Fault Diagnosis Rule Extraction for Distribution System Based on Rough Set Theory
Quan Zhou, Ehv Power · Gao dianya jishu · 2006
As the first step of service restoration, rapid fault diagnosis in distribution system is a significant task in order to shorten power outage time, decrease outage loss and improve supply reliability and safety. In this paper, rough set theory is applied into fault diagnosis of distribution system, and how to reduce the corresponding decision table fast and effectively is the core of this approach. Aiming at reduction of decision table, a reduction algorithm by using attribution length and frequency as heuristic information is proposed. In conjunction with this algorithm, the corresponding value reduction method is proposed and fulfills the reduction and diagnosis rules extraction. Meanwhile, the rule matching method based on Euclid distance is introduced when some information is lacking during fault diagnosis process. Principle of the whole algorithm is clear and diagnostic rules extracted from the reduction are concise. Moreover, it needs less calculation towards the discernibility matrix formed from the specific diagnosis problem, and thus avoids the traditional time consuming and calculation intensive approach which obtains the reduction of decision table by calculating the corresponding discernibility function. The whole process is realized by MATLAB programming. A simulation example shows that the method possesses a fast calculation speed, and the extracted rules can reflect the characteristic of fault with a concise form. The rule database formed by different reduction of decision table can diagnosis efficiently and give a satisfying results even when the information is incomplete. The proposed method has good error-tolerate ability and the potential for the on-line fault diagnosis.