Centrifuge modeling of energy foundations – effect of seasonal temperature fluctuation and soil state

Energy foundations can reduce carbon footprint and lead to energy savings. Understanding the process of heat migration in soil is therefore of great interest in the field of geotechnical engineering. However, limited literature exists on the thermo-dynamic interaction and structural performance of an energy foundation and the surrounding soil. Lab testing of energy foundations remains inexpensive compared to field tests; however lab tests cannot easily simulate representative in-situ stress conditions. This can be resolved by using a geotechnical centrifuge to correctly mimic the stress gradient of the self-weight of the soil. A series of centrifuge models of energy foundations in different soil states (dense and loose) have been tested. Representative seasonal temperature fluctuations in the UK are used as the benchmark in this study. A total of three years of heating/cooling cycles are modelled and foundation response is captured by means of embedded temperature sensors at different distances away from the thermal foundation. latant behaviour. Mitchell 1964, Plum & Esrig (1969), Habibagahi (1977), and Boudali et al. (1994) also reported that the compression curves obtained at different temperatures are parallel, with lower values of void ratio at higher temperatures. The aforementioned changes in stress and strength could have considerable implications on the thermo-mechanical response of foundations (i.e. energy piles) deployed as energy structures; such that, under working stress, deterioration of stability and serviceability could manifest leading to uncertainty in long term performance (Black et al. 2015). With this in mind, a series of centrifuge tests were conducted at the Centre for Energy and Infrastructure Ground Research (CEIGR) to investigate the heat migration phenomena in soils of different density and over 3 heating/cooling cycles for a single energy pile. 2 CENTRIFUGE EXPERIMENTS 2.1 Centre for Energy and Infrastructure Ground Research (CEIGR) The centrifuge used for this investigation was the newly established University of Sheffield 50gT geotechnical beam centrifuge located in the Centre for Energy and Infrastructure Ground Research. The centrifuge was designed and manufactured by Thomas Broadbent and Sons Limited, United Kingdom, and commissioned in 2014. The centrifuge beam has a radius of 2 m to the base of the swing platform, of plan area 0.8 m 2 , and can accelerate a 500 kg payload to 100 gravities. 2.2 Centrifuge scaling laws Geotechnical centrifuge modeling uses centrifugal acceleration to increase the self-weight stresses in a small model to equal the self-weight stresses in a large prototype. If a soil model containing the same soil as the prototype is spun at centrifugal acceleration of N times the prototype gravity, the vertical stress in a soil layer at depth h in model scale is identical to the vertical stress in the prototype soil layer of depth N*h. The length scaling factor (model: prototype) is therefore 1:N. Once the length scaling factor is determined, the scaling factor of volume, force, and strain can be calculated to be 1:N 3 , 1:N 2 and 1:1 correspondingly. Temperature does not vary with the increased body forces in a centrifuge; Krishnaiah and Singh (2004) confirmed that the centrifugation of a heatflow model does not change the heat flow process (Stewart and McCartney 2014). If the thermal conductivity of the soil model and prototype are assumed to be similar, and if dimensions associated to spatial distribution of heat flow are scaled from model to prototype, then the conduction time is N 2 times faster in the centrifuge model (Stewart & McCartney 2014, Savvidou 1988, Krishnaiah & Singh 2004, Haigh 2012). Since the D50 of the sands (shown in next section) used in the centrifuge tests were largely smaller than 1-2mm, the influence of convection on the heat flow can be assumed to be negligible and neglected (Krishnaiah & Singh 2004, Johansen 1975, and Farouki 1986).