Power Network Fault Diagnosis Combining Rough Set Theory And Evidence Theory

Based on the advantage of rough set theory and evidence theory to process uncertain and inaccurate problem, this thesis presents a new idea of distribution network fault diagnosis based on rough set theory and evidence theory. First, a fault diagnosis decision table considering many fault conditons is built, and an algorithm based on the frequency of discerning matrix attribute is used to reduce the decision table to obtain the simplest effective date combination. Then, using the relation between rough set theory and evidence theory, the effective data is translated to basic probability assignment. Lastly, the idea of attribute significance in rough set theory is introduced to fault diagnosis. In response to the Rule of Combination, the Combination of evidence is processed by two steps to realize the fault diagnosis. Through the combination of these two methods, the error tolerance of the proposed fault diagnosis algorithm is effectively improved. This algorithm has been verified by the application on a 66kV substation.