基于IFWA-PNN的变压器故障诊断研究

打开文本图片集

摘  要：为克服概率神经网络在判断变压器的故障类型方面的不足，文章采用改进的烟花算法优化概率神经网络（IFWA-PNN）的平滑因子。实例分析表明，提出的变压器故障诊断模型诊断准确率达到91.7%，相比传统的支持向量机诊断模型有较大提升，以此证明故障诊断模型的有效性和实用性。

关键词：概率神经网络;平滑因子;改进烟花算法;变压器;故障诊断

中图分类号：TP3      文献标志码：A          文章编号：2095-2945（2022）06-0086-04

Abstract： In order to overcome the deficiency of probabilistic neural network in judging the fault type of transformer， this paper adopts the smoothing factor of an improved fireworks algorithm optimized probabilistic neural network（IFWA-PNN）. The analysis of examples shows that the diagnosis accuracy of the proposed transformer fault diagnosis model reaches 91.7%， which is greatly improved compared with the traditional support vector machine diagnosis model. This proves the effectiveness and practicability of the fault diagnosis model.

Keywords： probabilistic neural network; smoothing factor; improved fireworks algorithm; transformer; fault diagnosis

变压器是电力系统中的关键设备[1]。（剩余4343字）