摘要
基于参数自适应差分进化算法(ADE),提出了堆石坝本构模型参数的反演分析方法,可有效提高收敛速度和避免陷入局部最优;并采用具有强大非线性映射能力的BP神经网络模型来近似模拟计算堆石坝的应力应变,提高了反演过程的效率。最后,以某抽水蓄能电站堆石坝为例进行应用研究,通过比较设计工况和反演工况下计算沉降值与实际沉降值之间的误差,验证了所提方法的有效性和可靠性,可为后续客观评价堆石坝安全性提供基础。
The paper presents a new approach for back analyzing the parameters of rockfill dam constitutive model based on adaptive differential evolution( ADE) algorithm,which can efficiently improve convergence speed and avoid being trapped into local optimal solution. And then the efficiency of inversion procedure is improved by using the BP neural network with strong nonlinear mapping ability to simulate the stress-strain of rockfill dam. Taking the rockfill dam of a certain pumped storage power station as an example,the errors between calculated settlement and actual settlement are compared under design condition and inverse condition. It verifies the effectiveness and reliability of the proposed approach,which can provide basis for safety analysis of rockfill dam objectively.
出处
《水电能源科学》
北大核心
2017年第6期62-66,共5页
Water Resources and Power
基金
国家重点基础研究发展计划(973计划)(2013CB035904)
国家自然科学基金项目(51479132)
关键词
堆石坝
本构模型
参数反演
自适应差分进化
神经网络
rockfill dam
constitutive model
back analysis of parameters
adaptive differential evolution
neural network
