Crack Detection in Machine-Structure by Evolutionary Algorithm

An evolutionary approach for fatigue crack detection in cantilever type of machine structure is investigated in this paper. Fault like fatigue crack detection has received a great attention in recent years. A surface transverse open edge crack is considered for this study. Actual fatigue crack may behave like a non-linear due to opening and closing under vibration. But due to small level of vibrations on structural beams it can be assumed that cracks remain open. To avoid non-linearity, and tackle most practical problems it is assumed that the crack is always open. Vibration parameters obtained from modal analysis as a criterion for crack detection has been extensively used in the last decade due to its simplicity. Most of the works in the application of evolutionary methods for this purpose are limited to numerical examples and small experimental examples only. However an efficient technique is necessary to obtain significant results. The first three natural frequencies were used for the fault detection. The crack parameters are accurately detected using the experimental natural frequencies as an input to genetic algorithm. To investigate the robustness of the proposed method, several experimental cases of cracked beam are considered. It is found that the method is capable of detecting crack in cantilever beam.