Time to failure prediction of a rotating beam-a case study

摘要

This paper describes a method to determine the extent of rotating team creack growth and to predict the residual life of the creacked beam. To detect and determine the size of a fatigue crack from measured pressure wave excited by the beam vibration, signal processing was used to determine the modal frequnecies of the beam. Diagnostic neural networks were utilized to map model frequencies to crack size and location. Inverse neural network models were used ot estiamte the impact load from the pressure wave. To predict the growth of a crack and hence the residual life, an FEM-based crack propagation model was used to extrapolate the current state into future based on the current crack size and the impact load. Experimental results show the usefulness of the proposed methodology for rotating beam crack diagnosis and prognosis.