Kinematic accuracy analysis of planar mechanisms with clearance involving random and epistemic uncertainty

• Kinematic accuracy of mechanisms with clearance involving random and epistemic uncertainty has been analyzed.
• A general methodology based on interval analysis and probability theory for kinematic accuracy analysis of a mechanism is presented.
• Surrogate models based on neural network approach are constructed.
• Examples are presented to illustrate the reliability and accuracy of the proposed method.

A general methodology for kinematic accuracy analysis of mechanisms with clearance, involving random and epistemic uncertainty, is presented. A continuous contact force model with consideration of the energy dissipation is applied to evaluate the contact force. Probability space and interval are used to characterize random uncertainty and epistemic uncertainty, respectively. Considering the inefficiency of uncertainty prediction, surrogate models based on neural network approach are constructed. The precision and reliability of the presented method have been validated by numerical simulation. It indicates that, comparing with probability theory, the reliability of the results obtained by the presented method is improved as well as the prediction accuracy, comparing with interval analysis.