A Knowledge-based System for Stencil Printing Process Planning and Control

Surface mount technology (SMT) is one of the most important developments in electronic industry. A surface mount assembly (SMA) has three consecutive manufacturing steps: solder paste stencil printing, component placement, and solder reflow. Stencil printing process (SPP) involves highly operation complexity and has multiple quality characteristics, and averagely accounts for 60% of soldering defects in SMA. This work presents a knowledgebased system for SPP planning and control to upgrade soldering quality level and system performance. A hybrid data set contains a 3 experimental design and statistical process control (SPC) records was collected firstly, and followed by data processing for removing conflicted, inconsistent, and redundant samples through a fuzzycluster algorithm. The output columns of the clustered data were then transformed using fuzzy quality loss function (FQLF) with respect to the dissatisfaction of printing performance. The neuro-fuzzy technique was adapted to model and learn SPP knowledge from the transformed data set into a SPP knowledge base. Finally, a GUI man-machine interface was developed to help engineers in predicting responses and evaluating the overall SPP performance. The empirical evaluations of soldering quality and productivity demonstrate the effectiveness and efficiency of this proposed system.