基于信息理论的电力系统在线故障诊断

该文引入信息理论来研究和处理电力系统故障过程中存在的不确定性，研究了电力系统故障中的信息运动过程，建立了故障诊断的信道模型；基于信息损失最小原理，提出了适用于大规模系统不确定性决策的故障诊断新方法，该方法将故障诊断问题转化为求取最小信息损失的组合优化问题。文中还研究了通过信息损失来计算故障诊断解空间的概率分布方法，并针对大规模系统给出了近似解空间概率分布的求解方法。应用该文方法进行电网故障诊断，能给出各种可能发生的故障组合以及每种故障组合发生的概率。算例表明，该方法的快速性能满足大规模电力系统在线故障诊断辅助决策的要求。

ONLINE FAULT DIAGNOSIS FOR POWER SYSTEM BASED ON INFORMATION THEORY

As the uncertainty exists in fault processing of power system, the Shannons information theory is introduced to deal with the uncertainty. In this paper, a thorough research on how the Shannons theory is applied to fault diagnosis is reported. An information channel model for fault diagnosis is built up firstly, and then by using a new optimal decision-making principle based on minimization of information loss, the fault diagnosis is mathematically converted to an optimal decision-making model. The probability models for all types of possible faults can be built up and calculated to get an index of information loss and finally a fast algorithm is developed to calculate the index. By using the proposed theory and algorithm, the probability indexes for all types of possible faults can be obtained and used to the fault diagnosis for power systems. Numerical tests have been done and the results have justified the theory proposed in this paper. Method in this paper is shown particularly suitable for use of online fault diagnosis for a large scale of power system.