Digital Twinning and ANN-based Forecasting Model for Building Energy Consumption

Digital twins are digital representations of realworld products, processes, or infrastructure that help us learn about and foresee how such things will function in the future. The development of digital twins for the energy sector improves energy efficiency, optimizes asset management, and lowers the environmental impact of energy use. The main objective of this study is to develop a Digital Twin of building energy consumption at the University of Sharjah. The technical approach used in this study includes: (1) collection of the energy consumption data from the Renewable Energy Research Laboratory at the University of Sharjah using smart sensors, (2) developing and validating a simulation model of the building energy consumption using Open Studio software and (3) development of artificial neural network (ANN)-based forecasting model to predict the building energy consumption 15 minutes ahead. The assessment of the accuracy of the developed Digital Twin was done using the correlation coefficients R. The results show that the building energy consumption simulation model R-values for three different timesteps (15 minutes, hourly, and daily) are 0.98106, 0.98651, and 0.99647, respectively. The R-value for the 15 minutes ahead forecasting model is 0.98667. These values indicate that the predicted energy consumption from the Digital Twin compares well with the experimental data from the smart sensors. To simplify the access to the developed Digital Twin, a dashboard application was created using MATLAB App Designer. The dashboard continuously and automatically monitors and forecasts energy consumption by showing the live, historical, and forecasted values in a variety of visual and graphical elements. This will help owners/operators and controllers to take actions regarding the energy consumption of the building by reducing the energy consumption and preventing sudden high peaks to occur.