Power quality improvement of the 33kv north-bank distribution network using artificial neural network based Dynamic Voltage Restorer (DVR)

Power quality issues such as voltage sags, swells, and harmonics contribute to over 90% of customer power interruptions in distribution networks, leading to increased downtime, equipment damage, and financial losses. The North-Bank 33kV distribution feeder in Makurdi experiences voltage fluctuations exceeding IEEE 519 and IEC 61000-3 standards, with a Total Harmonic Distortion (THD) of 6.67%, surpassing the recommended 3–5% limit. This study presents an Artificial Neural Network (ANN)-based Dynamic Voltage Restorer (DVR) to mitigate these disturbances and enhance power reliability. Using MATLAB/SIMULINK, the system was modeled and simulated under fault conditions, comparing the performance of Proportional-Integral (PI) and ANN controllers. Results show that while both methods mitigate voltage disturbances, the ANN-controlled DVR exhibits 15% faster response time, 99% classification accuracy, and reduces THD to below 5%. The DVR effectively compensates for voltage sags within 70 milliseconds, restoring voltage to the acceptable range of 0.95–1.05 p.u. across various fault scenarios, including line-to-ground and line-to-line-to-ground faults. The ANN-based approach outperforms conventional methods by dynamically adjusting to changing load conditions, ensuring a stable and reliable power supply. These findings validate the DVR as a viable and intelligent solution for improving power quality in modern distribution networks, reducing equipment failures, and minimizing operational losses.