Electrical Power Demand Forecasting of Smart Buildings: A Deep Learning Approach

In recent times, due to urbanization, the electrical power demand has increased rapidly, and that increases the greenhouse gases. So, there is a demand for optimum utilization of electrical power and that enforces to build a smart energy-efficient building. And to utilize the electrical power more optimally and efficiently, a more accurate and intelligent demand forecasting technique is also essential. In this work, a smart building energy management system (SBEMS) architecture is proposed. In this context, we also propose a deep learning model to forecast the electrical power demand based on weather conditions. The two most popular deep learning architectures such as 1D convolutional neural network (1D-CNN) and bidirectional long short-term memory (Bi-LSTM) recurrent neural network are merged to build the proposed forecasting model. The experimental results show that the proposed model produced more accurate results in comparison with other established forecasting models.