This paper takes the representative buried structure in permafrost regions, a transmission line tower foundation, as the research object. An inverse prediction is conducted in a scaled-down experimental system mimicking actual heat conduction of the frozen ground in a tower foundation. In permafrost regions, global warming and the heat transfer through the buried structures bring significantly adverse thermal effects on the stability of the infrastructures. In modeling the thermal effects, it has been a challenge to determine the ground surface boundary condition and heat source strength from the buried structures due to the complex climate and environmental conditions. In this work, based on the improved model predictive inverse method with an adaptive strategy, an inverse scheme is successfully implemented to simultaneously identify the time-varying surface temperature and the time-space-dependent heat source representing the buried structures. In this scheme, an adaptive time-varying predictive model is established by the rolling update of the sensitivity response coefficients according to the predicted temperature field to overcome the influence of nonlinear characteristics in the heat transfer process. The inverse method is verified by simulation and experimental data. According to the experimental inversion results, the reconstructed temperature distribution efficiently predicts the thermal state evolution of the permafrost foundation under seasonal freezing and thawing. It is found that, under the experimental conditions, the intensified thawing and freezing are significantly severe, e.g., the increased area ratio of active layer thickness is as high as 28% after building a tower, and the depth of permafrost table ranges from about 14 mm to about 38 mm, which could be detrimental to the stability and safety of the tower foundation. This study will provide valuable guidance for risk assessments or optimizing the design and maintenance of the real tower foundation and similar buried structures.

The assertion that pure conductive heat transfer always dominates in cold climates is at odds with decades of research in soil physics which clearly demonstrate that non-conductive heat transfer by water and water vapor are significant, and frequently are for specific periods the dominant modes of heat transfer near the ground surface. The thermal regime at the surface represents the effective boundary condition for deeper thermal regimes. Also, surface soils are going to respond more quickly to any climatic fluctuations; this is important to us because most facets of our lives are tied to earth's surface. To accurately determine the surface thermal regime (for example, the detection of climate change), it is important to consider all potential forms of heat transfer. Gradients that have the potential to alter the thermal regime besides temperature include pore water pressure, gravitational, density, vapor pressure and chemical. The importance of several non-conductive heat transport mechanisms near the ground surface is examined. Infiltration into seasonally frozen soils and freezing (release of latent heat) of water is one mechanism for the acceleration of warming in surficial soils in the spring. Free convection due to buoyancy-induced motion of fluids does not appear to be an important heat-transfer mechanism; estimates of the Rayleigh number (the ratio of buoyancy to viscous forces) are generally around 2, which is too low for effective heat transfer. The Peclet number (ratio of convective to conductive heat transfer) is on the order of 0.25 for snowmelt infiltration and up to 2.5 for rainfall infiltration for porous organic soils. In mineral soils, both vertical and horizontal advection of heat can be neglected (Peclet number is approximately 0.001) except for snowmelt infiltration into open thermal contraction cracks. The migration of water in response to temperature or chemical,gradients from unfrozen soil depths to the freezing front, and the redistribution of moisture within the frozen soil from warmer depths to colder depths, can also result in heat transfer although this effect has not been quantified here. Many of these processes are seasonal and effective only during periods of phase change when the driving gradient near the ground surface is relatively large. (C) 2001 Elsevier Science B.V. All rights reserved.