Metodické přístupy k analýzám systémů měření

Abstract

The thesis focuses on development of methodological approaches of the quality of the measurement system. In the theoretical part of this work are in detail presented the latest information about the recent approaches to the measurement systems analysis. Attention has been paid to the critical analysis of methodologies MSA (Measurement System Analysis) and VDA 5. In detail are specified methods of evaluation of particular statistical properties of the measurement system and the components of measurement uncertainty. The practical part is fully paid to achieving the goals of the work, which are defined in the second chapter of the thesis. Within the framework the fulfillment of first goal of this work was performed the multi-criteria evaluation of the most widely used methodological approaches to the measurement system analysis. On the basis of this evaluation were subsequently defined strengths and weaknesses of evaluated methodological approaches. Based on the real data in the next step has been explored the possibilities of interchangeability of the performance indicators used to the evaluation of the acceptability of the measurement systems. Next part of this work aims on a detailed analysis of the factors affecting the implementation and evaluation of the gage repeatability and reproducibility study. Using experiments conducted on both real and simulated data was evaluated the way in which defined factors affect the values of observed parameters. The results of these experiments were used to define the conditions under which are the results of the measurement system analysis as accurate as possible. Critical analysis has been also subjected the various graphical tool of the measurement system analysis. By using the experiments were defined weaknesses of each charts and diagrams. To eliminate these weaknesses, has been subsequently designed a completely new and upgraded graphical tools useful in the analysis of the causes of high variability of the measurement systems. In the last part of this work was designed the procedure of comprehensive measurement system analysis allowing evaluation of several statistical properties of the measurement system within a single analysis.