Qualitative interpretation of creep and relaxation experimental data

Automated interpretation of experimental data is a necessary component in automated model building environments, i.e. tools capable to select a model which explains the observed behavior of a physical system. This paper describes an algorithm for the qualitative interpretation of experimental data whose goal is to single out the qualitative properties which characterize the behavior of a real physical system, and, consequently, to make a guess for a set of plausible candidate models of the phenomenon under study. We consider observations which are gathered from creep and relaxation experiments on visco-elastic materials. In this specific context, the interpretation problem consists in determining rules for the conversion of the numerical data into the specific qualitative properties which characterize the creep and relaxation response of the material. Such properties allow us to select within a library of mathematical models [4] of ideal materials a set of candidate models which describe, at least qualitatively, the behavior of the material.