基于多种人工智能技术集成的电力变压器故障诊断

POWER TRANSFORMER FAULT DIAGNOSIS BASED ON INTEGRATED ARTIFICIAL INTELLIGENCE

本文提出的算法综合了进化规划、模糊理论、人工神经网络、范例推理四种人工智能技术的优点,弥补了单种人工智能技术的缺陷,缩小了样本间的数据差异,优化了网络初始权值,大大提高了人工神经网络的收敛性。人工神经网络的拓扑结构及训练样本经过大量的计算进行筛选确定,并通过基于范例推理的专家系统检索出最佳源范例并给出诊断结论。诊断结果表明,本方法能较准确地判别电力变压器故障。

An algorithm integrating the advantages of evolutionary algorithm, fuzzy theory, artificial neural network and case-based reasoning is presented. This algorithm improves the shortcoming of each single method of artificial intelligence, narrows the discrepancy among the sampled data, optimizes the initial weight value of network and greatly improves the convergence of artificial neural network. After great many calculations the topological structure and training samples are screened and chosen, through the expert system based on case-based reasoning the optimal source case is retrieved out and the diagnosis conclusion is given. Practical examples of power transformer diagnoses show that the presented method is effective.