Prediction of Strength Parameters of Thermally Treated Egyptian Granodiorite Using Multivariate Statistics and Machine Learning Techniques

The mechanical properties of rocks, such as uniaxial compressive strength and elastic modulus intact rock, must be determined before any engineering project by employing lab or in situ tests. However, there are some circumstances where it is impossible to prepare the necessary specimens after exposure high temperatures. Therefore, propensity estimate destructive parameters thermally heated rocks based on non-destructive factors a helpful research field. Egyptian granodiorite samples were temperatures up 800 °C being treated two different cooling methods: via oven (slow-cooling) using water (rapid cooling). condition, temperature, mass, porosity, absorption, dry density (D), P-waves used input predictive models for UCS E granodiorite. Multi-linear regression (MLR), random forest (RF), k-nearest neighbor (KNN), artificial neural networks (ANNs) create models. performance each prediction model was also evaluated (R2), (RMSE), (MAPE), (VAF). findings revealed that methods mass predict have minor impact In contrast, other had good relationship with E. Due severe damage samples, many output measure 600 °C. thus developed this threshold temperature. Furthermore, comparative analysis demonstrated ANN pattern predicting most accurate model, R2 0.99, MAPE 0.25%, VAF 97.22%, RMSE 2.04.