Prediction of Strength Properties of Soft Soil Using Machine Learning Techniques

In civil engineering projects, the strength of soil, particularly its cohesion, is pivotal for the stability of building foundations and slopes. Traditionally, determining cohesion (c) involves labor- intensive methods such as unconfined compression tests, direct shear tests, and triaxial tests, which require collecting soil samples. However, these methods are often constrained by time and cost considerations, exacerbated by the diverse nature of soil types. This research initiative aims to introduce a simplified approach for assessing the cohesion strength parameter of cohesive soil. Our proposal entails leveraging statistical correlations and machine learning techniques to establish connections between soil properties such as liquid limit, plastic limit, moisture content, % fine particle content, and the strength parameter. These laboratory tests are comparatively straightforward, rapid, and cost-effective when juxtaposed with conventional methodologies.