Machine Learning Techniques for Power Quality Enhancement of Power Distribution Systems with FACTS devices.

The power quality problem refers to the issues caused by the sudden rise of nonstandard voltage, current, or frequency. The problems that emerge from poor power quality due to non-linear loads are voltage sag, swell, interruptions, harmonics, and transients in distribution systems. Various compensation devices are used nowadays to improve power quality. The advances in power electronic technologies improve the reliability and functionality of power electronic based controllers, resulting in increased applications of FACTS devices like DSTATCOM and Dynamic Voltage Restorer (DVR) which are fast, flexible, and efficient solutions to the power quality problems. These devices are used to restore the source, load voltage, and current disturbances caused by different loads and faults. These devices were tested in a standard IEEE 14-bus system for Total Harmonic Distortion (THD) minimization while utilizing PI-based Artificial Neural Networks (ANNs) and Linear Regression (LR). The results were analyzed and compared. [ABSTRACT FROM AUTHOR]