Space/time Separation based Process Design for Distributed Cure Control

Many of industrial processes, like IC cure process in semiconductor packaging industry, are spatially distributed dynamic systems. This kind of process is very difficult to model, design and control due to the inherent spatio-temporal dynamics. On the other hand, process design and control have long been separated into two different disciplines in both academia and industry. This separation has made it difficult to achieve an optimal solution to industrial productions, and even more difficult and challenging for the spatially distributed dynamic process.We aim to develop a novel model-based optimal design for distributed control of the IC cure oven under the space/time separation framework. First, the design variable will be considered first time in the time-space separation to develop a parametric model framework of the cure process.Through the model-based optimal design, the process can be designed to have a linear dominant dynamics with the minimized nonlinearity. This design will also result in a good model for the subsequent decoupling control.Using the linear nominal model, the observer-based decoupling mechanism can be designed to suppress the model-process mismatch and the external disturbance, so that the process can be transformed approximately into a linear system. This approximate linear system may match better with time-space separation for the purpose of easy control.By fine-tuning the nominal modeland setpoints of control loops, the near optimal cure profile can be achieved and maintained over the entire spatial domain.