Software Tools for System Identification and Control using Neural Networks in Process Engineering

Neural networks offer an alternative approach both for identification and control of nonlinear processes in process engineering. The lack of software tools for the design of controllers based on neural network models is particularly pronounced in this field. SIMULINK is properly a widely used graphical code development environment which allows system-level developers to perform rapid prototyping and testing. Such graphical based programming environment involves block-based code development and offers a more intuitive approach to modeling and control task in a great variety of engineering disciplines. In this paper a SIMULINK based Neural Tool has been developed for analysis and design of multivariable neural based control systems. This tool has been applied to the control of a high purity distillation column including non linear hydrodynamic effects. The proposed control scheme offers an optimal response for both theoretical and practical challenges posed in process control task, in particular when both, the quality improvement of distillation products and the operation efficiency in economical terms are considered.

Authors:

• J. Fernandez de Canete
• S. Gonzalez-Perez
• P. del Saz-Orozco

Keywords:

• Distillation
• neural networks
• software tools
• identification
• control.

References:

[1] D. Fruehauf and D. Mahoney, Improve Distillation Control Design.
Chemical Engineering Progress, March 1994.
[2] M.A. Hussain, "Review of the applications of neural networks in
chemical process control. Simulation and on-line implementations",
Artificial Intelligence in Engineering, Vol. 13, pp. 55-68, 1999.
[3] H. Demuth, M. Beale and M. Hagan, M Neural Network Toolbox for
use with MATLAB. The Mathworks, 2006.
[4] M. Norgaard, O. Ravn and N. Poulsen, "NNSYSID and NNCTRL
tools for system identification and control with neural networks",
Computing and Control Engineering Journal, Vol. 23, pp. 29-36, 2001.
[5] J. Dabney, T. Harman. Mastering SIMULINK, Prentice Hall, 2004.
[6] G. Cybenko, "Approximation by superposition of sigmoidal
functions," Math. Contr., Signals, Syst., vol. 2, pp. 303-314, 1989.
[7] S. Haykin, Neural Networks: A comprehensive foundation, 2nd ed.
Prentice Hall, 1998.
[8] M.T. Hagan and M. Menhaj, "Training feed-forward networks with the
Marquardt algorithm," IEEE Transactions on Neural Networks, Vol. 5,
pp. 989-993, 1994.
[9] Norgaard, M, O. Ravn, N.K. Poulsen and L.K. Hansen. Neural
Networks for Modelling and Control of Dynamic Systems. Springer
Verlag, 2000.
[10] M. Diehl, I. Uslu, R. Findeisen.,"Real-time optimization for large scale
processes: Nonlinear predictive control of a high purity distillation
column", On Line Optimization of Large Scale System:State of the Art,
Springer-Verlag, 2001.