中文说明：在前人工作的基础上深入分析了探讨了神经网络和思维进化算法的思想精髓、主要算法及特点,并将其应用于变压器的故障诊断当中,取得了良好的诊断结果。 人工神经网络利用本身分布式并行处理、自学习、自适应、非线性映射以及联想记忆等优点,为解决传统方法的不足开辟了新途径。但是由于神经网络自身结构特点,这种方法的收敛速度低,且常常陷入局部极小点,在学习样本数量多、要求精度高以及输入输出关系较复杂时,神经网络的收敛速度比较慢,收敛精度不太理想,甚至不收敛。思维进化算法具有搜索全局寻优的能力,可有效的提高神经网络收敛速度和精度,提高故障诊断成功率,为弥补神经网络的不足创造了条件,根据变压器油中溶解特征气体和故障类型的特点,本文提出了利用思维进化算法对神经网络的权值和阈值进行优化方法,以避免神经网络陷入局部最小值,并且提高其收敛速度。 通过将经过思维进化算法优化的神经网络模型应用于变压器故障诊断,经过训练和诊断结果表明:系统采用的思维进化优化算法明显的比未经优化的神经网络收敛速度得到了大幅度提高。
English Description：
In previous works on the basis of an in-depth analysis of the discussion on the essence of mind evolutionary algorithm and neural networks, algorithms and features, and its application to fault diagnosis of transformers, and have achieved good results. Artificial neural networks use distributed parallel processing, self-learning, adaptive, nonlinear mapping as well as advantages of associative memory, a new way to address the shortcomings of traditional methods. But due to the neural network structure, this convergence of low and often trapped in local minima, more in the study sample size, high accuracy and when more complex input/output relationships, neural network convergence is slow, accuracy is not perfect, not even converge. Thinking evolution algorithm has search global found excellent of capacity, can effective of improve neural network convergence speed and precision, improve fault diagnosis success rate, for cover neural network of insufficient created has conditions, accord