论文部分内容阅读
BP神经网络(Back Propagation Neural Network,BP-NN)具有良好的自学习能力以及自适应和泛化能力,但运算过程中容易陷入局部极小值,同时隐含层节点数的选择也影响着诊断的效果;文中根据经验公式缩小隐层节点数范围,在小范围里寻找最优的隐层节点数;根据遗传算法(Genetic Algorithm,GA)具有全局寻优的特点,用遗传算法优化BP神经网络训练的初始权值阈值,可以避免BP神经网络陷入局部极小的问题;但是,传统遗传算法也有自身的缺点,其在全局寻优的过程中,易陷入“早熟”的问题;为了解决传统遗传算法“早熟”现象,文中提出了一种协同进化的遗传算法,即使用3个种群同时进化的遗传算法,协同进化遗传算法不但可以避免传统遗传算法的“早熟”问题,而且可以加强局部搜索提高运行效率;将协同进化遗传算法应用到BP神经网络中,仿真结果表明,该方法可以准确有效地诊断出变电站故障元件,提高变电站故障诊断过程中的容错性及效果。
Back Propagation Neural Network (BP-NN) has good self-learning ability and self-adaptability and generalization ability, but it is easy to fall into local minima in the process of computation. At the same time, the choice of hidden layer nodes also affects the diagnosis In this paper, the number of hidden nodes is reduced according to the empirical formula and the optimal number of hidden nodes is found in a small area. According to the global optimization of genetic algorithm (GA), genetic algorithm is used to optimize BP neural network However, the traditional genetic algorithm also has its own shortcomings. In the process of global optimization, it is easy to fall into the problem of “premature ”. In order to solve the problem that the initial value of training threshold can not get into the local minimum, The traditional genetic algorithm “precocious ” phenomenon, the paper proposes a co-evolutionary genetic algorithm, that is, using three simultaneous evolutionary population genetic algorithm, the co-evolutionary genetic algorithm not only avoids the “premature” problem of the traditional genetic algorithm, But also can enhance the local search to improve the efficiency of operation. By applying the co-evolutionary genetic algorithm to the BP neural network, the simulation results show that this method can diagnose the change accurately and effectively Station failure element, the effect of improving fault tolerance and fault diagnosis of the substation.