在电网的运行过程中,受到紫外线、电场、水分、温差等因素影响,不可避免地出现一些老化的现象,危害电网的正常运行,需要对复合绝缘子的寿命进行预测,以此更换老化和劣化的绝缘子.以某地区 78 串绝缘子为研究对象,通过泊松相关性进行分析,甄选出与绝缘子寿命相关性较强的性能指标,采用 BP 神经网络的方法对绝缘子寿命进行研究预测.结果表明,使用硬度、憎水性能、耐电起痕深度、水扩散泄漏电流、拉伸强度、抗撕裂强度、机械负荷值、机械破坏值、高压侧盐密 9 个性能指标来对复合绝缘子寿命进行预测,检验的绝缘子预测寿命与实际寿命差距在 5%以内.实际算例分析表明,所构建的预测模型可以实现对复合绝缘子运行寿命进行预测,方便高电压外绝缘设备的及时更新,减少了运行事故的发生.
BP Neural Network-based Insulator Lifetime Prediction
The life of composite insulators needs to be predicted in order to replace the ageing and deteriorating insulators.In this paper,78 strings of insulators in a certain region are studied,and the performance indicators with strong correla-tion with insulator life are selected by Poisson correlation analysis,and the BP neural network is used to study and predict the insulator lifetime.The results show that the predicted lifetime according to nine performance indicators,i.e.,hard-ness,water repellency,depth of electrical start resistance,water diffusion leakage current,tensile strength,elongation at break,tear strength,mechanical load value,and mechanical damage value,have an error of within 5%against the actual lifetime of the insulators.The analysis of the actual cases shows that the constructed model is adequate in predicting opera-tional lifetime of composite insulators,which facilitates the timely replacement of high-voltage external insulation equip-ment and reduces the occurrence of operational accidents.
万方数据
维普
