Using historical results obtained in the tensile tests for Type A evaluation of uncertainty

The specification of mechanical properties is essential for adequate choice of material, as well as for design and manufacturing of components and products. The main limitations in estimating uncertainty in testing are related to specimens destruction, sample size and existence of a wide range of influences quantities that are hard to quantify. This paper shows that it is possible to make use of historical data to minimize the effects of these limitations. A case study applied to the tensile test is presented, showing that Type A evaluation of uncertainty obtained from multiple homogeneous samples is more appropriate for the estimation of measurement uncertainty. And it is also demonstrated that the application of strength standard deviation has higher representativeness than the use of diameter and force standard deviations separately.