Occupant Behavior Modeling Research Articles

Research on air-conditioning usage behaviour in offices with different occupancy

Occupant behaviours in the buildings are not only random and uncertain but also related to each occupant’s habitual preference. This leads to the performance gap between actual and expected energy consumption in buildings. Therefore, accurate information and modelling with regard to occupant behaviour are important for reliable energy simulation and energy-saving optimization design. Existing studies on occupant behaviour models in office space usually focus on single-person offices or full-floor buildings, without considering the behavioural differences among offices with different occupancy. Therefore, this study established the air-conditioning usage behaviour models in offices with different occupancy based on questionnaires and measured data. The results show that occupant compromise and clustering effect will increase with the increase of occupancy. Using the established models as input, this study compared the simulation results with that under the standard schedule. The difference rate is as high as 32.19% in winter and 13.07% in the whole year. And for areas with high energy consumption in winter, the gap may be bigger.

Impact of Window-to-Wall Ratio on Heating Demand and Thermal Comfort When Considering a Variety of Occupant Behavior Profiles

Energy consumption and thermal comfort in residential buildings are highly influenced by occupant behavior, which exhibits a high level of day-to-day and dwelling-to-dwelling variance. Although occupant behavior stochastic models have been developed in the past, the analysis or selection of a building design parameter is typically based on simulations that use a single “average” occupant behavior schedule which does not account for all possible profiles. The objective of this study is to enhance the understanding of how window-to-wall ratio (WWR) of a residential unit affects heating demand and thermal comfort when considering occupant behavior diversity through a parametric analysis. To do so, a stochastic occupant behavior model generates a high number of possible profiles, which are then used as input in an energy simulation of the dwelling. As a result, one obtains probability distributions of energy consumption and comfort for different WWR values. The paper shows that the shape of the probability distributions is affected by WWR and dwelling orientation, and that the influence of different occupant behavior aspects on performance also varies with WWR. This work could help designers to better assess the impact of WWR for a large spectrum of possible occupant behavior profiles.

Utilizing commercial heating, ventilating, and air conditioning systems to provide grid services: A review

The modern power grid faces multiple challenges due to an increase in the adoption of renewable generation, such as dynamically balancing supply and demand at different time scales. Demand side management in buildings plays a vital role in achieving this balance because buildings can provide grid services through a variety of building assets. However, the development of grid-interactive, efficient buildings is still in its infancy, and a systematic and holistic understanding of grid service delivery strategies in terms of energy efficiency, load shifting, load shedding and load modulating is still limited. This paper is a comprehensive review of the development and application of building-level control strategies for utilizing heating, ventilating, and air conditioning systems to provide grid services. These strategies have been investigated through numerical and experimental studies. Control algorithms, such as heuristic rule-based control and model-based control, have been used to enable the automatic control delivery of grid services. The advantages and disadvantages of the strategies are summarized and discussed. Research trends are also identified, which include considering predicted mean vote-based and occupant-based thermal comfort, modeling of occupant behavior, integrating power grid operations with building control, and combining different demand flexibility modes in the control design.

Optimizing occupant-centric building controls given stochastic occupant behaviour

Occupant-centric control (OCC) strategies represent a novel approach for indoor climate control in which occupancy patterns and occupant preferences are embedded within control sequences. They aim to improve both occupant comfort and energy efficiency by learning and predicting occupant behaviour, then optimizing building operations accordingly. Previous studies estimate that OCC can increase energy savings by up to 60% while improving occupant comfort. However, their performance is subjected to several factors, including uncertainty due to occupant behaviour, OCC configurational settings, as well as building design parameters. To this end, testing OCCs and adjusting their configurational settings are critical to ensure optimal performance. Furthermore, identifying building design alternatives that can optimize such performance given different occupant preferences is an important step that cannot be investigated during field implementations of OCC due to logistical constraints. This paper presents a framework to optimize OCC performance in a simulation environment, which entails coupling synthetic occupant behaviour models with OCCs that learn their preferences. The genetic algorithm for optimization is then used to identify the configurational settings and design parameters that minimize energy consumption under three different occupant scenarios. To demonstrate the proposed framework, three OCCs were implemented in the building simulation program, EnergyPlus, and executed through a Python package, EPPY to optimize OCC configurational settings and design parameters. Results revealed significant improvement of OCC performance under the identified optimal configurational settings and design parameters for each of the investigated occupant scenarios. This approach would improve OCC performance in actual buildings and avoid discomfort issues that arise during the initial implementation phases.

On the impact of stochastic modeling of occupant behavior on the energy use of office buildings

The reliability of building performance simulation is hindered by several uncertainties, with aleatory uncertainty due to occupant behavior being one of the most critical. The present study aims to assess the propagation of uncertainty due to the adoption of stochastic models for modeling Occupant Presence and Actions (OPAs) available in the literature on the annual electric energy use of a reference office building. To this purpose, a global sensitivity analysis was designed and carried out by analyzing model inputs and energy outputs of 144 permutations of 15 different stochastic models for OPAs for a total of 7200 simulations. Building energy use computed considering stochastic OPAs modeling resulted in being sensibly higher than the reference value estimated assuming scheduled occupancy and rule-based occupant’s actions as suggested by reference standards. The median value of the electric energy use was 58.6% higher than the base case electric energy use. Furthermore, the stochastic models used to model window operation have the highest effect on energy output, followed by light switch-off, and occupancy models. Light switch-on models showed a lower influence on the overall building energy performance. Furthermore, the Generalized Estimating Equations method was adopted to assess the interdependence among stochastic models for OPA and showed that changing the stochastic model in window operation, occupancy estimation, and light switch-off behavior generates a considerable difference in building’s energy performance. Contrariwise, the available stochastic models for light switch-on and blind operation perform quite similarly among each other and have a limited impact on a building’s energy performance.

Occupant behavior modeling methods for resilient building design, operation and policy at urban scale: A review

Traditional occupant behavior modeling has been studied at the building level, and it has become an important factor in the investigation of building energy consumption. However, studies modeling occupant behaviors at the urban scale are still limited. Recent work has revealed that urban big data can enable occupant behavior modeling at the urban scale – however, utilizing the existing data sources and modeling methods in building science to model urban scale occupant behaviors can be quite challenging. Beyond building science, urban scale human behaviors have been studied in several different domains using more advanced modeling methods, including Stochastic Modeling, Neural Networks, Reinforcement Learning, Network Modeling, etc. This paper aims to bridge the gap between data sources and modeling methodologies in building science by borrowing from other domains. Based on a comprehensive review, we 1) identify the modeling challenges of the current approaches in building science, 2) discuss the modeling requirements and data sources both in building science and other domains, 3) review the current modeling methods in building science and other domains, and 4) summarize available performance evaluation metrics for evaluating the modeling methods. Finally, we present future opportunities in building science with enhanced data sources and modeling methods from other domains.

The Role of Occupants in Buildings’ Energy Performance Gap: Myth or Reality?

Buildings’ expected (projected, simulated) energy use frequently does not match actual observations. This is commonly referred to as the energy performance gap. As such, many factors can contribute to the disagreement between expectations and observations. These include, for instance, uncertainty about buildings’ geometry, construction, systems, and weather conditions. However, the role of occupants in the energy performance gap has recently attracted much attention. It has even been suggested that occupants are the main cause of the energy performance gap. This, in turn, has led to suggestions that better models of occupant behavior can reduce the energy performance gap. The present effort aims at the review and evaluation of the evidence for such claims. To this end, a systematic literature search was conducted and relevant publications were identified and reviewed in detail. The review entailed the categorization of the studies according to the scope and strength of the evidence for occupants’ role in the energy performance gap. Moreover, deployed calculation and monitoring methods, normalization procedures, and reported causes and magnitudes of the energy performance gap were documented and evaluated. The results suggest that the role of occupants as significant or exclusive contributors to the energy performance gap is not sufficiently substantiated by evidence.

Reinforcement learning of occupant behavior model for cross-building transfer learning to various HVAC control systems

Occupant behavior plays an important role in the evaluation of building performance. However, many contextual factors, such as occupancy, mechanical system and interior design, have a significant impact on occupant behavior. Most previous studies have built data-driven behavior models, which have limited scalability and generalization capability. Our investigation built a policy-based reinforcement learning (RL) model for the behavior of adjusting the thermostat and clothing level. Occupant behavior was modelled as a Markov decision process (MDP). The action and state space in the MDP contained occupant behavior and various impact parameters. The goal of the occupant behavior was a more comfortable environment, and we modelled the reward for the adjustment action as the absolute difference in the thermal sensation vote (TSV) before and after the action. We used Q-learning to train the RL model in MATLAB and validated the model with collected data. After training, the model predicted the behavior of adjusting the thermostat set point with R2 from 0.75 to 0.8, and the mean absolute error (MAE) was less than 1.1 °C (2 °F) in an office building. This study also transferred the behavior knowledge of the RL model to other office buildings with different HVAC control systems. The transfer learning model predicted the occupant behavior with R2 from 0.73 to 0.8, and the MAE was less than 1.1 °C (2 °F) most of the time. Going from office buildings to residential buildings, the transfer learning model also had an R2 over 0.6. Therefore, the RL model combined with transfer learning was able to predict the building occupant behavior accurately with good scalability, and without the need for data collection.

A framework for estimating the energy-saving potential of occupant behaviour improvement

Energy-related occupant behaviour in buildings has demonstrated considerable energy-saving potential. However, the current modelling method of occupant behaviour does not give sufficient considerations on the implementation difficulty of behaviour and provide a holistic map from survey data to various behaviour models. This article proposes a holistic survey-and-simulation-based framework for estimating the energy-saving potential of occupant behaviour improvement. In the framework, seven typical categories of occupant behaviour models are identified based on the survey results. According to the implementation difficulty, the models are integrated into four behaviour styles (baseline, wasteful, moderate and austere) to represent different levels of energy-saving consciousness of occupants. Based on a case study with a nationwide survey in Singapore, there are remarkable energy savings potential if occupant behaviour is improved; the building energy consumption can be reduced by up to 9.5% with the moderate behaviour improvement, and up to 21.0% with the aggressive behaviour improvement. The simulation results accord well with the measured results within a reasonable range of deviation. The framework can be applied to estimate the energy-saving potential of occupant behaviour improvement in a building with affordable cost, and the findings can inform a behaviour improvement programme with effective and efficient measures.

Development of a systematic procedure to establish customized shading behavior identification model

Occupant behavior is becoming a major factor impacting indoor environments and building operations. Individual differences in indoor environment preferences impact the implementation of occupant-centric building operations. Previous studies regarding occupant behavior have focused on improving the building simulation performance by establishing a representative occupant behavior model to reduce the gap between the real performance and simulation results. In this study, a customized shading behavior identification model (CSBIM) is developed to learn and identify the personalized shading behavior habits and indoor environmental preferences of targeted occupants during the building operation phase. The CSBIM can be used to extract personalized behavioral preferences and embedded into the optimization controls of occupant-centric buildings. The proposed CSBIM is developed based on the cognitive concept of subliminal perception. Monitored data from an office on campus is used to select the optimal classification algorithm and evaluate the effectiveness of the CSBIM. The results indicate that using CSBIM – Random Forest improved capturing the personalized shading behavior by 12.56%. Finally, the advantages and limitations of the CSBIM are also considered.

Influence of Occupant Behavior for Building Energy Conservation: A Systematic Review Study of Diverse Modeling and Simulation Approach

Energy consumption in buildings depends on several physical factors, including its physical characteristics, various building services systems/appliances used, and the outdoor environment. However, the occupants’ behavior that determines and regulates the building energy conservation also plays a critical role in the buildings’ energy performance. Compared to physical factors, there are relatively fewer studies on occupants’ behavior. This paper reports a systematic review analysis on occupant behavior and different modeling approaches using the Scopus and Science Direct databases. The comprehensive review study focuses on the current understanding of occupant behavior, existing behavior modeling approaches and their limitations, and key influential parameters on building energy conservation. Finally, the study identifies six significant research gaps for future development: occupant-centered space layout deployment; occupant behavior must be understood in the context of developing or low-income economies; there are higher numbers of quantitative occupant behavior studies than qualitative; the extensive use of survey or secondary data and the lack of real data used in model validation; behavior studies are required for diverse categories building; building information modeling (BIM) integration with existing occupant behavior modeling/simulation. These checklists of the gaps are beneficial for researchers to accomplish the future research in the built environment.

Investigating the Impact of Actual and Modeled Occupant Behavior Information Input to Building Performance Simulation

Occupant behaviors are one of the most dominant factors that influence building energy use. Understanding the influences from building occupants can promote the development of energy–efficient buildings. This paper quantifies the impact of different occupant behavior information on building energy model (BEM) from multiple perspectives. For this purpose, an occupant behavior model that uses agent–based modeling (ABM) approach is implemented via co-simulation with a BEM of an existing commercial building. Then, actual occupant behavior data in correspondence to ABM output, including operations on window, door, and blinds in selected thermal zones of the building are recorded using survey logs. A simulation experiment is conducted by creating three BEMs with constant, actual, and modeled occupant behavioral inputs. The analysis of the simulation results among these scenarios helps us gain an in–depth understanding of how occupant behaviors influence building performance. This study aims to facilitate robust building design and operation with human–in–the–loop system optimization.

A simulation-based method to investigate occupant-centric controls

Occupant-centric control (OCC) strategies rely on different algorithms to learn and predict occupants’ patterns and preferences, then utilize these predictions to optimize building operations. However, testing different OCC algorithms or fine-tuning their configurations in real buildings can be a lengthy process. To this end, we present a framework for testing OCCs in a simulation environment prior to field implementation. The proposed workflow entails using synthetic occupant behaviour models and simulating OCC strategies to learn their preferences. The goal is to enable quick comparison of different OCC configurations under various scenarios by modifying occupant behaviour assumptions, as well as climate and design parameters. For proof-of-concept, the proposed method was applied in a case-study to simulate OCCs for lighting and heating/cooling setpoint adjustments in a single office under various occupant types, as well as OCC settings and design configurations. Results demonstrated the benefits of the proposed framework and its potential for providing a more holistic evaluation of OCCs under different scenarios. Using the proposed framework, building designers and operators can identify potential issues with OCCs and fine-tune their settings prior to field implementation.

Establishment of a personalized occupant behavior identification model for occupant-centric buildings by considering cost sensitivity

Modern automation systems play an important role in improving building energy efficiency. However, some evidence shows that occupants are often dissatisfied with automation systems and may intervene, e.g., by manually adjusting automated shading or air conditioning. Therefore, the concept of occupant-centric buildings has been proposed by some scholars to solve the abovementioned problem. As the basis for realizing occupant-centric buildings, personalized occupant behavior should be identified dynamically. Thus, this study proposes a systematic procedure for establishing a personalized occupant behavior model with cost sensitivity analysis (POBM–CSA). The CSA can improve the identification performance of the model while considering the personalized preferences and attitudes of target occupants. By taking shading behavior identification as an example, the total cost and area under the cost curve were used to evaluate this consideration. The results showed that the POBM–random forest with cost-sensitive classification has the best cost sensitivity performance. By establishing a customer rule library, personalized behavior patterns with reduced misclassifications can be output by the POBM–CSA to improve the operation of occupant-centric buildings.

Building Performance Evaluation Using Coupled Simulation of EnergyPlus™ and an Occupant Behavior Model

Building energy simulation programs are used for optimal sizing of building systems to reduce excessive energy wastage. Such programs employ thermo-dynamic algorithms to estimate every aspect of the target building with a certain level of accuracy. Currently, almost all building simulation tools capture static features of a building including the envelope, geometry, and Heating, Ventilation, and Air Conditioning (HVAC) systems, etc. However, building performance also relies on dynamic features such as occupants’ interactions with the building. Such interactions have not been fully implemented in building energy simulation tools, which potentially influences the comprehensiveness and accuracy of estimations. This paper discusses an information exchange mechanism via coupling of EnergyPlus™, a building energy simulation engine and PMFServ, an occupant behavior modeling tool, to alleviate this issue. The simulation process is conducted in Building Controls Virtual Testbed (BCVTB), a virtual simulation coupling tool that connects the two separate simulation engines on a time-step basis. This approach adds a critical dimension to the traditional building energy simulation programs to seamlessly integrate occupants’ interactions with building components to improve the modeling capability, thereby improving building performance evaluation. The results analysis of this paper reveals a need to consider metrics that measure different types of comfort for building occupants.

An international review of occupant-related aspects of building energy codes and standards

In light of recent research, it is evident that occupants are playing an increasingly important role in building energy performance. Despite the important role of building energy codes and standards in design, the occupant-related aspects are typically simple and have not kept up with the leading research. This paper reviews 23 regions’ building energy codes and standards by first comparing their quantitative aspects and then analyzing their mandated rules and approaches. While the present paper focuses on offices, general recommendations are applicable to other building types as well. The review revealed a wide range of occupant-related values, approaches, and attitudes. For example, code-specified occupant density varies by nearly a factor of three between different codes. This underlines the need for development of advancement in occupant behavior modeling approaches for future occupant-centric building performance codes and standards. Moreover, occupants are often referred to only implicitly; underlying expectations about energy-saving occupant behavior from building occupants varies greatly; and, only a few codes address occupant feedback and system usability. Based on the findings of the review, a set of initial recommendations for future building energy codes is proposed.

A stepwise approach for assessing the appropriate occupant behaviour modelling in building performance simulation

ABSTRACTOccupant behaviour (OB) is recognized as a leading source of uncertainty in building performance predictions. Neglecting the potential influence of uncertainties on building performance could result in erroneous decision-making during the design phase. Therefore, it is essential that uncertainties are appropriately considered within building performance simulation (BPS) models. As for OB, there are various approaches to model occupant presence and actions in BPS tools. Literature shows that the appropriate modelling approach depends on the object and purpose of the simulation, which makes it difficult to favour a method over another a priori. Moreover, there is very little support for selecting the most appropriate modelling approach. As a result, OB is modelled in practice in various ways, mostly dictated by intuition and habit. This study builds on previous literature to introduce and test a complete approach for appropriate OB modelling. The method can be generalized and readily applied to design questions.

Context-specific urban occupancy modeling using location-based services data

Energy-related occupant behavior is a major source of uncertainty in building and urban energy performance simulations. Standardized assumptions, published by ASHRAE and others in the form of occupancy schedules, are widely used in research and practice, especially on the district-scale. In this work, we gathered location-based services data to create context-specific, data-driven occupancy schedules. Using a web mapping service, we collected data for retail and restaurant uses in the downtown neighborhoods of 13 different U.S. cities to create data-driven schedules for each context. The schedules were compared to ASHRAE standard assumptions using the earth mover's distance approach and the schedules' energy-related features. We found that standard schedules seem to significantly overestimate weekly building occupancy, although the shapes of the schedules are generally similar. The use of standard schedules could therefore, have significant impacts on district-scale energy demand simulations, as the overestimation will be cumulative.As compared to the differences between data-driven and standard schedules, the differences between different locations are significantly smaller. However in extreme cases, the weekly cumulative occupancy and the number of occupied hours differ by more than 30% between locations, which means that context-specific differences together with climatic differences might also impact building performance simulation results. Furthermore, we found differences in daily data between the different days of the week. In particular, the observed behavior on Fridays is significantly different from other weekdays for both considered use-types. This indicates that the conventional categorization of occupant behavior models into three day-types: weekday, Saturday, and Sunday, should be reconsidered.

Energy Saving at Work: Exploring the Role of Social Norms, Perceived Control and Ascribed Responsibility in Different Office Layouts

Reducing energy consumption in office buildings is critical for improving energy efficiency and decarburization at the large scale. This study (N=854) investigated the influence of social-psychological factors on energy-saving intention and behaviors in single-person versus shared offices based on the extend model of Theory of Planned Behavior (TPB). We found that ascription of responsibility, a variable added to the TPB, is the strongest predictor of energy-saving intentions for both single-person and shared offices. Interestingly, while injunctive norms are an important predictor of behavioral intention for single-person offices, descriptive norms are an important one for shared offices. Energy-saving intention mediates the influences of the aforementioned variables on energy-saving behaviors. Contrary to our hypotheses, perceived control over energy-saving and perceived ease of access to building control features have no direct impacts on energy-saving behaviors in single-person offices, but they have impacts on energy-saving behaviors in shared offices. This study provides useful insights for building designers and occupant behavior and energy modeling researchers.

Predicting the demand flexibility of wet appliances at national level: The case of France

The manual shift of residential wet appliances (dishwashers and washing machines) in French households that have subscribed to a Time of Use (ToU) tariff represents a flexibility potential available already today and at zero-cost. In this paper we quantify, for the first time, this potential by building a national representative synthetic population for France and using a novel bottom-up model. This includes a stochastic occupant behaviour model for predicting the use of the wet appliances and an overlying agent-based model for predicting their manual shifting under a ToU tariff system. The two models are calibrated and validated using empirical data collected over two recent monitoring campaigns, each carried out in more than 100 French households. By using the new model we observe that a 100% penetration rate of the ToU tariff together with an improved (but practically feasible) households’ flexibility could lead, between 19:00 and 20:00, to a mean hourly load reduction of 0.48 GW, which represents around 0.5% of the typical French annual peak load. This flexibility potential, which is today only partially exploited, could be achieved with zero additional costs for new infrastructures, by encouraging eco practices and/or making the ToU tariff more economically attractive for the households.