Wyniki wyszukiwania - BazTech

1

Repeatability and reproducibility studies for non-replicable tests

Mikulová Pavlína, Plura Jiří, Knop Krzysztof

System Safety : Human - Technical Facility - Environment

|

2020

|

Vol. 2, nr 1

275--284

EN

The paper presents several approaches to gauge repeatability and reproducibility (GRR) analysis regarding non-replicable measurements. Measurement systems have to deal with processes in which, by the nature of the measured object or by the type of measurement itself, measurements are not repeatable. In these cases, each sample unit can be measured only once. Such situations are referred as non-replicable measurement systems. The aim of the paper is to map out the current approaches being used in GRR analysis in various cases of non-replicable tests and compare each other in order to find out the suitable use of analysis application. Approaches used are subject to critical analysis so that its review can serve a useful base for analysis of different non-replicable tests. At present, it is desirable to bring the improving actions in order to obtain the results of high quality from such kind of measurements. Since different non-replicable tests can measure a different quality characteristic, it is valuable to bring the appropriate designs for various tests. Subsequently, this review will serve an outline how to proceed in analyzing the results obtained by non-replicable tests. Specifically, GRR analysis works with two known de-signs named as “Crossed” and “Nested” design, which statistical software normally use. Doubtfully, crossed design is suggested to use at certain cases and nested at other specific cases. This is assessed and improving actions designed.

2

Effect of non-fulfillment of assumptions on gage repeatability and reproducibility study evaluation

Klaput Pavel, Vykydal David, Plura Jiří

Quality Production Improvement - QPI

|

2019

|

Vol. 1, Iss. 1

472--478

EN

The evaluation of the measurement system quality has already become an integral part of quality planning activities in both the automotive and metallurgical industries. An important assumption for obtaining the most relia ble results is compliance with the basic assumptions for evaluating the variability of the measurement system. The main goal of this paper is to analyze, how the failure to meet the basic assumptions influences the evaluation of the measurement system's statistical properties. This goal is achieved by performing a detailed analysis of the latest developments in the field of measurement systems analysis aimed at verifying the assumptions of normality and uniformity. The evaluation of the effect of non-fulfillment of both assumptions on the values of the most important statistical properties of the measurement system is performed using simulated data. Suitable graphical tools are used for practical verification of both assumptions.

3

Detailed analysis of grr study results and their visualization

Mikulová Pavlína, Plura Jiří

Quality Production Improvement - QPI

|

2019

|

Vol. 1, Iss. 1

479--486

EN

The paper presents the possibilities of analyzing the measurement system repeatability and reproducibility (GRR) analysis results in more detail and their visualization. Based on real data, the influence of appraisers’ behavior on GRR analysis results is evaluated. The measured data obtained in the GRR study by three operators are analyzed in more detail for cases involving only two operators. Comparison of the behavior of individual operators is performed using Gaussian curves, which allow to evaluate graphically the repeatability of individual operators and the reproducibility of measurement system. This approach makes possible to visualize the GRR analysis results with regard to behavior of individual operators. Obtained results make possible to improve the interpretation of GRR analysis results and are useful for measurement system improvement.