Improving inverse control on real time by means of neural networks

Although nonlinear inverse control techniques based on artificial neural networks have been extensively studied, their use in real time applications is generally limited. In this paper, that disadvantages are analysed and an inverse neural system to the real-time control of the heat transfer fluid temperature in a pilot chemical reactor is proposed. The training of the neural controller is carried out using both, generalised and specialised learning. This allows to obtain a set of weights to initialise the controller which acts in real time. Appropriate performances of inverse neural controller can be achieved. The behaviour of that controller is compared against the PID controller currently installed in the plant. The results show that neural-based controllers improve the performance of the real plant.