Volume 66

Research on Multi-parameter Cooperative Adjusting Method for Electric Submersible Pump Layered Production Wells Based on PID-Bayesian Optimization Qi Wei1, Chaodong Tan 1,2,*, Xudong Guan 3, Yuxun Sun2, Yaoyao Hua1

Abstract

The production adjustment of ESP layered production wells in offshore oilfields relies on manual measurement and adjustment. The intricate interplay of layered completion systems, lifting equipment, and formation energy results in nonlinear production allocation. This complexity impedes multi-layer collaborative control, extends operation time, and presents challenges in maintaining optimal ESP efficiency.Therefore, this study takes maximizing ESP operating efficiency and minimizing production allocation errors as the cooperative decision objectives. It utilizes position PID to explore the target distribution space, and based on the Bayesian probabilistic information inference algorithm(PID-BO), introduces an uncertainty filtering strategy to establish a multi-parameter cooperative adjusting method for ICV openings and ESP operating frequency. Taking offshore layered production well A-01 as the experimental object, PID-BO can quickly narrow the search space based on limited historical production data, identifying the optimal joint adjustment strategy of ESP frequency and ICV openings within only 20 exploration rounds. The maximum production allocation error of the three production layers does not exceed 5.31%, with the pump efficiency of ESP reaching 63.15%. The comprehensive index of the strategy is significantly higher than that of traditional methods. Verification results demonstrate that this method ensures fine production and high-efficiency operation of ESP layered production wells, effectively improves the level of fine oilfield development, reduces energy consumption losses in oil and gas exploitation, and holds important practical significance for the efficient and energy-saving development in the field of energy exploitation.

Keywords Offshore oilfield; Layered oil production; Cooperative Decision-making; Bayesian optimization; Energy consumption reduction