<![CDATA[The electricity monitoring system that implemented by PLN currently still uses conventional technology, so that requires officers to record energy usage manually and periodically from house to house. Seeing these conditions, authors wants to take the great potential of this problem. Smart energy meters are designed to automatically record electricity consumption and estimate costs to be paid, so that electricity usage can be controlled in real time. The components used are NodeMCU ESP8266 as a data processor, PZEM-004T as a current and voltage measuring sensor while calculating the energy used in a certain time, RTC DS1307 as a time module, LCD2004 as a display, SD card module as a data storage backup, push buttons as accessibility switching modes, and power supply as a power supply. In delivering data, the website is used as a user-friendly interface media. There also additional features that predict the power and cost that will use by user in the future. The forecasting uses the Encoder-Decoder Long Short-Term Memory (LSTM) Neural Network method. This smart energy meter prototype has high accuracy. Measurements show results similar to energy calculations on PLN meters, with an error percentage of 0.032%. In addition, the forecasting feature also has accurate results. Shown by the RMSE value of 0.49 and MAPE of 4%.]]>
Copyrights © 2022
