ANFIS based Dual Unified Power Quality Conditioner to Mitigate Voltage Sag/Swell and to Improve the Power Quality of the System

The project mainly describes the utilization and the advantages of the adaptive neural fuzzy inference system while the application of the same is done at the dual unified power quality conditioner. The basic idea of compensating for the power quality issues like voltage sag and swell will remain same but the controlling part that was involving with a PI controller initially will be replaced with an adaptive neural fuzzy inference system. The ANFIS system will be trained using the same PI controller output only and it will be designed or more specifically it will set its own rules based on the output it is getting from the PI controller. The two different inputs that this ANFIS will take up are one from the same steady state error evaluation block and the other the derivative block involved which is for the change in the error value with dynamic changes in the system behaviour. The same conditions like as of the PI controller is checked with the ANFIS and the output graphs are plotted for various cases. The output compensation graph is seen for the case and from that it could easily be make out that the voltage compensated waveform is having a greater system response compared to the PI controller and the output voltage starts hitting the magnitude which was given as the reference in the parallel active filter quite earlier than the other controller. One more advantageous feature of this same controller could be noted down while be calculating for the THD. The THD values variation could easily be seen for the ANFIS. A MATLAB/SIMULINK concept will be proposed here along with the comparison study for the both the controllers and various cases of the power quality issues will be checked.