Akustická diagnostika procesu laserového čištění

Abstract

This bachelor’s thesis deals with the design of metrics for analysing the laser cleaning process. The introductory part focuses on the motivation behind the addressed topic and on the use of acoustic diagnostics in industry, from which the objectives of the thesis are subsequently derived. The research section addresses technical diagnostics and modern possibilities for evaluating the condition of industrial machines and equipment. The literature review also focuses on methods used for preprocessing, analysis, and extraction of diagnostic features required for processing by means of machine learning and deep learning. In the second part of the research section, experiments found in scientific sources were described and compared. In the practical part of the thesis, two metrics for evaluating the laser cleaning process are developed. The most conclusive of these is time-domain analysis in the form of RMS acoustic pressure, while a supplementary indicator is created in the frequency domain in the form of the ratio of low-frequency to high-frequency amplitudes. This is followed by the processing of measured data for individual experimental measurements and their evaluation using the proposed metrics in the form of normalized data and graphical processing that captures trends and time intervals used for evaluating laser cleaning.