Electronic Thesis/Dissertation
 

A Decision Support Tool for Designing Energy-efficient Residential Buildings at the Early Planning and Design Stage

Open Access

Predicting building energy consumption is essential at the design stage as it assists in estimating the costs of building operation. Today, most of the current building energy predictions are conducted on a building energy simulation software. Such simulation software is commonly preprogrammed to perform detailed engineering calculations. However, due to the unavailability of detailed building factors at the early design stage, the building energy simulation software tends to yield a large discrepancy between predicted results and actual consumption.Machine learning is a promising technique in the domains of classification and prediction. Algorithms like Artificial Neural Network (ANN), Random Forest (RF), and Support Vector Machine (SVM), have been extensively adopted as predictive tools in various areas such as logistics, retail, financial services, and telecommunication. This research presents a new approach based on machine learning algorithms to reduce buildings’ energy costs during operation. This proposed energy predictive model combines four algorithms, including multivariate regression (MVR), sequential minimal optimization (SMO), ANN, and RF. The model built with hybrid algorithm provides an accurate and rapid forecast on energy consumption. With this model, designers can quantify the energy consumption of different design alternatives at the early design stage. Building energy costs can be reduced by selecting the design with the minimum energy consumption, among other options. The predictive model is developed and tested based on a large set of data collected by the Office of Energy Efficiency, Natural Resources of Canada, the Government of Canada. The performance of the proposed model is compared with the models developed in the previous literature, and the application of the model is demonstrated by using the building design examples.

Author Language Keyword Date created Type of Work License
  • All rights reserved
Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL

Notice to Authors

If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.

Thumbnail Title Date Uploaded Visibility Actions
Preview of Sun_gwu_0075A_15356.pdf Sun_gwu_0075A_15356.pdf 2020-12-17 Open Access