Neuro-fuzzy modeling of dynamic systems in energetics using pruning methods

The presented paper deals with the modeling principles of nonlinear dynamic systems and with the design of optimal model structures created by artificial neural networks and fuzzy theory. Four models were created for modeling of selected system: a neural network model, an adaptive neuro-fuzzy model and two neural networks with optimal structure obtained using Optimal Brain Damage and Optimal Brain Surgeon algorithms. The quality of proposed solutions were compared and verified on a real system of steam turbine in thermal power plant and in nuclear power plant.