DATAEMS: Design and development of a data analysis-based energy monitoring system

Electrical load forecasting is now a key component of planning and running a smart grid due to the ongoing integration of green energy sources (RES) into the energy mix and the transformation of the traditional electric grid into a more intelligent, flexible, and interactive system. Either when it comes to the distribution system or a single household, predicting the electric load is a difficult task due to its high volatility and uncertainty. In this paper, the study suggests a method for monitoring electrical energy that is based on long short-term memory (LSTM). The research for the suggested method is based on real data of electricity use from four distinct clients during a week. The analysis is made more effective and produces better findings owing to the use of a customized meter that transmits data to the server every minute. Modeling irregular trends in electric power consumption is made possible by the LSTM's aptitude for examining recurrent patterns in time series components. The suggested method calculates the dataset on individual household power usage with the lowest root mean square error (RMSE) as compared to existing forecasting methods. The simulation results highlight the parameters that have the biggest impact on forecasting power consumption.The suggested LSTM model substantially reduced time series forecasting errors when compared to previous techniques.