Research on a Shale Gas Fracturing Casing Deformation Early Warning Model Based on Multi-Parameter Fusion of Microseismic Data: A Case Study of the Luzhou Block, Sichuan

To effectively predict casing deformation during shale gas hydraulic fracturing, this study leverages microseismic monitoring data from 85 hydraulically fractured wells in the Luzhou block, Sichuan, to investigate the relationship between microseismic activity characteristics and casing deformation. The results reveal that well segments with casing deformation are consistently associated with microseismic events exhibiting higher magnitudes and pronounced spatial clustering. Based on these findings, a random forest classification model was developed, integrating multidimensional microseismic features, including event magnitude, event density, spatial clustering situations, and angle to the fracturing well trajectory, to identify high-risk well segments. Validation demonstrates that the model achieves a predictive accuracy of 89.8% on the test dataset. Furthermore, well segments subjected to intervention based on the model’s predictions and adjusted operational parameters exhibited significantly improved wellbore integrity, with preservation rates reaching up to 98%. This study provides theoretical support and field-validated evidence for the application of microseismic monitoring in risk management during shale gas fracturing, offering robust adaptability and promising potential for broader implementation.