Miniature Circuit Breaker Based Non-intrusive Power Monitoring and Load Classification System with AIoT Technology

Hou Kit Mun,Tzen Ket Wong,Kai Lok Lum,Zhen Ye Tham,Lee Chern Han Sean,Wei Qiang Tan,Zhi Sam Lee
DOI: https://doi.org/10.1007/978-3-030-92604-5_29
2021-12-11
Abstract:Power monitoring is a way to help optimize energy usage in a home or factory. Existing intrusive power monitoring systems are not cost-effective and require a high number of sensors and wireless modules for monitoring different types of loads in a building. In this work, we have developed a miniature circuit breaker (MCB)-based power monitoring system and load classification system with Artificial Intelligence of Things (AIoT) technology to monitor and classify the loads non-intrusively and present the data to users remotely through a web application in a dedicated server. The competent integration of NodeMCU, PZEM-004T v3.0 module and web server allows the system to accurately and non-intrusively measure power consumption at MCB and perform the load classification using the artificial neural network. Performance of the developed system in load monitoring was characterized by current, power factor and power measurement of four different loads. The absolute errors, given by the difference between data measured (current, power factor and power) by clamp meter and the developed system were in the range of 0.02–0. 11 A, 0–0.06 and 1–3 W respectively. Besides, the average accuracy of the system in load classification for nine combinations of loads was 94.7%. The testing results endorse the accuracy and reliability of the power monitoring system, on top of its easy-accessibility gifted by IoT.
What problem does this paper attempt to address?