Abstract: The dispensing robot uses piezoelectric stack to drive the impact needle up and down reciprocating high speed motion to generate pressure difference to achieve spraying dispensing, so as to achieve the purpose of semiconductor and electronic device surface packaging. One of the most important failure modes of dispensing robot is the reduction of service accuracy life due to the wear distortion of core parts impact needle. The impact needle material is the WC-Co cemented carbide prepared by powder metallurgy technology, and there are microstructure differences of materials such as voidage and grain size. It is difficult to build a physical model for impact needle wear, which involves the difference of microstructure, the impact frequency and the axial travel. At the same time, existing wear models were constructed based on macroscopic impact conditions, and rarely consider heterogeneous properties of cemented carbide materials. In this paper, a multi-factor coupling impact fretting wear model is established considering the microstructure difference of the impact needle material and the uncertainty of operating conditions. The uncertainty modeling of the wear parameters involved in the impact needle is carried out by the quadratic derivation λ-PDF method, and the uncertainty propagation of the wear is analyzed by the dimensionality reduction integral method. The results show that the model and the proposed method can well reflect the impact fretting wear of the impact needle, and provide a new perspective and a new method for the quality protection design of domestic long-life, high-precision and high-reliability dispensing robot.

Key words: impact-fretting wear, multi-source uncertainties, wear model, propagation of uncertainty