Daubechies Wavelet Finite Element Method and Genetic Algorithm for Detection of Pipe Crack

Y. M. Wang,X. F. Chen,Z. J. He
DOI: https://doi.org/10.1080/10589759.2010.521826
2011-01-01
Nondestructive Testing And Evaluation
Abstract:A model based on forward and inverse problems of pipe transverse crack detection is studied. First, Rayleigh beam elements of Daubechies wavelet are constructed to solve the forward problem. The first three frequencies influencing functions of normalised crack location and size are approximated by means of surface-fitting techniques. Then, the first three measured natural frequencies are employed as inputs of the functions. A genetic algorithm is used to identify the pipe crack location and size, solving the smallest sample points of the difference between the outputs of structural vibration and the pipe crack model. Experimental results demonstrate that the proposed method is an accurate and effective method for the detection of pipe crack.
What problem does this paper attempt to address?