Deep Learning Model for Forecasting Institutional Building Energy Consumption

South Africa is currently facing an on-going energy crisis that seems to persist every year. Load shedding has become one of the country’s biggest challenge. This is because of energy consumption being at an all-time high and inconsistent in terms of supply. In this paper, we propose a deep learning framework (called Dense Neural Network) for the prediction of energy consumption for University buildings. The deep learning model is evaluated on an energy dataset (collected from the University of KwaZulu-Natal), to forecast the energy consumption of the buildings in the University of KwaZulu-Natal. Furthermore, we compared the performance of the proposed model with two classical algorithms (Support Vector Machine and Multiple Regression), and the deep learning model outperformed the classical algorithms. The forecasted energy consumption can be used by various University managements to assess where most of the energy is being consumed. It can provide an opportunity to devise strategies for optimal utilization of energy in Universities.

[1]  Betul Bektas Ekici,et al.  Prediction of building energy needs in early stage of design by using ANFIS , 2011, Expert Syst. Appl..

[2]  Jui-Sheng Chou,et al.  Forecasting energy consumption time series using machine learning techniques based on usage patterns of residential householders , 2018, Energy.

[3]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[4]  Won-Hwa Hong,et al.  Energy Consumption and the Power Saving Potential of a University in Korea: Using a Field Survey , 2011 .

[5]  B. P. Numbi,et al.  Energy cost saving potential in educational buildings-case study of MUT campus , 2017, 2017 International Conference on the Industrial and Commercial Use of Energy (ICUE).

[6]  Wes McKinney,et al.  Python for Data Analysis , 2012 .

[7]  Do-Hyeun Kim,et al.  A Prediction Methodology of Energy Consumption Based on Deep Extreme Learning Machine and Comparative Analysis in Residential Buildings , 2018, Electronics.

[8]  Yueren Wang,et al.  A novel ensemble learning approach to support building energy use prediction , 2018 .

[9]  B. Dong,et al.  Applying support vector machines to predict building energy consumption in tropical region , 2005 .

[10]  Tiberiu Catalina,et al.  Multiple regression model for fast prediction of the heating energy demand , 2013 .

[11]  Philippe Poizot,et al.  Clean energy new deal for a sustainable world: from non-CO2 generating energy sources to greener electrochemical storage devices , 2011 .

[12]  Frédéric Magoulès,et al.  A review on the prediction of building energy consumption , 2012 .

[13]  Zhengwei Li,et al.  Using Support Vector Machine to Predict Next Day Electricity Load of Public Buildings with Sub-metering Devices☆ , 2015 .

[14]  Yg Sandanayake,et al.  Building energy consumption factors : a literature review and future research agenda , 2012 .

[15]  Rizwan Ahmad,et al.  Intelligent techniques for forecasting electricity consumption of buildings , 2018, Energy.

[16]  R. Faranda,et al.  Load Shedding: A New Proposal , 2007, IEEE Transactions on Power Systems.

[17]  Daniel L. Marino,et al.  Building energy load forecasting using Deep Neural Networks , 2016, IECON 2016 - 42nd Annual Conference of the IEEE Industrial Electronics Society.

[18]  Wei-Bin Wu,et al.  Energy Consumption Level Prediction Based on Classification Approach with Machine Learning Technique , 2018, Proceedings of the 4th World Congress on New Technologies.

[19]  Nikhil Ketkar,et al.  Introduction to Keras , 2017 .

[20]  Cecilia Punt.  The potential impact of productivity losses as a result of the electricity shortages in South Africa , 2008 .