Prediction Model of the Remaining Useful Life of the Drill Bit during Micro-Drilling of the Packaging Substrate

Abstract

The packaging substrate plays a significant role in electrical connection, heat dissipation, and protection for the chips. With the characteristics of high hardness and the complex material composition of packaging substrates, drill bit failure is an austere challenge in micro-drilling procedures. In order to monitor the health state of the drill bit and predict its remaining useful life (RUL) in micro-drilling of packaging substrate, an improved RUL prediction model is established based on the similarity principle, degradation rate, and offset coefficient. And then, a micro-drilling experiment on packaging substrate is carried out to collect the axial drilling force through the precision drilling force measurement platform. Axial drilling force signals, which are processed via the Wiener filtering method, are used to analyze the effectiveness of the improved RUL prediction model. The experiment results indicate that, compared to the curves of the traditional RUL prediction model, the curves of the improved RUL prediction model present a higher fitting degree with the actual RUL curves. The average relative errors of the improved RUL prediction model are small and stable in all groups; all of the values are less than 15%, while the fluctuation of the average relative errors of the traditional model is greatly large, and the maximum value even reaches 74.43%. Therefore, taking the degradation rate and offset coefficient into account is a proper method to enhance the accuracy of the RUL prediction model. Furthermore, the improved RUL prediction model is a reliable theoretical support for the health state monitoring of drill bits during the micro-drilling of packaging substrates, which also acts as a potential method to improve micro hole processing efficiency for packaging substrates.