Fault detection and severity classification based on adaptive filter and fuzzy logic

Rotating machines faults are the most common faults in the industry. Thousands of faults detection techniques are widely used to identify the faults in the rotating machines. However, severity classification of the fault is more important to prevent the breakdown of the system as well as save the properties even human causality. The aim of this paper is to determine the fault signatures to identify the status of the rotating machines. This paper proposed a fault detection and criticality classification (FDCC) method for rotating machines based on an adaptive filter, fuzzy logic and computed order tracking that not only detects the faults but also classifies the severity of the faults. At first, the adaptive filter is used in proposed FDCC method to reduce the noises as well as artificial artifacts from the faulty signal. After that, order tracking is used to remove the speed variation of the rotating machine. Then, fault detection is done by envelope analysis. Finally, fuzzy logic is used to classify the fault severity. Experimental results indicate that proposed FDCC technique effectively detects the faults as well as classifies the severity of faults.