Time-Dependent Reliability Analysis for Function Generator Mechanisms

Abstract

A function generator mechanism links its motion output and motion input with a desired functional relationship. The probability of realizing such functional relationship is the kinematic reliability. The time-dependent kinematic reliability is desired because it provides the reliability over the time interval where the functional relationship is defined. But the methodologies of time-dependent reliability are currently lacking for function generator mechanisms. We propose a mean value first-passage method for time-dependent reliability analysis. With the assumption of normality for random dimension variables with small variances, the motion error becomes a nonstationary Gaussian process. We at first derive analytical equations for upcrossing and downcrossing rates and then develop a numerical procedure that integrates the two rates to obtain the kinematic reliability. A four-bar function generator is used as an example. The proposed method is accurate and efficient for normally distributed dimension variables with small variances.