摘要
为提高裂隙岩体渗透张量的计算精度,探讨裂隙规模对渗透张量的影响。通过比较遗传算法(Genetic Algorithm, GA)和象群优化算法(Elephant Herding Optimization, EHO)计算时间和精度,提出了象群优化算法(Elephant Herding Optimization, EHO)与模拟退火算法(Simulated Annealing, SA)相结合的反演方法来优选离散裂隙网络(discrete fracture network, DFN)管单元管径,建立了DFN-EHO-SA模型,分析了不同裂隙缩减规模对应的管单元尺寸和表征单元体(representative elementary volume, REV)在反演算法条件下裂隙岩体渗透张量的精度,通过实例和工程应用验证了所提出模型和算法的优越性。结果表明,象群模拟退火算法(Elephant Herding Optimization-Simulated Annealing, EHOSA)收敛速度快且不易受困于局部最优解,基于该算法建立的DFN-EHO-SA模型确定渗透张量,计算精度高;缩减裂隙规模可以极大减少计算时间,特别是缩减规模为0.7时,模型满足精度同时大幅度提升时间效率,在工程中具有较强的实用价值,为裂隙岩体渗透张量计算提供一定的参考。
To improve the calculation accuracy of permeability tensor in fractured rock masses and investigate the influence of fracture scale on permeability tensor,this study compares the calculation time and accuracy of Genetic Algorithm(GA)and Elephant Herding Optimization(EHO).An inversion method combining Elephant Herding Optimization and Simulated Annealing is proposed to optimize the diameter of pipe elements in the discrete fracture network.The DFN-EHO-SA model is established to analyze the pipe element size and representative elementary volume under different fracture reduction scales and to evaluate the accuracy of the permeability tensor in the inversion algorithm.Examples and engineering applications verify the superiority of the proposed model and algorithm.The results show that the Elephant Herding Optimization-Simulated Annealing has a fast convergence speed and is less likely to become trapped in the local optimal solution.The DFN-EHO-SA model based on this algorithm can accurately determine the permeability tensor.Reducing the fracture scale can greatly shorten calculation time,especially at a reduction scale of 0.7,where the model maintains accuracy while significantly improving computational efficiency.This has strong practical value in engineering and provides a reference for the calculation of the permeability tensor in fractured rock masses.
作者
王俊奇
韦小婷
王子颜
WANG Junqi;WEI Xiaoting;WANG Ziyan(School of Water Resources and Hydropower Engineering,North China Electric Power University,Beijing 102206,China)
出处
《华北电力大学学报(自然科学版)》
北大核心
2025年第6期133-142,共10页
Journal of North China Electric Power University(Natural Science Edition)
基金
水资源与水电工程科学重点实验室开放基金资助项目(2016SGG03)
国家自然科学基金资助项目(52279065)。
关键词
离散裂隙网络
渗透张量
象群模拟退火算法
管单元
表征单元体
反演分析
discrete fracture network
permeability tensor
Elephant Herding Optimization-Simulated Annealing
pipe element
representative elementary volume
inversion analysis
