Abstract: In the context of the global energy transition, renewable energy has become a key driver of the energy system transformation, leading the traditional oil and gas industry to explore pathways for deep integration with new energy sources. However, the intermittency of wind and solar power presents a poor match with the energy demand profiles across the entire shale oil development cycle, which limits the large-scale application of clean energy. This paper proposes an energy management optimization method driven by the integration of wind, photovoltaic, geothermal, and grid energy, which accurately characterizes the energy consumption variations throughout the shale oil development cycle. A coupled wind-photovoltaic-geothermal power generation model based on real-time meteorological data is developed, while establishing a multi-energy collaborative optimization system and a real-time dynamic scheduling method based on deep reinforcement learning. The results show that:(1)Wind, photovoltaic, and geothermal power forecasts exhibit excellent resource complementarity. Wind and solar resources exhibit distinct seasonal characteristics, while geothermal power provides a stable and reliable baseload supply, with low fluctuation and gradual degradation. (2)The system implements stage-specific strategies:Leveraging active grid power supply during the short-duration, high-energy consumption fracturing phase; utilizing time-of-use pricing mechanisms to export energy during high-price periods in the shut-in stage; prioritizing energy storage during low-price periods to reserve low-cost electric power for the subsequent high-load stages in the depletion development stage; optimizing real-time supply and storage dispatch by deep reinforcement learning to achieve economic operations through peak-shaving and valley-filling in the long-term high-energy consumption artificial lift phase. (3)Taking a Gulong shale oil block in Daqing as an example, the optimized wind-photovoltaic-geothermal-grid integrated system not only meets over 85 % of the total electricity demand but also reduces overall costs by 24.1 % compared to traditional methods. This approach achieves the strategy of "local adaptation, intelligent allocation, and on-site consumption" for clean energy.

Key words: shale oil development, renewable energy, integration of wind, photovoltaic, geothermal and grid energy, deep reinforcement learning, integrated energy system

摘要： 在全球能源转型的背景下,可再生能源已成为引领能源系统变革的重要力量,传统油气行业正积极探索与新能源深度融合的发展路径。但风、光等可再生能源的间歇性特征与页岩油开发全周期各阶段能耗需求适配差,制约了清洁能源的规模化应用。通过准确描述页岩油开发全周期能耗差异特点,构建了耦合实时气象信息的风-光-热发电模型,提出了风-光-热-电融合驱动的页岩油开发能源管理优化方法,形成了基于深度强化学习的多能协同优化系统与实时动态调度方法。研究结果表明:①风力、光伏和地热发电预测结果展现出良好的资源互补特性,风、光资源具有显著季节性特征,地热发电热源稳定,输出功率波动性小且衰减缓慢,可为系统提供可靠基荷电源供应。②针对页岩油压裂阶段"短时、高能耗"的特点借助电网主动供电;焖井阶段能耗极低,系统利用分时电价机制在高电价时段积极对外供能;衰竭开发阶段持续较低能耗,系统根据能耗预测优先在低价时段储能,为高负荷阶段储备低成本电力;人工举升阶段长期高能耗运行,通过深度强化学习算法优化供-储实时调配,实现"削峰填谷"经济运行。③以大庆古龙现场区块页岩油为例,优化配置的风-光-热-电协同系统可满足超85%的总用电需求,相比传统方法降低综合成本24.1%,实现清洁能源的"因地制宜、智能分配、就近消纳"。

关键词: 页岩油开发, 可再生能源, 风-光-热-电融合, 深度强化学习, 综合能源系统