Smart Tenaga: IoT Backboned and AI Integrated Energy Efficient and Carbon Footprint Alert System

Abstract

Smart Tenaga is an IoT (Internet of Things) enabled energy analytics framework for live monitoring, short term forecasting and high consumption detection in shared environments. This system consists of the integration of ESP32 microcontroller and PZEM-004 sensor to check electrical parameters and transfer the same to the cloud base for analysis. 14280-time stamped observations were recorded at 20 second intervals, typical of real-world usage conditions. Noise, missing data, and irregular sampling intervals were dealt with using a strong preprocessing pipeline and feature engineering to include temporal and device level patterns. Both regression and classification models were used for machine learning. Linear Regression performed the best in forecasting with an MAE of 0.001335, RMSE of 0.8217, and an $\mathbf{R}^{\mathbf{2}}$ of 0.000692, while the Random Forest was the best in accuracy among the classifiers with 0.9522 with a precision of 0.8712, recall of 0.8452, and F1-score of 0.8580 in high consumption detection. The energy consumption was between 0.007 and 0.574 kWh with the total energy consumed being 4.257 kWh which is equivalent to $3.2949 \mathbf{k g C O}_{\mathbf{2}}$ emissions. The findings highlight the potential of combining IoT and machine learning to provide accurate monitoring and predictive analytics of energy use, thereby contributing to energy efficiency and sustainability.