Performance Evaluation of Seven Level Reduced Switch ANPC Inverter in Shunt Active Power Filter with RBFNN Based Harmonic Current Generation

In this project, a modified seven level boost Active-Neutral-Point-Clamped (7LB-ANPC) inverter is utilized as a Shunt Active Power Filter (SAPF). Another vital aspect of this work is to retain the link voltage across the capacitor, which is accomplished through a PI controller tuned with an Adaptive Neuro-Fuzzy Inference System (ANFIS). 

An adaptive instantaneous p-q theory is instigated in the direction of extracting reference current and the harmonic extraction is carried out by using Radial Basis Function Neural Network (RBFNN). Gating sequence of inverter is generated for the outputs, which are attained from ANFIS and RBFNN and thus the opposite harmonics are injected to the Point of Common Coupling (PCC) by which current harmonics are eliminated with reactive power compensation. 

The 7Lb-ANPC inverter has a minimized number of switching devices with low switching losses and high boosting ability. By RBFNN based reference current generation, the source current THD of 0.89% is achieved. The proposed methodology is simulated through MATLAB/SIMULINK Software.

Keywords: Power Quality, Shunt Active Power Filter, Multi-Level Inverters, Active-Neutral-Point Clamped Inverter, Radial Basis Function Neural Network, Adaptive Neuro-Fuzzy Inference System.