Title :
Evaluation of crack caustics by neural networks
Author :
Ajtonyi, Istvan ; Czap, Laszlo
Author_Institution :
Dept. of Autom., Univ. of Miskolc, Miskolc, Hungary
fDate :
Aug. 31 1999-Sept. 3 1999
Abstract :
The shadow optical method of caustics which is also known as the shadow spot method is a relatively new experimental technique to determine stress intensity factors in crack problems under static and dynamic loading. The evaluation of caustics needs automatic measurement of the characteristic dimensions. This needs the detection of edges and smoothing of the curve or curve fitting before counting the diameter of caustics. It enables more precise measurement than the resolution of the image capturing system. An Artificial Neural Network was trained to smooth the caustics.
Keywords :
cracks; edge detection; internal stresses; mechanical engineering computing; neural nets; artificial neural network; characteristic dimension; crack caustics evaluation; curve fitting; curve smoothing; dynamic loading; edge detection; image capturing system; shadow optical method; shadow spot method; static loading; stress intensity factor; Image edge detection; Image resolution; Optical imaging; Optical polarization; Optical reflection; Smoothing methods; Stress; crack analysis; crack caustics; image processing; neural networks; stress intensity factor;
Conference_Titel :
Control Conference (ECC), 1999 European
Conference_Location :
Karlsruhe
Print_ISBN :
978-3-9524173-5-5
