Feed-Forward Neural Network with Backpropagation Training Based Fault Detection in Rolling Element Bearing using Time Domain Features and Frequency Domain
Bearings are one of the most usual elements in rotating machinery, and as a consequence, bearing failure is likewise one of the main causes of breakdown in rotating equipment. The robustness and durability of the rollers are essential qualities for a machine's safety. Deficiencies in bearings may occur during use or during manufacturing. The identification of these defects is therefore critical for condition monitoring as well as for inspecting the quality of the bearings . This paper exhibits a strategy for flaw discovery in moving component bearing utilizing time-space highlights and recurrence area highlights of vibration signals. This system involves two sequential processes: feature extraction and decision-making. Vibration signals were recorded in this process. Neural Network Feed Forward Back Propagation was used for the classification. The 12 extracted features such as Mean, Peak, Mean Square, Variance, Standard Deviation, RMS, Shape Factor, Skewness, Kurtosis, Impulse Factor, Clearance Factor, Crest Factor were used to train and check the neural network for four bearing conditions namely: healthy, outer race fault, inner race fault and defective ball & outer race fault condition.