Energy costs are the second highest operational expense for K-12 schools in the United States. Improving energy efficiency and moving towards sustainable school buildings not only result in substantial cost savings and reduction of environmental emissions, but also provides an opportunity to enhance students’ awareness regarding energy, environment and sustainability. Effective tools and techniques that provide thorough understanding of energy consumption in school buildings are valuable to school districts by helping them with prioritizing energy efficiency projects. In the present paper, a multi-layer perceptron (MLP) neural network model is developed for estimating monthly energy consumption of K-12 schools in Brevard County, Florida. The inputs to the network are considered as number of occupants, days of operation per months, building’s area, average monthly outdoor dry-bulb temperature and relative humidity, month’s number as well as the year built of the building and the output from the network is monthly energy consumption. Various network topologies are considered and tested to achieve the optimal configuration for the network. The selected network is successfully trained using two years of energy consumption data for 30 schools in Brevard County, FL (high schools, middle schools and elementary schools). The results showed that the developed neural network model is capable of estimating monthly energy consumption of schools with a good accuracy. The network tested and validated using the data from schools which were not included in the training dataset and the errors between the known values and estimated values for monthly energy consumptions are evaluated and discussed using statistical concepts such as percent error and confidence intervals. The network is then saved and incorporated in an energy assessment tool developed in MATLAB. The tool uses the same inputs that were used for the neural network, but the outputs not only includes the output from the network, but also reports the energy use intensity (EUI) of the school building and how it compares with the average EUI of buildings from the same category within the same climate zone. Although the current study covers one particular school district (Brevard county) in a given climate zone (2a-hot and humid), the developed approach can be extended to incorporate various climate zones and serve as an effective tool for school energy conservation managers. The end user may obtain a clear idea of the energy consumption of the school building and how it compares against other buildings within the same category and climate zone, with minimum input data required.


A Machine Learning Based Approach for Energy Consumption Forecasting in K-12 Schools

Division / Sponsors: 
Aerospace, Bioengineering, Applied Mechanics, Design Engineering Division, Advanced Energy Systems, Manufacturing Engineering Division, Noise Control and Acoustics, Nondestructive Evaluation, Diagnosis, and Prognosis (NDPD), Nuclear Engineering Division, Process Industries Division, Fluids Engineering, Heat Transfer, Materials, Microelectromechanical Systems, Safety Engineering and Risk Analysis, Computers & Information in Engineering


Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2020/volume/84478

IMECE2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.