Prediction of Bearing Capacity of Thermally Treated Soft Clay Using Artificial Intelligence and Heated Borehole Method

Soft clay soils are known as problematic geomaterials due to their low bearing capacity, high compressibility, and large amounts of settlement, which greatly affect the stability and performance of civil engineering structures. This research explores an experimental and data-driven framework for investigating the strength of soft clay using the heated borehole method to improve soft clay’s bearing capacity by using artificial intelligence (AI) based modeling. To create thermal treatment within the soil mass, a field-simulated heating system was developed using locally available cooking gas. The effects of borehole spacing, borehole depth, borehole heating duration, and bores' pattern were systematically examined using laboratory model footing tests. The results show that thermal treatment improved the strength of the soil and the bearing capacity of the soil. The optimal conditions were found to be 3D spacing, 2b depth, and approximately 8 hours for heating duration. Compared to arrangements in a circular or triangular shape, the arrangement that is square produced the maximum improvement. Furthermore, predictive models such as GLMs, SVR, and polynomial were used in relation to the experimental dataset. Of these three predictive models, polynomial had the best results with a coefficient of determination greater than 0.98. Therefore, the predictive equations developed are a reliable method for estimating the bearing capacity of thermally treated soft clay, providing geotechnical engineers with a useful and efficient method of accomplishing their work.