A deep Q-network (DQN) was applied for model-free optimal control balancing between different HVAC systems. The DQN was coupled to a reference office building: an EnergyPlus simulation model provided by the U.S. Department of Energy. The building was air-conditioned with four air-handling units (AHUs), two electric chillers, a cooling tower, and two pumps. EnergyPlus simulation results for eleven days (July 1–11) and three subsequent days (July 12–14) were used to improve the DQN policy and test the optimal control. The optimization goal was to minimize the building’s energy use while maintaining the indoor CO₂ concentration below 1,000 ppm. It was revealed that the DQN—a reinforcement learning method—can improve its control policy based on prior actions, states, and rewards. The DQN lowered the total energy usage by 15.7% in comparison with the baseline operation while maintaining the indoor CO₂ concentration below 1,000 ppm. Compared to model predictive control, the DQN does not require a simulation model, or a predetermined prediction horizon, thus delivering model-free optimal control. Furthermore, it was demonstrated that the DQN can find balanced control actions between different energy consumers in the building, such as chillers, pumps, and AHUs.

This article compares two modeling approaches for optimal operation of a turbo chiller installed in an office building: (1) a machine learning model developed with artificial neural network (ANN) and (2) a hybrid machine learning model developed with the ANN model and available physical knowledge of the chiller. Before developing the ANN model of the chiller, the authors used Gaussian mixture model in order to check the validity of measured data. Then, the hybrid model was developed by combining the ANN model and physics-based regression equations from the EnergyPlus engineering reference. It was found that both the ANN and hybrid ANN model are satisfactory to predict the chiller’s power consumption: mean bias error (MBE) = −2.63%, coefficient of variation of the root mean square error (CVRMSE) = 8.05% by the ANN model; MBE = −3.99%, CVRMSE = 11.98% by the hybrid ANN model. However, the hybrid model requires fewer inputs (four inputs) than the ANN model (eight inputs). The energy savings of both models are similar coefficient of performance (COP) = 4.32 by the optimal operation of the ANN model; COP = 4.44 by the optimal operation of the hybrid ANN model. In addition, the hybrid ANN model can be applied where the ANN model is unable to provide accurate predictions.

Blind systems have been introduced to provide visual and thermal comfort, as well as to reduce energy use in buildings. A wide variety of such systems exist in terms of thermal and optical properties, location and physical configuration. The current problem with blinds is that their operation is not based on the dynamics of the room but on the static or manual control operated by occupants, although many studies have recognized that dynamic control can far outperform static control. One reason for the lack of dynamic control is that it is not easy to combine the room dynamics with any possible optimization algorithm. Hence, in this study, a whole building simulation tool, EnergyPlus, was integrated with the MATLAB optimization toolbox to solve for optimal control of blind systems. This paper addresses the difference between static versus optimal control of interior and exterior blind systems in office buildings.

While facility design is increasingly playing a role addressing strategic organizational objectives, issues pertaining to facility maintenance have typically been left out of the decision-making process. Reasons include the traditional disconnection between facility design and facility maintenance, mainly originating from the lack of a modality to meaningfully represent facility maintenance information during design decision making. This paper discusses two sets of facility maintenance indicators that have the potential to bridge this traditional divide. The objective of this paper is threefold: (1) to illustrate how maintenance performance indicators can support higher-level decision making; (2) to explain how the methodology could be mapped across building sectors; and (3) to show how facility maintenance could play a crucial role in informing strategic decision making. Two types of indicators are introduced based on: (1) normative models in biophysics and physiology and (2) empiricist models of environ...

This paper presents a simulation study to reduce heating and cooling energy demand of a school building in Seoul Metropolitan Area, Korea. The aim of this study was to estimate the impact of passive vs. active approaches on energy savings in buildings using EnergyPlus simulation. By controlling lighting, the energy saving of the original school building design was found most significant, and increased by 32% when the design was improved. It is noteworthy that energy saving potential of each room varies significantly depending on the rooms’ thermal characteristics and orientation. Thus, the analysis of energy saving should be introduced at the individual space level, not at the whole building level. Additionally, the simulation studies should be involved for rational decision-making. Finally, it was concluded that priority should be given to passive building design strategies, such as building orientation, as well as control and utilization of solar radiation. These passive energy saving strategies are related to urban, architectural design, and engineering issues, and are more beneficial in terms of energy savings than active strategies.