Research on Drillability Prediction of Shale Horizontal Wells Based on Nonlinear Regression and Intelligent Optimization Algorithm

Shale oil and gas reservoirs are characterized by low porosity and low permeability. The development of ultra-long horizontal wells can significantly increase reservoir contact area and enhance single-well production. Shale formations exhibit distinct bedding structures, high formation pressure, high rock hardness, and strong anisotropy. These characteristics result in poor drillability, slow drilling rates, and high costs when drilling horizontally, severely restricting efficient development. Therefore, accurately predicting the drillability of shale gas wells has become a major challenge. Currently, most scholars rely on a single parameter to predict drillability, which overlooks the coupled effects of multiple factors and reduces prediction accuracy. To address this issue, this study employs drillability experiments, mineral composition analysis, positional analysis, and acoustic transit-time tests to evaluate the effects of mineral composition, acoustic transit time, bottom-hole confining pressure, and formation drilling angle on the drillability of horizontal well reservoirs, innovatively integrating multiple parameters to construct a nonlinear model and introducing three intelligent optimization algorithms (PSO, AOA-GA, and EBPSO) for the first time to improve prediction accuracy, thus breaking through the limitations of traditional single-parameter prediction. Based on these findings, a nonlinear regression prediction model integrating multiple parameters is developed and validated using field data. To further enhance prediction accuracy, the model is optimized using three intelligent optimization algorithms: PSO, AOA-GA, and EBPSO. The results indicate that the EBPSO algorithm performs the best, followed by AOA-GA, while the PSO algorithm shows the lowest performance. Furthermore, the model is applied to predict the drillability of Well D4, and the results exhibit a high degree of agreement with actual measurements, confirming the model’s effectiveness. The findings support optimization of drilling parameters and bit selection in shale oil and gas reservoirs, thereby improving drilling efficiency and mechanical penetration rates.