Neural network control of static synchronous compensators

Neural Network Control of Static Synchronous Compensators In last years, increasing energy demands and widespread use of precision equipment has led to concepts of energy efficiency and power quality. Voltage drops, surges, harmonics and reactive power are the most important quality problems of complex power system. To minimize this problem, reactive power compensation is an effective method. With developments in power electronic devices, inverter-based static compensators have become popular rather than conventional compensation techniques based on mechanically switched shunt reactor or capacitor groups. Static Synchronous Compensator (STATCOM) is the most prominent devices, known as Flexible AC Transmission Systems. In this thesis, reactive power compensation is studied by using STATCOM which is connected to load fed from power system. Reactive power compensation can be accomplished by controlling the inverter voltage in the inverter based STATCOMs. Hence, DC supply voltage or pulse with modulation index should be controlled for the reactive power compensation. In this study, reactive power compensation is performed by controlling the modulation index with Neural Networks (NN). Simulation results of neural network based STATCOM are obtained from the prepared MATLAB *.m files and the designed MATLAB / Simulink environment also. Simulation results showing the reactive power compensation performance of PI and NN controlled STATCOM represent the effectiveness of the NN controlled STATCOM. Keywords: Power Quality, Reactive Power Compensation, STATCOM, PI Controller, Neural Network Controller.