Rancang Bangun Sistem Monitoring dan Prediksi Konsumsi Listrik Menggunakan Metode Simple Exponential Smoothing serta Kontrol Lampu

Abstract

Penelitian ini bertujuan merancang dan mengimplementasikan sistem monitoring, prediksi konsumsi energi listrik, dan kontrol lampu berbasis Internet of Things (IoT). Sistem menggunakan ESP32, sensor PZEM-004T dengan current transformer (CT) PZCT-02, Solid State Relay (SSR), protokol MQTT, serta web dashboard berbasis Laravel dan MySQL. Sistem mampu melakukan monitoring parameter kelistrikan secara real-time, kontrol lampu melalui web dashboard, dan prediksi konsumsi energi listrik harian menggunakan metode Simple Exponential Smoothing (SES) dengan smoothing factor (a) 0,3. Hasil pengujian menunjukkan seluruh fungsi sistem berjalan dengan baik. Sensor PZEM-004T memiliki rata-rata error 1,94% pada pengukuran tegangan dan 3,17% pada pengukuran arus. Evaluasi metode SES menggunakan 30 data historis harian menghasilkan MAD sebesar 1,0400 kWh, MSE sebesar 2,1894 (kWh)², dan MAPE sebesar 6,4327%, sehingga metode SES tergolong sangat akurat untuk memprediksi konsumsi energi listrik harian.

This study aims to design and implement an Internet of Things (IoT)-based system for electrical energy monitoring, forecasting, and lamp control. The system employs an ESP32 microcontroller, a PZEM-004T sensor with a PZCT-02 current transformer (CT), a Solid State Relay (SSR), the MQTT protocol, and a Laravel- and MySQL-based web dashboard. The system provides real-time electrical monitoring, web-based lamp control, and daily energy consumption forecasting using the Simple Exponential Smoothing (SES) method with a smoothing factor (a) of 0.3. The results show that all system functions operated successfully. The PZEM-004T sensor achieved average errors of 1.94% for voltage and 3.17% for current measurements. Evaluation of SES using 30 daily historical data points produced MAD, MSE, and MAPE values of 1.0400 kWh, 2.1894 (kWh)², and 6.4327%, respectively, indicating that SES is highly accurate for forecasting daily electrical energy consumption.