Energy optimization of a residential building using occupancy prediction via sensor fusion and machine learning algorithms

• Author / Creator

  Heidarifar, Hamed

• Occupancy-based control systems for floor heating can lead to energy saving in a residential building. This study focuses on the energy consumption by space heating system in a half-duplex residential house located in Edmonton, Alberta. In the first part, a sensor fusion model was designed to predict the occupancy status in the residential building. To predict the occupancy, data including room temperature, relative humidity, CO2 concentration, and day of the week were collected. The actual occupancy status of the room was collected using a Passive Infrared (PIR) motion sensor and the data sheets filled by the occupants. The study considered four different machine learning algorithms including K-nearest neighbors (KNN), Gaussian Support Vector Machine (SVM), Artificial Neural Network (ANN), and Decision Tree (DT)
  to predict the state of occupancy for a room. The results showed KNN method outperforms the other methods by reaching the Geometric Mean (GM) accuracy of
  92% for occupancy prediction.
  In the second part, a 3D energy model was developed for the entire house. The energy consumption was simulated with and without considering occupancy information
  using EnergyPlus software. The actual energy consumption of the house was obtained from the gas meter. Using the developed model and methods of machine learning, a virtual sensor was proposed to define the thermostat temperature according to the desired temperature in the occupied rooms to reduce energy consumption and improve the comfort level for the occupants. The results showed 9.5% to 30.7% energy saving, depending on the occupancy-based control methods. In addition, the results of the operating analysis showed a possible reduction of the yearly floor heating cost of the testbed up to 329 CAD which is about 30% reduction in the yearly heating cost of the building. Overall, this study quantized the importance of considering
  occupancy information in reducing energy consumption in residential buildings, while maintaining or improving the occupants’ comfort.

• Subjects / Keywords

  • Energy optimization
  • Sensor fusion
  • Simulation
  • Occupancy detection
  • Machine learning

• Graduation date

  Fall 2023

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-xjv5-b413

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.