Hybrid Deep Learning and Internet of Things based Building Energy Management System: A Case Study on Energy Conservation in a University
Building energy consumption monitoring has huge untapped potential, especially in a developing economy like India. Current solutions have not focused on a cost-effective, small-scale, accurate and open-source data-driven methodology. Further, several obstacles hinder research in this domain, such as security and privacy concerns, high investment and uncertainty of research outcomes. Therefore, we have implemented an automated hybrid deep learning and Internet of Things (DL-IoT) based Building Energy Management System (BEMS) for energy conservation. The DL-IoT integrates a hybrid of DL and fuzzy logic to handle the uncertainty and noise in electrical features. Our DL- IoT regression model achieves low mean absolute error and mean squared error as well as a coefficient of determination of 0.99 out of sample. Twenty-seven input electricity consumption features have been selected for extraction from the raw data for training the model. Experimental results indicated a linear relationship between the features and the energy consumption. The proposed model successfully predicted certain features that could lead to energy savings, viz., Power Factor and Power in Y Phase. The model estimated that a one-unit increase in Power in Y Phase and Power Factor will lead to a reduction in the energy consumption. The findings of the experiment indicated that the model captured the variability of the data better than other models. The results demonstrated the superiority of the proposed model over other mainstream existing models. Through the results of this paper, a more efficient energy data management and consumption plan can be established.