Abstract
Under the "dual carbon" goal, the mismatch between the intermittent nature of wind-solar power generation and the stable energy demand of oil wells hinders efficient green energy utilization in oilfields, leading to low green electricity consumption, high curtailment rates, and poor economic benefits. To address this challenge, an optimization approach for oil well operation scheduling is proposed, which couples photovoltaic power fluctuations with the characteristics of intermittent pumping technology. Firstly, a multiobjective optimization model was developed to minimize grid electricity consumption per unit of liquid production and maximize the share of green electricity. The on-off schedule was mapped into a constrained binary sequence using a run-length encoding to reducing the solution space. Non-dominated Sorting Genetic Algorithm II (NSGA-II) was improved to increase both accuracy and convergence speed by introducing: (1) dual-mode initialization strategy guided by historical operating schedules and photovoltaic fluctuations. (2) a key-gene-preserving crossover operator to retain high-matching time segments; and (3) a peak-valley-guided mutation strategy to enable dynamic pruning of the solution space and goal-directed optimization. Case studies showed that the proposed method doubled green electricity consumption under stable production conditions, reduced grid electricity consumption per unit liquid by 41.67%, improved computational efficiency by two to three orders of magnitude, and enhanced the solution accuracy by 26.69%, indicating strong practical applicability.