A connectionist approach to hydroturbine speed control parameters tuning using artificial neural network

The adjustment of speed control parameters is crucial for stability and satisfactory transient behaviour of hydroturbine units. Turbine nonlinearities, changing dynamic nature of electric loads and hydraulic supply system, are inherent to operations of hydroelectric power plants and have to be considered in the design of the speed controller. This work describes a phase of a collaborative research between the University of Para and the North-Electric Power Company (Eletronorte) to update and improve speed control technology in hydrogenerators. A connectionist approach is proposed to auto adjusting a PI digital speed control parameters of an hydroelectric isolated power plant to changing operating conditions by an Artificial Neural Network (ANN). Preliminary testing using a simulator of the hydroelectric power plant demonstrated the effectiveness of the proposed scheme under disturbances conditions.