Artificial neural network models for predicting electrical resistivity of soils from their thermal resistivity

The knowledge of soil electrical and thermal resistivities is essential for several engineering projects such as laying of high voltage buried power cables, nuclear waste disposal, design of fluidized thermal beds, ground modification techniques etc. This necessitates precise determination of these resistivities, and relationship between them, which mainly depend on the soil type, its origin, compaction density and saturation. Such a relationship would also be helpful for determining one of these resistivities, if the other one is known. With this in view, efforts were made to develop artificial neural network (ANN) models that can be employed for estimating the soil electrical resistivity based on its soil thermal resistivity and the degree of saturation. To achieve this, measurements of electrical and thermal resistivities were carried out on different types soils compacted at different densities and moisture contents. These models were validated by comparing the predicted results vis-à-vis those obtained from experiments. The efficiency of these ANN models in predicting the soil electrical resistivity has been demonstrated, if its thermal resistivity is known. These ANN models are found to yield better results as compared to the generalized relationships proposed by the earlier researchers.