Multi-objective optimization of mechanisms with clearances in revolute joints

Abstract

Despite extensive work on the evaluation of the dynamic performances of various mechanisms and multi-body mechanical systems with revolute joint clearances, limited work has been conducted in optimizing the performances of these systems. A multi-objective optimization technique is presented to examine and to quantify the effect of combined objective functions with several design variables on the response of the systems. In particular, a Kriging meta-model based on the Design-of-Experiment method is utilized to optimize the systems’ performance. The reason for implementing this meta-model is to replace the computational intensive simulations with a more efficient mathematical model. In this study, a simple slider-crank mechanism with a revolute clearance joint at the slider pin is modeled, and its dynamic response is analyzed using the multi-body dynamic software, MSC ADAMS. The revolute joint clearance is modeled as a pin-in-hole dry contact utilizing the Hertzian contact force model with hysteresis damping. Response surfaces are generated on the prediction of the system’s performances for three different objective functions and for different range of design variables. The objective functions are combined to develop a single response surface characterizing the dynamic responses of the system at different range of design variables in order to optimize its performance.