基于MPGA-BP的重力坝变形预测研究

发布时间：2018-05-11 06:02

  本文选题：重力坝变形预测 + 多种群遗传算法　； 参考：《兰州理工大学学报》2016年05期

【摘要】：位移是重力坝变形监测的重要物理量,对其进行准确预测是确保大坝安全运行的前提.目前已经有许多预测方法,但是大部分方法都存在易落入局部极小、收敛速度慢和收敛对初值敏感等问题.为解决或减小这些问题,提高预测精度,将多种群遗传算法(MPGA)与反向传播(BP)神经网络算法结合起来,提出一种适用于重力坝变形预测的多种群遗传神经(MPGA-BP)网络算法.实例计算证明,该算法能够有效克服BP神经网络收敛速度慢、易出现局部极小值的缺点和遗传算法的早熟收敛问题,在进行重力坝变形预测中具有更高的收敛性和精度.
[Abstract]:Displacement is an important physical quantity of gravity dam deformation monitoring, and accurate prediction is the prerequisite to ensure dam safe operation. At present, there are many prediction methods, but most of them are easy to fall into local minima, slow convergence speed and convergence sensitivity to initial values. In order to solve or reduce these problems and improve the prediction accuracy, a multi-population genetic neural network (MPGA-BPN) algorithm, which is suitable for gravity dam deformation prediction, is proposed by combining multi-population genetic algorithm (MPGA) with backpropagation (BP) neural network algorithm. Examples show that the algorithm can effectively overcome the shortcomings of slow convergence speed of BP neural network, local minimum and premature convergence of genetic algorithm. It has higher convergence and accuracy in gravity dam deformation prediction.
【作者单位】： 兰州理工大学能源与动力工程学院;
【基金】：国家自然科学基金(51069004)
【分类号】：TV642.3;TV698.11
，

本文编号：1872710