Power Transformer Intelligence Monitor System

An auto-monitoring system has been created that could monitor the main parameter of power transformers. It reads data of V＆I by precision rectifier circuit and amplifier, and converted electric energy into pulse signal by means of sensor technique and pulse converter circuit, it is sent into microcomputer to process, completed monitoring to power transformer parameters and running state by software system. It can register all kinds of key data, and provide all kinds of virtual technical data for managing sections and can auto-form all kinds of electric annual reports and monthly reports analysis graphs such as managing graphs. It notably improves the safety and reliability of transformer, with strong anti-jamming technique and unique function, it can be used on the transformers in railway＇ s self-closing lines and all kinds of transformer stations.     

基于人工智能的电力电子电路故障诊断

采用基于波形直接分析的神经网络故障诊断方法实现电力电子电路在线故障诊断。以三相整流电路为例,对电路发生故障时输出的波形进行分析,用故障波形的采样数据制作的样本对神经网络进行训练,将训练好的神经网络用于故障诊断。仿真实验表明该方法是有效的。     

可编程逻辑器件在定时控制电路中的应用

介绍了用ＩＳＰ可编程逻辑器件设计的一种器件少、可靠性高、实用性强的广告灯箱定时控制电路,详细阐述了其设计思路、硬件原理及ＩＳＰ可编程逻辑器件的使用技术。     

基于WPD-PSO算法的整流电路故障诊断

针对电力电子电路故障类型多、诊断正确率低的问题，提出基于小波包分解和粒子群算法优化概率神经网络的方法。建立三相桥式全控整流电路仿真模型，利用小波包分解技术对故障电压信号进行三层小波包分解与重构，提取特征值，并对数据进行归一化处理；用粒子群算法优化概率神经网络寻找合适的平滑因子，对数据进行训练和诊断；将该方法与未优化的概率神经网络作对比。仿真结果表明，该方法在训练效果和诊断正确率上都要优于未优化的概率神经网络。