Machine Learning Assisted Power Factor Correction Using Capacitor Banks

Energy utilization and efficiency are currently recognized as the main goals in the electrical power industry. Power factor is frequently used to determine whether a system's power supply is good or bad. A drop in power factor may result from the need for electrical power system to power inductive loads in homes and businesses. The AC power system's power factor is a concern because AC power is widely used and there are many different types of loads, including inductive loads. A lower power factor in electric machines translates into reduced power consumption and efficiency. For efficiency to be higher, the power factor value must be near unity. A power factor correction system can increase and maintain the power factor value close to unity by utilizing a variety of compensation strategies. The system's power factor measurement, compensation, and output display were all managed using an Raspberry Pi. The input to the system was an AC power supply which was given to different types of loads. Capacitor banks were controlled with a relay module to solve the power factor issue with the help of machine learning algorithms. Because of this, the system's lagging inductive KVAR caused by lagging loads was eliminated, resulting in better power factor. The suggested solution will improve the electrical power system's efficiency and lower the losses. The study's conclusions demonstrate the considerable gains in power factor, energy loss reduction, and overall system performance optimization made possible by the suggested approach, which calls for the development of several capacitor bank stages in order to achieve the intended KVAR and guarantee the best possible use of capacitor banks.