By exploiting the input dependent nature of the parameters of a discrete-time bilinear model structure, the paper combines knowledge of both discrete and continuous model parameters to form a hybrid approach to self-tuning controller design. With consideration given to an identified bilinear model of a high temperature heating plant application the appropriateness of the hybrid approach is demonstrated. In particular it is shown that the hybrid approach may be used to advantage within a Kalman filtering prediction/correction parameter estimation scheme. Practical issues relating to implementation of the approach are also discussed.
Non-linear systems are often represented by a linear model around an operating point. However in many industrial processes large variations in operating conditions make this representation inaccurate. A bilinear model representation is often used as a compromise between a linear model and the severe non-linearity of the process. In this paper modelling of a bilinear system using frequency response methods is considered. Conventional frequency response methods cannot be applied to bilinear systems because of the harmonics produced by the multiplicative term in the bilinear representation. This paper obtains theoretical relationships between the complex amplitudes of the Fourier series components present at the output of a bilinear system when the input is a sinusoid. By applying these relationships to measured Fourier components the parameters of the system can be identified and a bilinear model obtained. The method is applied to the modelling of a temperature control system.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.