论文部分内容阅读
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。
Based on parameter adaptive differential evolution algorithm (ADE), the inversion analysis method of rockfill dam constitutive model parameters is proposed, which can effectively improve the convergence rate and avoid falling into local optimum. Using BP neural network with strong nonlinear mapping ability, Model to simulate the approximate stress and strain of rockfill dam and improve the efficiency of the inversion process. Finally, a case study of a pumped-storage power station rockfill dam is carried out. The error between the settlement value and the actual settlement value is calculated by comparing the design conditions and the inversion conditions, which verifies the effectiveness and reliability of the proposed method It can provide a basis for subsequent objective evaluation of the safety of rockfill dam.