DEVELOPMENT AND INVESTIGATION OF AN OPTOELECTRONIC SYSTEM FOR AUTOMATED INSPECTION OF MICRODEFECTS IN LAYERS OF ADDITIVELY MANUFACTURED PARTS

Abstract

The study employs a comparative analysis of non-destructive testing (NDT) methods, the theory of acoustic wave propagation to examine the limitations of ultrasonic testing (UT), and computer vision algorithms for surface defect identification. The mathematical framework is based on the evaluation of ultrasonic wave propagation velocity in anisotropic media and correlation analysis of visual descriptors associated with defects. Critical limitations of conventional ultrasonic testing for porous structures fabricated via fused deposition modeling (FDM) have been identified, primarily due to significant signal attenuation and temperature-induced drift. An integrated approach to optical inspection is proposed, enabling differentiation of defect types (warping, under-extrusion, stringing) based on their morphological characteristics without mechanical interaction with the specimen. A hardware–software complex for optical inspection has been developed and experimentally validated. The system provides automated defect detection based on geometric and textural features, enabling a reduction in production losses within digital manufacturing processes based on additive technologies. It has been demonstrated that the implementation of an optical monitoring system allows for the identification of precision defects with a characteristic size of 0.1 mm or greater, resulting in a 15–20% reduction in production losses through real-time adjustment of printing parameters.