Thermal Conductivity of Cuddapah and Kurnool Limestone Rocks: Prediction and Analysis Using MVR, Sensitivity Analysis, and K-Fold Cross Validation

Thermal Conductivity (TC) is a critical physical property of rocks, essential for applications such as geothermal modelling, radioactive waste disposal, building insulation, underground construction, and various civil and mining operations. This paper presents an experimental study of the measurements of thermal conductivity and other physical properties of limestone rock samples of the Cuddapah and Kurnool regions. The physical properties, including TC, density, porosity, and P-wave velocity, were measured for each sample. A significant correlation was identified between TC and the intrinsic rock properties of porosity, density, and P-wave velocity. This paper focuses on developing and validating predictive models for TC using Multivariate Regression (MVR) analysis and K-fold cross-validation. The determination coefficient (R²) for the MVR model was 0.95, while the K-fold cross-validation model showed a performance R² of 0.92, indicating the high accuracy of the MVR model in predicting TC. The thermal conductivity data obtained from this research establishes an initial database for rocks from the Cuddapah and Kurnool regions, serving as a valuable resource for designing thermal management systems and engineering projects in these areas.