A megfelelőség értékelésének átalakítása a bizonytalanságok és kockázatok figyelembevételével

Absztrakt

The fierce competition for customers gives a double pressure onto companies: to reduce costs, to keep or to improve quality. Right decisions on product conformity are vital for profitability. Widely used classical conformity assessment and process control techniques alone are not able to meet these requirements. One of their drawbacks is that they neglect the uncertainty of measurements. Variance of measure is added to the variance of the process resulting a greater fluctuation of data. In this case variance related calculations lead to poor acceptance decisions, especially close to the tolerance limits. In this paper a new risk-based approach is presented and particular techniques are proposed for taking measurement uncertainty and consequences of the decision into consideration hence critical values for acceptance decisions can be optimized. The proposed method gives the optimal results even if the normality criteria of the probability distribution is violated. Application of the technique is illustrated on a small practical problem.