System Framework and Configuration for On-line Quality Assurance System Framework and Configuration for On-line Quality Assurance

It is now widely accepted within the industrial community that quality assurance is essential for future survival and competitiveness. However, most quality improvement activities have been performed off-line without being integrated into the shop floor control system (SFCS) that focuses on efficient production planning and scheduling. An integrated systematic framework for both enhancing productivity and assuring quality in real-time has yet to appear. The goal of the paper is to propose the functional and implementational architecture necessary for developing a quality-oriented shop floor control system (Q-SFCS). The on-line quality assurance module in the control system is functionally composed of 2 auxiliary components, data matching and knowledge acquisition, and 3 primary components, model design, feedforward adjustment, and online adjustment. Each component is responsible for matching sensory data and quality data (data matching), knowledge definition and generation (knowledge acquisition), prediction and control model construction (model design), monitoring, diagnosis, and optimization of the quality (feedforward adjustment, on-line adjustment). To consider the relationship between multiple process variables and multiple quality variables effectively, a partial least square (PLS) model is formulated and then implemented. To obtain the set of optimal process parameters, model predictive control models are formulated and implemented. The conceptual framework proposed in the paper has been applied to the shadow mask manufacturing process. The experimental results showed that quality has been significantly improved and further, the handling expenses associated with quality faults have been dramatically decreased.