Non-probabilistic time-dependent reliability analysis of mechanism motion accuracy based on extreme value response

In this work, a novel non-probabilistic time-dependent reliability analysis method is proposed for evaluating the motion accuracy of a mechanism. Previous studies have usually assumed that uncertain variables satisfy a certain probability distribution and have given insufficient consideration to time-dependent variables. In practical engineering problems, however, there are also uncertain variables that need to be expressed in non-probabilistic models. The novelty of the proposed method is that the nonprobability interval variables and time-dependent variables are considered together in the modeling, and extreme value response analysis is applied to the reliability analysis. As a result, the time-dependent reliability problem is transformed into an instantaneous reliability problem for simplified calculation. The reliability indices obtained can be used to evaluate the reliability of the motion accuracy of the mechanism. Finally, the accuracy and efficiency of the local optimization algorithm in the non-probabilistic model are obtained through the analysis of two application examples.