基于贝叶斯原理的高电压绝缘配合实现方法

High-voltage Insulation Coordination Implementation Method Based on Bayesian Theory

提出目前对电气设备非自恢复型内绝缘采用惯用法进行绝缘配合,而对于绝缘子和空气间隙这类恢复型绝缘而言,其绝缘闪络电压是一个随机变量,而且作用于绝缘的过电压也是一个随机变量。当获知它们的信息之后,通过概率理论应用统计手段进行绝缘配合设计,这些绝缘配合方法仅考虑单独类型的过电压与电气装置耐压之间的关系,事实上,各种类型过电压与电气装置耐压性能之间是复杂的网状因果关系结构。尝试用贝叶斯网络结构来表征过电压与电气装置绝缘性能之间的关系,利用C语言集成产生式系统(CLIPS)实现该网状关系的结构与推理分析,为绝缘配合设计与分析提供定量依据。

At present, conventional method is used for insulation coordination of electric equipment non-self-restoring inner insulation. As to self-restoring insulation like insulator and air gap, insulation flashover voltage is a random variable, as well as overvohage affecting insulation. After information of these variables is received, insulation coordination design is made through statistical method using probability theory. However, for this kind of insulation coordination method, only the relationship between single-type overvohage and electric device withstand voltage in taken into consideration. In fact, the relationship between muhi-type overvohage and electric device voltage-withstand function is complicated reticular causality. This article uses Bayesian network structure to simulate the relationship between overvohage and electric device insulation function. C language integrated production system （CLIPS） is introduced to realize the structure of this reticular relationship and make inferences and analysis, providing quantitative evidence for insulation coordination design and analysis.