Conference Agenda

The supply air temperature (SAT) of an air handling unit in multi-zone variable air volume systems could impact the energy use significantly. Formerly, buildings used a constant SAT which resulted in high energy consumption due to the increased load on perimeter heaters. Lately, ASHRAE guideline 36 introduced the trim and respond logic (Taylor, 2015) as an improved SAT reset strategy which depends on the feedback of the cooling requests of the zones. However, the trim and respond logic might fail to provide comfort and/or energy savings in case of higher demands, conflicting thermal preferences at different zones and varying occupancy patterns. This study investigates four SAT reset strategies: 1) constant 13 (55℉) SAT, 2) SAT reset based on outdoor air temperature (OAT), 3) trim and respond, and 4) trim and respond combined with OAT reset; with different cases of varying zone setpoints. It also introduces a deep Q-network (DQN) reinforcement learning (RL) algorithm for SAT reset and compares its performance with the other strategies. All the cases are simulated using EnergyPlus. The objective is to address the shortcomings of the currently adopted methods in industry and to show the potential of reinforcement learning in HVAC controls. The results show that the common SAT reset strategies do not perform well with cases of varying setpoint leading to either higher energy cost or decrease in occupant comfort, while the DQN-based method provided a better alternative. These findings establish a basis for future work that would focus on developing a multi-agent occupant centric control (OCC) method that takes energy and occupant comfort into account by utilizing RL methods.

Maintaining an optimal operating environment within manufacturing facilities is crucial for enhancing energy efficiency, boosting manufacturing productivity, and ensuring occupant comfort and health. With the increasing adoption of Internet of Things (IoT) sensors and cloud-based data acquisition systems in manufacturing facilities, a wealth of IoT operational data is available. This abundance of data empowers data-driven energy analytics and facilitates intelligent control for building operations. Deep Reinforcement Learning (DRL) control is an emerging intelligent control that leverages building big data and artificial intelligence algorithms to optimize operational efficiency and environmental conditions. In this work, operational data is collected and streamed from a manufacturing building equipped with IoT sensors. A customized environment for the DRL is constructed using the operational data. Within the environment, building characteristics and heat transfer process are modeled by a Resistor-Capacitor network. The DRL model is subsequently trained with the Proximal Policy Optimization algorithm to find an optimal control policy. Results show that the DRL building control framework effectively maintains desired indoor conditions in conditioned zones with reduced fluctuation. Moreover, there is a notable decrease in energy consumption, with a demonstrated 33.8% reduction in energy cost savings over a two-month testing period. The implementation of the DRL method also leads to an estimated annual reduction of 4.80 kg/m2 in carbon emissions for the entire building, therefore contributing to environmental impact mitigation.

Many simulation studies have suggested that advanced control strategies for heating, ventilation, and air-conditioning (HVAC) equipment in commercial buildings can reduce energy costs and greenhouse gas emissions. However, despite these potential benefits, adoption of advanced HVAC control remains limited, due in part to a lack of confidence in the technology among decision-makers in business and government. Field demonstrations of advanced HVAC control can build confidence in the technology by demonstrating its effectiveness and economic value in the real world. This paper accordingly reviews field demonstrations of advanced commercial HVAC control strategies, such as Model Predictive Control (MPC) and Reinforcement Learning Control (RLC). This paper discusses building types, control methods, test durations, measurement and verification procedures, control objectives, and reported benefits. It further provides a critical assessment of the state of the technology and highlights research opportunities that could accelerate real-world adoption of advanced commercial HVAC control strategies. The literature review confirms that advanced HVAC controls can significantly enhance energy efficiency and occupant comfort. However, most field studies cover relatively short durations and control small spaces within larger buildings. Longer-duration studies frequently report lower savings, suggesting that short-duration studies may overestimate potential benefits. Similarly, whole-building control studies typically report lower savings than smaller-scale studies, likely because the latter tend to overlook thermal coupling between controlled zones and adjacent zones. Finally, data and discussions concerning deployment costs and challenges are almost nonexistent. This suggests an important area for future research, as achieving adoption at scale will require demonstrating not only reliable benefits but also manageable deployment costs.

Efficient electrification of space heating/cooling presents the most viable pathway to GHG emissions reduction, and heat pumps (HPs) remain the dominant alternative for replacing gas/oil-based space heating systems. To achieve widespread adoption of HPs, it is imperative to improve their energy efficiency and operational cost. In this paper, a scalable and computationally inexpensive advanced predictive rule-based-control (PRBC) strategy for HPs is presented. The controller is tested on an EnergyPlus prototype model of a single-family detached house within the building optimization testing framework (BOPTEST). The HVAC system consists of a single-speed HP, inclusive of a single-speed DX heating coil, a single-speed DX cooling coil, and a constant-speed fan. The PRBC model uses the current indoor air temperature inside the building, day-ahead ambient air temperature, and hourly electricity price (HEP) forecasts to preheat/precool a building, with the final goal of HVAC cost/energy reduction without a noticeable increase of indoor thermal discomfort. The ambient air temperature and HEP forecasts are integrated into the PRBC model by: (i) assigning proportional weight to the forecasted values, prioritizing closer time steps to the present, due to the intuitive principle that forecasting accuracy diminishes with greater temporal distance from the present, (ii) modulating the amount of precooling/preheating based on weighted ambient air temperature and HEP forecasts to not only shift HVAC energy usage from high to low HEP periods but also avoid excess precooling/preheating. Results show the advanced PRBC of being able to identify and quickly respond to finer trends in HEP and ambient temperature than the other controllers resulting in cost/energy savings. The thermal discomfort of the advanced PRBC is comparable to the other controllers, proving the efficacy of the proposed PRBC in judiciously preheating/precooling the building. The advanced PRBC performs significantly better in the cooling season than the heating season, achieving as high as 14%, 9%, and 8% in monthly cost savings, and 11%, 6%, and 8% in monthly HVAC energy savings, as compared to the industry standard, relaxed baseline and literature inspired controllers respectively.

Proactive building demand response has been proved to be a cost-effective approach for power system ancillary services. Air-conditioning systems are one of the significant and preferable demand response resources in large commercial buildings. Demand limiting control by shutting down part of operating chillers is fast enough to achieve load curtailment for providing valuable spinning reserve. However, it runs with the sacrifices of indoor thermal comfort due to the insufficient cooling supply during the demand limiting period. In this situation, the proper management of the limited cooling supply remains to be a challenge considering the trade-off of different factors affecting the indoor thermal comfort. Conventional building automation system fails to achieve rational use of determined power supply considering this trade-off when performing fast demand response after shutting down operating chillers. This paper therefore develops a model-based control optimization of supply air flow rates in the air-conditioning system to address this problem. A multi-objective problem is formulated to quantify this problem. Incremental building thermal balance models are employed to predict the indoor environment responses for optimization. Test results reveal that the proposed control strategy can determine the rational delivery of the cooling supply to the zones considering appropriate importance of objective for indoor air temperature and relative humidity.

The energy demand of buildings within the hospitality sector, notably hotels, is high and diversely distributed. Furthermore, it is rising due to growing needs of convenience per guest. Determination and analysis of electrical and thermal energy is necessary to achieve reduced energy loads. This is hindered by missing concrete data on loads. Therefore, 24 buildings linked to the hospitality sector are equipped with devices for gathering electrical and thermal energy load data in detail from January 2019 to January 2021, a period markedly affected by the COVID-19 pandemic. The data is used for identifying possible energy enhancements and developing a model predictive control (MPC) algorithm for adapting efficiency improvements. A new methodology is introduced by a classification system for categorizing the hotel buildings in terms of utilization and energy efficiency. The buildings are further differentiated into the four application categories conference, journey, event and wellness. This study also explores the integration of renewable energy sources, including photovoltaic (PV) and block-type thermal power stations (BTTP), combined by district heating (DH), besides conventional oil/gas boilers. A specific concept for collecting data in every category is developed and instrumentation is installed to measure energy demand spatially and temporally resolved for a period of two years. The unique context of the pandemic provides insights into the resilience and adaptability of energy demand patterns in the hospitality sector. By analyzing the collected data, key energy improvement opportunities are identified in relation to the acquired data for base loads, typical peaks and mid-term impacts like holidays.