This paper mainly targets the nonlinear characteristics in industrial heating furnace control and uses the multimodel method to decompose the global dynamics of the process into a series of local model sets. The corresponding predictive functional controller is designed using the local fractional order models of the set within their working range. The weight coefficients of the different models in the model sets are obtained based on the error of the different model sets at current moment, and the system performance at current moment is obtained by weighting between many models using the weight coefficient. Since the accuracy of the local models is increased, the error between the local model set and the process output is reduced, which decreases the influence of the system model inaccuracy on the system performance. Finally, the strategy is verified through the experiment on the temperature of a heating furnace.
A new genetic-algorithm (GA)-based infinite horizon linear quadratic (IHLQ) control method is developed for injection molding batch processes under partial actuator faults and unknown disturbances. To obtain improved system performance, an extended state space model, which combines the process state and output error information, has been adopted, where extra tuning freedom is gained for the corresponding control system design by adjusting the extended weighting matrix. However, no specified rules are established for the choices of these weighting coefficients. To cope with such situations, GA is employed to optimize these weighting factors in this paper. Furthermore, the robust stability of the control system, which can be regarded as the process under uncertainty, is addressed. Finally, the validity of the proposed approach is tested on the injection velocity regulation process.
This paper addresses the genetic algorithm (GA) optimization and the linear quadratic (LQ) structure based predictive functional control (PFC) for batch processes under non-repetitive unknown disturbances and partial actuator faults. First, by adopting the extended non-minimal state space (ENMSS) model in which the state variables and the tracking error are united, the new state vector with more degrees is provided for the controller design. In order to enhance the ensemble control performance under the PFC structure, GA is adopted for the optimization of the weighting matrix in the controller. The case study on the injection velocity control in an injection molding machine demonstrates the effectiveness of the proposed PFC scheme against various disadvantages.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.