Neurofuzzy modelling of the reflow thermal profile for surface mount assembly

Abstract Troubleshooting and production set-up time are often the major contributors for productivity loss in surface mount assembly. The solder reflow process is one of the least understood process segments. The research proposes a neurofuzzy reflow thermal profile control system for a convection oven. An incompatible temperature profile can result in a poor soldering quality and reduced system throughput. This paper proposes a neurofuzzy-based solder reflow thermal profile control system that responds to facilitate the production set-up process. The model is trained and constructed using data from both an experimental design and from historical production records. Customized computer code is used to generate a user-friendly human–machine interface and to link between neurofuzzy-reasoning rules and reflow oven set-up parameters for thermal profile planning and control. Empirical results illustrate the effectiveness and efficiency of the proposed system to solve a practical application. Acknowledgement This work is supported in part by the National Science Council of Taiwan, Republic of China under Grant No. NSC90-2212-E006-076.