Design and Implementation of Data Analysis-Based Energy Monitoring System

Abstract

Rapid population and technological growth in today’s world have resulted in a dramatic increase in energy demand. Because electricity is used while it is being generated at a power plant, accurate energy consumption prediction has become critical. The study proposed an LSTM (Long ShortTerm Memory) based electrical energy monitoring technique in this paper. In comparison to other algorithms, the proposed LSTM model significantly reduced time series forecasting errors. The proposed system’s research is based on actual electricity consumption data from four individual households over a one-week period. A customized meter is used, which sends data to the server every minute, allowing the analysis to produce better results and be more efficient. The LSTM is well suited for analyzing recurring patterns in time series components, making it ideal for modeling irregular trends in electric power consumption. When compared to existing forecasting approaches, the proposed technique computes the lowest root mean square error (RMSE) for the dataset on individual household power usage. The variable findings show which factors have the greatest influence on power consumption forecasting.