Bayesian approach for fault treatment

Abstract

After the successful fault diagnosis, the faulty system should be recovered. A technical system is made up of many different interconnected components, where the symptoms of a specific problem may have different causes. Furthermore, a failure in a component may result in several consequences. Thus, detection of the root causes of alarms and applying the correct maintenance operation are always crucial tasks in a complex industrial system. Therefore, it is necessary to have a decision support system which helps the plant engineer to discover the causes of the alarms through the troubleshooting process and provide a systematic search for the root cause of a problem in a process to make the process operational again.