Automatic Defect Classification in Ultrasonic NDT Using Artificial Intelligence
- 871 Downloads
A methodology is developed to detect defects in NDT of materials using an Artificial Neural Network and signal processing technique. This technique is proposed to improve the sensibility of flaw detection and to classify defects in Ultrasonic testing. Wavelet transform is used to derive a feature vector which contains two-dimensional information on various types of defects. These vectors are then classified using an ANN trained with the back propagation algorithm. The inputs of the ANN are the features extracted from each ultrasonic oscillogram. Four different types of defect are considered namely porosity, lack of fusion, and tungsten inclusion and non defect. The training of the ANN uses supervised learning mechanism and therefore each input has the respective desired output. The available dataset is randomly split into a training subset (to update the weight values) and a validation subset. With the wavelet features and ANN, good classification at the rate of 94% is obtained. According to the results, the algorithms developed and applied to ultrasonic signals are highly reliable and precise for online quality monitoring.
KeywordsUltrasonic testing TIG welding Defect classification Wavelet transform Artificial neural networks
Unable to display preview. Download preview PDF.
- 3.Veiga, J.L.B.C., de Carvalho, A.A., da Silva, I.C., Rebello, J.M.A.: The use of artificial neural network in the classification of pulse-echo and TOFD ultra-sonic signals. J. Braz. Soc. Mech. Sci. Eng. 27(4), 394–398 (2005) Google Scholar
- 10.Burrus, C.S., Gopinath, R.A., Guo, H.: Introducing to Wavelets and Wavelet Transformation. Prentice-Hall, Englewood Cliffs (1998) Google Scholar
- 11.İşcan, Z., Kurnaz, M.N., Dokur, Z., Olmez, T.: Ultrasound image segmentation by using wavelet transform and self-organizing neural network. Neural Inf. Process. 10(89), 183–191 (2006) Google Scholar
- 13.The Math Works, Inc. 1998: Neural Network Toolbox for use with Matlab 5.3/Version 2007, Natick, MA Google Scholar