期刊
JOURNAL OF APPLIED GEOPHYSICS
卷 58, 期 2, 页码 112-129出版社
ELSEVIER SCIENCE BV
DOI: 10.1016/j.jappgeo.2005.05.002
关键词
thermal conductivity; opalinus clay; numerical simulation; optimization; neural network
Storage in deep geological formations is a potential solution for the management of high-level radioactive wastes. In this context, different types of rocks such as argillite are extensively studied. In the Mont Terri underground laboratory (Switzerland), several experiments have been performed in order to characterize the properties of the opalinus clay. One of these experiments, called HE-C, has consisted in measuring in situ the time evolution of the rock temperature submitted to a heating source. Experimental measurements have shown that the thermal behaviour of the clay was not homogeneous around the borehole where the heater was installed. Furthermore, 3D direct numerical simulations of this experiment performed with the code Cast3M have proved that it was necessary to introduce a new parameter alpha to model the amount of electric power lost in cables and by air convection inside the metallic tube containing the heater. A numerical simulation-optimization technique has been used to estimate the thermal longitudinal and transverse conductivities (lambda(//) and lambda(perpendicular to)) of the host rock. It consists in minimizing an objective function that is the sum of the squared differences between measured and calculated temperatures. But this method induced a lot of Cast3M simulations. In order to drastically reduce the CPU time, we used a neural network approximation built from a sample training of I 100 Cast3M simulations. It allowed us to calculate the objective function for 500000 different values of the triplet (lambda(//),lambda(perpendicular to),alpha). Finally, we obtained the following values for the thermal conductivities on one side of the borehole, lambda(//) = 1.84 +/- 0.04 W m(-1) K-1 and lambda(perpendicular to) = 0.55 +/- 0.03 W m(-1) K-1; on the other side, lambda(//) = 1.90 +/- 0.07 W m(-1) K-1 and lambda(perpendicular to) = 1.07 +/- 0.09 W m(-1) K-1. The estimated thermal conductivities lambda(perpendicular to) perpendicular to the bedding plane are quite different. It is perhaps caused by the presence of an intensive fractured zone on one side of the borehole, due to bentonite swelling. it can also be due to the presence of a bed of carbonated rock. (C) 2005 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据