A PFC-based Hybrid Approach for Control of Industrial Heating Furnace

Abstract

Background and Objectives: In this paper, a predictive functional control based on Laguerre functions is designed for control of an industrial heating furnace. The fractional order model (FOM) of the heating furnace is assumed as the plant model.Methods: For designing the predictive functional controller (PFC), a reduced integer order approximation of the fractional order heating furnace model is derived. The order of the reduced integer model is determined based on Hankel singular values of the original system. Coefficients of the reduced integer model are assumed to be unknown. Unknown parameters are then obtained by minimizing a many-objective fitness function including weighted summation of differences of step responses, steady state errors, maximum overshoots as well as magnitude of frequency responses of the original and reduced systems. Routh-Hurwitz criteria are used as stability criteria and added to optimization problem as its constraints. The optimization tool is Genetic algorithm.Results: Advantages of the proposed method are preserving stability and focusing on various important features of both time and frequency responses of system. In addition, it uses a direct order reduction method without the need to intermediated approximations such as Oustaloup approximation.Conclusion: Laguerre-based PFC controller has been evaluated via two scenarios and the obtained results represent the satisfactory performance of the proposed controller.

Keywords

• fractional order system
• genetic algorithm
• model predictive control
• model order reduction
• predictive functional control