Research on the Prediction Model of Pressure Crossover in Changqing Chang-7 Shale Oil Horizontal Wells Based on BP Neural Network

In addressing the challenges of predicting channeling in horizontal wells of Chang 7 shale oil reservoirs in the Changqing Oilfield, traditional prediction methods often fall short. These methods encounter issues such as inadequate accuracy and limited applicability, making it challenging to meet the intricate geological and engineering requirements specific to the Chang 7 shale oil reservoirs in the Changqing Oilfield. This study integrates key geological-engineering parameters, including nine indicators such as reservoir permeability, porosity, shale content, fracturing interval spacing, and total fluid volume. By doing so, we establish an optimized network structure featuring a momentum factor and adaptive learning rate. This approach enhances clarity and precision in our academic discourse. Following the training of 26 datasets related to fracturing, the model accurately identified five non-training cases during verification. This highlights the model's superior prediction accuracy. Consequently, it serves as an intelligent decision-making tool, offering both high prediction accuracy and practical applicability, thereby facilitating the effective development of shale oil resources.