Predictive machine learning model for the future trend of energy consumption in fully electricity homes considering occupancy status of the building

HVAC unit is one of the most power-intensive loads in buildings. It is also very significant for residential customers because indoor temperature must be maintained within an acceptable range of occupant’s comfort. To minimize energy consumption while providing a comfortable environment for the occupants, Building Energy Simulation (BES) gained considerable attention in recent years. Available BES calculates building energy consumption during the design phase and therefore, they optimize building energy consumption in this stage. However, there are still deficiencies that prevent BES from achieving higher efficiencies. Using a fixed occupancy schedule, not considering complexities in the occupant’s interactions with indoor appliances and HVAC unit are some of the setbacks that reduce the accuracy of the BES tools. Introduction of smart thermostats, made it possible for researchers to study the trend of changes in different measurable variables in the indoor environment like temperature, humidity, the runtime of the HVAC system, occupancy schedule, cooling and heating set point temperatures, etc. Data obtained from smart thermostat can be used to build a predictive model using a machine learning technique. Machine learning techniques help to estimate future trends of indoor variables like occupancy schedule, set point temperature and building energy consumption. Feeding the predicted variables by machine learning to the BES software helps to create a more accurate model for the energy simulation of buildings. This study presents a novel predictive model based on a co-simulation method using EnergyPlus and machine learning technique to better manage the energy consumption in the residential buildings. The proposed approach, which combines neural network and physic-based energy modeling, successfully estimates the total energy consumption in buildings with CV(CV(RMSE)) of 2.22% and NMBE of 5.65% on hourly basis.

Author

• Hosseini, Amin

Identifier

hdl:10057/25701

OAI identifier oai:identifier

oai:soar.wichita.edu:10057/25701