Introduction

Before new buildings are constructed, or existing buildings are retrofitted, their energy performance is generally estimated for various purposes, such as design optimisation, system configuration, compliance documentation, and the issuing of certificates. To do this various calculation methods and simulation tools can be used. However, the actual energy use of buildings during the operational phase often deviates from expectations that are based on the results of such computational estimations. Various factors can contribute to this difference, which is broadly referred to as the Energy Performance Gap (EPG) [1-5]. Occupants' control-oriented behaviour is a frequently cited item among such factors, as occupants make use of control opportunities to bring about favourable indoor-environmental conditions. Instances of such control opportunities include the opening and closing of operable windows, operating electrical lighting, and adjusting thermostat settings for space heating and cooling [6-8]. However, review of the literature in this area suggests that the evidence regarding the role of occupants and their control-oriented behaviour in buildings' EPG is not conclusive [9,10]. In this context, this article explores the ways occupants' control-oriented behaviour (e.g., interactions with buildings' control components and systems) can influence buildings' EPG and revisits the empirical evidence for the extent of occupant-driven EPG. Moreover, strategies are discussed that can provide control opportunities to the occupants while reducing the risk of compromising energy efficiency targets.    

Overview of the principal contributors to buildings' EPG

The main reasons behind the deviation of buildings' actual energy performance from the pre-construction estimates may be classified in terms of the following categories (Figure 1):

• There can be differences between as-designed versus as-realised construction features (e.g., building envelope, fabric, components).
• Specified building systems (for heating, cooling, ventilation, lighting) and their intended operation regime in the design phase may differ from the actually installed and operated system elements.
•  Predictions of energy performance must make assumptions about buildings' contextual conditions, most importantly, the future weather conditions. However, long-term weather forecasts can be highly unreliable. Likewise, predicting future urban conditions (e.g., urban microclimate) is not a trivial task.
• Design-phase assumptions about buildings' future occupants cannot be expected to accurately capture occupants' control-oriented behaviour, that is their interactions with building environmental control elements and systems.
• Selected calculation methods and tools may involve limitations in view of validity, fidelity, and adequacy for the purpose. Moreover, errors could be made in the deployment of computational tools (mistaken input data, improper settings, etc.).
• Energy metering infrastructure may have insufficient resolution and errors may occur while monitoring and documenting the magnitude of consumed energy.  

​

Figure 1. Schematic illustration of the contributing factors to buildings' EPG.

Before addressing the role of occupants on the EPG in more detail, it is important to address a common misunderstanding about the purpose of energy calculations and performance simulation. Given the stochastic characteristics of both weather conditions and occupants' behavioural patterns, energy calculations should not be expected to provide exact predictions of buildings' future energy requirements. Rather, the validity of computational results can be only judged with respect to the model input assumptions. This implies that occupants' role in EPG can be isolated only if the influences of other potentially contributing factors can be eliminated via a normalisation process [11]. In other words, the initial model input assumptions need to be modified after the fact to reflect the i) realised building construction, ii) installed systems, iii) prevailing weather conditions, and iv) monitored occupant behaviour.  

Are occupants the main cause of the buildings' energy performance gap?

A recent review of literature on the building-related energy performance gap assembled after a multi-step process 144 relevant technical articles that were subjected to a detailed quality and content assessment [1]. The articles entailed studies from different continents relating to EPG in different building types. The reported energy performance gaps in these studies are summarised in Figure 2, where positive and negative numbers indicate underestimation and overestimation of the actual energy use respectively. The EPG across all studies amounts, on average, to 55% (±90%). In this Figure, the mean and median EPG magnitudes are shown separately for residential and non-residential buildings. The mean and median of EPG magnitudes are +37% and +30% in the case of residential buildings and +16% and +14% in the case of non-residential buildings.  

Close to two-thirds of the reviewed papers report some form of identified or presumed occupant-driven EPG related to building envelope, mechanical equipment, plug-loads, lighting, and internal heat gains. The building envelope category relates mainly to operating windows and blinds. Window operation frequency appears in more than one-third of the studies as a recurrent occupant-driven EPG cause (e.g., increase in actual heating demand). Similarly, some studies reported higher or lower solar gains than expected due to occupants' operation of shading systems. As to the mechanical systems category, a number of studies reported that, when compared to model assumptions, actual indoor temperatures and heating/cooling durations can be higher/lower and shorter/longer respectively. Some studies report higher actual indoor temperatures than assumed, which may have contributed to the EPG magnitude. As-planned versus actual plug-loads, lighting usage, and internal heat load categories may also have contributed to the EPG. The most frequently reported occupant-related causes of EPG were plug-load schedules, window operation, and set-point temperature in 40%, 36%, 33% of the reviewed studies respectively. 

A relatively large fraction of the studies report some kind of occupant-related EPG, but they do not uniformly provide strong evidence for the decisiveness of the occupants' role. In fact, only 14% of the studied articles entailed quantitative data covering both energy use and occupant behaviour. These circumstances triggered a further investigation, which assigned quality labels to the review studies [12]. The quality labels were determined based on data quality, extent of normalisation, and the method used to infer the cause of the gap. The results of this latter investigation suggested that the reported magnitudes of general EPG was considerably smaller in high quality studies versus studies with lower quality (Figure 3).  

​

 Figure 2. Results of the statistical analysis of EPG magnitude (in %) in the review studies shown separately for residential and non-residential buildings (based on Mahdavi, Berger, et al., 2021).

Figure 3. Distribution of EPG as a function of the quality of the respective studies 
(based on Amin et al., 2022).

These observations underline the importance of improving methods and processes for the assessment of buildings' energy demand. Specifically, better models and perhaps more importantly, higher quality empirical data regarding occupants' indoor-environmentally relevant behaviour in buildings are needed [13,14]. The consideration of the demographic, socio-economic, and cultural issues, and their influence on the variance of occupants' behaviour would be valuable. Model calibration based on actual energy use data could also contribute to improving the reliability of energy demand projections, however, this possibility applies mainly to scenarios involving building retrofit or building operation.  

Avoiding potentially negative energy implications of occupants' behaviour

Even if past research may not have provided a conclusive and quantitatively substantiated empirical evidence for the importance of the occupants' role in EPG, their potentially considerable influence on buildings' energy performance should not be dismissed out of hand. In fact, there are plausible indications and partial data to support the supposition that, in specific circumstances, occupants' behaviour may have negative implications for the ultimate energy performance of buildings. Certain strategies and measures could help mitigate such negative impacts. Two broad categories can be distinguished, namely measures within the realm of building design and technology, and measures dealing with the provision and dissemination of information to occupants. A number of items in the former category (design and technology features) can be listed as follows:   

• When mapping the spatial reach of environmental control systems (for heating, cooling, ventilation, lighting) to different areas of the building, achieving a high level of zonal granularity would be advantageous. For instance, if a single control opportunity (e.g., a thermostat) is allocated to an entire floor of an office building, uniform conditions must be maintained across the entire area independent of the occupancy level. High zonal granularity of the systems' coverage allows for differentiated provision of services depending on the presence or absence of occupants.  
• The fine-grained zoning can be combined with individual control opportunities for the occupants to achieve higher levels of satisfaction with indoor-environmental conditions without undue ramifications for energy performance. Preferred indoor-environmental conditions vary significantly among different individuals, and over time. Personal control opportunities enable occupants to maintain preferable conditions in their immediate surroundings (e.g., at their workstations), without imposing major energy consequences upon larger areas in the building.  
• Smart zoning further allows for the effective utilisation of occupants' presence detection towards energy-efficient systems operation during the non-occupancy periods. Instances of such applications include switching off (or dimming down) luminaires, enacting thermostat setbacks, and the operation of air supply and solar shading elements.  
• To be operated properly, buildings control components, from simple windows and blinds to complex HVAC (Heating, Ventilating, Air-Conditioning) systems, must feature intuitive, transparent, and responsive user interfaces. Truly occupant-centric buildings and system designs offer modes of operation that can be understood without extensive instructions and are accessible to occupants of different cognitive and physical capabilities.   
• Intelligent automated operation of windows, blinds, and luminaires can use monitored indoor-environmental conditions (e.g., ambient temperature, illuminance level, CO2 concentration, or incident irradiance on the façade) to anticipate and accommodate occupants' requirements and preferences and thus reduce the probability of counterproductive actions.  

Some key items in the latter category (information) are briefly presented as follows:   

• There appears to be a need for better sources of information for occupants about the essential features of buildings' systems and equipment (and their interfaces) and how they could be most appropriately operated. The layout, user-friendliness, and timely provision of relevant instructions can help occupants to meet their needs and avoid potential pitfalls relating to energy-wasting behaviour.  
• If thoughtfully conceived and implemented, both general information campaigns and dynamic (real-time) energy-centric feedback mechanisms can be affective in encouraging and accommodating more energy-conscious user behaviour. General information dissemination initiatives and campaigns have the potential to raise occupants' level of awareness, both of environmental issues in general and opportunities (such as those entailed in adaptive behaviour) to act in an energy-conscious manner without compromising personal comfort. Moreover, informative feedback to occupants can include the track record of their past behavioural patterns and the corresponding consequences for the building's energy performance.  

As highlighted in the literature, and independent of the persuasiveness level of the available empirical evidence on the exact magnitude of occupant-driven EPG, the bulk of the above-mentioned measures and efforts ‘represent rationally arguable and common-sense options’ and have the potential to contribute to both the occupants' satisfaction and improving the buildings' energy efficiency.  

Conclusion

The projection of buildings' future energy demand frequently deviates from their actual use, a phenomenon referred to as the EPG. As buildings' envelope and systems have become increasingly efficient, it has been suggested that the occupants' behaviour is the main factor behind building-related EPG. However, EPG may also stem from performance modelling and monitoring deficiencies, as-designed versus as-implemented building construction and systems installation, and the stochastic nature of weather conditions. Even though the supposition that occupants are the main contributors to EPG has not been conclusively substantiated, it stands to reason that interactions of occupants with buildings' control components and systems do have the potential to undesirably influence buildings' energy performance. As such, it is important to pursue several measures, both in terms of building technology and occupants' information, that create win-win situations regarding both indoor-environmental quality and energy efficiency. Moreover, the positive impact of these measures would perhaps be even larger if our discussion does not assign the occupants as the main culprits responsible for the EPG, but as partners in a collective endeavour to enhance the energy performance of the built environment. 

References

[1] Mahdavi, A., Berger, C., Amin, H., Ampatzi, E., Andersen, R. K., Azar, E., Barthelmes, V. M., Favero, M., Hahn, J., Khovalyg, D., Knudsen, H. N., Luna-Navarro, A., Roetzel, A., Sangogboye, F. C., Schweiker, M., Taheri, M., Teli, D., Touchie, M., & Verbruggen, S. (2021). The Role of Occupants in Buildings’ Energy Performance Gap: Myth or Reality? Sustainability, 13(6), 3146. https://doi.org/10.3390/su13063146

[2] Mahdavi, A., & Berger, C. (2019). Predicting Buildings’ Energy Use: Is the Occupant-Centric “Performance Gap” Research Program Ill-Advised? Frontiers in Energy Research, 7, 124. https://doi.org/10.3389/fenrg.2019.00124 

[3] Menezes, A. C., Cripps, A., Bouchlaghem, D., & Buswell, R. (2012). Predicted vs. actual energy performance of non-domestic buildings: Using post-occupancy evaluation data to reduce the performance gap. Applied Energy, 97, 355–364. https://doi.org/10.1016/j.apenergy.2011.11.075 

[4] Andersen, R. K., Fabi, V., & Corgnati, S. P. (2016). Predicted and actual indoor environmental quality: Verification of occupants’ behaviour models in residential buildings. Energy and Buildings, 127, 105–115. https://doi.org/10.1016/j.enbuild.2016.05.074 

[5] Hansen, A. R., Gram-Hanssen, K., & Knudsen, H. N. (2018). How building design and technologies influence heat-related habits. Building Research & Information, 46(1), 83–98. https://doi.org/10.1080/09613218.2017.1335477 

[6] Mahdavi, A. (2023). The one true sustainable building [TED Talk]. TEDxTUWien, Vienna, Austria. https://www.youtube.com/watch?v=4GgkDMxVG7w 

[7] D’Oca, S., & Hong, T. (2014). A data-mining approach to discover patterns of window opening and closing behavior in offices. Building and Environment, 82, 726–739. https://doi.org/10.1016/j.buildenv.2014.10.021 

[8] Sadeghi, S. A., Karava, P., Konstantzos, I., & Tzempelikos, A. (2016). Occupant interactions with shading and lighting systems using different control interfaces: A pilot field study. Building and Environment, 97, 177–195. https://doi.org/10.1016/j.buildenv.2015.12.008 

[9] Mahdavi, A., Berger, C., et al. (2021a). A Pragmatic Theory of Occupants’ Indoor-Environmental Control Behaviour. Frontiers in Sustainable Cities, 3, 748288. https://doi.org/10.3389/frsc.2021.748288 

[10] Mahdavi, A., Berger, C., et al. (2021b). An occupant-centric theory of building control systems and their user interfaces. Energies, 14. https://doi.org/10.3390/en14164788 

[11] Berggren, B., & Wall, M. (2017). Two Methods for Normalisation of Measured Energy Performance—Testing of a Net-Zero Energy Building in Sweden. Buildings, 7(4), 86. https://doi.org/10.3390/buildings7040086 

[12] Amin, H., Berger, C., & Mahdavi, A. (2022). A structured approach to the evaluation of evidence for the purported role of occupants in energy performance gap. Proceedings of the 36th PLEA Conference 2022 - Sustainable Architecture and Urban Design.  

[13] Mahdavi, A. (2020). In the matter of simulation and buildings: Some critical reflections. Journal of Building Performance Simulation, 13(1), 26–33. https://doi.org/10.1080/19401493.2019.1685598 

[14] Dong, B., Liu, Y., Mu, W., Jiang, Z., Pandey, P., Hong, T., Olesen, B., Lawrence, T., O’Neil, Z., Andrews, C., Azar, E., Bandurski, K., Bardhan, R., Bavaresco, M., Berger, C., Burry, J., Carlucci, S., Chvatal, K., De Simone, M., … Zhou, X. (2022). A Global Building Occupant Behavior Database. Scientific Data, 9(1). https://doi.org/10.1038/s41597-022-01475-3