RCPTU Evaluation and Artificial Neural Network Prediction of Diesel Contamination in Silty Clay: A Study of Resistivity Response

Abstract To validate the application of in-situ Resistivity Cone Penetration Testing (RCPTU) in assessing organic-contaminated soil sites, correlation analysis was conducted between RCPTU and the Wenner Four-Electrode Method. Using typical silty clay from Nanjing, soil samples with varying moisture and oil contents were prepared using 0# diesel for both RCPTU tests and Wenner Four-Electrode To validate the application of in-situ Resistivity Cone Penetration Testing (RCPTU) in assessing organic-contaminated soil sites, correlation analysis was conducted between RCPTU and the Wenner Four-Electrode Method. Soil samples with varying moisture and oil contents were prepared using typical silty clay from Nanjing and 0\# diesel for both RCPTU tests and Wenner Four-Electrode resistivity experiments. The results demonstrate that resistivity decreases with increasing moisture content under constant oil conditions and increases with rising oil content under constant moisture conditions. Diesel-contaminated soil exhibited higher resistivity than uncontaminated soil. The strong linear relationship between RCPTU and indoor four-electrode test results confirms the effectiveness of RCPTU in assessing contaminated sites. Further validation through oil-water adsorption experiments and microscopic analyses revealed that increasing oil content enhances particle encapsulation by diesel, thereby impeding cation transport and altering resistivity. Integrating RCPTU data with Artificial Neural Network (ANN) models enabled more accurate predictions of diesel concentration in soil, demonstrating the potential of machine learning to enhance prediction precision and efficiency.