A Decision Network Model of Vibration Fault Diagnosis for Steam Turbine Unit Based on Rough Set Theory
Wang Xiu-jiang · Energy Technology · 2010
Redundancy and inconsistency are universal characteristics for the issues of vibration fault diagnosis for steam turbine unit.Based on a novel model proposed in this paper,decision rules of fault diagnosis still can be extracted from data including redundant and inconsistent information.The model mentioned above was composed of decision network of fault diagnosis for different simplified levels.According to each node of different network levels,a tidy rule set of diagnosis decision can be obtained.The concept of rule coverage was also introduced on the basis of given rule confidence,which can remove noise effectively and improve the extraction efficiency of diagnosis rules.Thus,relatively satisfied diagnosis conclusions can be draw in the case of having incomplete fault diagnosis information.