The COVID-19 impact on air condition usage: a shift towards residential energy saving

Abstract

The enforcement of the Movement Control Order to curtail the spread of COVID-19 has affected home energy consumption, especially HVAC systems. Occupancy detection and estimation have been recognized as key contributors to improving building energy efficiency. Several solutions have been proposed for the past decade to improve the precision performance of occupancy detection and estimation in the building. Environmental sensing is one of the practical solutions to detect and estimate occupants in the building during uncertain behavior. However, the literature reveals that the performance of environmental sensing is relatively poor due to the poor quality of the training dataset used in the model. This study proposed a smart sensing framework that combined camera-based and environmental sensing approaches using supervised learning to gather standard and robust datasets related to indoor occupancy that can be used for cross-validation of different machine learning algorithms in formal research. The proposed solution is tested in the living room with a prototype system integrated with various sensors using a random forest regressor, although other techniques could be easily integrated within the proposed framework. The primary implication of this study is to predict the room occupation through the use of sensors providing inputs into a model to lower energy consumption. The results indicate that the proposed solution can obtain data, process, and predict occupant presence and number with 99.3% accuracy. Additionally, to demonstrate the impact of occupant number in energy saving, one room with two zones is modeled each zone with air condition with different thermostat controller. The first zone uses IoFClime and the second zone uses modified IoFClime using a design-builder. The simulation is conducted using EnergyPlus software with the random simulation of 10 occupants and local climate data under three scenarios. The Fanger model’s thermal comfort analysis shows that up to 50% and 25% energy can be saved under the first and third scenarios.

Validating the Simulation Model for the Investigation of Applying Saudi Building Code 601 to Najran Governmental Office Building in Saudi Arabia

“…DesignBuildder is among the best and most crucial tools for thermal comfort and energy performance analysis as discussed in [22,23]. These research demonstrate the flexibility and simplicity of DesignBuilder to examine various building envelope parameters including thermal comfort, energy, solar shading and support simulation from renewable technologies with the aid of graphical interface options to make adjustments as necessary.…”

“…It is important to note that the result of our finding can be extended in the development of an HVAC system controller to ensure air conditioners are utilized only when the room is occupied. The results also can be used to improve the thermal comfort performance and energy efficiency of the simulation study proposed in [22,23] under different scenarios of occupant perception of indoor temperature.…”

“…This methodology is widely applied in similar studies and proves solid performance. For example, a study [22,23] employs a quantitative approach in residential buildings achieving high-performance accuracy. Our study specifically used a similar approach to investigate the implementation of the Saudi building code in government office buildings to provide insight into addressing the issues of thermal performance and energy consumption in government office buildings in Saudi Arabia, both in the present and future climate.…”

“…The quality of the dataset is one of the major factors that contribute to the performance of the model. Previously in [23], an interactive technique was applied to provide a classifier to ensure room temperature reading tally with occupancy numbers during data collection to deal with incorrect values used in model training. In the same vein, this study used a similar approach (see Figure 10) to record data during data collection enabling self-estimation or labeling.…”

Systematic Review Analysis on Smart Building: Challenges and Opportunities

“…The predictive approach uses automatic systems to control appliances energy consumption based on real-time data collected by sensors known as dynamic approach or appliance energy consumption profile data such as occupancy schedules, fixed setpoint, etc. On the other hand, the non-predictive approach is mostly remote controlled via mobile applications and web services through a central control system such as a smart meter [8,[31][32][33]. In this context, this study classified selected literature into predictive and non-predictive approaches, as presented in Figure 6.…”