Understanding shading through home-use experience, measurement and modelling

1. Introduction

The well-documented two- to threefold energy performance gap in low-energy homes (Gupta et al. 2018) may be underpinned by the less recognised impact of the overly optimistic design stage assumptions in relation to solar gains. Could building performance models and simulations be improved to reflect more accurately solar gains and inhabitants’ shading behaviours? Would this optimise heating and (summer overheating) cooling demand over the annual cycle of seasons?

The share of solar and internal gains in the annual heating load is particularly high for energy-efficient and passive houses located in climate zones with a heating season lasting roughly half a year (Feist & Schnieders 2009). Hence, solar and internal gains need to be well understood at the design stage. The impact of internal gains depends on the inhabitants, their lifestyle and occupancy patterns (Dar et al. 2012), and is estimated at 3.8–6.6 W/m² (Elsland et al. 2014; Bandurski 2021). The impact of solar gains is more significant and dynamic than that of internal gains (Feist & Schnieders 2009; Toledo et al. 2024) because it depends on solar irradiance intensity and also on the shading practices of the inhabitants. In the temperate Cfb climate zone (Köppen–Geiger classification), gains from solar radiation on an autumn afternoon for a south-facing facade can be estimated at ~25 W/m². The relevance of solar gains on the heating load design stage calculation increases for buildings where glazed areas are intended to maximise their impact.

Energy certificates for European Union housing are typically based on the monthly calculation method (van Dijk et al. 2005). Dynamic building energy analysis tools allow for a less simplified and potentially more precise understanding of the impact of solar gains on annual energy balance. However, in practice the robustness of the results can be questioned. According to Strachan et al. (2015) and Roberts et al. (2019), even experienced modellers may struggle to model an unoccupied building, or a building occupied by synthetic users, emulating the assumed human behavioural patterns. When user behaviour has been predefined, the focus is on the definition of the thermal processes of the modelled building. The specification of heat exchange and thermal inertia may prove insufficient, and a model of mass exchange, the so-called airflow network (AFN), within the interior may prove necessary for reliable results. As Eguía-Oller et al. (2021) have shown, indoor air movement has a significant impact on a building’s thermodynamic state.

Estimating the contribution of solar gains becomes even more challenging if the impact of real-life user-related uncertainties is to be accounted for, e.g. varied window treatments (Bellia et al. 2014), differently adjusted (Day et al. 2020), and used according to individual preferences (Verbruggen et al. 2020). Shading practices-monitoring in occupied houses is challenging due to privacy concerns and a lack of off-the-shelf monitoring equipment acceptable for voluntary participants, and most empirical shading studies focus on non-residential environments (Stazi et al. 2017; Vasquez et al. 2022), with no evidence for in-use acceptance of ‘optimised’ solutions such as automated shading in housing (Firląg et al. 2015).

In the residential sector, few studies explore shading practices at the occupancy stage, and those few are mostly based on online surveys (Verbruggen et al. 2020; Andersen et al. 2009). Meteorological data linked with a survey of 900 residents in Denmark (Andersen et al. 2009) suggested that shading behaviour is correlated with triggers such as solar irradiance, thermal sensation and perceived indoor air quality, but also with home ownership type and respondents’ age. A field study in nearly zero-energy building housing in Belgium (Verbruggen et al. 2020) did not confirm the influence of environmental factors on shading activity, pointing towards individual preferences as the key trigger.

This suggests there is a research gap in the understanding and categorisation of factors that underpin the individual preferences. Due to the lack of generally agreed triggers of shading activity in both housing and offices, varied thresholds are adopted in shading simulation studies, mostly arbitrary or based on simple surveys (Yao 2014; Nicoletti et al. 2020; Karjalainen 2019; Niknia & Rashed-Ali 2024; Bavaresco & Ghisi 2018).

Despite the varied thresholds, simulation studies have consistently shown the potential of shading to optimise heating and cooling demand (Karjalainen 2019). However, most housing studies explore either summer overheating (Jensen Skarning et al. 2017; Brown 2023; Alrasheed & Mourshed 2023) or winter heating (Kharchi et al. 2012), not both. The previously unaccounted-for proliferation of air-conditioning into housing in a transitional temperate climate (Ford et al. 2022) adds urgency to the challenge of understanding the actionable opportunities to limit thermal comfort-related energy use through all seasons. Limited understanding exists of the in-use quantified impact of shading compared with other energy relevant behavioural aspects of domestic energy–comfort nexus.¹ As for long-term (all seasons) empirical monitoring of shading in a housing context, the authors are unaware of such studies. Furthermore, shading options, unified within office environments, can be highly diverse, not only within a housing development (Baborska-Narożny et al. 2017) but also in a single dwelling. This is due to varied room functions, user preferences or capabilities to adjust the interior, e.g. owner-occupiers are assumed to have more control than renters. Therefore, there is a need for more research to understand the drivers of inhabitants’ shading-type choices, contextualised shading triggers and simulation approaches capable to predict reliably the related thermal-energy consequences.

The aim of the study is to address this gap in the knowledge about inhabitants’ control of shading and its impacts on thermal conditions and energy demand. To achieve that end, in-depth building performance evaluation (BPE) findings are used to inform a dynamic thermal model focused on context-specific shading strategies adopted in energy-efficient homes.

The objective is to explore simulation scenarios and modelling approaches leading to results relatable to the inhabitants’ first-hand experiences. Several research questions (RQs) arise:

• RQ1: How do the inhabitants understand the relevance of shading in their homes across the seasons based on their experience?

• RQ2: Is the impact of varied shading strategies adopted by households visible in their respective energy use and indoor thermal profiles across the seasons?

• RQ3: How does airflow model complexity (inclusion of AFN analysis across the building or only in the inner high zone) influence building performance simulation results for the isolated impact of shading strategies on the overall building energy balance and overheating extent?

• RQ4: What is the impact of varied shading strategies adopted by households on energy use and indoor thermal profiles across the seasons?

The study reported here is part of a larger research project designed to provide robust understanding about how the practices of inhabitants of low-energy houses contribute to energy use and thermal comfort. To achieve that aim, a case study approach was adopted (Yin 2013) involving two stages. In-depth BPE results were collected to provide insight into inhabitants’ practices in order to inform and calibrate dynamic thermal modelling (Hensen & Lamberts 2019). Modelling was used to test the impact of alternative scenarios for isolated aspects of home use on energy and thermal comfort.

2. Methods

2.1 Case study characteristics

Six new-built low-energy houses located in Wrocław, Poland, were selected based on an occupancy period of no less than two years.² The six participating houses comprise a small developer-led investment of three semi-detached homes (Figure 1) built one next to the other with the same orientation and specifications by a single contractor in 2013. One of the six houses remained uninhabited throughout data collection; therefore, it was excluded from the study. A total of five dwellings constitute this study.

Figure 1

Several solutions unique in single-family housing construction in Poland were applied, such as solid wood technology (cross-laminated timber—CLT) for the envelope, air-to-water heat pumps, and whole-house mechanical ventilation with heat recovery (MVHR). Comparable heating, ventilation and air-conditioning (HVAC) systems became mainstream after 2021, when more stringent regulations became mandatory (Polish Government 2017). Each household individually specified and contracted the installation of HVAC units and control systems, while ventilation ducting and floor heating loops were included in the initial shared construction process. Therefore, the envelope energy standard and mechanical systems distribution provide a unified background for comparisons, but the specific HVAC characteristics and its controls, as well as interior design, including window treatments (i.e. shading), are part of the different physical context for practices unique for each household.

Data for building and systems characteristics are presented in Table 1, while detailed envelope specification has been previously published together with co-heating test results (Baborska-Narożny et al. 2023a; Grudzińska et al. 2025). Almost half of the volume of each house is a double-height living room connected to a kitchen and open towards the circulation space. The living room has a south-oriented floor-to-ceiling window with an area of 20 m². Different strategies of its shading are the main focus of the analysis.

Table 1

Building and systems characteristics.

	BUILDING ID
	WA1	WA2	WZ1	WZ2	WZ3
Floor area (m2)			150
Measured total heat loss coefficient (W/K)			116.1
Wall thermal transmittance (W/(m2K))			0.1
Windows thermal transmittance (W/(m2K))			0.8
Mechanical ventilation (MVHR)			Yes
Heat source for space heating and domestic hot water	AWHP	AWHP + wood stovea	AWHP	AWHP	AWHP + wood stove with a water jacket (back boiler)
Space heating			Underfloor

[i] Note: AWHP = air–water heat pump; MVHR = mechanical ventilation with heat recovery. For more details, see Baborska-Narożny et al. (2023a).

^aWood stove for space heating only.

2.3 building performance simulation

2.3.1 Building energy model

One semi-detached house was modelled in TRNSYS 18 (Figure 2). The violet box represents the adjacent building, which is the only source of shading at the site (Figure 1 and Table 2). The assumptions regarding transparent and opaque envelope properties were based on technical documentation and on-site surveys. To investigate the impact of model complexity on the results, two modelling approaches were implemented. Both are based on building energy simulation (BES) and AFN (Hensen & Djunaedy 2005; Hensen & Lamberts 2019). BES models calculate heat transfer through opaque and transparent envelopes caused by temperature difference and solar radiation. It is possible to add to the BES model heat transfer caused by airflow, but only as directly assumed airflow values. The AFN models calculate mass (air) transfer caused by pressure difference. Two sources of pressure difference are considered in this approach: the buoyancy effect (stack effect) and wind velocity.

Figure 2

Table 2

Parameters of the shading types for a TRNSYS model representing actual shading contexts in the case study homes.

TRNSYS MODEL	SHADING TYPE
	A	B	A (LIVING ROOM)	C/C1	D
tsh down (°C)	–	26	–	26/24	26
tsh up (°C)	–	24	–	24/23	24
Shading in other rooms	No	Yes	Yes	Yes	Yes
Indoor/outdoor shading	–	Indoor	Outdoor	Outdoor	Outdoor
Shading factor (%)	0%	90%	0%	70%	70%
Case study homes	WZ1, WZ2, WA2	WZ3	WA1

• Living room air circulation (simplified approach)

  The simplified approach is based mainly on the BES model, while AFN is used to calculate air mixing between the lower and upper parts of the double-height living room space. Additionally, airflows caused by mechanical ventilation are defined as the input to the model. In line with on-site observations and MVHR airflow check, it is assumed that the MVHR is used continuously, with 180 m³/h of the supplied and exhausted air. An 80% efficiency rate of heat recovery from exhausted air was adopted. A similar approach was used in a preliminary simulation study focused on shading (Baborska-Narożny et al. 2023c), where surprisingly extreme overheating was suggested, contrasting with the measurement results. The discrepancy of the modelled and measured thermal profiles prompted the need for a more advanced modelling approach proposed here.

• Whole-building air circulation (advanced approach)

  In this case, AFN is also used to calculate airflow in the whole open space of the house. This approach considers mass and heat transfer across the whole building through the loop: living room–corridors–staircase. Other assumptions of the BES and AFN models are the same as in the simplified approach.

2.3.2 Occupant behaviour model

For building performance analysis, it is very important to properly model not only the building physics but also the human–building interaction (HBI). The objective here is to consider shading usage; therefore, models focus exclusively on exploring varied shading types and triggers that prompt action within the predefined occupancy schedule.

Shading/unshading triggers

The analyses are based on four HBI models for shading triggers: a default software setting as a benchmark for three models developed in line with the BPE findings (see Section 3.1) (Figure 3). The default software model for shading regulation is defined as follows:

• TRNSYS 18 (TRNSYS18)

  Irrespective of occupancy profile, shading activity is triggered based on radiation thresholds onto the shaded window’s external surface: radiation > 140 W/m² for shading down and < 120 W/m² for shading up. The default software radiation thresholds are low, therefore the shading is mostly assumed ‘down’ on sunny days.

Figure 3

Shading trigger models informed by user feedback and observations are as follows:

• Temperature based (temp)

  The shading trigger applies to occupancy hours only. Its objective is to limit solar radiation penetration in the case of indoor overheating risk, defined as a static threshold of 26°C for shading and 24°C for unshading activities. Also, a variant of the temp model is proposed to test the impact of assuming lower threshold temperatures of 24 and 23°C (Figure 4 and Table 2).

• Temperature and radiation based (temp-rad)

  Shading is used as in the previous model and additionally in case of direct high solar radiation into a room which can cause glare (visual and summer thermal discomfort prevention) (Figure 5), unlike in the TRNSYS18 model where radiation is considered internally: radiation into the room related to floor area. In this model, the shading is ‘down’ more frequently on sunny days in swing and winter seasons, limiting solar gains that could diminish the need for heating.

• Temperature and radiation based with passive solar heating anticipation (temp-rad-solH)

  Shading is used as in previous models, but in the case of a foreseen absence during sunny daylight hours the occupants care to unshade the windows to provide passive solar heating (visual and thermal discomfort prevention plus heating load optimisation) (Figures 6 and 7).

Figure 4

Figure 5

Figure 6

Figure 7

Figure 3 illustrates the inhabitants’ shading objectives and how these relate to the TRNSYS model for shading. This is further explained in Section 3.1.2.

The temp-rad and temp-rad-solH models are implemented only for rooms with south-oriented windows: the living room and first floor bedroom. In other rooms, only the temp model is applied, as the risk of visual discomfort is minor due to positioning of the windows.

Shading types and triggers modelling

Based on walk-through observations, five shading types are defined for further investigation (Table 2). The baseline scenario A, prevalent in most houses, assumes a lack of shading. Scenario B represents a curtain attached to the beam at the height of about 2.2 m, shading the lower part of the window, as in WZ3. Scenarios C and D represent two main-use cases of an external roller blind, as in WA1. The roller blind is either fully open (equivalent to scenario A), lowered to the horizontal beam, shading the upper part of the window (scenario C), or lowered to the floor level (scenario D). Scenario C1 is a variant of scenario C, assuming different indoor temperatures as triggers for shading closing or opening.

Occupancy schedule

The occupancy schedule was simplified as the internal gains and occupancy are not the main focus of this paper. Although repeated on-site visits and interviews revealed that each household is characterised by different occupancy hours, a single occupancy schedule was adopted for all simulation scenarios to clearly distil the role of shading (see the supplemental data online). Also, occupancy is assumed to be constant over the year in line with a prevailing pattern determined through work, school and other recurring commitments schedule established through occupant feedback and which differs between workdays and weekends. Seasonal adjustments were not introduced to explore the impact of predefined shading scenarios for all weather conditions.

3. Results

3.1 RQ1: How do the inhabitants understand the relevance of shading in their homes across the seasons based on their experience?

3.1.1 Walks-through

Differences were observed between households, and within homes between rooms, in terms of the presence of window treatments (Table 3). Most significantly, the double-height south-facing window with a glazed area > 20 m² had window treatments installed in two of the five households. All the residents, unprompted, referred to their experiences about the impact of that window on the indoor thermal environment. However, the shared experiences triggered different responses in terms of the affordances of shading options and contributed to varied home-use practices. Exploring the meanings and functionalities contributing to shaping the material context of these practices is the focus of the following section.

Table 3

Window treatments.

FLOOR	ROOM AND ORIENTATION OF WINDOW			WINDOW TREATMENTS
				WA1	WA2	WZ1	WZ2	WZ3
First	Entrance area		N	–	–a	–	–a	–a
	Living room/kitchen		S	External roller blinds	–a	–	–a	Curtain
	Room		N + E/W	Curtain	–	Curtain	Pleated shades	Curtain
Second	Upstairs corridor		N	Curtain	–	–	Pleated shades
	Bedroom	Master	N + S/E	Curtain	Curtain	–	Curtain	Curtain
		Child	N + E/W	Curtain	–	Pleated shades	Pleated shades	Curtain
		Teenager	S + E/W	Roman shades	–	Curtain	Pleated shades	Curtain
Interviewees + walk-through contributors				P	P + T	P + S + YA	P + S	P + S

[i] Note: ^aOther.

Orientation: a solidus ‘/’ indicates either side of the semi-detached houses.

Interviewees: main participant (P), spouse (S), teenager (T) and young adult (YA).

3.1.2 Interviews

Overall, the large window within a double-height living room is perceived across all households as a unique and highly appreciated feature, allowing for a visual connection with the green outdoors while retaining privacy. The quality of spaciousness and light it offers is praised. Moreover, some residents see it as part of a thoughtful holistic design, with an overhang at the roof level and walls extended towards the garden by 2.5 m. The overhang is seen as intended precisely to address the need for shading in summer (P_WZ1). For those still willing to introduce window treatments, the unique glazing size and construction technology of the houses has become a practical challenge. As P_WA1 recalls, finding a contractor willing and capable of installing large external roller blinds on a wall with 30 cm insulation was a difficult challenge.

Shading functionality

Inhabitants’ feedback indicates that the introduction of window treatments in their homes is mainly driven by two functionalities.

• The desire to control privacy meant limiting the visibility of the indoor spaces from the outside. In most houses the curtains or pleated shades were installed in the bedrooms and upper corridor linking the bathroom with the bedrooms, on windows facing the semi-public entrance area. The living room/kitchen area windows face gardens and are protected from public gaze by planting and the projecting walls framing the facade (Table 2). Therefore, securing privacy is not perceived as applicable to this window.

• The desire to create visual comfort meant controlling the amount of direct sunlight and daylight. The diversity of occupancy patterns for different rooms is seen as linked to household members’ demographics and preferences. The two households with shading in the living room both regularly spent several afternoon hours there, engaging in activities that require control of glare, such as using the television or playing board games. The households with teenagers and young adults reported that more time was spent in individual bedrooms. The lack of control over direct sunlight entering the living room reflects the different pattern of use. The bedrooms of teenagers or young adults all had window treatments installed allowing for solar control.

Thermal environment and shading

Attitudes towards the thermal aspects of direct solar radiation control are nuanced, with preferences varying between individuals and seasons. There was also a lack of clarity about strategies for improvement.

For most of the year, when overheating is not an issue, direct solar radiation is perceived as thermally pleasing. Furthermore, in households heated intermittently (WA2), where the inhabitants sometimes work from home, the observed tangible impact of direct sunshine through unshaded windows on indoor temperature is perceived as being highly beneficial, sufficient to postpone the need for active heating (i.e. a wood burner) to the late afternoon hours:

When the sun is shining, I don’t need the heating, it gets so warm.

(P_WA2)

The effect is much less visible to the inhabitants of houses that remain mostly unoccupied during peak daylight hours, with heating controlled via programmable thermostats (WZ2, WA1) or through a heating curve adjusting the heat pump’s output based on outdoor temperature to ensure predefined indoor temperatures (WZ1). Any savings on heating resulting from solar heat gains were impossible to discern from the total energy bill.

In summer, across all households, unprompted inhabitants mentioned that the houses became hot:

I prefer this house in winter rather than in the summer. In winter it’s great, in summer it’s too hot.

(P_WA1)

However, the installation of window shading was not considered by all the inhabitants desiring lower indoor temperatures. Instead, households introduced temporary solutions, set up specifically for heatwaves, such as a garden umbrella placed next to the window, moveable screens next to a couch or local shades applied directly onto the glazing (Figure 8).

Figure 8

It gets so hot in the summer, but I manage. I place these screens here. I am considering an awning, however. Would you think this would work?

(P_WA2)

Another house adopted a long-term shading strategy and planted deciduous trees, adjusting their expectations to reality:

It’s hot but we don’t mind. We use linen sheets and sleep well. They have a cooling effect.

(S_WZ1)

Shading was perceived as one of many diverse practices intended to keep their homes cool in summer, such as the use of fans or fans with moisturisers, night-time purge cooling by one household or the use of air-conditioning in another. The diversity of means used in parallel contributes to uncertainty about the impact of a single factor in isolation. Also, guidance on overheating prevention was difficult to put into practice.⁴

3.2 RQ2: Is the impact of varied shading strategies adopted by households visible in their respective energy use and indoor thermal profiles across the seasons?

The measured summer thermal profiles (Figure 9) correlate with shading capability. In two households with window treatments on the living room windows (WA1, WZ3) the indoor temperatures were lower than in the unshaded houses (Table 3).

Figure 9

Moreover, the adaptive comfort (Nicol & Humphreys 2002) calculated for the kitchen/living room area (Baborska-Narożny et al. 2024) confirms that only the households with installed shading did not exceed the upper thermal comfort threshold throughout the summer (Figure 10). However, as the measured window-opening timing and duration reveals, these two households also kept their large window open for longer in summer (Figure 11). In particular, windows open in the morning and evening, as observed in WZ3, have ventilative cooling potential. Thus, the contribution of natural ventilation or shading in overheating prevention is not possible to discern clearly based on temperature measurement results.

Figure 10

Figure 11

Overall, the indoor temperature profiles for seasons other than summer in WA1 and WZ3 reveal only single overheating hours, suggesting effective shading strategies. However, the thermal profiles for the non-shaded houses vary significantly, with some also experiencing few overheating hours. The effects of shading or window operation are thus difficult to isolate, confirming the difficulties expressed by the inhabitants.

For the swing seasons and winter, the hourly energy consumption profiles varied between households and did not allow for a clear interpretation of the potentially blocked solar gains (Figure 12). The estimated heating energy consumption, derived from monitoring data, is difficult to correlate with the observed shading practices. As the final heating energy consumption is influenced by household-specific processes other than solar gains control, e.g. thermostat set points, simulation experiments are useful to distil the impact of the observed shading scenarios.

Figure 12

One of the two houses (WZ3) with the large shaded windows regularly used a wood burner for space heating, and its electrical energy consumption was stable over the year, with the highest consumption in the later afternoon hours. The other shaded household (WA1) showed a significant diurnal variability of electricity consumption patterns. Overall, the midday consumption was the lowest across the seasons, including in summer, which seems driven rather by occupancy patterns than midday solar gains lowering the heating load. The three unshaded houses each showed a unique diurnal energy-use profile, characteristic across the seasons, with peak consumption during the early morning (WZ2) or evening hours (WZ1) not clearly related to the impact of solar heat gains on heating demand. Also, the use of biofuel in one house (WA2) impeded direct comparisons as the impact of solar gains and the use of wood burners overlapped and were not clearly visible in the presented data.

Total and space heating energy use for the case study households were analysed in (Baborska-Narożny et al. 2023b). Monthly analysis suggests that wood burners, when used regularly to complement heat pumps, actually double the energy consumption when compared with households relying solely on heat pumps (WZ3, WA2). The overall consumption in households with only heat pumps differs significantly, i.e. 29, 21 and 18 kWh/m²/yr for WA1, WZ1 and WZ2, respectively.

3.3 RQ3: How does air-flow model complexity influence building performance simulation results for the isolated impact of shading strategies on the overall building energy balance and overheating extent?

Model complexity influences the results: significantly for the living room temperatures (Figure 13) and slightly for overall heating energy needs (Figure 12).

Figure 13

Air circulation across the house, modelled in the advanced approach, indicates about half of the overheating hours, compared with the simplified approach, across all the shading scenarios.

In case of usable energy demand for space heating purposes (Figure 14), the difference between the two modelling approaches is less pronounced, but still visible. The simplified model suggests usable heating energy to be lower by about 1–2 kWh/m²/yr, i.e. about 3–5%, for all the shading scenarios than the advanced model. This is a direct consequence of the higher living room temperatures shown by the simplified approach.

Figure 14

3.4 RQ4: What is the impact of varied shading strategies adopted by households on thermal and energy outcomes across the seasons?

The results of the advanced model are more reliable and can be used to answer the key question about the quantified impact of varied shading strategies across the seasons. Results are relatable to the measurements observed in summer (Figure 12) when the potential of ventilative cooling is low (see the supplemental data online) due to high outdoor air temperatures throughout the day. In autumn, when low outdoor temperatures increase the ventilative cooling potential, the discrepancy between simulation and measurement is higher, as the model does not include the effect of window opening, assuming a continuous reliance on MVHR (Figure 15).

Figure 15

The simulations confirm that shading is overall a valid strategy to mitigate overheating in these specific climatic conditions. However, it can also contribute to an increased demand for space heating (Table 4). The modelled shading and usages result in at least 200 fewer annual overheating hours (CIBSE 2017) than in the unshaded scenario. However, even with window treatments in place, the thermal outcomes of some of the modelled usage scenarios prove unsatisfactory: significant exceedances occur over the 3% threshold of occupancy hours > 26°C. Satisfactory outcomes occurred for the indoor curtains and outdoor roller blinds: they were close to meeting the stringent threshold or were below the threshold. More specifically, simulation shows shading has the potential to cut overheating from 615 occupancy hours > 26°C in an unshaded living room to 28 occupancy hours > 26°C annually if the whole glazed area were to be covered with external roller blind (shading type D) driven by the visual or thermal triggers (temp-rad-solH). However, that scenario causes an increase in the usable heating energy demand from 33.7 to 45.2 kWh/m²/yr due to the blocked solar gains. If, on cooler days, the users acted to unshade and allowed the solar gains to enter, then the heating demand would drop to 38.4 kWh/m²/yr while keeping the number of occupancy hours > 26°C well below CIBSE’s threshold.

Table 4

Simulated annual impacts of different shading options and usage scenarios.

	LACK OF SHADING	CURTAINS				EXTERNAL ROLLER BLINDS
	A	B, TRNSYS18	B, temp	B, temp-rad	B, temp-rad-solH	C, TRNSYS18	C, temp	C, temp-rad	C, temp-rad-solH	C1, temp	C1, temp-rad	C1, temp-rad-solH	D, TRNSYS18	D, temp	D, temp-rad	D, temp-rad-solH
Simulated useful energy for heating (kWh/m2/yr)	33.7	36.4	34.4	40.2	35.4	40.7	33.9	34.9	34.8	34.4	35.1	35.0	50.4	34.9	45.2	38.4
Annual total hours in living room > 26°C (h)	1,693	1,214	449	119	366	822	1,424	1,353	1,402	1,237	1,220	1,254	11	242	67	157
Annual occupancy hours in living room > 26°C (h)	615	433	166	66	131	325	513	488	503	456	453	460	6	52	28	47

[i] Note: Based on useful energy for heating, total hours in the living room > 26°C and occupancy hours in the living room > 26°C. Numbers shown in bold exceed the Chartered Institution of Building Services Engineers’ (CIBSE) threshold for occupancy hours (102 h/yr for the simulated occupancy profile; see the supplemental data online).

The established range of energy and thermal outcomes of varied shading strategies confirms that usage patterns are highly relevant in both practice and when defining shading scenarios for simulations.

Care needs to be taken when relying on the default TRNSYS 18 shading trigger model. Comparing the results for the default model with those developed based on observations and interviews indicates the biggest discrepancies are for B and C shading types which assume control for only part of the glazed surface. The reason is that the default shading trigger is radiation on the shaded window area. The default model does not account for two different states for a single window. If only half the window is shaded, then solar gains through the unshaded glazing are neglected in the shading activation, thus the reliability of the model drops. The BPE-driven shading triggers proposed in this study are related to indoor parameters, i.e. indoor temperatures and indoor solar gains, in line with inhabitants’ objectives. Seasonal dynamics of shading practices are underpinned by the need to secure visual comfort and either optimise heating or mitigate overheating. Also, the proposed scenarios consider only selected occupancy hours, when shading activity is likely, leading to less extreme results than the default scenario, e.g. for shading type D.

4. Discussion

For energy-efficient housing located in a Cfb climate zone, south-oriented glazing covering 38–42% of the facade has been recommended as a passive solar design measure to optimise annual heating and cooling energy use (Obrecht et al. 2019). The recommendation was based on a modelling study calculated using the monthly method in a steady-state spreadsheet of the Passive House Planning Package (Passive House Institute 2012). However, Taylor et al. (2023) suggest that static models are unfit for reliable overheating predictions. Also, the model assumed ‘optimised’ performance of external shading. Such an approach in energy modelling indicates a focus on precision in technical aspects, while using standard ‘optimised’ user behaviours with low predictability and low influence (Yan et al. 2017). However, the inhabitants’ shading choices and practices need to be recognised as having a high influence on the reliability of the predictions.

Through repeated on-site visits and interviews, the present study confirmed that manual shading control is context specific and dynamic. Shading practices depend on inhabitants’ expectations of what can be achieved by activating shading or unshading. As shown, in-depth BPE can meaningfully inform modelling experiments. Such a process simplifies the dynamics of real-life situations; however, it is effective in isolating and quantifying the impacts of a range of shading scenarios plausible within a given context. Verification of the simulation results for HBI scenarios with measurement data from occupied homes means comparing simplified versions of isolated behaviours with real-life complexity. Therefore, the same results should not be expected, but there are recognisable patterns for those who understand the context.

Further effort is needed to quantify the contributing factors of comfort and energy use in energy-efficient homes. For the growing number of households equipped with heat pumps, energy use is perceived as ‘opaque’ and not relatable to specific home-use patterns. It is only understood at a high level of overall acceptability of utility bills. Reported in the literature, and confirmed in this study, lower priority given to energy than visual or privacy drivers of shading may be linked with a lack of a tangible feedback loop between practices and energy use.

For an increased overlap of building modelling and measurement results, not only is a better representation of HBI factors is needed but also the building physics representation is gradually being improved. Hensen & Djunaedy (2005) present an algorithm for modelling a decision on how detailed an airflow model should be: fully determined by input assumption versus AFN versus computational fluid dynamics. Their approach is based on sensitivity analysis to check how the more complex approach decreases simulation results’ uncertainty in the context of the required performance criteria. The present study shows that effective use of AFN can account for convection in double-height room and also for the different thermal loads of connected spaces with different solar or internal heat gains. In such cases, AFN is essential to mimic convection heat transfer both vertically (Lu et al. 2020; Eguía-Oller et al. 2021) and horizontally (Choi et al. 2023).⁵ Taylor et al. (2023) note that the challenge of overheating assessment and prevention is not yet standardised, and the many modelling approaches available give different solutions. Therefore, more case study analyses are needed to provide a foundation for standardisation. This study contributes to addressing this need and provides a more comprehensive insight into HBI in low-energy single-family buildings.

The uncertainties related to the utilisation of solar heat gains in energy balance captured in this study were also analysed by Bandurski (2021). Informed by measurements made in multifamily buildings, an estimated heat gain (solar and internal) utilisation factor was lower than if calculated by the formula presented in standards for the monthly energy balance model (EN ISO 52016; ISO 2017). The standard approach assumes that for a low heat-balance ratio (heat gains-to-heat losses ratio) the gains utilisation is 100% (van Dijk et al. 2005). An estimation based on energy-use monitoring data suggests that only part of the heat gains decreases heating needs, even for the coldest part of the year. Solar radiation may cause transient thermal or visual effects perceived as adverse, triggering shading or window-opening. While building modelling should aim to better understand and account for the user’s role, the prescriptive energy codes should enforce usable shading options in low-energy houses to limit the proliferation into housing of more energy-consuming technologies such as air-conditioning.

5. Conclusions

This paper explored the challenge of understanding the objectives of inhabitants’ shading practices when defining shading model-input specification to obtain reliable predictions.

The combined annual impact of shading practices on overheating mitigation and annual heating energy demand for thermal comfort in low-energy-occupied homes proves challenging to capture based on home-use experience or monitoring. Its contribution is difficult to isolate, as shading is one of several home-use practices relevant in terms of the comfort–energy nexus, while it is particularly difficult to monitor across seasons in a real-life housing context.

A dynamic thermal modelling experiment focused on quantifying the impact of varied shading practices was developed, informed by in-depth building performance evaluation (BPE) of energy-efficient houses. Using a comparison of two modelling approaches, varied in terms of airflow network (AFN) complexity, this research confirms that considering heat distribution based on indoor airflow between spatially connected zones is crucial to reliably estimate building performance in the case of open spaces with varied orientations. It is recommended to engage AFN to model convection heat transfer across the connected indoor spaces for indoor temperature analysis. However, if only energy consumption is estimated, a simple one-zone model with well-considered assumptions about shading seems suitable.

The modelling confirmed a mixed potential role of shading across seasons, not only positively contributing to overheating mitigation, mostly in summer, but also risking a significant increase in the annual heating energy demand. The results of a calibrated simulation, representing the key characteristics of the observed shading practices, indicate that for the shading scenarios proposed, the usable space heating energy demand varies within a range of 34–45 kWh/m²/yr and 67–1693 h/yr of indoor temperature > 26°C in the living room. A comparison of the simulation results with measurements suggests that the developed methodology of modelling is a useful framework with which to capture the role of shading in low-energy houses. The broad range highlights the crucial role of user practices that involve setting the scene early in the occupancy, i.e. making decisions about the shading control capacity and its subsequent deployment. Significantly, a default TRNSYS 18 shading-usage scenario simulates control triggers only partly related to those observed as relevant for the building users. The default setting does not account for (1) shading control limited to a portion of a large window plane, (2) leaving the room unshaded when away to maximise heat gains or (3) using shading in response to indoor thermal discomfort. Unsurprisingly, for the same physical model of a building and shading types, the default shading simulation delivered results different from those representing the triggers found relevant for the inhabitants.

In housing, window treatments that enable shading control are considered an interior design feature and beyond the scope of interest of building design. The significant impact of shading on the annual energy balance suggests it is worthy of attention as an energy performance design issue. This is also a research issue to untangle the energy-use uncertainties linked with the broad category of ‘suboptimum user behaviours’ contributing to the performance gap. Large glazing, characteristic for new-built housing, combined with the recent proliferation of air-conditioning in this building sector add weight and urgency to that challenge.

The research findings challenge the current approach of energy codes: a ‘single value’ certificate, which delivers building energy performance data for a given building for an average occupant. This is particularly relevant for highly energy-efficient buildings where the relative impact of occupant behaviour is more pronounced in the annual energy balance. The unresolved problem is the definition of an ‘average’ occupant, as exemplified by the exploration of the human–shading interaction in this study, as human–building interaction is not just one variable to be averaged over the year. If the main task of the energy code was to encourage designers and users to make energy-efficient decisions, then more appropriate, robust tools are needed to account for the performance outcomes of alternative usage scenarios.

Competing Interests

The authors have no competing interests to declare. Magdalena Baborska-Narożny is a member of the journal’s editorial board, but had no role in the review and decision-making process for this manuscript.

Data accessibility

To protect the privacy of voluntary participants, the building performance data for this study cannot be shared.

Ethical approval

Voluntary households were recruited following best-practice ethical process. Written informed consent was obtained for the activities foreseen in the study and an unconditional right to withdraw their data at any point.

Supplemental data

Two files containing supplemental data for this article can be accessed at: https://doi.org/10.5334/bc.568.s1