On the possibility of sign-perturbed sums method application for the processing of dynamic tests data

Abstract

Methods of measurement and prediction of materials dynamic strength are complicated and non-standardized. The way of experimental data treatment based on incubation time criterion is considered. In practice, we only have a finite number of noisy measurements and limited statistical knowledge about the noise since dynamic tests are labor consuming and each assay demands a lot of time. This strongly restricts the number of available data treatment methods unless we are satisfied with approximate, heuristic solutions. Under the assumption of noise symmetries, we use Sign-Perturbed Sums (SPS) approach for constructing finite-sample confidence regions which have exact confidence probabilities, i.e. they contain the true parameter with a user-chosen exact probability for any finite data set. It is shown that a few experimental points are enough to determine strength parameter with accuracy proper for engineering. The proposed method is demonstrated through a number of practical experiments.