Objective Given that the rate of penetration (ROP) serves as a key indicator of drilling efficiency, constructing an accurate ROP model holds great significance for optimizing drilling processes and reducing drilling costs. However, deep geological drilling faces challenges such as nonlinearity, non-convex optimization, multiple operating conditions, and temporal variations. Consequently, traditional modeling methods are difficult to adapt to complex geologic environments.

Methods To address these challenges, this study proposed a fusion method combined with temporal regulation for ROP modeling: the SVR-MDBO method. Initially, a basic ROP model was constructed using support vector regression (SVR) to solve the nonlinear problem caused by ROP changes. To solve the non-convex optimization problem encountered in model parameter design, a modified dung beetle optimizer (MDBO) algorithm was designed through weight fusion, modified echolocation, modified iterated local search, and the re-updating strategy of the optimal solution. To adapt to the temporal variations of the ROP, a temporal regulation method based on fuzzy C-means clustering and the Mann-Kendall trend test was employed to conduct the temporal regulation of the model output.

Results and Conclusions  The results indicate that the MDBO algorithm yielded satisfactory results in the tests of 11 benchmark functions, suggesting that the MDBO algorithm can effectively solve the problem encountered in model parameter design. The simulation results based on actual drilling data demonstrate that the ROP model constructed in this study achieved optimal results in two well sections. Post-temporal regulation, the ROP model yielded more accurate predicted trends for both well sections, with respective prediction accuracy reaching up to 80% and 87.5%. The tests of the microdrilling experimental system reveal that the constructed ROP model yielded the highest accuracy under different rock samples. Overall, the constructed ROP model can effectively cope with changes in complex geologic environments, laying a solid foundation for controlling the process of deep geological drilling.