基于BP神经网络的增强型地热系统产热性能:以伊通盆地花岗岩体热储层为例  

Heat Production Performance of Enhanced Geothermal System Based on BP Neural Network:Taking the Granite Thermal Reservoir in the Yitong Basin as an Example

作　　者：周玲 郭州朋 类红磊 张延军[4] 陈云娟[1,2] ZHOU Ling;GUO Zhou-peng;LEI Hong-lei;ZHANG Yan-jun;CHEN Yun-juan(School of Civil Engineering,Shandong Jianzhu University,Jinan 250101,China;Key Laboratory of Building Structural Retrofitting and Underground Space Engineering(Shandong Jianzhu University),Jinan 250101,China;Shandong Jianda Engineering Appraisal and Reinforcement Design Co.,Ltd.,Jinan 250014,China;College of Construction Engineering,Jilin University,Changchun 130026,China)

机构地区：[1]山东建筑大学土木工程学院,济南250101 [2]山东建筑大学建筑结构加固改造与地下空间工程教育部重点实验室,济南250101 [3]山东建大工程鉴定加固设计有限公司,济南250014 [4]吉林大学建设工程学院,长春130026

出　　处：《科学技术与工程》2024年第21期8950-8961,共12页Science Technology and Engineering

基　　金：山东省自然科学基金(ZR2021QD123)。

摘　　要：探究增强型地热系统(enhanced geothermal systems,EGS)工程采热性能,对提高系统产热效率、延长储层寿命是至关重要的。以伊通盆地北部昌27井实际地质构造、岩性、地热特征及热物性测试为基础,以地下4054~4168 m花岗岩为热储层建立三水平井水热耦合数值模型,并在此水平井模型基础上建立BP神经网络热储产热能力预测模型,研究了不同影响因素对系统换热过程的影响。结果表明:BP神经网络预测生产温度、焓值差和压力差与水平井数值模型的相对误差基本不超过±1.0%、±2.0%和±3.0%,预测准确性较高。水平井地热开采中,对系统生产温度造成影响较大的因素是井间距和注入速度,对系统发电效率造成影响较大的因素是注入速度和裂隙渗透率。综合考虑各影响因素交互作用,采用三水平井开采地热能可选择开采方案井间距600 m,注入温度60℃,注入速度23 kg/s,裂隙渗透率1×10^(-10 )m^(2)。在该工况下系统的经济寿命可达32年,累积发电量为605.9 GW·h,比基础模型提高7.4%。Exploring the heat recovery performance of enhanced geothermal systems(EGS)engineering is essential to improve system heat production efficiency and extend reservoir life.Based on the geological structure,lithology,geothermal geological data and laboratory data of Chang 27 well in the northern Yitong Basin,the hydro-thermal coupling numerical model of three horizontal wells and the prediction model of thermal reservoir capacity based on BP neural network were established with 4054~4168 m granite as the thermal reservoir,and the main factors affecting the performance and efficiency were studied.The results show that the relative errors of temperature,enthalpy difference and pressure difference predicted by BP neural network model and the numerical model are basically no more than±1.0%,±2.0%and±3.0%,and the prediction accuracy is high.The factors that have a greater impact on the system production temperature are well spacing and injection rate,and the factors that have a greater impact on the system power generation efficiency are injection rate and fracture permeability.Considering the interaction of all influencing factors,the extraction scheme can be chosen for geothermal energy extraction using three horizontal wells with a well spacing of 600 m,injection temperature of 60℃,injection rate of 23 kg/s,and fracture permeability of 1×10^(-10) m^(2).The economic life of the system under this working condition can reach 32 years,and the cumulative power generation capacity is 605.9 GW·h,which is 7.4%higher than the base model.

关 键 词：花岗岩热储 EGS 水热耦合模型 BP神经网络 产热性能 

分 类 号：TK529[动力工程及工程热物理—热能工程]