Naslov
The Role of the Measurement Uncertainty in the Decision Making Process of Conformity Assessment

Autori
Bošnjaković, Alen ; Karahodžić, Vedran ; Jugo, Ehlimana ; Džemić, Zijad

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, prošireni sažetak, znanstveni

Izvornik
Mathematical and Statistical Methods for Metrology (2019) / - , 2019, 20-22

Skup
Mathematical and Statistical Methods for Metrology Workshop (MSMM 2019)

Mjesto i datum
Torino, Italija, 30.05.2019. - 31.05.2019

Vrsta sudjelovanja
Radionica

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
conformity assessment, decision making, measurement uncertainty, distribution function, error function, software, ISO/ IEC 17025

Sažetak
Measurement uncertainty plays a very important role in most industries related to the conformity assessment of products. Therefore, the measurement uncertainty has an impact on the decision making process of the conformity assessment of the product. Without adequate estimates of measurement uncertainty or its neglecting in the process of making a decision on conformity assessment, the cost of production increases, and production mistakes are guaranteed. Therefore, measurement uncertainty is an indispensable factor for proper making of decisions and risk assessment in the conformity assessment procedure [1], [2], [3]. Most of the industries make a decision on product conformity on the basis of a report that contains measurement results issued by certain laboratory. If, for some reason, the measurement results are not appropriate, the risk of the wrong decision is increased. Also, the selection of the unincorporated laboratories, or laboratories without proven competencies, also increases the risk of the wrong decision. Therefore EN ISO / IEC 17025 [4] standard sets requirements for the qualification of test and calibration laboratories. The first edition of the mentioned standard dates from 1999, and from this period a large number of accredited laboratories, as well as laboratories of the national metrology institutes, base their quality management system precisely on this standard. The new edition of the standard, besides the structural changes, also introduces new requirements for the training of the laboratories. Major changes in the new standard edition include risk-based thinking, flexibility in process-related requirements, documented information and organizational abilities, and more. This work does not intend to give instructions for interpreting certain standard points, nor how to implement them in the laboratory. Taking into account the importance of the documented decision making rules, this work aims to demonstrate, through practice examples and through the application of mathematical models, that it facilitates implementation of decision making requirements for the laboratory. The rule for making the decisions brings a new requirement that appears in the new edition of the standard, and it occurs when there is the need for laboratories to evaluate the conformity of results with pre- set criteria taking into account the measurement uncertainty, ”...the laboratory shall document the decision rule employed, taking into account the level of risk associated with the decision rule and apply the decision rule” [4]. Throughout the paper, models related with the process of documented decisions, decision-making rules, and conformity assessment based on two scenarios (with and without the use of guard bands) will be presented [5]. All these analysis will be accompanied by the mathematical model which will facilitate their application in practice. Conformity assessment will be determined by using knowledge of distribution functions (for calculating probabilities), and displayed by using the error functions (Q and erfc function) whose values are taken from the tables. The results mentioned above will be confirmed using appropriate software tools. The aim of this paper is to make the code for conformity assessment publicly available to reduce the risk of making wrong decisions in practice. It is planned to implement code in MATLAB/ OCTAVE, using functions and m-files. Output of every code execution will be appropriate probability of acceptance what can serve as acceptance or rejection criteria. In this way, the laboratories that plan the transition from an old edition to the new one, or the laboratories that have already made that transition, will have open access to the program code. Laboratories will have the ability to modify the code for their needs later, or if necessary, to compare them with their existing solutions. Also, the impact of measurement uncertainty on the acceptance or rejection of product conformity with product specification will be analyzed in the work. Practical example on which results of code will be tested is regarding pressure gauges [6].

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo

POVEZANOST RADA