Heat Production Performance of Enhanced Geothermal System Based on BP Neural Network:Taking the Granite Thermal Reservoir in the Yitong Basin as an Example
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 4 054～4 168 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 x 10-10 m2.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.
granite heat storageEGSwater-heat coupling modelBP artificial neural networkheat production performance
NETL
NSTL
万方数据
