A Comparative Experimental Study on Fault Diagnosis of Rolling Element Bearing Using Acoustic Emission and Soft Computing Techniques

In engineering processes, Health condition Monitoring (HCM) is a fault-finding task to guarantee consistency of rotating machinery. Rolling Element Bearings (REBs) are the main origin of damage in such equipment. They are the key components used in most of the rotating devices. These faults are mainly caused due to premature failure or improper installation of the machine element. Finding and analysis of defects is very vital in rotating machinery for its optimality. A test-rig was established to analyze the various line defects in REBs under different speeds and load conditions. The Acoustic Emission (AE) signatures responses are obtained and analyzed. Soft computing methods, especially Artificial Neural Network (ANN) and Support Vector Machine (SVM) used to compare with experimental results for fault diagnosis of bearings. The AE statistical features were fed as inputs in ANN and SVM. A comparative experimental study on prediction of seeded defect size is carried out using soft computing methods. This study concludes that these methods can be used for prediction of rolling element bearings seeded faults.