The free-model based neural networks are implemented in the model reference adaptive inverse control scheme for a boiler-turbine plant. Since the mathematical model for the plant is assumed to be unknown, neural networks and the model reference adaptive inverse control scheme are applied in this paper. In order to obtain smaller error in the initial phase of training, free-model based neural networks are applied as the neuro-identifier and neuro-controller in the control scheme. The functional mapping is applied for the drum pressure set point from the power demand in order to be more economical on the fuel burning at the throttle valve. How to choose the free-model order for the unknown plant is discussed. Different rates for the change of power set points can be attempted in order to achieve more satisfactory performance in the controlled plant outputs.
展开▼
