Experimental Investigations on Crack Detection Using Modal Analysis and Prediction of Properties for Multiple Cracks by Neural Network

摘要: In the present study, a method is proposed for detection and prediction of properties multiple transverse cracks on simply supported stepped rotor shaft. Two cases are considered. Initially, both perpendicular to axis. Later, inclined vertical plane also with each other. Modal analysis performed extract natural frequency mode shapes. Finite element (FEM) treated as basis numerical analysis. For validation, experimentation using fast Fourier transform analyzer. Based frequency, detected. The results FEM found in agreement. Crack predicted forward technique artificial neural networks (ANN). database frequencies used train network ANN predict crack properties. Applicability verified by comparing predictions experimentation. given It envisages that competent, suitable would be alternate existing methods. enhances performance structural integrity assessment online conditioning monitoring.