摘要
地热作为一种清洁能源具有广阔的应用前景,可持续地开发和利用地热资源中地热水的温度评估是重要的研究课题。人工智能技术已成为矿产和油气资源勘探开发研究的热点和前沿方向,然而在地热资源开发方面相关研究和应用较少。剖析了油气资源开发中大数据与人工智能应用的重要价值,对当前地热资源开发中人工智能技术的应用与探索进行了介绍。以陕西咸阳地热田为例,采用长短期记忆(long short-term memory,简称LSTM)神经网络构建了以灌定采模式下单井水温的时间序列模型;采用随机森林和XGBoost算法,建立了多个井地热水温度的预测模型。研究结果表明,建立的机器学习模型在地热水温度预测方面表现优秀,模型准确度均在95%以上,且速度快。该地区地热水温度的首要影响因素是取水段顶深,模型验证了渭北断裂带对热储的重要作用。实例应用验证了机器学习模型在解决地热资源开发复杂难题中的优越性,人工智能技术的合理应用能够为地热资源的高效开发和科学降本提质增效提供更多有效的决策依据。
Geothermal energy,as a clean energy source,has broad application prospects.The temperature assessment of geothermal water is a key topic in the sustainable development and utilization of geothermal resources.[Objective]Artificial intelligence(AI)technology has become a hotspot and frontier direction in the exploration and development of mineral,oil,and gas fields.However,studies on its application in geothermal field development are limited.This paper first discusses the significant value of big data and AI in oil and gas field development,and then reviews the current applications of AI in geothermal field development.[Methods]Taking the Xianyang geothermal field in Shaanxi Province as a case study,a time series model for the temperature of geothermal water in a single well was constructed using a long short-term memory(LSTM)neural network under the predetermined production mode.Additionally,the random forest and XGBoost algorithms were used to predict the groundwater temperature of multiple geothermal wells.[Results]The accuracy of the three models was above 95%,and their running speed was fast.The depth of the water intake section's top was found to be the primary influencing factor on geothermal water temperature in the area.The model also confirmed that WeiBei fault zones play an important role in heat storage.[Conclusion]The application of these models demonstrates the superiority of machine learning in addressing complex problems in geothermal field development.The reasonable application of AI can provide a more effective decision-making basis for the efficient development of geothermal fields,as well as for scientific cost reduction,quality improvement,and efficiency enhancement.
作者
董珮瑶
杜利
赵磊
包一凡
尹茂生
DONG Peiyao;DU Li;ZHAO Lei;BAO Yifan;YIN Maosheng(State Key Laboratory of Deep Geothermal Resources,Beijing 100083,China;SINOPEC Star(Beijing)New Energy Research Institute Co.,Ltd.,Beijing 100083,China;School of Environmental Science and Engineering,Southern University of Science and Technology,Shenzhen Guangdong 518055,China)
出处
《地质科技通报》
北大核心
2025年第3期388-398,共11页
Bulletin of Geological Science and Technology
基金
中国石化集团科研项目(JP23087,JP23178)。
关键词
地热开发
机器学习
模拟
地下水
水温预测
geothermal development
machine learning
modeling
groundwater
groundwater temperature prediction
