Neuro fuzzy approaches for fault Detection in rotating machines

Abstract

Rotating machines are essential components in most of the newlinemanufacturing and production industries They are exposed to a variety of newlineenvironmental conditions These operating conditions coupled with natural newlineageing cause incipient faults in the machines The most common incipient newlinefaults are winding insulation failure and bearing wear With proper newlinemonitoring scheme if these faults were detected at their early stages the newlinemaintenance cost and down time can be reduced Many of the conventional newlinefault detection methods require the need of an expert to evaluate the machine newlinecondition The development of soft computing techniques in the area of newlinecomputer science motivated the researchers to use these techniques for newlineintelligent problem solving which exhibits the characteristics of human newlineintelligence The soft computing tools like neural networks and fuzzy system newlinehave been used in many engineering applications such as fault identification newlineand control of dynamic systems It is well known that a feed forward neural newlinenetwork is capable of approximating any continuous functions closely The newlineuse of fuzzy logic in control applications has tremendously increased over the newlinelast decades newline newline