A Radial Basis Function Network Model for the Adaptive Control of Drying Oven Temperature
Artificial neural networks are new modelling tools for process control, especially in non-linear dynamic systems. They have been shown to successfully approximate non-linear relationships. This paper describes a neural control scheme for the temperature in a drying oven. The control strategy used is internal model control, in which the plant is modelled by a radial basis function network. The process was identified in an off-line phase —training the network while determining its optimal structure— and an on-line phase —adapting the neural model to any change in the process dynamics. The control strategy was tested experimentally in a series of trials with the oven empty and loaded.
KeywordsHide Layer Model Predictive Control Nonlinear Dynamic System Radial Basis Function Neural Network Radial Basis Function Network
Unable to display preview. Download preview PDF.
- Hermes, Paris. Ljung, L.: System identification — Theory for the user. (1987) Prentice-Hail, Englewood Cliffs, London.Google Scholar
- Lowe, D., Webb, A.R.: Time series prediction by adaptive networks: a dynamical systems perspective. IEE Proceedings-F 138 (1991) 17–24.Google Scholar
- Powel, M.J.D.: Radial Basis functions for multivariable interpolation: a review. In J.C. Mason and M.G. Cox (Eds) Algorithms for Approximation (1987) 143–167 Clarendon Press, Oxford.Google Scholar