Soft computing supervised smart energy-efficient harmonic current compensation for internet of things devices

The commercial & industrial internet of things are characterized by the quality of services (QoS) of energy transmissions such as accessibility, data transfer rate, delay and precise control. The smart class of power converters are the essential entities of the sensor-based technology in power management unit (PMU) in the internet of things based devices. Insertion of higher-order harmonics causes associative hazards and a rapid increase in machine-to-machine communication (MMC) generates a high impact on energy efficiency. This paper contributes to the analytical identification and self adaptive learning-based technique to rectify the harmonic challenges in smart devices to improve energy efficiency. The Adaptive Neuro Fuzzy interface system (ANFIS) is the main control technique to deal with the uncertainty of the load-variation and eventual effects of harmonics within the power management unit of IoT devices. The shunt active power filters are upgraded with smart computing techniques and hysteresis current control is optimized for minimization of the error to mitigate harmonic distortions. The neural network is trained through the ANFIS model using MatLab & Python programming and validated by the experimental setup. The proposed model offers a significant reduction in a harmonic-distortion from 72% on nonlinear load to 0.81% and elicits high feasibility for commercial and industrial applications as-per the IEEE519 standards.