Pattern-based adaptive control for sintering process

摘要

This paper presents a new pattern-based adaptive control method for sintering process. It is the actual run status data driven directly and pattern classes are divided. The system model is established by the changing between pattern classes in a new space called “pattern moving space”. Pattern class variable has been defined in this space. This modeling method is different from state equation and differential equation. The adaptive controller is designed by this new statistical variable. In the new controller, the output is pattern class variable (this process is equivalent to soft sensor measurement process) and system model parameters are estimated by pattern class variable. The adaptive control method proposed here provides a good alternative to sintering process and is readily applicable to other similar complex industrial production processes.