A Real-Time Comparative Study on the Performances of UPQC with Different Control Algorithm

In this paper, a comparative study of neural network based adaptive least mean square (ALMS) learning mechanism with traditional instantaneous active and reactive (p-q) control technique is proposed for 3-φ, 3-wire distribution system. The voltage source converter (VSC) based proposed Unified power quality conditioner (UPQC), aims to manage the power quality (PQ) problems. The ALMS-NN controller is implemented by its own learning mechanism using embedded system toolbox which combines various weighting parameters such as learning rate, step size to obtained compensating reference current and voltage for the triggering of VSC based UPQC. Similarly, the p-q control method is implemented for UPQC for comparison purpose. Among these two control method, the ALMS provides better voltage regulation, suppression of total harmonic distortion, load balancing, power factor correction and rating of UPQC is also reduced. The effectiveness of ALMS controlled UPQC over traditional p-q control method is portrayed by vigorous simulation using MATLAB/Simulink software in parallel with real-time implementation by RTDS (real time digital simulator).