Hybrid dimension reduction for mechanism reliability analysis with random joint clearances

Randomness in mechanism dimensions and joints makes the mechanism motion deviate from its designed motion. The probability (reliability) that such deviation is within an error tolerance limit should be invariably large. This study shows that the accuracy of the reliability analysis for dependent joint clearances is insufficient by existing kinematic reliability methods, such as the First Order Second Moment (FOSM) Method and First Order Reliability Method (FORM). We therefore propose a Hybrid Dimension Reduction Method (HDRM) to better handle the dependent joint clearance variables. With the first order Taylor expansion for independent dimension variables and bivariate dimension reduction for dependent joint clearance variables, HDRM produces more accurate solutions than the FOSM and FORM while maintains higher efficiency than FORM and Monte Carlo simulation. A slider-crank mechanism is used as an example for the methodology demonstration and validation.

► We demonstrate the inaccuracy of existing methods for mechanism reliability with dependent joint clearances.
► We propose a hybrid dimension reduction method for mechanism reliability with random joint clearances.
► The new method is accurate and efficient.