Energy monitoring system for textile industry

Abstract

This project introduced energy monitoring system for textile industry in Bangladesh. It combines IoT hardware with NILM methodology to carry out the real-time monitoring of individual machine level energy consumption. Two methods i.e. (Intrusive Load Monitoring) ILM and (Non Intrusive Load Monitoring) NILM were compared among which NILM was chosen for its cost effectiveness, scalability and ease of adaptability. The network monitors the total voltage and current with sensors and sends the data to the cloud server with ESP8266. A deep learning model MATNilm was developed to disaggregate the total energy into machine wise power consumption with acceptable accuracy (the average F1-score: 0.73). Real time display via Google Sheets with Blynk app is enabled. The product is conducive for remote monitoring, energy audit and adheres to national and international efficiency goals. Upcoming work will help refine model accuracy, on edge deployment and predictive maintenance. The system provides an industrial application for energy optimization.