Condition Monitoring Setup for Long Term Reliability in Fault Recognition
Non-Confidential description of the invention
Present invention provides a condition monitoring system that improves the long term reliability of fault recognition in machines. The system comprises a feature selection method that is able to select good, reliable and consistent features, such that these features can perform fault recognition. The process begins with a sensitive position analysis where optimal sensor position(s) for placing the sensor devices is located. Further, the data is acquired from the sensor devices as datasets while machine state is switched between healthy and faulty state in cyclic order such that the long term trends of machine are captured effectively. Once the data recordings are acquired, they are preprocessed and desired features are extracted. Further, the feature selection procedure ranks the features based on their ability to confidently identify faults. The whole process is experimentally found to significantly reduce the downtime of machine, and also improve the life of condition monitoring setup.
Dr.Nishchal Kumar Verma(EE), Mr.Rahul Kumar Sevakula(Student,EE) and Mr.Thirukovalluru Raghuveer(Student,EE)
Date of Filing