Semiconductor Packaging Equipment Maintenance Based On Time Series Studies And Optimization Of The Method

In industrial production, equipment will break down inevitably. The high frequency of failures can reduce the reliability and productivity of equipments greatly. It also has different effects on products'quality and the accomplishment of production plan on schedule. Timely and appropriate testing and maintenance can lower equipments'failure rate, prolong service life, and ensure practical operation in good condition.The maintenance management of equipment is a domain associated with practicality closely. Its development includes four main stages, namely, breakdown maintenance, preventative maintenance, productive maintenance, and side-by-side development of various equipment management modes. Condition-based maintenance has become a developing trend in maintenance management of equipment. Researchers home or abroad have made many achievements in one of its branches—failure prediction technology based on time series.This paper is mainly about the optimization of preventative maintenance of a set of semiconductor assembly equipment. It includes several parts as follows.1. Review the development, theory and trend of equipment maintenance management.2. Introduce such backgrounds of a certain machine as basic mechanism, working principle, position in the streamline and fulfilled functions, in one semiconductor assembly and test factory. At the same time, present the current preventative maintenance system of this machine, and indicate the shortcomings and aspects needed improvement for this system.3. Present the theoretical method of failure prediction of the equipment based on time series, and choose the most appropriate theoretical model, that is, a theoretical method based on the classical time series analysis.4. Collect related information concerned about the machine in practical production, and get sample data of the machine's Throughput by calculation, and establish time series ARIMA model. Then by comparison with the data in practical production, obtain ideal predictive result comparatively.5. Based on effective failure analysis of this machine in a certain phase, complete the mission of collection and arrangement maintenance logs and batch logs about some type failures of this machine in practical production. Analyze and calculate to obtain the sample data of operation time between failures. Establish time series ARIMA model, and then realize the optimization of the original preventative maintenance system finally.This paper is based on time series ARIMA model, and actualizes modeling and prediction of the equipment productive throughput and operation time between failures. It achieves the goal of the original preventative maintenance system's optimization, which provides relevant foundations for the technicians to maintain the equipments timely and accurately. In addition, equipment management and maintenance in other spheres of industrial production also can take this method for reference.