Bayesian Networks Based Distributed Fault Diagnosis Approach for Power Grids
Shu Dao Zhou, Xiaoru Wang, Qian Qing-quan · Power System Technology · 2010
A Leaky Noisy-Or Bayesian network-based dynamic modeling approach,in which actuating signal of measuring element of network protection is taken as diagnosis evidence,is presented for distributed fault diagnosis in power system.In allusion to the probability parameter calculation in diagnosis model,an a priori probability calculation method for event sampling is proposed.For the problem of using local information to implement distributed fault diagnosis of power network,the Leaky Noisy-Or nodal model is derived and two ratiocinations are given to prove the feasibility of this distributed method.By use of gradual exclusion,the element diagnosis model and combined diagnosis model are dynamically built,and through reasoning computation the fault diagnosis of power network is implemented.Calculation examples prove that the proposed method is correct and available.