Suboptimal Nonlinear Predictive Control with MIMO Neural Hammerstein Models

This paper describes a computationally efficient (suboptimal) nonlinear Model Predictive Control (MPC) algorithm with neural Hammerstein models. The Multi-Input Multi-Output (MIMO) dynamic model contains a steady-state nonlinear part realised by a set of neural networks in series with a linear dynamic part. The model is linearised on-line, as a result the MPC algorithm solves a quadratic programming problem. The algorithm gives control performance similar to that obtained in nonlinear MPC, which hinges on non-convex optimisation.

[1]  W. Greblicki.  Non-parametric orthogonal series identification of Hammerstein systems , 1989 .

[2]  Michael A. Henson,et al.  Nonlinear model predictive control: current status and future directions , 1998 .

[3]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[4]  Jay H. Lee,et al.  Model predictive control: past, present and future , 1999 .

[5]  Piotr Tatjewski,et al.  Advanced Control of Industrial Processes: Structures and Algorithms , 2006 .

[6]  Sirish L. Shah,et al.  Constrained nonlinear MPC using hammerstein and wiener models: PLS framework , 1998 .

[7]  Piotr Tatjewski,et al.  Soft computing in modelbased predictive control footnotemark , 2006 .

[8]  M. Ayoubi.  Comparison between the dynamic multi-layered perceptron and the generalised Hammerstein model for experimental identification of the loading process in diesel engines , 1997 .

[9]  Andrzej Janczak,et al.  Neural network approach for identification of Hammerstein systems , 2003 .

[10]  S. Joe Qin,et al.  A survey of industrial model predictive control technology , 2003 .

[11]  I. Mareels,et al.  Dead-beat control of simple Hammerstein models , 1998, IEEE Trans. Autom. Control..

[12]  Stephen A. Billings,et al.  Identification of systems containing linear dynamic and static nonlinear elements , 1982, Autom..

[13]  Jan M. Maciejowski,et al.  Predictive control : with constraints , 2002 .

[14]  Maciej Ławryńczuk,et al.  A Family of Model Predictive Control Algorithms With Artificial Neural Networks , 2007, Int. J. Appl. Math. Comput. Sci..

[15]  V. Chandrasekar,et al.  Hammerstein model identification by multilayer feedforward neural networks , 1997, Int. J. Syst. Sci..

[16]  Daniel E. Rivera,et al.  Nonlinear black-box identification of distillation column models - design variable selection for model performance enhancement , 1998 .

[17]  Stanley H. Johnson,et al.  Use of Hammerstein Models in Identification of Nonlinear Systems , 1991 .