12. Balancing of Planar Mechanisms Having Imperfect Joints Using Neural Network-Genetic Algorithm (NN-GA) Approach

As a result of design, manufacturing and assembly processes or a wear effect, clearances are inevitable at the joints of mechanisms. In this study, dynamic response of mechanism having revolute joints with clearance is investigated. A four-bar mechanism having two revolute joints with clearance is considered as a model mechanism. A neural network was used to model several characteristics of joint clearance. Kinematic and dynamic analyses were achieved using continuous contact mode between journal and bearing. A genetic algorithm was also used to determine the appropriate values of design variables for reducing the additional vibration effects due primarily to the joint clearance. The results show that the optimum adjusting of suitable design variables gives a certain decrease in shaking forces and their moments on the mechanism frame.