An ANN Based Capicitor Voltage Balancing Method For Neutral Point Clamped Multi-Level Inverter

摘要

Multi-level inverters are became popular for usage in medium voltage, low voltage power applications due to flexibility in control and better performance characteristics in terms of harmonic regulation. Neutral point clamped are popular as they require less number of sources as their input when compared with their counter parts i.e. cascaded multi-level inverters and found to be reliable when compared with flying capacitor based multi-level inverters. But when neutral clamped technologies are used for generation of three-phase voltages, the capacitors that are connected at input side experiences imbalance in their voltages, this makes neutral point clamped multi-level inverters less reliable. In the proposed work an attempt is made to study circuits that balances the capacitor voltages and a scheme is investigated for balancing the capacitor voltages. Method proposed in [1] uses PID controller for balancing the capacitor voltages. In this project PI based control scheme and artificial neural network (ANN) based control scheme for the front end circuit shown in [1] are designed for achieving balance among the capacitor voltages. The proposed control scheme is simulated with the help of Simpowersystems block set and neural network toolbox of MATLAB software for different load conditions. Results obtained from ANN based controller and PI controller are presented.