour at the soil-pipe interface duringtemperature changes in the medium. This will give insight into the interac-tion at the interface and provide a basis to explain changes in pullout resis-tance.• Provide field data on the behaviour of DH pipes. This could include temper-ature, strain and displacement measurements. Multiple, long lasting thermalcycles under real-life conditions could be studied in the field that would bevery difficult in a laboratory under controlled conditions due to logistic con-straints.An experimental program was developed to conduct research to pursue someof the needs listed above using the ASPIRe™ facility at The University of BritishColumbia (UBC). This thesis presents details of the facility, the experimental pro-gram, the research findings and conclusions in the next chapters.21Chapter 3Experimental Aspects3.1 Full-scale tests in soil chamberBased on the research needs identified, it was decided to carry out an experimentalprogram mainly comprising axial and lateral pullout tests on DH pipes using the ex-isting soil chamber at the Advanced Soil Pipe Interaction Research™ (ASPIRe™)facility at UBC. Since some of the key axial tests had to be conducted with waterinside the pipe along with different heating histories, a heating system had to bedesigned, fabricated and commissioned, including water heaters and temperaturemonitoring. Additionally, strain measurements were also conducted during one ofthe tests to improve understanding of the behaviour at the soil-pipe interface.Experimental aspects including relevant technical details of the aforementionednew testing methodologies and setup are included in this section.223.1.1 EquipmentSoil chamberThe soil chamber used in this project was initially developed by Anderson (2004)and later modified by Karimian (2006), Weerasekara (2007) and Monroy-Concha(2013) to meet specific testing requirements. The chamber has been designed toprovide rigid boundaries for the soil mass, as well as flexibility for reconfigura-tion for different applications. The skeleton of the walls is made of vertical steelH-sections (W150x37). They are welded to C250x30 U-sections as base, with di-agonal square beams (HSS 89x89x3.8). This steel frame is connected with 90 mmx 140 mm timber blocks, with a vertical distance of 300 mm from centre to centre.On the inner (soil) side of the chamber, the structure is completed with 19 mmthick plywood sheets or respective plexiglass sheets. The overall dimensions ofthe chamber are 3.8 m L x 2.5 m W x 2.5 m H. An overview picture is shown inFigure 3.1. Additional details of the latest updates to the chamber can be found inMonroy-Concha (2013) and are not repeated herein for brevity.In this project, a flexible divider wall was introduced to the chamber to adaptthe size to the specific needs for each test. For axial tests, only half of the availablewidth of the chamber was used as shown in Figure 3.2a. Karimian (2006) con-firmed previous expectations that the width of the shear zone around a pipe underaxial relative movement would be approximately 10 · d50. For Fraser River sandwith d50 = 0.23 mm this width would be 2.3 mm. The resulting distance from pipeto boundary in the chosen configuration was two orders of magnitude larger, andthus assumed to be sufficient. Karimian (2006) also showed that end wall effectsare negligible in a chamber with a length for 3.8 m for pipes up to a diameter of23Figure 3.1: Overview of the ASPIRe™ soil chamber. (1) A DH pipe witha diameter of 520 mm is buried in the chamber, (2) the red conveyorbelt used for removing soil is located on the right, (3) an empty bulkbag beside it in white. (4) another DH pipe with an outer diameter of315 mm is stored on the left side.0.5 m.For lateral tests, only two thirds of the length were used as shown in Fig-ure 3.2b, as this was estimated to be sufficient for the chosen pipe sizes withoutdisturbing the wedge formation in front of the pulled pipe. This expectation wasconfirmed by visual evidence during the tests (see Chapter 4) and the findings frominvestigations on the effects of chamber boundary conditions conducted during pre-vious research using numerical and experimental analysis by Karimian (2006).Adapting the size of the soil chamber by installing the newly developed dividerwall reduces the amount of soil backfill material required to be handled, while24(a) Configuration for axial tests. (b) Configuration for lateral tests.Figure 3.2: Layouts of the flexible divider wall. Divider wall in red, pipe inyellow, areas filled with soil in brown, pulling direction indicated withblue arrows.maintaining appropriate distance from the pipe to chamber boundaries. Other im-provements were made to make material handling more efficient. The position ofthe band conveyor was modified to minimize the distance that the soil material hasto be moved manually to the feeder of the conveyor. By introducing the dividerwall and streamlining the setup in the space above the soil chamber, forklift accesscould be improved significantly, making installing and removing the pipe as wellas unloading soil material more efficient. These improvements not only made thenumber of tests undertaken in this project possible, they will also contribute to theefficiency of future research projects.Loading systemThe pipe was loaded using two hydraulic actuators with a digital control systemand a capacity of 418 kN each at 21 MPa working pressure. The actuators weremanufactured by Royal Cylinders Inc., New Westminster, BC, Canada, and have a200 mm bore diameter, with a full stroke of ± 305 mm and a 90 mm rod diameter.The mounting arrangement provided a connection height of 740 mm above the steelplate at the floor of the soil chamber. Displacement was controlled by an externallymounted Temposonic Synchronous Serial Interface (SSI) feedback system with a25Figure 3.3: Schematic layout of the loading mechanism used in axial pullouttests including hydraulic actuators, load cells and pulling cables.captive sliding magnet and a resolution of 2 microns. A controller compared themeasured position with the target position and regulated the hydraulic pressureaccordingly using a servo valve, 10 GMP, for each actuator and speeds of up to25 mm/sec. The RMC controller system model RMC100-S2-ENET was made byDelta Computer Systems Inc., Vancouver, WA, USA. Commands can be sent to itfrom a PC using the RMCWin software, also made by Delta Computer SystemsInc. Further details can be found in Karimian (2006).For axial tests, a horizontal steel beam was attached to both actuators, as de-picted in Figure 3.3. The pipe was then in turn attached to this steel beam via ashackle, allowing connecting the pipe exactly in line with its axis to the actuators.For lateral tests, a cable connected each actuator to a hook on the pipe. This layoutis shown in Figure 3.4.26Figure 3.4: Schematic layout of the loading mechanism used in lateral pull-out tests including hydraulic actuators, load cells and pulling cables.Load and displacement measurementEach actuator was equipped with a load cell between the cylinder and the hook.The measured loads from both cells were added together and reported as the totalload. The load cells were made by MTS Systems Corporation, Eden Prairie, MN,USA Type 661.22 and had a maximum load of 222 kN.For lateral tests, one string potentiometer was used to measure displacement ofeach cable on either side of the pipe. For axial tests, as the pipe was sticking outof the chamber, a string potentiometer could be directly connected to the pipe. Allstring potentiometers were of type SP1-50 made by Celesco Transducer ProductsInc., Chatsworth, CA, USA.Loads and displacements were recorded by a data acquisition system that wasbuilt in-house at the Department of Civil Engineering at UBC. A PC running Lab-VIEW SignalExpress was used to log the data during the pullout tests.27Heating systemFor previous research carried out using the ASPIRe™ soil chamber, the pipes werenot filled with fluid. To simulate field conditions with changing water temperatures,it was necessary to fill the pipe with water and be able to control the temperatureof the water mass. Because of that, it was necessary to develop a new system tocontain water mass in the pipe, regulate the water temperature and measure theeffects of thermal changes. At first, different approaches including electric blanketheating, immersed electric water heaters and propane heaters were evaluated forcriteria like uniformity of temperature distribution and time required for each heat-ing phase. Blanket heaters were ruled out for the difficulty to guarantee uniformtemperature distribution. Trial tests showed that available electric heaters could notprovide enough power to increase the temperature of the water mass in a reasonabletime frame. Eventually a combined system of propane and electric water heaterswas chosen and implemented that makes use of the high heating power of propaneheaters and the precise temperature regulation of microprocessor-controlled elec-tric heaters.A schematic overview of the heating system developed as part of this projectcan be found in Figure 3.5.The pipe had to be sealed on both ends to provide a contained space for water.This was achieved by welding a steel end plate and machining a removable endplate specifically for the DH pipe used in this study to provide access from thenorth end of the pipe. A valve for outflowing water was provided on top of thesteel pipe, while a valve for inflowing water was fitted on the removable end cap.The latter valve was connected to a hose inside the pipe, that would wind from28end to end in the pipe three times. This was to ensure that the hot inflowing waterevenly heats the water mass in the pipe.The water mass was heated by circulating water from the pipe through propaneand electrical heaters. Water flowed from the pipe into a small water tank, whichfed into a pump. The pump was made by SHURflo Inc., Costa Mesa, CA, USAand had power of 104 W.Two heating immersion circulators were placed in the small water tank. Theywere made by Haake, Karlsruhe, Germany and VWR International LLC., Radnor,PA, USA and had a total power of 2.3 kW. The required target temperature could beset on a digital interface. Each had a thermal sensor that allows the microprocessorto adjust the power of the integrated heating elements accordingly.From the pump, water was pushed through a propane heater and back into themain pipe. The propane heater was made by Eccotemp, Summerville, SC, USA,and had a rated heat input of 20 kW. This heater delivered most of the energy to heatthe water mass, while the electric heaters were mainly used to maintain a steadytemperature once the target was reached. All heaters are shown in Figure 3.6.Temperature measurementA total of seven thermal sensors were used in the tests with heating cycles. The sen-sors were built in house and calibrated for the expected temperature range. Threeof them were located at the soil-pipe interface of the buried pipe: two at a equidis-tant diameter from both chamber walls in the axial direction, on top and side ofthe pipe. The other was located at the crown between the aforementioned diameterand the southern chamber wall. An additional thermal sensor was buried in the soilmass at a distance of 150 mm from the crown of the middle diameter. The layout29Figure 3.5: Schematic layout of the heating system. (1) Welded steel endcap (2) Removable aluminium end cap (3) Hose connection for outflowfrom pipe (4) Water tank (5) Electric water heater (6) Pump (7) Propanewater heater (8) Hose loops.of all buried sensors is summarized in Figure 3.7. Ambient temperature was mea-sured with one sensor at a shaded location outside the soil chamber. The two lastthermal sensors were used to measure water temperature. One at the inflow valvewas used to control the heating process. The other at the outflow valve was used toestimate the current average temperature of the water mass.The sensors located at the soil-pipe interface were covered in a 100 x 100 x30(a) Electric heaters (b) Propane heaterFigure 3.6: Different heaters used in this project.Figure 3.7: Schematic layout in longitudinal section (distances in mm) ofthermal sensors buried in the soil chamber for axial tests with heatinghistory.31100 mm block of insulation to avoid heat transfer into the soil mass, which wouldlead to underestimation of interface temperature. To increase heat flow from theHDPE interface, silicone grease was applied to the tip of the sensor before attachingit to the interface. Any stronger bonds were avoided to keep the change in frictionduring the pullout phase to a minimum.Resistance temperature detectors model F3101 made by Omega EngineeringInc., Stamford, CT, USA were built in house into waterproof casings for thisproject. They were connected along a 100 W resistor to a data acquisition systemtype SCXI manufactured by National Instruments Inc., Austin, TX, USA. Data wasread and logged on a computer via the commercially available program DASYLab.All thermal sensors were calibrated using a water bath and the microprocessor-controlled electrical heaters described earlier in this section. A unique relationshipfor each sensor was then stored in DASYLab to directly log temperature data.Strain measurementOne test was performed with strain gauges mounted in two directions in six loca-tions on the pipe surface. In detail, two locations along the pipeline length werechosen: one on the buried part of the pipe, equidistant from both chamber wallsin axial direction. The other is located outside the soil chamber, far enough awayfrom the soil chamber to keep the gauges outside during pullout. At each diam-eter, three pairs of strain gauges were installed: one at the crown, the others onthe sides of the pipe (0°, 90° and 270° from the crown). The layout is depicted inFigure 3.8. In each pair, one gauge was installed in axial direction and the other inradial direction.The procedure described by Groves (2014) was followed to securely mount the32Figure 3.8: Schematic layout in longitudinal section with embedded crosssections (distances in mm) of strain gauges attached to the HDPE layerof the pipe for one test.gauges on the HDPE surface while minimizing local stiffness increase due to glue.In summary, the pipe surface was first cleaned and coated with a primer. Loctite414 adhesive was applied sparingly to the gauge for bonding. Cables were sol-dered to the strain gauge, with stress relief loops to avoid rupture of the connectionduring soil movement during pullout. Special precautions were taken to protectthe gauge from moisture from the surrounding soil. Two coats of M-Coat a, anair-drying acrylic coating made by Intertechnology Inc., were applied to the straingauge and connection points to the wire. Rubber tape was attached as outer layer ofthe installation both to protect it from mechanical influence as well as to provide asimilar surface to the original HDPE pipe. Pictures of the strain gauge installationcan be found in Figure 3.9.The gauges used were of type KFEL-5-120-C1 made by Kyowa, Japan. Theirtemperature coefficient was +0.015 %/°C.33(a) Attached to HDPE surface. (b) Covered in protective layers.Figure 3.9: Strain gauges installed at the HDPE surface of the pipe. (1)Metallic foil pattern, sensing strain (2) Stress relief loop (3) Solderingpad connecting strain gauge to cable (4) Cable to data acquisition sys-tem (5) Rubber tape protecting gauge setup (6) Cable mounted on pipewith zip tie looped through holes in HDPE layer.3.1.2 Properties of materialsFraser River sandAll tests in this project were conducted using Fraser River sand, obtained fromAT&H Industries Inc., Maple Ridge, BC, Canada. This material was used in previ-ous full-scale pipe research projects at UBC and was therefore the preferred bench-mark material. Additionally, sand is the recommended backfill material for DHpipes. The average particle size D50 of Fraser River sand is 0.23 mm, its minimumparticle size 0.074 mm, its coefficient of uniformity Cu 1.5 and its specific gravityGs 2.7 (Anderson, 2004). The grain size distribution for the sand used in this studycan be found in Figure 3.10. Karimian (2006) showed that repeated material han-dling, i.e. placement and removal of sand, does not significantly affect its particles.Accordingly, the same sand mass was reused for every test.34Figure 3.10: Grain size distribution of Fraser River sand used in this study.District heating pipesPre-insulated bonded pipes of the “Insulation series 1”, made by LOGSTOR A/S(Denmark), were used in this project. The service pipe transmitting the heat car-rier (usually water) is made of steel with a yield stress of 235 N/mm2. It wascovered by polyurethane foam that was prepared with cyclopropane as the blow-ing agent. Thermal conductivity of the insulating foam was 0.023 W/(m K) for8” pipes and 0.026 W/(m K) for 16” pipes. Pipes with both diameters are depictedin Figure 3.11a. This difference was due to different production methods (axialconti versus spiral conti). HDPE formed the outer casing that protects the foamlayer from physical damage and water ingress. As the HDPE layer was producedby wrapping an infinite band as a spiral around the pipe, a small rib can be found35Table 3.1: Layers of pre-insulated 8” pipes made by LOGSTOR A/S used inthis projectPurpose Material Innerdiameter(mm)Thickness(mm)Service pipe Steel 210.1 4.5Insulation Polyurethane foam 219.1 43.9Outer casing High-density polyethylene 306.8 4.1Table 3.2: Layers of pre-insulated 16” pipes made by LOGSTOR A/S usedin this projectPurpose Material Innerdiameter(mm)Thickness(mm)Service pipe Steel 393.8 6.3Insulation Polyurethane foam 406.4 51.1Outer casing High-density polyethylene 508.6 5.7on the pipe surface every 60 mm in the axial direction. It was found in direct sheartests described later in this chapter that the influence of these ribs is small. A sum-mary of layer purpose, material and dimensions can be found in Tables 3.1 and 3.2.Two copper wires were embedded in the polyurethane foam during produc-tion for use in a leak detection system. These wires are not used in this researchproject. Finished pipes comply with European Standard (EN) 253 but have a widertemperature range for application than specified in this standard.The pipes were delivered at longer lengths than needed. Those of smaller di-ameter could be cut using a stationary band saw. For the larger pipes a differentsolution had to be found due to their size. It was decided to plasma cut themto length. However, high temperatures are imposed on the steel with this cutting36(a) Pipe ends with both diameters (315and 520 mm).(b) Pipe ends with restored insulation,diameter of 520 mm.Figure 3.11: DH pipes manufactured by Logstor (Denmark) (1) ProtectiveHDPE layer (2) Polyurethane foam as insulation (3) Transport pipemade of steel (4) Steel end cap welded in (5) Restored insulation rings(6) Continuous insulation in original condition.technique, which meant that the insulation had to be removed to avoid the releaseof toxic gases. Three rings of roughly 100 mm width were removed on all pipeends that were to be cut. To restore the initial condition of the pipe surface asclosely as possible the rings were attached after cutting using special foam glue.The resulting surface can be seen in Figure 3.11b.Surface roughnessIn order to provide an estimation of the influence of repeated tests on the same pipespecimen, measurements of surface roughness were taken after the testing phaseand compared with a piece of unaffected HDPE surface. A portable roughnesstester type SJ-210 made by Mitutoyo was used to test in accordance with Interna-tional Organization for Standardization (ISO)4287-ISO/TC5 (1997).The control specimen had an average roughness of 0.74 µm (σ = 0.08 µm, n =8). Roughness on the pipe after the testing phase was measured in two areas: nor-mal pipe surface and distinct scratches, which most likely were caused by damage37during material handling. In the first group, the average roughness was 0.82 µm (σ= 0.10 µm, n = 10), and 2.78 µm (σ = 3.52 µm, n = 21) in the second.In an ideal situation a new pipe would be used for every test, which is notfeasible due to practical constraints. The results from pullout tests did not suggestany significant increase in pipe resistance (compare for example Tests No. A-520-H0-1 and A-520-H1C1). Together with the reasonably small difference in surfaceroughness between control specimen and normal pipe surface this could indicatethat reusing the pipe specimen is justified within the bounds of this project despitethe sizable scratches.3.1.3 ProceduresSpecimen preparation and samplingThe sand was loaded in lifts with a thickness of 200 to 250 mm into the chamber.Between the tests, the sand mass was stored in bulk bags, each holding approx-imately 1000 kg of soil. The bags were lifted over the chamber walls using aforklift, then the flap on the underside of the bags was opened and sand fell intothe chamber. As the sand in this project had a moisture content of around 4 %, noissues with formation of dust were encountered.Each lift was compacted using a 100 kg static roller with 16 passes back andforth, as shown in Figure 3.13a. At locations where the roller could not fit betweenpipe and chamber wall, a tamper plate was dropped 30 times from a height ofapproximately 200 mm to achieve a similar compaction.During the preparation of each test, density was measured at least six timesusing “buried bowls”. Herein, bowls of known volume were buried and then re-38Table 3.3: Moisture content measurements from axial testsTest No. Average moisture content (%) σ (%) n (-)A-520-H0-1 4.5 0.6 6A-520-H0-2 3.9 0.5 6A-520-H1-1 4.0 0.2 12A-520-H1-2 3.7 0.9 10A-520-H1C1 3.6 0.2 12A-520-H3C3 4.2 0.6 11A-520-H0-A 3.6 0.2 12A-520-H0-A after 45 day period 2.9 0.6 12A-520-H1C1-S 2.9 0.4 12trieved, to measure the soil mass in the volume. Care was taken to place bowlsat different locations throughout the soil chamber. As the sand’s moisture contentchanges its thermal conductivity, the number of samples taken for this purpose wasincreased to at least ten per test that involved heating, as opposed to at least six forother tests.Results from moisture content and density sampling for all axial tests can befound for reference in Tables 3.3 and 3.4. For each test, average values are listedalong with standard deviation σ and number of samples n.There is a notable decline in moisture content over the entire testing phase.For the ageing test A-520-H0-A, moisture content was measured both at specimenpreparation and after pullout due to the long period in between. Moisture contentdecreased from 3.6 to 2.9% over the 45-day period. The lowest average moisturecontent of the soil mass at specimen preparation of 2.9% was measured for testA-520-H1C1-S.Results from moisture content and density measurements from soil specimenprepared for lateral pullout tests are summarized in Tables 3.5 and 3.6.39Table 3.4: Density measurements from axial testsTest No. Average Density (kg/m3) σ (kg/m3) n (-)A-520-H0-1 1636 31 6A-520-H0-2 1576 43 6A-520-H1-1 1599 48 6A-520-H1-2 1591 28 6A-520-H1C1 1605 18 6A-520-H3C3 1588 32 6A-520-H0-A 1627 60 6A-520-H1C1-S 1589 38 6Table 3.5: Moisture content measurements from lateral testsTest No. Average moisture content (%) σ (%) n (-)L-315-1 4.5 0.3 6L-315-2 4.4 0.1 6L-520-1 4.6 0.4 6L-520-2 4.6 0.3 6Once the sand level was raised to the bottom of the pipe, as shown in Fig-ure 3.12a the pipe was brought in through a hole in the plywood wall, as shownin Figure 3.12b. For axial tests, at the holes where the pipe was sticking out ofthe chamber, a rubber gasket was installed to prevent soil loss. A small distancebetween gasket and pipe was maintained to prevent unwanted additional friction.The pipe was aligned with a level, and then buried in further sand layers untilthe desired soil cover was reached. The last sand layer was smoothened with aTable 3.6: Density measurements from lateral testsTest No. Average Density (kg/m3) σ (kg/m3) n (-)L-315-1 1559 38 4L-315-2 1571 42 6L-520-1 1537 45 8L-520-2 1536 31 640(a) Sand level raised to bottom of pipe,notice the rubber gasket around the hole.(b) Pipe introduced into chamberthrough hole.Figure 3.12: Hole in plywood wall of soil chamber with sand level raised tobottom of pipe, before and after pipe is introduced.(a) Sand compaction with a static roller (b) Leveling of sand surfaceFigure 3.13: Sand preparationwood block and a level on top, as depicted in Figure 3.13b. Thereby it was ensuredto have the same overburden of sand at every location and as a result the samevertical stress throughout the chamber.After completion of a test, the chamber was emptied through an access holeand the sand was transferred back into bulk bags via a conveyor belt, both depictedin Figure 3.14.Heating historyThis section only applies to axial tests.41(a) A conveyor band transports sandfrom the access hole into bulk bags.(b) Sand is stored between tests in bulkbags, holding approximately 1000 kg each.Figure 3.14: Material handlingA schematic overview of the heating system is shown in Figure 3.5. Both pipeends were closed to create a hollow, watertight space in the pipe. A steel plate waswelded into the pipe at the south end (in pulling direction). To close the oppositenorth end, a removable cap was machined of aluminum, as shown in Figure 3.15b.On the same end of the pipe, two valves, one at the top of the pipe and the other inthe end cap, allowed water to circulate through the pipe. Filling of the pipe did notstart until all soil was placed for the respective test. A hose that was connected tothe valve in the end cap ran to the far end of the pipe, as shown in Figure 3.15a.The water mass was heated in a circulatory system, as indicated in Figure 3.5.Water from the main pipe flowed into a small water bath, from which a pump wouldpush it through a propane water heater and back into the pipe. The water bath wasequipped with two electric water heaters that were used in some tests to maintainthe temperature. Water temperature was measured at both valves leading into themain pipe. Once the target temperature was reached, the propane heater was turnedoff and only turned on as required to maintain the temperature. Target temperatureswere generally held within ± 0.5 °C until no further change in temperature at the42(a) View into open pipe with hose actingas heat exchanger connected to valve.(b) Removable end cap machined of alu-minium.Figure 3.15: Installations at the north end of the pipe.soil-pipe interface (defined as± 0.1 °C/h) was observed. Bare pipe ends and waterhoses were covered in fibreglass insulation to minimize heat loss, as shown inFigure 3.16.As each heating and cooling phase took roughly one day, Test No. A-520-H3C3 involving three full cycles could not be completed within one work week.This is why the cooling phase after the second heating phase was planned over theweekend, so that this delay caused as little disturbance as possible.Pipe pullout testAfter the specimen was prepared and the heating history applied (where applica-ble), the pipe was pulled. The pipe was connected to the actuators using shacklesand steel cables. Photos of the cable arrangements for lateral and axial tests areincluded in Figure 3.17. The string potentiometer(s) were attached to the cables(for lateral tests) or to the end of the pipe (axial tests). After the proper functional-ity of all systems was checked, the pulling mechanism was started. All tests werecarried out at a constant speed of 0.25 mm/s. This speed was chosen to resembleconditions previously used in studies on steel pipes.43Figure 3.16: Bare pipe end covered in fibreglass insulation to minimize heatloss.(a) Lateral tests (b) Axial testsFigure 3.17: Cable and shackle setup to transfer the pulling force from theactuators to the pipe.443.2 Direct shear testing for interface frictionThe interface friction angle between the soil and the pipe material is of interestin order to evaluate the results of full-scale tests in this project. For that purpose,direct shear tests were carried out with the materials used in this study, namelyFraser River sand and HDPE coupons cut from the DH pipes.3.2.1 ApparatusThe direct shear tests were conducted on a device that is in use for undergraduatelevel courses at UBC. It consists of a two-piece frame with inner dimensions of100 x 100 mm, a loading frame, a stepper motor to displace the lower half of theframe with a load cell attached to it and two linear variable differential transform-ers (LVDTs). The upper half of the frame is kept in place while the lower halfis moved, generating relative displacement. One linear variable differential trans-former (LVDT) measures this horizontal displacement, the other measures verticaldisplacement of the top cap on the soil sample. This top cap is slightly smallerthan the frame and serves as connection to the loading frame from top. The desiredvertical stress is applied by this frame using air pressure, reduced by the mass ofthe top cap. An overview of the device is shown in Figure ProceduresIn order to test the HDPE layer from the DH pipes in this device, a coupon hadto be cut from the full pipe. One sample from each pipe diameter (due to theribbed surface in larger pipes) was then glued to a plywood sheet. A flat surfacewas produced by clamping the plywood sheet with HDPE coupon between two flatplates during the curing of the glue. Pieces of 100 x 100 mm were cut to fit into45Figure 3.18: Overview picture of the direct shear device used in this project.the lower half of the direct shear box frame. The result can be seen in Figure 3.19.Fraser River sand with a moisture content closely resembling average conditionsin the full-scale tests was then filled into the upper half using a spoon. The soilwas compacted manually using a wooden tamper to achieve the target density toresemble full-scale tests.3.2.3 Testing programThree tests were performed to investigate the interface friction between the high-density polyethylene (HDPE) layer of the district heating (DH) pipes and FraserRiver sand. A coupon cut from the outer HDPE jacket of the smaller diameter pipewith a generally smooth surface was used in two tests conducted under identical46(a) A piece of HDPE glued to plywood (b) Sample placed in direct shear boxFigure 3.19: Preparations for direct shear tests on soil-HDPE interface fric-tion.vertical stress levels (DS-315-1 and DS-315-2) to show repeatability of the results.A third test was conducted on a coupon of the HDPE jacket obtained from the largerpipe to provide information on the influence of the spiral rib on the larger pipe’ssurface described in Chapter 3. The sample with ribs is shown in Figure 3.19a, thesmooth sample is similar but lacks ribs.The soil material for all direct shear tests was Fraser River sand taken fromthe soil mass in use at the large soil chamber. Moisture content of the sand wasmeasured at 4.0 %, density at 1600 kg/m2.The amount of vertical stress was chosen to resemble anticipated vertical stresslevels at the spring line of the larger pipe, (i.e. at a soil depth of 0.78 m) of about12 kPa.3.2.4 Test resultsA summary of the results from the interface direct shear tests is given in Table 3.7.The interface friction angle δ for the flat sample was measured at 26.4 °and 26.1 °.The value was slightly higher for the ribbed sample from the larger pipe with an47Table 3.7: Results of direct shear tests between HDPE pipe and Fraser RiversandTest No. Material tan(δ ) (°) δ (°)DS-315-1 Flat sample from 8in pipe 0.497 26.4DS-315-2 Flat sample from 8in pipe 0.491 26.1DS-520-1 Sample with ribbs from 16in pipe 0.506 26.9interface friction angle of 26.9 °. Considering the relatively close values derivedfrom the three tests it was therefore considered reasonable to assume an averagetan(δ ) of 0.50 for interpretation purposes.O’Rourke et al. (1990) reported a tan(δ ) of 0.35 for interface friction betweenHDPE pipe and Ottawa sand. Weidlich (2008) measured a value of 0.44 for DHpipes used in his study. Given the average value tan(δ ) of 0.5 derived from thepresented tests seems to compare well with the interface friction values obtainedby the above two researchers.3.3 LimitationsAs the soil chamber was located outside, it was exposed to changing ambient tem-peratures - both during changing seasons and daily cycles. The effects of thislimitation during individual tests with durations of a few days were mitigated bythe thermal capacity of the soil mass. For some longer tests, however, the baselinetemperature in the soil mass underwent changes exerting influence on the temper-ature at the soil-pipe interface. Because of this, results should be taken with careand focus on qualitative results. Logs of the ambient temperature are reported inthe respective section for each test with heating history applied to the soil mass.The selection of the distance to boundaries always creates potential limita-tions for tests in soil chambers. Previous experience, most notably from Karimian48(2006), could be evaluated to inform the decision for the chamber’s dimensions.The boundaries were chosen such as no significant influence on the shear zonearound the pipe (axial tests) or formation of the soil wedge ahead of the pipe (lat-eral tests) was expected, as described in more detail earlier in this chapter.For practical reasons, soil and pipes were reused in multiple tests. As previ-ously shown by (Karimian, 2006), Fraser River sand does not suffer significantalteration from material handling in the extent of the present project. Test resultsshowed reasonably consistent values for axial resistance, indicating a negligibleinfluence of surface alteration of the pipe.49Chapter 4Experimental ResultsIn this chapter, results from the experimental program undertaken are presentedand compared to previous work. As the core of the project, axial pullout testswill be presented first, followed by lateral pullout tests. Results from the small-scale interface friction test in the direct shear device are included at the end of thechapter.4.1 Soil-pipe interaction testing programThis section is intended to present an overview of the full-scale soil pipe interac-tion tests conducted for this project and explain the motivation behind the chosenschedule. A summary of the testing schedule can be found in Table 4.1. All testswere carried out with an overburden ratio of 1.5 pipe diameters to the spring lineof the pipe.A total of eight (8) pullout tests were conducted to assess the development ofsoil loads when the pipe is subjected to relative axial soil movement. All thesetests were conducted using DH pipes having an outer diameter of 520 mm (i.e. the50larger diameter pipe were used). The decision to assess the behaviour using onlyone pipe size was made so that a reasonable number of tests with different heatinghistories could be conducted while keeping the test program within the budget andschedule requirements. It is of relevance to note that full-scale testing of pipes isa highly involved undertaking which requires about 2 to 4 days (along with labourand cost) for setting up and dismantling a given test. Additional modifications likethe installation of strain gauges significantly increase the required time.As may be noted in Table 4.1, the first two axial tests were conducted withoutany heating to provide a baseline for the following data (Tests No. A-520-H0-1and -2). The first letter of the Test No. indicates the pullout direction (A for axial,L for lateral), followed by the outer pipe diameter in millimetres. A followingcombination of the letters ’H’ and ’C’ indicate the number of heating and coolingphases. An added ’A’ denotes soil ageing during the tests, while ’S’ represents atest where strain gauges were attached to the pipe surface. Trailing numbers finallyindicate that multiple tests with the same configuration were conducted, providingnumbering.In order to assess the effect of different water heating histories on the soil-pipeinteraction, it was determined to use an identical temperature increase of 50 °Cfrom installation temperature to maximum temperature in the water mass insidethe pipe for all heating phases. To keep the cooling phase in reasonable bounds, atemperature decrease of around 43 °C was applied.Two additional axial pullout tests were conducted after one heating phase toinvestigate the change in resistance at the initial temperature increase (Tests No.A-520-H1-1 and -2). After this, one test was performed with full thermal cycleconsisting of one heating and one cooling phase prior to pullout with cold water51mass (Test No. A-520-H1C1). This should show the change in axial soil resistanceduring pullout due to potential arching after decrease in temperature of the watermass.One test was conducted to provide data on the influence of cyclic thermalchanges, and was thus performed after three full cycles of heating and coolingbefore pulling the pipe with cold water mass (Test No. A-520-H3C3). In order toprovide more data on previous observation on the influence of soil ageing aroundthe pipe, one pullout test without heating of the water mass was performed after45 days from installation of the pipe in the chamber (Test No. A-520-H0-A). Forthe last test, the pipe was equipped with strain gauges to provide data on the expan-sion and contraction of the pipe during one heating and cooling cycle of the watermass before pulling the pipe with cold water mass (Test No. A-520-H1C1-S).Lateral tests were planned to provide data to compare the behaviour of DHpipes with data on previous studies on pipes made of different materials. Twotests with an identical configuration (L-315-1 and -2 as well as L-520-1 and -2)were scheduled to gain confidence about the repeatability in the early stage of theexperimental phase. Two different diameters were chosen to shed light on potentialscale effects.4.2 Axial pullout testsThis section holds results from axial pullout tests with different water heating his-tories, that are the core of this research project. For each aspect including testswith full thermal cycles, soil ageing and strain measurements, detailed results arepresented first, followed by a discussion.At the end of this section, results are compared with previous work and inter-52Table 4.1: Testing schedule for full-scale tests in soil chamberTest No. Direction Outer pipe diameter (mm) Heating historyOther configurationA-520-H0-1 axial 520 -A-520-H0-2 axial 520 -A-520-H1-1 axial 520 HeatingA-520-H1-2 axial 520 HeatingA-520-H1C1 axial 520 Heating and coolingA-520-H3C3 axial 520 Heating and cooling (3x)A-520-H0-A axial 520 45 day agingA-520-H1C1-S axial 520 Heating and coolingStrain gauges installedL-315-1 lateral 315 -L-315-2 lateral 315 -L-520-1 lateral 520 -L-520-2 lateral 520 -face friction from direct shear tests.4.2.1 Baseline and repeatability testsThe results from two axial pullout baseline tests conducted without heating thewater inside the pipes (Tests No. A-520-H0-1 and A-520-H0-2) are presented inthis section. The pipes were filled with water at ambient temperature so that thestresses around the pipe due to self-weight of pipe and water mass are as repeat-able as possible throughout the testing program across all temperatures and heatinghistories.Average densities for the soil backfill in both of the tests are 1636 and 1576kg/m3 respectively, and average moisture contents 4.5 and 3.9%. A summary of alldensity and moisture content measurements is presented in Tables 3.4 and 3.3.Results are presented in Figure 4.1 in terms of axial soil resistance per unitlength of pipe FA (kN/m) and axial pipe displacement Y (m). As may be noted,5300. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFA’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1A-520-H0-2Figure 4.1: Results of two axial pullout tests on pipes with an outer diameterof 520 mm without heating, Tests No. A-520-H0-1 and A-520-H0-2.the mobilisation of loading is initially very steep, and it reaches peak values of 8.3and 8.6 kN/m within some 30 mm of shearing. Large-strain resistance is about 6.6kN/m at the maximum displacement of about 440 mm. The practically identicalresponse observed between the tests suggests very good test repeatability, and thusindicates appropriate quality control with respect to specimen preparation and dataacquisition.Visual observationsPhotographs of the soil surface were taken before and after every pullout test bothfor quality control and to show any visible changes. Figure 4.2 compares the stateof the surface before and after the pullout of Test No. A-520-H0-1. The surfaceseems to be unchanged, which was also observed for all other axial pullout tests.Another area of interest for visual observation is where the pipe leaves thechamber at the gasket on the actuator side of the chamber. It is preferable to have54(a) After specimen preparation (b) After pullout testFigure 4.2: Visual comparison of the soil surface before and after the pullouttest A-520-H0-1.as small a soil loss as possible to mimimise the disturbance of the soil mass. Photosof the actuator end of the pipe after pullout of Test No. A-520-H0-1 are shown inFigure 4.3. There is minimal soil loss both on the top and on the bottom of the pipenoticeable, and it seems small enough not to interfere with the overall behaviour ofthe soil mass.Both the change of the soil surface and the soil loss on the leading end of thepipe are deemed negligible for test A-520-H0-1. Similar insignificant soil loss andsurface changes were encountered in all the subsequent tests conducted, hence onlythe photos for Test No. A-520-H0-1 are presented to maintain brevity.4.2.2 Tests with increased water temperature inside the pipeResults from tests with the water inside the pipe heated to a certain temperatureby one heating phase before pulling the pipe (Tests No. A-520-H1-1 and A-520-H1-2) are presented in Figure 4.4. Average soil densities of the backfill in thesetests are 1599 and 1591 kg/m3, and average moisture contents are 4.0 and 3.7 %,respectively.The test results in Figure 4.4 show steep initial phases in the load-displacement55(a) Top of pipe (b) Bottom of pipeFigure 4.3: Soil loss at the gasket in pulling direction during test A-520-H0-1. Note: photo taken of the visible part of the pipe as it comes out of thesoil chamber.relationship as in Figure 4.1 for the tests conducted with water at ambient temper-ature inside the pipe. However, it can be noted that the value of the peak loads areclearly higher than those in Figure 4.1, i.e. values of 8.3 and 8.6 kN/m comparedto about 8.9 and 9.4 kN/m in Figure 4.1. Likewise, large-displacement loads inthe tests conducted with the heated water are also higher (at 7.1 and 7.5 kN/m)compared to those from Figure 4.1.While the individual results between the two heated tests are still comparable,repeatability is not as good as observed between the results from identical non-heated baseline tests. This difference observed in the heated tests could be dueto different ambient temperature profiles throughout both tests, which might affecttests that involve heating more than those without. Temperature histories for thetwo tests are given in Figures 4.5 and 4.6. Detailed temperature histories of thewater mass for all tests are presented in Appendix A. Data from the three thermalsensors installed at the soil-pipe interface is shown in dark green, light green andred. Purple indicates data from the sensor buried in the soil mass 150 mm awayfrom the pipe, while blue shows ambient temperature reading from a shaded loca-5600. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFA’(-)Normalized axial pipe displacement Y’ (-)A-520-H1-1A-520-H1-2Figure 4.4: Results of two axial pullout tests on pipes with an outer diameterof 520 mm with heating and pullout with hot water mass, Tests No.A-520-H1-1 and A-520-H1-2.tion behind the soil chamber. This colour convention is the same for temperaturehistory plots with multiple lines from all tests.As may be noticed in Figure 4.6, data acquisition was interrupted in the timeperiod between 14 to 18 hours. The data acquisition system stopped workingovernight during that period and had to be manually restarted the next morning.It is expected that this does not affect the usability of the data set negatively, as thetemperature change in the soil mass can be linearly approximated based on mea-surements before and after. The system was not modified during that time, so noheat source or heat bridge could alter the thermal flow.During Test No. A-520-H1-1, ambient temperature was generally increasingduring the duration of heating. Average temperatures at the soil-pipe interface wasincreasing by 4.7 °C (i.e. from -0.4 to 4.3 °C). In contrast, during A-520-H1-2ambient temperatures decreased, and the increase at the interface was 4.3 °C (i.e.57-2-10123450 5 10 15 20 25 30Temperature(◦C)Time (h)ambient temp.soil-pipe-interface-temperaturesoil mass temp.Figure 4.5: Temperature history during A-520-H1-1, at soil-pipe interface,150 mm away from the pipe in the soil mass and ambient temperature.from 7.0 to 11.3 °C), which is a slightly smaller change compared to the previoustest. It is also possible that the temperatures just below zero at the beginning of A-520-H1-1 may have influenced the overall behaviour. However, as the temperatureswere above zero at the time of pullout, it was decided to continue with the test.The results from tests can be compared to the baseline tests to assess the effectof increasing the temperature of the water mass inside the pipe on the developmentof axial resistance. In order to compensate for the small differences in densitybetween the different tests when attempting to compare values, the results are pre-sented in the form of normalized axial soil resistance and normalized displacement,as described in Equations (2.3) and (2.4). Additionally, the results were smoothedto support the visual comparison. Furthermore, this allows to easily compare withresults from different sources originating from experimental work using differentpipe diameters, densities and overburden ratios.58246810120 5 10 15 20 25 30Temperature(◦C)Time (h)ambient temp.soil-pipe-interface-temperaturesoil mass temp.* data not collected due to technical problems***Figure 4.6: Temperature history during A-520-H1-2.Figure 4.7 shows the comparison of axial soil resistance from baseline tests andthose from tests where pullout was conducted after the water temperature insidethe pipe was increased. (Note: both tests had a water mass inside the pipe duringtesting.) Results show consistently that, the tests conducted with heated waterinside the DH pipe displayed a noticeable higher pullout load than that observedfrom the tests conducted on the pipe with unheated water mass. Both peak andlarge-strain load are roughly 5% higher than in tests without heating but otherwiseidentical configurations.4.2.3 Tests with water mass inside the pipe subjected to full thermalcyclesResults from tests with one and three full heating and cooling cycles will be pre-sented in this subsection. For Test No. A-520-H1C1 involving water inside thepipe subjected to one heating cycle, the backfill soil had an average density of1605 kg/m3, and an average moisture content of 3.6 %. For A-520-H3C3, that was5900. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFR’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1+2A-520-H1-1+2Figure 4.7: Comparison of average results from pullout tests without heatingand with pulling with hot water mass, Tests No. A-520-H0-1 and -2 aswell as A-520-H1-1 and -2.conducted after three thermal cycles, the backfill soils had an average density of1588 kg/m3 and average moisture content of 4.2 %.Each full thermal cycle stands for a heating phase, where the maximum tem-perature is maintained long enough until the temperature in the chamber reacheda close-to-steady-state (i.e. only small changes of temperature would occur at agiven location with time), followed by a cooling phase, where the minimum tem-perature is maintained long enough, again to reach a close-to-steady-state. Thisheating cycle application is discussed in Chapter 3.The full temperature histories for locations in the soil chamber and ambienttemperature during Tests No. A-520-H1C1 and A-520-H3C3 are presented in Fig-ure 4.8 and Figure 4.9, respectively. While the ambient temperature stayed withina narrow band (about 2 to 6 °C) during Test No. A-520-H1C1, the fluctuations ofthe ambient temperature during the test with three thermal cycles (Test No. A-520-600246810120 20 40 60 80 100Temperature(◦C)Time (h)ambient temp.soil-pipe-interface-temperaturesoil mass temp.* data not collected due to technical problems*********Figure 4.8: Temperature history during Test No. A-520-H1C1.H3C3) were significant (i.e. between -8 and +6 °C). However, the absolute value ofambient temperature was highest during pullout at the end of the test, and compara-ble with those ambient temperatures measured during the other tests. Consideringthat the temperatures at the measuring locations in the soil chamber stayed abovezero, the temperature history is assumed to be acceptable.Normalized soil resistance versus normalized axial pipe displacement relation-ships for the above mentioned cyclic thermal tests as well as the average fromthe baseline tests are superimposed in Figure 4.10. Results for one thermal cycleclosely resemble those from baseline tests for most of the displacement path. Af-ter a normalized displacement of 0.45 they even exceed baseline results, howeveronly in an amount that is within the expected fluctuation between tests. On theother hand, normalized axial resistance for the test conducted after three full ther-mal cycles is consistently lower than baseline results, with a greater difference ofroughly 5% at large displacements. Examination of Equation (2.1) suggests that, if61-10-505100 50 100 150 200 250Temperature(◦C)Time (h)ambient temp.soil-pipe-interface-temperaturesoil mass temp.Figure 4.9: Temperature history during Test No. A-520-H3C3.the interface friction is unchanged, any reduction in axial pullout resistance couldlikely be a result of lower effective normal soil stresses on the pipe (i.e. a reduc-tion of equivalent K0). The pipe would expand during heating and contract duringcooling. The soil that would have experienced some outward plastic deformationduring pipe expansion would not likely rebound back by the same amount (i.e. thesoil will be held back due to arching) when the pipe retracts during cooling. Theabove results for the case with three thermal cycles seem to support the presence ofsuch an arching mechanism. However, based on presented data, the arching mech-anism does not seem to have been mobilized in a significant way for the case thatinvolved only one full cycle of thermal loading.4.2.4 AgeingThe results from Test No. A-520-H0-A, where the soil specimen was allowed to agefor 45 days before the pipe was pulled, are presented in Figure 4.11. For this test,6200. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFR’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1+2A-520-H1C1A-520-H3C3Figure 4.10: Comparison of results from two axial tests with one (Test No.A-520-H1C1) and three (Test No. A-520-H3C3) full thermal cycles aswell as results from baseline tests.the pipe was again filled with water, so that it has comparable normal stresses tothe other tests presented herein. However, there was no heating or cooling appliedto the water mass. This test took place from mid-February until the end of March,with ambient temperatures in Vancouver between -2.4 and 12.6 °C according toEnvironment Canada (2014).As noted from Figure 4.11, the peak normalized axial resistance is noticeablylower for the pullout after the 45-day period. A difference in the range of about5% remains until roughly 0.3 of normalized displacement. Thereafter, both curvesconverge, to approximately have the same normalized axial resistance from about0.5 normalized pipe displacement level to the point until the test was terminated.6300. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFR’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1+2A-520-H0-AFigure 4.11: Comparison of results from a test with 45 days of soil agingbefore pullout (Test No. A-520-H0-A) with results from baseline tests.4.2.5 Strain measurementsFor Test No. A-520-H1C1-S, the DH pipe was equipped with twelve strain gaugesin axial and radial direction as described in Chapter 3. The temperature historyat locations throughout the soil mass as well as ambient temperature is shown inFigure 4.12. Additionally, Figure 4.13 presents the temperature of the water massfrom the start of the heating period until pullout.Axial and radial strain versus time graphs obtained from strain gauges mountedat a location within the buried part of the pipe are presented in Figures 4.14 and4.15 respectively, It is noted that strain measurements were taken manually at var-ious times during this test, as indicated by the symbols in the plots. In a similarmanner, Figures 4.16 and 4.17 present data from strain gauges mounted on the pipe,but at a location outside the soil chamber. As the temperature at the strain gaugelocations changes, it is important to consider corrections for the thermal change646810121416180 10 20 30 40 50Temperature(◦C)Time (h)ambient temp.soil-pipe-interface-temperaturesoil mass temp.* data not collected due to technical problems****Figure 4.12: Temperature history during Test No. A-520-H1C1-S.0102030405060700 10 20 30 40 50Temperature(◦C)Time (h)water tempFigure 4.13: Temperature history of the water mass during Test No. A-520-H1C1-S.65-10001002003004005006000 10 20 30 40 50Strain(10-6)Time (h)lefttoprightFigure 4.14: Axial strain history from strain gauges mounted on the DH pipeat a diameter buried in the soil chamber during Test No. A-520-H1C1-S, on top of the pipe and on the sides.in resistance. The thermal coefficient of the strain gauges was +0.015 %/°C. Thelargest temperature difference measured at the soil-pipe interface was 6.3 °C (from10.6 to 16.9 °C, see Figure 4.12). Consequently, the maximum error due to ther-mal effects was 0.095 %. It was decided to use the data without correction fortemperature, as these would be minor in relation to other sources of error.It appears that the measured radial strains can help to explain the mechanismsinvolved in the change in axial soil resistance of DH pipes with temperature of thewater mass. During Test No. A-520-H1C1-S, the average peak radial strain at theend of the heating phase is 860 microstrain (10-6) at the location on the buried pipe,as shown in Figure 4.15. This translates to an increase in circumference of 1.4 mm.This noted expansion of the HDPE layer on the soil-pipe interface can be ra-tionally considered due to two mechanisms. First, strains can be caused by thethermal expansion due to the temperature increase at the HDPE layer. This tem-66020040060080010000 10 20 30 40 50Strain(10-6)Time (h)lefttoprightFigure 4.15: Radial strain history from strain gauges mounted on the DH pipeat a diameter buried in the soil chamber during Test No. A-520-H1C1-S.-500050010001500200025000 10 20 30 40 50Strain(10-6)Time (h)lefttoprightFigure 4.16: Axial strain history from strain gauges mounted on the DH pipeat a diameter outside the soil chamber during Test No. A-520-H1C1-S.67-500050010001500200025000 10 20 30 40 50Strain(10-6)Time (h)lefttoprightFigure 4.17: Radial strain history from strain gauges mounted on the DH pipeat a diameter outside the soil chamber during Test No. A-520-H1C1-S.perature increase is of significantly smaller extend than the temperature increase ofthe water mass inside the pipe due to the insulation layer in between. Secondly, theexpansion of the steel pipe at the core could mechanically propagate through theinsulation layer, pushing the outer layers, figuratively speaking, and also contributeto the strains of the HDPE layer.In order to evaluate the potential influence of both factors, it seems appropri-ate to investigate the chronology of strains at the HDPE layer, as full propagationof the temperature increase to the soil-pipe interface appears to occur significantlylater than temperature increase at the steel layer. Steel has a relatively high thermalconductivity, thus it is fair to assume that the steel pipe would adopt the tempera-ture of the water mass almost in harmony with the increase in water temperaturewithin the pipe. However, at the HDPE layer a delay in the order of several hourswas measured (see Figure 4.12), until a noticeable temperature increase could be68observed.After the initial fast heating phase of the water mass in this test, between hour0 and 4, strains measured at the HDPE layer on the soil-pipe interface have alreadyreached roughly 60% of the maximum strain (see Figure 4.13). The average tem-perature at the soil-pipe interface, however, has only increased by some 30% of themaximum temperature increase during the same time (see Figure 4.12). and it tookanother 4 hours for the temperature at the interface to reach 50% of the maximumtemperature increase. Because of this, it is fair to argue that this additional strainon the HDPE layer is possibly due to expansion of the steel pipe at the core, thataffects the outer HDPE layer through the polyurethane foam.Examining axial strain (i.e. axial elongation of the pipe) at the same locationin Figure 4.14 reveals a similar pattern as for radial strain. Peak axial strain val-ues herein are somewhat lower than for radial strain, with an average of roughly500 microstrain. The ratio of instant strain after the first heating phase of the watermass is even higher for the case of axial strain than to that for radial strain dis-cussed in the above paragraph. More than 70% of the total axial strain is reachedafter only 30% of temperature increase has taken place at the soil-pipe interface. Apossible explanation is an even better bonding between the HDPE layer and thosebelow in the axial direction.It is of interest to compare the strain measurements from the gauges mountedon the pipe portion that is located outside the soil chamber with those from straingauges mounted on the buried part of the pipe (see Figures 4.16 and 4.17 for axialand radial strains, respectively). It can be noted that results in Figures 4.16 and4.17 are not as consistent as those noted in Figures 4.14 and 4.15 displaying thestrain gauge readings from the buried portion of the pipe. It has to be noted that69the gauges on top of the pipe (green data in the figures) were covered by fibreglassinsulation used to cover the bare pipe ends. This was located near the top of thepipe shown in Figure 3.16. These gauges report a similar strain history over timesas those buried, but with a larger amount of strain. Peak strains for both radialand axial directions reach 2200 microstrain. This could be explained by the muchhigher temperatures that would prevail at the location due to better insulation fromthe fibreglass than the soil mass surrounding the buried portion of the pipe, poten-tially higher than at the soil-pipe interface inside the soil chamber, and the lack ofsoil resistance to pipe expansion. Strain readings from the gauges that are exposedto the atmosphere also reach significant levels exceeding 500 microstrain, but gen-erally seem to follow the trend of the ambient temperature (see Figure 4.12). Suchbehaviour suggests that any heat that reaches the HDPE layer is quickly transferredto the air around the pipe. In turn, this highlights the importance to study the pipeand surrounding soil as one system when investigating the thermal behaviour ofDH pipes, and not simply by investigating the behaviour of the heated pipe alone.Comparing the temperature history of Test No. A-520-H1C1-S with the similarTest No. A-520-H1C1 in Figure 4.8 gives the opportunity to assess the repeatabil-ity of the tests conducted with similar temperature increases at the soil-pipe inter-face, but not necessarily conducted at the same ambient temperatures. In this test,Test No. A-520-H1C1-S, the average temperature at the interface was increasedby 6.1 °C from 10.4 to 16.5 °C. During the comparable Test No. A-520-H1C1,the increase was 6.0 °C from 3.3 to 10.3 °C. The fact that the temperature dif-ference measured at the soil-pipe interface between start and end of the heatingphase was almost identical for Tests No. A-520-H1C1-S and A-520-H1C1 (i.e. 6.1and 6.0 °C, respectively) suggests that the constant temperature increase of the70water mass is most important for repeatable tests, and relatively independent ofthe starting temperature (i.e. 10.4 and 3.3 °C for Tests No. A-520-H1C1-S andA-520-H1C1, respectively).Axial soil resistanceAlthough not the main objective of this particular test, the Test No. A-520-H1C1-Sgave an insight into the influence of the chosen installation method for gauges anddata acquisition cables on axial soil resistance.Pullout resistance from Test No. A-520-H1C1-S is compared with a test ofthe same heating history (i.e. Test No. A-520-H1C1) but without strain gaugesinstalled as well as baseline results in Figure 4.18. Peak resistance is notably lowerfor the pipe with strain gauge installations compared to the test without. Large-displacement resistance is higher for the test where strain gauges were attached tothe pipe.In summary, the chosen installation method seems to affect the pullout resis-tance in the range of the influence of different heating histories applied in thisproject. Further investigation seems necessary before data on soil resistance fromtests with strain gauges could be used in reliable analysis.4.2.6 Comparison with previous workIt is of relevance to compare the results from the present program with those fromaxial pullout tests on steel pipes buried in Fraser River sand that were previouslyconducted at by Wijewickreme et al. (2009) using the UBC ASPIRe™ facility. Thediameter of the steel pipes tested by Wijewickreme et al. (2009) of 457 mm wassimilar to the DH pipes having an outer diameter of 520 mm. Overburden ratio7100. 0.2 0.4 0.6 0.8 1NormalizedaxialsoilresistanceFR’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1+2A-520-H1C1A-520-H1C1-SFigure 4.18: Comparison of results from the test with strain gauges attachedafter one full thermal cycle (Test No. A-520-H1C1-S) with the testwith one cycle without gauges (Test No. A-520-H1C1) as well as re-sults from baseline tests.varied from 1.7 up to 2.6 pipe diameters for the cases reported by Wijewickremeet al. (2009), again similar to those for the present testing with DH pipes.In the present study, the pipe was filled with water, while the pipe tested by Wi-jewickreme et al. (2009) was empty. Therefore the axial loads have to be correctedin order to compare results between both studies.Equation (2.1) presented a method to estimate axial resistance due to normalsoil pressure acting on the pipe:FA = pi ·D ·σ ′n,av · tan δ = pi ·D ·1+K02· γ ·H · tan δwhere pi ·D ·σ ′n,av equals the total normal force due to normal soil pressure actingon the pipe per unit length. Similarly, the added axial resistance per unit weight due72to weight of the water mass inside the pipe can be estimated using the followingapproach:FA,water =Ww · tan δ = pi · (Di2)2 · γw · tan δ (4.1)with Ww = weight of water mass inside the pipe per unit length, γw = unit weight ofwater and Di = inner diameter of the pipe. Axial resistance per unit length correctedfor weight of the water can now be defined as:FA,corr = FA−FA,water (4.2)A comparison of results from the present study and Wijewickreme et al. (2009)can be found in Figure 4.19. It can be noted that the axial loading response issimilar with a distinct peak after relatively small displacements. The amount ofaxial displacement to reach the peak soil resistance is significantly smaller for steelpipes (10 mm) compared to those for DH pipes (30 mm). This is very likely due tothe relatively less overall stiffness of HDPE felt at the soil interface for DH pipes,however, this notion needs to be confirmed with further study.It is also notable that both peak and large-displacement normalized axial re-sistance is some 20% higher for DH pipes, even when pulled with the cold watermass. A number of reasons may be contributing to this difference. The differ-ent material at the soil interface with varying stiffness and surface hardness mightagain contribute to this difference. It also has to be noted that the DH pipes werefilled with water and thus subjected to higher average normal stresses.7300. 0.2 0.4 0.6 0.8 1CorrectednormalizedaxialsoilresistanceFA’(-)Normalized axial pipe displacement Y’ (-)A-520-H0-1+2A-520-H1-1+2(a) DH pipe, pulled with hot and cold water mass, D = 520 mm, H/D = 1.5; Axial resistancewas corrected for weight of water mass.(b) Steel pipe, D = 457 mm, AB-6 was pulled after 45 days of ageing, H/D = 1.7 to 2.6,from Karimian (2006).Figure 4.19: Comparison of normalized axial pullout resistance from thisstudy with previous work on steel pipes.744.2.7 Comparison with interface friction from direct shear testInterface friction angles between Fraser River sand and samples from the DH pipesurface were obtained specifically for this study and presented earlier in this chap-ter. This serves as another tool to evaluate the plausibility of results from the large-scale tests.Equation (2.1) can be used with the friction angle obtained from direct sheartests for an estimation of axial resistance for a given geometry. With an average in-terface friction angle of 26.5 °, an average density of 1600 kg/m3 and a lateral earthpressure coefficient K0 of 0.46 for Fraser River sand this yields an axial resistanceof 6.8 kN/m. The lateral earth pressure coefficient K0 for normally consolidatedsoils can be obtained from the following equation commonly used in soil mechan-ics:K0 = 1− sin(φ ′cs) (4.3)with φ ′cv = effective friction angle at critical state, a fundamental soil friction con-stant, its value being approximately 33° for Fraser River sand.As Figure 4.20 illustrates, this slightly underestimates the force at large dis-placements in soil chamber tests. This could be due to a different stress state thanexpected, both as a result of medium-dense packing and soil dilation. The rela-tion fails to predict the initial peak at smaller displacements, which could be againcaused by dilation.7502468100 100 200 300 400 500AxialsoilresistanceFR(kN/m)Axial pipe displacement Y (mm)A-520-H0-1stA-520-H0-2ndASCEFigure 4.20: Comparison of baseline Tests No. A-520-H0-1 and A-520-H0-2with estimations based on Equation (2.1) with δ from direct shear test.764.3 Lateral pullout tests4.3.1 Test resultsFigure 4.21 shows the load-displacement curves derived from two identical lateralpullout tests (i.e. Tests No. L-315-1 and L-315-2) conducted on a pipe with anouter diameter of 315 mm. As may be noted, the load-displacement response dur-ing Test No. L-315-1 could not be effectively obtained due to some intermittentproblems with the load cell readings that arose as a result of a loose connection inthe cable. These can be seen as spikes on the graphs. However, when connectionwas established during different times of the test, data collected during those timeswas considered reasonable to be used for interpretations. Considering the extremetime, effort and expenses required in testing, it was difficult to discard the wholetest and re-conduct the tests in its entirety.A comparison between the load-displacement curves from the two tests sug-gests that the results are in good agreement in the initial loading phase until around120 mm displacement and in the very last 20 mm. In between, significant parts areaffected by the aforementioned artefacts, making a comparison difficult. In two ofthe other likely unaffected sections at 180 mm and 280 mm the load values differslightly, while the other sections around 300 mm are again in good agreement. Al-though not perfect under post peak condition, it is judged that the data from thetwo tests are in agreement suggesting reasonable repeatability.Load-displacement curves of the two tests (i.e. Tests No. L-520-1 and L-520-2) on a pipe with an outer diameter of 520 mm are presented in Figure 4.21.Both curves are in good agreement after both reach a plateau starting at 150 mmall the way to 500 mm displacement. Two sections of the test L-520-1 had some770510152025300 100 200 300 400 500ForceNh(kN/m)Displacement y (mm)L-315-1L-315-2Figure 4.21: Results of two lateral pullout tests on pipes with an outer diam-eter of 315 mm, Tests No. L-315-1 and L-315-2.unanticipated load cell connection problems around 300 mm and between 380 and450 mm, however, to a significantly lesser extent than the case with Test No. L-315-1 conducted with the smaller size pipe having a diameter of 315 mm.There is, however, disagreement between both curves in the initial loadingphase, with one reaching maximum load after around 80 mm, the other after 150 mm.One possible explanation is that the position of both actuators had a small offsetduring pullout in test L-520-1 due to technical difficulties with the controller. An-other possibility is the variation of density of soil backfill immediately around thepipe of a given test.To provide insight into scale effects, results of tests with both diameters arecompared in Figure 4.23 using normalized values of horizontal soil restraint Nqh,as described in Equations (2.6) and (2.4). Only one test for each diameter is shown,L-315-2 and L-520-2 were chosen to avoid the artefacts due to the loose connection7801020304050600 100 200 300 400 500ForceNh(kN/m)Displacement y (mm)L-520-1L-520-2Figure 4.22: Results of two lateral pullout tests on pipes with an outer diam-eter of 520 mm, Tests No. L-520-1 and L-520-2.to the load cell. It is of interest to note that the tests share almost the same pathduring the initial loading path. However, the load in the test with smaller diameterreaches a higher and more distinct peak. The load at large displacements is stilllower with larger diameter, but the difference is smaller.4.3.2 Comparison with previous workMonroy-Concha (2013) conducted lateral pullout tests on steel pipes for his re-search project. Both the pipe diameter of 406.4 mm and overburden ratio of 1.6were similar to parameters chosen in this study. Fraser River sand was used asbackfill material for both studies. Figure 4.24 includes normalized force versusdisplacement plots for comparison between DH and steel pipes.Peak lateral force appears to be approximately comparable for both pipe typeswith a normalized force of around 8. This might indicate that pipe geometry hasa larger impact on peak lateral resistance than type of material at the soil inter-7902468100 0.5 1 1.5 2NormalizedForceNqh(kN/m)Normalized Displacement, y/D (-)L-315-2L-520-2Figure 4.23: Comparison of normalized lateral resistance of pipes with anouter diameter of 315 and 520 mm, Tests No. L-315-2 and L-520-2.face. Other factors that influence lateral resistance include surface roughness, soildensity - especially in the vicinity of the pipe - and soil overburden. Figure 4.24indicates that the loading behaviour for the steel pipe is steeper than for the DHpipes. The softer loading behaviour could maybe be caused by the polyurethanefoam layer around the DH pipes tested, which could potentially allow some move-ment of the inner pipe before the full soil resistance is mobilized. More testingseems necessary to provide reliable conclusions on the comparability of lateral soilresistance of DH pipes.4.3.3 Visual observationsAll lateral tests were documented photographically before and after the pulloutphase. To enhance clarity, stripes of white sand were placed facing the plexiglasssheets in a spacing of roughly 100 mm. Photos from a test with the smaller diameterpipe (L-315-2) are shown in Figure 4.25, and those from a test with the larger8002468100 0.5 1 1.5 2NormalizedForceNqh(kN/m)Normalized Displacement, y/D (-)L-315-2L-520-2(a) DH pipe, D = 315 and 520 mm, H/D = 1.5.(b) Steel pipe, D = 406.4 mm, H/D = 1.6, from Monroy-Concha (2013).Figure 4.24: Comparison of normalized lateral pullout resistance from thisstudy with previous work on steel pipes81diameter (L-520-2) are presented in Figure 4.26. The second tests were chosenas the method to place sand stripes was improved after the first tests, providing aclearer result.In the photos taken after the pullout test, the wedge formation ahead of the pipein pulling direction can be observed. For both pipe diameters, two distinct wedgescan be identified, with a shearing zone in between. This might be caused by the factthat displacement is applied to the pipe in horizontal direction via cables connectedto the actuators. Consequently, the pulling is always directed towards the actuators,creating a small vertical force component in a downward direction as the pipe riseswith the initial wedge formation. This downward pull inhibits the pipe from risingstrictly along the initial wedge surface, potentially creating another wedge as thepipe “leaves” the initial wedge surface and moves further horizontally.In Figure 4.26b the wedge formation ends roughly 100 mm in front of thechamber wall.A typical view of the soil surface behind the pipe (opposite of pulling direction)after the pullout of test L-520-2 is included in Figure 4.27. The soil is vastlydisturbed and displaces in chunks that predominantly form in the axial direction.82(a) Before pullout(b) After pulloutFigure 4.25: Side view of the soil specimen during test L-315-2.83(a) Before pullout(b) After pulloutFigure 4.26: Side view of the soil specimen during test L-520-2.84Figure 4.27: Typical view of the soil surface behind the pipe after pullout inL-520-2.85Chapter 5Summary and ConclusionsDespite widespread use of district heating (DH) pipe networks in Europe for decades,published research on the soil-pipe interaction response related to buried DH pipesystems is limited compared to those available for other types of pipe such as steelpipes. This research work was undertaken with the focus of generating initial ba-sic experimental data on the reponse of buried DH pipes subjected to relative soilmovement. The work mainly involved soil-pipe interaction physical modelling us-ing equipment and procedures established at The University of British Columbia(UBC). As this is an initial exploratory undertaking, the research work was de-signed to obtain a basic understanding of the key soil-structure mechanisms in-volved with respect to DH pipes. Since DH pipes have a composite material cross-section, it could prove to be beneficial to have a platform to link results from cur-rent research with those findings that have been obtained with various other pipematerials, and contribute to some of the research needs identified in Chapter 2.865.1 Contributions to experimental equipmentIn order to undertake the planned experiments for this research project, the existingsoil chamber that is part of the ASPIRe™ facility at UBC had to be modified to fillpipes with water that would be heated. Additionally, upgrades were made to makehandling of soil backfill material more efficient.As core component, a heating system had to be developed that would allowheating the required amount of water effectively. Different approaches were eval-uated for criteria like uniformity of temperature distribution and time required foreach heating phase. After trial tests, a combined system of propane and electric wa-ter heaters was chosen and implemented that makes use of the high heating powerof propane heaters and the precise temperature regulation of microprocessor-controlledelectric heaters. Furthermore, a water circulation system was developed that en-sures potent use of the heaters as well as uniform temperature distribution in thewater mass inside the pipe. To enable this, end caps were specifically manufacturedfor the DH pipes in this project that allow filling the pipes with water and connect-ing the circulation system to the water mass inside the pipe. Finally, thermal sen-sors were built and calibrated before being placed at carefully chosen locations tomonitor temperature relevant for heating process control as well as gathering datafor scientific investigation. The development of this heating system contributes tofuture research by expanding the ASPIRe™ platform with a thermal component,that is ready to use in future projects.A flexible divider wall was designed and built that can be installed in the AS-PIRe™ soil chamber to change its size accordingly to the requirements for theproject. Improvements for soil backfill material handling were developed and im-87plemented, including more efficient positioning of the band conveyor and betterforklift access to the soil chamber. The number of tests conducted in this researchproject would not have been possible without these upgrades, that drastically re-duced the time required for handling backfill material. The improvements willequally contribute to future research projects by providing efficient material han-dling.5.2 Axial pullout testsA series of axial pullout tests were conducted on buried DH pipes to study the soil-pipe interaction with different heating histories applied to the water mass insidethe pipe. This work formed the main component of this research project. To theauthor’s knowledge, this is the first time that experimental data is published on thissubject.Baseline tests with and without heating show good repeatability of the resultsamong tests with the same parameters and allow a comparison with steel pipes orsubsequently other types. Axial soil resistance of DH pipes was corrected for theweight of the soil mass inside the pipe, as soil-pipe interaction tests are mostlyconducted without product inside the pipe. For this project it was, however, nec-essary to fill the pipe with water mass in order to apply the heating and coolingphases. Tests with one and three full cooling and heating cycles were conducted toprovide data on potential arching effects reducing axial resistance. Strain gaugeswere attached to the HDPE surface of the pipe in the final test to provide data onthe expansion and contraction of the pipe at the soil-pipe interface with changingtemperatures of the water mass.It could be shown that changes in the temperature of the water mass have a88significant influence on axial pullout resistance of the DH pipe. After heating thewater mass by ∆T = 50 °C, large-strain resistance increased by roughly 15 % com-pared to the control tests. While only one full cooling and heating cycles did nothave a noticeable effect, three cycles reduced the axial soil resistance of the pipe,potentially due to aforementioned arching. Arching means the plastic deformationin the surrounding soil mass, which causes a decrease in normal stress on the pipe.Monitoring strain during one test gave a first insight into the mechanical mech-anisms involved at the soil-pipe interface. Considerable strain was measured bothin axial and radial direction, the latter translating to an increase in pipe circumfer-ence of 1.4 mm. Based on the development of strain with the heating history it wasinferred that the expansions at the pipe surface result from a combination of strainsfrom both the steel pipe at the core and the HDPE cover. Consequently, DH pipeshave to be treated as a complete system in order to accurately model their mechan-ical behaviour under thermal load. Results from strain gauge measurements showfurthermore that the soil mass is an integral part of this system as it substantiallyinfluences the heat transfer from the soil-pipe interface. This important findinghighlights the importance of conducting full-scale axial pullout tests with buriedDH pipes.The test results presented illustrate the significance of full-scale testing in re-gard to understanding the soil-pipe interaction of DH pipe systems under cyclicthermal loading. The findings can be used to explain and predict axial loads expe-rienced by buried DH pipes, and consequently contribute to designing more cost-effective and durable district heating (DH) pipe networks.895.3 Lateral pullout testsLateral resistance is especially relevant in corners of DH pipe networks, whereaxial expansions are translated to lateral movement of the adjacent pipe section.The tests of lateral resistance prepared in this study should also contribute to thedatabase of normalized resistances of different pipe materials.A small number of lateral pullout tests were conducted with DH pipes of twodifferent diameters. It was shown that the maximum resistance is comparable withsteel pipes, but with a flatter initial loading curve for DH pipes.5.4 Direct shear tests for interface frictionDirect shear tests were conducted to obtain data on the interface friction betweenthe district heating (DH) pipes used in this study and Fraser River sand, the chosensoil backfill material, as no published data was available on this. The test resultswere used to evaluate results from the full-scale tests in this project.5.5 Recommendations for future researchThe conclusions described are only based on a limited number of tests due to theexploratory nature of this project. It is expected that further investigations couldcontribute to higher confidence and bring more insights into the various mecha-nisms involved in soil-pipe interaction of DH pipes under cyclic thermal load. Ad-ditional tests under controlled conditions in a soil chamber can confirm the resultsof this first project and expand the test database with different parameters like pipediameter, texture of the pipe surface, temperature increase, number of cycles orduration of cycles. Measurements of strain seem particularly promising to provideexplanations of pipe behaviour in the light of experiences gained in this project.90Production systems buried in the ground could be instrumented to gather data thatwould not be possible to be obtained from laboratory testing as a result of limita-tions in time and space. Strains and movements in pipe networks with complicatedgeometries could be monitored during several full seasonal temperature cycles.The work presented demonstrates the value of full-scale testing in understand-ing the soil-pipe interaction mechanisms due to cyclic thermal load. Future re-search would further improve the knowledge base available for predictions of loadsand movements of pipes. Accurate predictions of behaviour in various situationsthroughout the lifespan are essential to improve the way pipe networks are de-signed. Better cost-effectiveness and durability can help to make district heating(DH) as a very sustainable option more competitive in the market of heating sys-tems.91BibliographyAchmus, M. (1995). Zur Berechnung der Beanspruchungen und Verschiebungenerdverlegter Fernwa¨rmeleitungen. PhD thesis, Institut fu¨r Grundbau,Bodenmechanik und Energiewasserbau (IGBE) Leibniz Universita¨t Hannover.→ pages 14AGFW - Der Energieeffizienzverband fu¨r Wa¨rme, Ka¨lte und KWK e. V. (1983).Kunststoff-Verbundmantelrohre fu¨r Fernwa¨rmeleitungen. → pages 16American Lifeline Alliance (2001). Guidelines for the Design of Buried SteelPipe. → pages 7, 11American Society of Civil Engineers (1984). Guidelines for the seismic design ofoil and gas pipeline systems. → pages 4, 7, 10, 11Anderson, C. (2004). Response of buried polyethylene natural gas pipelinessubjected to lateral ground displacement. Master’s thesis, The University ofBritish Columbia. → pages 10, 23, 34Anderson, C., Wijewickreme, D., Ventura, C., and Mitchell, A. (2005). Full-scalelaboratory testing of soil-pipe interaction on branched polyethylene pipelines.Experimental Techniques, 29(2):33–37. → pages 10Audibert, J. M. and Nyman, K. J. (1977). Soil restraint against horizontal motionof pipes. Journal of the Geotechnical Engineering Division,103(10):1119–1142. → pages 11Bilgin, O¨., Stewart, H. E., and ORourke, T. D. (2007). Thermal and mechanicalproperties of polyethylene pipes. Journal of Materials in Civil Engineering,19(12):1043–1052. → pages 14Coˆte´, J. and Konrad, J.-M. (2005). A generalized thermal conductivity model forsoils and construction materials. Canadian Geotechnical Journal,42(2):443–458. → pages 1992D. G. Honegger Consulting, C-CORE and SSD Inc. for the Design, Construction& Operations Technical Committee of Pipeline Research Council InternationalInc. (PRCI) (2009). Guidelines for constructing natural gas and liquidhydrocarbon pipelines in areas subject to landslide and subsidence hazards.Catalog No. L52292(V). → pages 4Dalla Rosa, A., Li, H., and Svendsen, S. (2011). Method for optimal design ofpipes for low-energy district heating, with focus on heat losses. Energy,36(5):2407–2418. → pages 6Duncan, J. M. and Chang, C.-Y. (1970). Nonlinear analysis of stress and strain insoils. Journal of the Soil Mechanics and Foundations Division,96(5):1629–1653. → pages 15Environment Canada (2014). Daily data report for february 2014. → pages 63Groves, J. (2014). A field investigation to study the response of buriedpolyethylene natural gas pipelines subjected to ground movement. Master’sthesis, The University of British Columbia. → pages 32Guo, P. and Stolle, D. (2005). Lateral pipe–soil interaction in sand with referenceto scale effect. Journal of geotechnical and geoenvironmental engineering,131(3):338–349. → pages 12Hansen, J. B. and Christensen, N. (1961). The Ultimate Resistance of Rigid PilesAgainst Transversal Forces; Model Tests with Transversally Loaded Rigid Pilesin Sand. Geoteknisk Institut, Denmark. → pages 11Itasca, FLAC (2007). Fast lagrangian analysis of continua (flac). ItascaConsulting Group Inc., Minneapolis. → pages 9Kalinci, Y., Hepbasli, A., and Tavman, I. (2008). Determination of optimum pipediameter along with energetic and exergetic evaluation of geothermal districtheating systems: modeling and application. Energy and Buildings,40(5):742–755. → pages 6Karimian, S. A. (2006). Response of buried steel pipelines subjected tolongitudinal and transverse ground movement. PhD thesis, The University ofBritish Columbia. → pages 8, 10, 12, 23, 24, 26, 34, 48, 49, 74Li, X.-l., Duanmu, L., and Shu, H.-w. (2010). Optimal design of district heatingand cooling pipe network of seawater-source heat pump. Energy and Buildings,42(1):100–104. → pages 693Monroy-Concha, M. (2013). Soil restraints on steel buried pipelines crossingactive seismic faults. PhD thesis, The University of British Columbia. → pages23, 79, 81O’Rourke, T., Druschel, S., and Netravali, A. (1990). Shear strengthcharacteristics of sand-polymer interfaces. Journal of GeotechnicalEngineering, 116(3):451–469. → pages 13, 48Ovesen, N. K. (1964). Anchor slabs, calculation methods and model tests.Geoteknisk Institut, Denmark. → pages 11Smits, K. M., Sakaki, T., Limsuwat, A., and Illangasekare, T. H. (2010). Thermalconductivity of sands under varying moisture and porosity in drainage–wettingcycles. Vadose Zone Journal, 9(1):172–180. → pages ix, 19, 20Stewart, H., Bilgin, O¨., ORourke, T., and Keeney, T. (1999). Technical referencefor improved design and construction practices to account for thermal loads inplastic gas pipelines. Final Rep. No. GRI-99/0192, Gas Research Institute,Chicago. → pages 11Trautmann, C. H. and O’Rourke, T. D. (1985). Lateral force-displacementresponse of buried pipe. Journal of Geotechnical Engineering,111(9):1077–1092. → pages 11von Wolffersdorff, P.-A. (1996). A hypoplastic relation for granular materialswith a predefined limit state surface. Mechanics of Cohesive-frictionalMaterials, 1(3):251–271. → pages 17Weerasekara, L. (2007). Response of buried natural gas pipelines subjected toground movement. Master’s thesis, The University of British Columbia. →pages 10, 12, 23Weidlich, I. (2008). Untersuchung zur Reibung an zyklisch axial verschobenenerdverlegten Rohren. PhD thesis, Institut fu¨r Grundbau, Bodenmechanik undEnergiewasserbau (IGBE) Leibniz Universita¨t Hannover. → pages 16, 17, 18,21, 48Weidlich, I. and Achmus, M. (2006). Reduction of friction forces between soiland buried district heating pipes due to cyclic axial displacements. 10thInternational Symposium on District Heating and Cooling, Hannover,Germany. → pages ix, 16, 1894Weidlich, I. and Wijewickreme, D. (2012). Factors influencing soil friction forceson bruied pipes used for district heating. 13th International Symposium onDistrict Heating and Cooling, Copenhagen, Denmark. → pages 3, 14Wijewickreme, D., Karimian, H., and Honegger, D. (2009). Response of buriedsteel pipelines subjected to relative axial soil movement. CanadianGeotechnical Journal, 46(7):735–752. → pages 8, 9, 18, 71, 72, 7395Appendix ATemperature Histories of theWater Mass9601020304050600 5 10 15 20 25 30Temperature(◦C)Time (h)water tempFigure A.1: Temperature history of the water mass during Test No. A-520-H1-1st.01020304050600 5 10 15 20 25 30Temperature(◦C)Time (h)water tempFigure A.2: Temperature history of the water mass during Test No. A-520-H1-2nd.9701020304050600 20 40 60 80 100Temperature(◦C)Time (h)water tempFigure A.3: Temperature history of the water mass during Test No. A-520-H1C1.01020304050600 50 100 150 200 250Temperature(◦C)Time (h)water tempFigure A.4: Temperature history of the water mass during Test No. A-520-H3C3.980102030405060700 10 20 30 40 50Temperature(◦C)Time (h)water tempFigure A.5: Temperature history of the water mass during Test No. A-520-H1C1-S.99


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