A new online fault diagnosis algorithm based on likelihood ratio and Tabu search in distribution networks

Abstract

A new online fault diagnosis algorithm based on likelihood ratio and Tabu search (TS) is proposed for the uncertainties caused by many factors, such as mal-operation and non-operation of circuit breaker or relay, data-transmission error of loss and the inaccurate time of the protective operation in distribution networks. First, the likelihood ratio is represented and the model for fault diagnosis is built up in this paper, and then by quantifying uncertain information, the fault diagnosis is mathematically converted to an optimal decision-making model. Finally, Tabu search (TS) approach is presented to solve this problem. The proposed algorithm is not only used to deal with the uncertainties encountered in the fault diagnosis problem, but also useful for multiple faults. Test results have shown that the developed fault diagnosis method is correct and efficient, and is of promise for online application in actual distribution networks.