РАЗВИТИЕ МЕТОДА КОРРЕЛЯЦИИ ЦИФРОВЫХ ИЗОБРАЖЕНИЙ ДЛЯ ИЗУЧЕНИЯ ПРОЦЕССОВ ДЕФОРМАЦИИ И РАЗРУШЕНИЯ КОНСТРУКЦИОННЫХ МАТЕРИАЛОВ

Abstract

A number of problems related to the digital image correlation method from the standpoint of both the hardware and software development and testing, as well as solving problems in the field of fatigue fracture mechanics are described. Testing of the developed computer stereo vision system was carried out using a series of stereo pairs reflecting the change in the location of the object in space as well as plane and out-of-plane deformations. It is shown that the error in determining spatial coordinates does not exceed 0.75 units, while the error in computing the strain tensor components in case of a system with a single camera is two orders of magnitude larger than that at using the stereo machine vision system. An algorithm for automatic crack detection on optical images and calculating its tip coordinates was proposed and tested. It is shown that when the frame size is 2000×1000 pixels the coordinates of the crack tip might be determined with an average error of about 56 pixels, while the average error of the crack area determination does not exceed 1.93%. A modified incremental algorithm for calculating displacements on a series of stereo pairs is proposed, which allows one to estimate large magnitude displacements during serial processing of images. An algorithm for measuring the J-integral using the digital image correlation method has been developed. It is shown that the deviation of the calculated J-integral values from the model ones is on average 1.75 %. A quantitative characterization of the fatigue crack growth process in metal alloys was carried out using the technique based on the digital image correlation (in terms of fracture mechanics) including the calculation of the fatigue crack growth rate da / dN , maximum strain (emax) and effective cycle asymmetry ( R eff).