Naslov
Data Modelling in Risk Assessment

Autori
Božić, Dubravka ; Runje, Biserka

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
LABORATORY COMPETENCE 2022, PROCEEDINGS / Grgec Bermanec, Lovorka ; Ljubas, Davor - Zagreb : CROLAB - Hrvatski laboratoriji, 2022, 24-30

ISBN
978-953-7329-26-6

Skup
17. međunarodna konferencija, Kompetentnost laboratorija - 2022

Mjesto i datum
Cavtat, Hrvatska, 09.11.2022. - 12.11.2022

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Recenziran

Ključne riječi
Consumer's risk, producer's risk, tolerance interval, acceptance interval, guard band

Sažetak
There are two sources of information used in the Bayesian approach to risk modelling: measurement data and a priori beliefs of an experienced measurer about the possible distribution of parameters that describes the measurement data. With each measurement, there is a problem of compliance of the measurement results with the manufacturer's specifications. When the measurement result is outside the tolerance interval and the actual value within the tolerance interval, we are talking about the producer's risk. If the measurement result is within the tolerance interval, and the actual value is outside that interval, we are talking about consumer's risk. The consumer's risk can be reduced by setting acceptance interval. The upper acceptance limit of roundness deviation for the inner bearing ring has been evaluated in this paper. The assessment was carried out for consumer risks of 0.1 % to 0.5 %. The measurement data are assumed to belong to a normal distribution. Since the roundness deviation of the inner bearing ring is always a positive value, the following prior distributions were selected: gamma distribution, beta distribution, one-parameter Rayleigh distribution, and two uniform distributions on different intervals.

Izvorni jezik
Engleski

Znanstvena područja
Matematika, Interdisciplinarne prirodne znanosti, Strojarstvo

POVEZANOST RADA