Crack detection in beam-like structures using a power series technique and artificial neural networks

MARTA B. ROSALES; CARLOS P. FILIPICH; FERNANDO BUEZAS,

Vienna, Austria

Congreso; Thirteen International Congress on Sound and Vibration (ICSV 13); 2006

Vienna University of Tech., Austrian Academy of Science, IIAV

The analysis of experimentally measured frequencies as a criterion for crack detection has been extensively used in the last decades for its simplicity. However the inverse problem of the crack parameters (location and depth) determination is not straightforward. An efficient numerical technique is necessary to obtain significant results. Two approaches are herein presented. One is the solution of the inverse problem with a power series technique (PST) and the other is the use of artificial neural networks (ANN?s). The free vibration problem of a Bernoulli-Euler beam with an intermediate spring is stated and then solved with the PST. The inverse problem is tackled as follows. The first three flexural frequencies are measured and input in the algorithm. Three location vs. spring constant curves are obtained and their intersection yields the ?detected parameters?. At this stage a numerical experiment on a 2D beam is employed. The crack depth is derived from a Mechanics of Fracture?s relationship. The ANN?s technique is a different approach since it needs a training set of data. A single hidden layer back-propagation neural network is trained with data found with 2D finite element models with more than four hundred scenarios. These data are also analyzed and some curves are depicted to show the variables influence. The first methodology is very simple and straightforward though no optimization is included. It yields negligible errors in the location and very small ones in the depth values. The ANN?s give acceptable values though handling one network for the detection of the two parameters. However better results are found when an ANN is used for each parameter. Finally a combination of the two techniques improves the results. The location is found with the first technique and the depth with the second one.