A New Method for Intelligent Prediction of Drilling Overflow and Leakage Based on Multi-Parameter Fusion

The technical focus of drilling operations is changing to oil and gas reservoirs with higher difficulty factors such as low permeability fracture. During the process, in deep complex formations are prone overflow leakage complications. Leakage problems will change performance fluid wellbore, impacting wellbore pressure, causing accidents stuck collapse. In order improve level control over risk leakage, it necessary predict mud situation arrange that may occur advance ensure safety drilling. By using a genetic algorithm optimize multi-layer feedforward neural network, this paper establishes GA-BP Neural Network Drilling prediction model based on multi-parameter fusion. Through optimization training 14 parameters affect occurrence downhole accidents, predicted. results compared conventional BP verified by real data. show MAE, MSE, RMSE network improved 2.91%, 4.48%, 10.93%, respectively, model, quality higher. Moreover, amount predicted matches well pattern data drilling, which proves effectiveness accuracy prediction.