Machine learning enhanced formation pressure prediction using integrated well logging and mud logging

The difficulty of accurately predicting abnormally high-pressure formation pressure is one of the critical challenges in the field of petroleum engineering. Due to the low accuracy of formation pressure prediction and the narrow drilling safety density window, accidents such as leakage and blowout occur frequently. To address this issue, improving the accuracy of pore pressure predictions is essential. The well logging and mud logging data were combined to analyze the correlation between various parameters. Analysis using the Spearman correlation coefficient revealed that pore pressure exhibits varying correlation relationships with different parameters. Pore pressure is closely related to factors such as depth, weight of hanging, and mud weight. Pore pressure has a medium to high correlation with the rate of penetration, weight on bit, torque, slurry pump pressure, acoustic time difference, density, and volume of clay. Pore pressure has a medium to low correlation with the rotation per minute. Based on machine learning algorithms and a large amount of known data, a machine learning formation pressure model with integrated well logging and mud logging data (IWM) was established. The prediction results of traditional models and IWM models were compared using neighboring wells as the prediction targets. The results indicate that the backpropagation neural network model based on a genetic algorithm and IWM (IWM-GABP) achieves the highest prediction accuracy, with an average prediction accuracy greater than 96%. When predicting formation pressure, it is advisable to use the back propagation neural network model based on IWM or the IWM-GABP model, rather than the radial basis function neural network model based on IWM. The IWM model significantly reduces the prediction error of formation pore pressure, achieving an average improvement of 8.32% enhancement in prediction accuracy compared to traditional data models. The research method effectively improves the accuracy of formation pressure prediction and provides support for efficient on-site development.