Evaluating the Resilience of Ventilation Strategies in Low-Energy Irish Schools

Abstract

In the face of increasing global temperatures, this study aims to explore ventilation strategies that could provide passive cooling to mitigate overheating in studied low-energy school buildings, in particular those that use ventilative cooling. This study utilises building modelling calibrated with field data to tackle the challenge of maintaining indoor thermal comfort and cognitive performance levels during increasingly warm seasons. The calibrated building model is used to evaluate the vulnerability of classrooms, identifying and addressing risks based on standardised overheating and resilience criteria. Two primary school classrooms were simulated in three main cities across Ireland to assess the possibility of natural ventilative cooling for maintaining indoor thermal conditions without sacrificing energy efficiency. The study highlights the critical need to enhance natural ventilation strategies to protect against projected future overheating, with peak indoor temperatures reaching 29 °C to 31 °C during May, June, and September. Implementing a maximum natural ventilation strategy during occupied times, with a 9.6% opening-to-floor area ratio, can reduce peak indoor temperatures by up to 2.5 °C. Findings show Irish classrooms in low-energy buildings equipped with hybrid ventilative cooling can act as potential climate shelters during July and August under extreme weather conditions, underlining their capacity to provide a comfortable environment for vulnerable people during heatwaves and reduce overheating risk by 42–48% compared to natural ventilation. Additionally, projections show that cognitive performance loss in students may rise to 23% by 2071 due to raised indoor temperatures; however, this can be reduced to below 10% in 2021 and 2041 with maximum natural ventilation. The novelty of this work lies in its systematic evaluation of ventilative cooling resilience under future climate scenarios across multiple Irish city contexts, providing a robust evidence base for designing climate-resilient, energy-efficient learning environments.

1. Introduction

As global temperatures continue to rise due to climate change, indoor thermal conditions are increasingly affected, particularly during summer and the non-heating season [1,2]. The increasing frequency and intensity of heatwaves highlight this concern. For instance, Europe experienced record-breaking temperatures during the summer of 2019 [3], while the devastating 2003 heatwave resulted in approximately 70,000 excess deaths [4]. More recently, in the summer of 2022, heatwaves were linked to 3271 additional deaths in England and Wales [5]. In Ireland, Met Éireann reported 2023 as the hottest year on record, with several unprecedented climatic events for the country [6]. Recent study highlights significant regional variations in temperature-related mortality risk across Europe, which are projected to grow further due to the combined impacts of climate change and an ageing population [7].

In response to these climate extremes, the concept of climate resilience in buildings has become a critical concern, particularly the role of non-residential buildings, such as schools, as potential climate shelters [8]. Therefore, these buildings need to be prepared not only to maintain thermal comfort and protect health, but also to provide refuge during heat emergencies in summer. This positions educational buildings at the intersection of several critical performance pillars: thermal comfort, health, well-being, productivity, and cognitive performance [6,7,8].

In modern societies, where people spend over 90% of their time indoors, students spend more time at school than in any other building [9]. This fact emphasises the importance of maintaining comfortable indoor thermal conditions in educational buildings [8,10]. However, designing and operating school buildings is a complex task, due to the necessity of meeting often-conflicting parameters related to indoor environmental quality (IEQ). These parameters include thermal comfort conditions, indoor air quality (IAQ), acoustics, and visual comfort [11]. Overheating in school environments has been shown to increase discomfort, elevate stress levels, and negatively affect both the health and academic performance of students [12,13].

Cognitive performance is closely tied to indoor environmental conditions, particularly temperature. Children, who are still physically and cognitively developing, are more vulnerable than adults to heat stress and poor air quality [14]. Studies have shown that raising temperatures can negatively impact concentration, memory, and academic performance, especially in densely occupied and poorly ventilated classrooms. As a result, the resilience of school buildings is not only a matter of comfort and energy use, but also a public health and educational concern [11,15,16].

Considering the anticipated significant increase in cooling demands for buildings, all buildings are expected to comply with the standards of nearly zero emission buildings (NZEBs) [17,18]. NZEBs or low-energy buildings typically feature a well-insulated, thermally decoupled envelope designed to minimise heat loss and energy consumption. However, the high levels of insulation, airtightness, and thermal mass promote heat retention within the fabric, which can lead to elevated indoor air temperatures and an increased risk of summer overheating [11,19,20]. The increasing cooling demand for thermal comfort may undermine NZEB goals, increasing the use of energy-intensive mechanical cooling systems [21]. According to the 2025 SBEM/BER database, approximately 1250 primary schools exist in Ireland, of which about 55% are A-rated and newly built. This highlights the growing importance of evaluating the thermal resilience of low-energy school buildings to ensure that current design practices effectively deliver thermal comfort and climate resilience, while older schools remain widely recognised as vulnerable to overheating.

To address these challenges, NZEBs have increasingly turned to passive cooling strategies, often adopting a ventilative cooling (VC) approach [22]. Previous studies indicate that the potential reduction in cooling energy consumption through natural ventilation (NV) ranges from 8% to 78%, depending on local weather conditions and outdoor air quality [23]. VC is widely recognised as a climate adaptation strategy aimed at mitigating the impacts of climate change and enhancing the resilience of buildings [24]. Despite the growing importance of climate resilience in education, limited research has explored adaptation strategies for school buildings. Building simulation techniques offer a rapid and reliable method for testing the effectiveness of these interventions [25]. To ensure the accuracy of these models, calibration is essential. Calibration involves adjusting the model parameters to match the observed data, enhancing the model’s predictive capabilities [26,27].

VC control strategies play a significant role in influencing both the cooling effectiveness and energy consumption of buildings. The adoption of VC as a heat dissipation technique is often complemented by passive interventions that both prevent and/or modulate the indoor heat build-up; this has led to the use of the term “VC+”, based on a recent definition where the plus refers to the combination of VC and these additional passive strategies such as solar shading, cool roofs, and insulation [28]. This integration seeks to leverage the strengths of both hybrid and passive cooling strategies to create a more robust and sustainable solution [29]. When the cooling potential of outdoor air is insufficient, a mechanical ventilation system can be employed as a supplementary cooling source, commonly known as hybrid ventilation [30,31].

VC use has grown over recent decades, with hybrid systems combining natural and mechanical VC proving most effective. Ramakrishnan et al. [32] revealed that heatwaves significantly reduced the effectiveness of night ventilation and negatively impacted thermal comfort in free-running buildings. To address this, another study [33] suggested that pre-cooling buildings during off-peak hours with high-capacity cooling systems reduced overheating by approximately 60% compared to the baseline. This highlights the need to assess diverse hybrid VC systems based on varying control strategies [30]. Despite promising case studies, VC remains neglected due to design variability across climates, O’Sullivan et al. [34] highlighted the importance of aperture sizing in VC performance by analysing the percentage of opening-to-floor area (POF) across 14 international case studies. They found that 65% of buildings had POF values below 4%, with NV systems averaging 3.6% and hybrid systems ranging between 4.6% and 6.0%. Interestingly, no clear correlation emerged between building category and POF, but the highest values were found in cooler-summer Mediterranean climates. The study recommends a POF range of 2–4% as a practical guideline, despite many building regulations requiring a minimum of 5%, emphasising that selecting a higher POF value during early design stages may improve VC effectiveness and resilience [35].

While VC can support thermal comfort in schools, practical challenges often limit its effectiveness. Breesch et al. [35] found that night ventilation in a NZEB lecture hall was highly dependent on weather conditions, with airflow rates varying due to wind speed, direction, and small temperature differentials. System performance was further affected by user behaviour. These findings highlight the need for improved control strategies and user engagement to ensure reliable VC performance in educational buildings. Similarly, Martin and Fletcher [36] suggest that hybrid systems with night ventilation do not necessarily outperform simpler NV, recommending their use only under extreme weather conditions. This insight highlights the importance of context-specific design [37]. Mohamed et al. [11] found that many classrooms with mixed-mode ventilation frequently overheated during more than 60% of their occupied hours due to airtight envelopes. This study highlights the need for more robust VC design strategies, including improved cross-ventilation, responsive controls, and user training, to meet both thermal comfort and air quality standards in educational buildings.

Resilience to overheating is critical for climate adaptation, while energy efficiency remains a core component of climate change mitigation. Burman and Mumovic [38] examined two schools in the UK and concluded that passive design alone may not prevent future overheating, whereas a mechanical system offers better resilience but increases energy use. Their study highlights the need for integrated design strategies from the early stages. Laouadi et al. [39] introduced health-based overheating thresholds for school buildings, emphasising the vulnerability of younger students. This study proposed age-specific limits for overheating, based on physiological responses such as dehydration. The findings indicate that existing comfort-based standards, such as BB101, may underestimate overheating risks, particularly in primary schools. Kohl et al. [40] found a comfort performance gap in a NZEB Austrian school: indoor temperatures of 23.1–25.1 °C and a CO₂ > 1000 ppm for 20% of school hours, with teachers reporting discomfort and calling for better ventilation and thermal control to protect childrens’ comfort and health.

In an Irish educational building, O’Donovan et al. [41] highlighted the need for a resilience assessment of overheating risk under future climate scenarios, along with the localised weather files that account for urban heat islands (UHI). Their findings suggest that the most effective control strategies for balancing comfort and energy efficiency involve solar shading and night ventilation.

Although numerous studies have examined VC and overheating risk, few have focused on enhancing the resilience of NV strategies in low-energy or NZEB-compliant school buildings. This study addresses this gap by evaluating the thermal resilience of typical Irish primary school classrooms, representative of hundreds of schools built under the Department of Education’s Technical Guidance Documents (TGD-022) [42], using Met Éireann Design Summer Year (DSY) weather files, including future climate projections [43]. Based on these observations and gaps, the main objectives of this study are as follows:

1.

To calibrate a building simulation model to accurately predict indoor thermal conditions and assess the comfort levels of occupants under future extreme weather conditions.

2.

To investigate the potential resilient and low-cost ventilative strategies in mitigating overheating risks in the studied NV low-energy school building through a methodical approach.

3.

To assess the impact of future extreme weather conditions on the cognitive performance of children in case study classrooms and explore the effectiveness of VC in mitigating the adverse effects.

4.

To evaluate the potential of the studied school to serve as a climate shelter during the summer, providing a comfortable environment for vulnerable people.

The novel contribution of this study is that it is the first to assess the thermal resilience of NV low-energy schools in Ireland under current and future climate scenarios using newly released Met Éireann DSY weather files, which have not previously been applied in building performance studies. Through calibrated simulation and multi-criteria assessment, the work examines low-cost implementable VC strategies, their effectiveness in mitigating overheating and cognitive performance loss, and the potential for schools to operate as climate shelters during the summer. These contributions directly support the four objectives outlined above.

The forthcoming sections include a detailed methodology in Section 2, which presents the case study used for calibration and validation, the future building model, and the simulation methodology. Section 3 will present the studied ventilation strategies. In Section 4, the simulation results are presented and they include the risks of overheating vulnerability, the resilience assessment results under extreme weather files, the potential natural and hybrid ventilation strategies, the potential adaptations to create a more resilient and cognitive built environment for students’ learning performance, and the potential of using the studied low-energy school in Ireland as a climate shelter during extreme summer events. Section 5 will discuss the effectiveness of NV strategies in mitigating overheating and cognitive performance loss, highlighting key challenges like the night-purge strategy (NpV), automated windows (AutV), and the need for hybrid ventilation (HybV) solutions. In Section 6, this study aims to derive conclusions contributing to the limitations and future research directions related to climate resilience in educational buildings.

2. Materials and Method

The methodological approach follows previous studies [11,41] and is based on a structured sequence that links the research objectives to the required analytical steps. First, the case study context is defined to establish the physical and operational characteristics of the school. This is followed by a calibration and validation stage to ensure that the model accurately reflects measured performance. Finally, the validated model is used to conduct dynamic simulations under current and future climate scenarios. This framework provides a transparent and replicable basis for assessing overheating risk and the performance of NV strategies.

2.1. Case Study

The studied primary school is a two-story building with 16 classrooms in Cork, Ireland, constructed in 2020. The selected classrooms follow the TGD-022 Primary School Design Guidelines, representing the typical layouts and construction used in new Irish schools [42]. For this study, Classroom East-GF (located on the ground floor with an eastern orientation) and Classroom South-FF (situated on the first floor with a southern orientation) were chosen as case studies. Each classroom’s area is approximately 72 m². These classrooms, as shown in Figure 1, are rectangular, measuring approximately 7.6 m (w) × 8.0 m (l) × 3.2 m (h) (width × length × height), linked by semi-open corridors that are about 2.0 m wide. The height of the classrooms from floor to false ceiling is 2.95 m. Both classrooms have openings along their long sides, featuring slight overhangs. A total of 35% of the external wall area includes windows and the POF is 7/72.7 = 9.6%. Each classroom has ten windows, four of which cannot be opened, but six windows are openable for NV, as shown in Figure 2 (left-top). External windows (including frame) are double-glazed with a total glazing thickness of 28 mm, a U-value of 1.2 W/m²K, and a g-value of 0.40, no external shading devices were installed, and internal block-out blinds/curtains were used. The construction of the school utilised concrete blocks, and the U-values for the floors, walls, and roofs were determined as 0.21 W/m²K, 0.21 W/m²K, and 0.6 W/m²K, respectively [44].

Heating for the school is supplied by centralised gas boilers, distributed through a pressurised and temperature-controlled hot water system with a local control system. Radiators are located under the windows in both classrooms. Manual-openable windows with Low-E double glazing are utilised for classroom ventilation. Each classroom includes toilet facilities, a storage wall, an information technology (IT) area, and a sink for educational purposes. The school building is open from 08:00 a.m. to 04:00 p.m. daily during weekdays. However, class runs on weekdays from 09:00 a.m. to 2:30 p.m. These operational hours as a normative profile were applied for future scenario simulations to ensure consistency across climate files and locations. In Ireland and other EU countries, the beginning and end of summer periods is based on the school and regional circumstances. Typically, schools remain closed throughout July and August for the summer holidays [45,46]. In line with Irish school calendars, classrooms are typically closed during July and August for the summer holidays. This was reflected in the simulation schedule, which excluded these months from occupied scenarios. For the purpose of modelling, calibration, and simulation, the occupancy rate was estimated to be 21 students for each classroom, with two short breaks and a lunchtime period.

2.2. Model Calibration and Validation

The simulation model in this study was created using Integrated Environmental Solutions Virtual Environment (IESVE-Version 2023) [47] software. The calibration process was carried out stepwise manually, as shown in Figure 3, to accurately predict indoor temperature conditions and assess occupant comfort levels under different weather conditions, using field data measurements collected from April 2023 to the end of June 2023 to capture the effects of NV, occupant–window interaction, and thermal mass on indoor air temperatures. Model parameters including the construction details, glazing transmittance, thermal mass, infiltration rate, and internal gains, were adjusted step-by-step based on measured indoor temperature data. Calibration relied on visual comparison of time-series data and statistical benchmarks, following ASHRAE Guideline 14 and the established literature [27,48].

To investigate the outdoor and indoor air temperatures and evaluate the indoor thermal comfort of two classrooms during non-heating seasons, excluding July and August, data collection for Classroom East-GF and Classroom South-FF commenced in April 2023 and continued until June 2023 using a calibrated LSI-LASTEM (Milano, Italy) air temperature sensor with an accuracy of ±0.1 °C @0 °C logged at 1 min intervals, which were sited at 1.1 m. The field data measurements (Figure 2) were collected over 22 days, including 14 days in unoccupied conditions (seven days with scheduled window interaction and seven days with closed windows) and eight days in occupied conditions, surveying occupant–window interaction. A detailed thermal model (Figure 2) of two classrooms in the studied primary school building and its operation, physical dimensions, layout, and construction details was built using IESVE [44,47]. The model was simulated based on the EPW weather file, generated using Big Ladder (“Elements”, Version 1.0.6) [49] software and the synoptic Met Éireann Cork airport weather station data [43].

The model was calibrated incrementally across nine adjustment scenarios (AD1–AD9), with each step refining a specific set of parameters to progressively improve predictive accuracy.

Table 1. The list of ADs made to the initial model.

ID	Name	Procedure
AD1	AD-ASB	As-built scheduled openings + airport weather data
AD2	AD-EPW	AD1+ local weather data in April
AD3	AD-CON	AD2+ construction’s thickness and U-value
AD4	AD-GTR	AD3+ glazing transmittance
AD5	AD-MAS	AD4+ thermal mass
AD6	AD-INF	AD5+ infiltration rate (final model from April)
AD7	AD-APR	AD6+ local weather data in June
AD8	AD-OPI	AD7+ openings interaction
AD9	AD-ING	AD8+ internal gains (final model from June)

summarises the adjustments applied. AD1 introduced the as-built opening schedules and airport weather data under unoccupied conditions, while AD2 replaced this with local April weather data to better represent site-specific conditions. AD3 incorporated detailed construction properties, including material thicknesses and U-values, followed by the addition of glazing transmittance in AD4 to improve the representation of solar heat gains. AD5 refined the thermal mass to capture the building’s heat-storage behaviour and AD6 adjusted the infiltration rate, completing the unoccupied April calibration. For the occupied June period, AD7 updated the model with local June weather data, AD8 integrated surveyed occupant–window interaction patterns, and AD9 added internal gains from occupants and equipment. This structured progression allowed the influence of each parameter group to be isolated and minimised discrepancies between measured and simulated indoor temperatures. Calibration performance was evaluated using established criteria from the literature [27].

The potential impact of a mismatch in weather conditions between the building site and the weather station location type (urban/rural) is considerable. Previous studies [50,51] have recommended adjusting the weather file to mitigate any resulting discrepancies in building performance simulations. By aligning the building simulations with local climate conditions, these adjustments ensure a more accurate representation of the environmental factors influencing the simulation findings. The weather data, including the outdoor temperature, relative humidity, global radiation, wind speed, and direction, were collected from an LSI-LASTEM [52] weather station located in the yard at a height of 8 m, as shown in Figure 2. The correlation between local and airport weather station data during the days studied is presented in Figure 4. The correlation between airport and local air temperature was evaluated, with a high R² value of 0.95, wind speed (R² = 0.81), and wind direction (R² = 0.68).

The thermal mass of the construction materials was defined based on Part L-2019 and for furniture, the furniture mass factor was calculated based on the existing items of furniture in the classrooms, as shown in Figure 2. The thermal characteristics of the building materials, such as the heat capacity and thermal conductivity, were attained from [53,54]. The calculation results show that the furniture mass factor is around 6; therefore, once drawers, chairs, books, and boards are considered, a furniture mass factor of 10 was considered appropriate. During the studied time range, both classrooms had 21 people, contributing maximum sensible and latent gains of 57 and 43, respectively, with occupancy times from 9 a.m. to 2:30 p.m. during occupied days. Each classroom contained one computer, adding a maximum sensible gain of 100 W and a power consumption of 135 W. The opening and closing of openings was scheduled manually during the April date range (DR1). The occupants’ opening interactions were gathered through a survey among the teachers (DR3, DR5).

This study employed the calibration criteria benchmarks in the literature [27,55,56], as shown in

Table 2. The calibration metrics results of the indoor air temperature in the final model.

Criteria	Benchmark	Ref.	AD6-Unoccupied		AD9-Occupied
			Classroom 7	Classroom 11	Classroom 7	Classroom 11
RMSE (°C)	<1.5	[57]	0.7	0.7	0.9	0.9
CV(RMSE) (%)	<±20	[57]	4.7	5.0	4.2	4.1
NMBE (%)	<±5	[58]	4.3	3.6	3.5	1.3
Pearson’s	+0.5 < r <+1	[27]	0.5	0.7	0.4	0.6
GoF	<+5	[55]	3.4	3.0	3.0	2.9

, to compare simulation results with measured parameters. The Root Mean Square Error (RMSE) value, as shown in Figure 5, quantifying the overall differences between the simulated and measured data, was less than 1.5 °C. This RMSE signifies a substantial enhancement in the model’s predictive capabilities. Furthermore, other criteria, such as the Goodness of Fit (GoF), Coefficient of Variation of the Root Mean Square Error (CVRMSE), and Normalised Mean Bias Error (NMBE), were used to assess the relative error, systematic bias, and association between the simulated and measured indoor air temperature. The results, as shown in Table 2, indicated that the CV(RMSE) for both classrooms is less than 20%, the NMBE is less than 5%, Pearson’s correlation is between 0.4 and 0.7, and the GoF is less than 5%, which shows good agreement between the calibrated model and the measured data.

According to previous studies [27,59], changes of 2% in the CV(RMSE) hourly are considered significant in calibration. This study adjusts the local weather data (AD2), construction thickness, and U-value. (AD3) and accurate infiltration rate were influential when predicting indoor air temperatures. Other studies [27,48] used the same method used in this study based on the ASHRAE Guide14 indices of the RMSE, CV(RMSE), and NMBE to calibrate educational buildings. The literature showed that the main challenge for calibration studies is the presence of occupants [60] and the NV rate [61]. This study also observed the significance of adjusting the occupancy schedule and infiltration rate and the calibration process in unoccupied conditions was more accurate than in occupied conditions.

The predicted indoor air temperature was compared with the corresponding measured indoor and outdoor air temperatures during eight occupied June days as line graph in Figure 6. This comparison represents an acceptable calibration, with good agreement (RMSE < 1.5 °C) between the simulated and measured data. Figure 6 shows that simulated indoor temperatures slightly overestimate measured values, particularly during peak afternoon hours. The average difference between monitored and simulated temperatures when comparing occupancy and non-occupancy periods was a mean difference increase to about 0.9 °C during school hours and a decrease to 0.7 °C over unoccupied hours. The slightly larger difference observed during occupied hours suggests that occupants’ window-opening behaviour, use of blinds, and change in activity level may have increased real ventilation rates and internal gain beyond those assumed in the model, leading to marginally lower measured indoor temperatures. Despite this, the model demonstrates strong predictive reliability for overheating risk assessment.

2.3. Future Building Modelling and Simulation

The dynamic calibrated model was created to replicate actual buildings, providing a robust foundation for accurately assessing overheating risks within the studied school building. The methodology adopted for this analysis integrated a building model, utilising the IESVE software (Version 2023) [47], with future climate projections from Met Éireann’s DSY weather files. The methodology for the future building simulation model and an analysis of the results is shown in Figure 7.

The internal gains specifications and opening schedules during heating season are based on typical ventilation control strategies in previous studies [41,62]. Based on the Rules for National Schools under the Department of Education and the Irish school calendar, the classrooms were occupied between 09:00 and 14:30 [63]. A heating system set point of 20 °C was used during the occupied days, 08:00 a.m. to 01:00 p.m., from 6 November to 31 March. Natural ventilation was modelled using IESVE’s MacroFlo module, which calculates airflow based on external wind speed, direction, and indoor–outdoor temperature differentials. No constant airflow rates were assumed. Instead, dynamic weather conditions were used to simulate realistic ventilation performance. Ventilation rates varied across scenarios and locations, and were influenced by window opening area, orientation, and local climate inputs.

This study conducted a comparative analysis involving three distinct future weather files (DSY1, DSY2, and DSY3) across three specific years (2021, 2041, and 2071) at six selected locations in Ireland (Belmullet, Cork, Birr, Clones, Dublin, and Limerick) [64] sourced from Met Éireann, a reputable Irish meteorological agency

Table 3. Maximum air temperature over future weather files for 6 locations in Ireland.

DSYs	Year	Belmullet	Cork	Birr	Clones	Dublin	Limerick
		Max Ta (°C)	Max Ta (°C)	Max Ta (°C)	Max Ta (°C)	Max Ta (°C)	Max Ta (°C)
DSY1	2021	27.9	27	29.9	27.4	27.2	31.2
	2041	28.8	28.2	30.6	28.4	27.6	32
	2071	30.7	30.3	32.7	30.5	29.7	33.9
DSY2	2021	28.8	29.9	31.8	31.5	28.7	31.5
	2041	29.4	30.6	32.4	32.2	29.5	32.2
	2071	30.9	32.2	34	34.2	31.4	34.4
DSY3	2021	28.7	29.2	31.1	29.7	29	31.4
	2041	29.3	30	31.8	30.4	29.7	32.1
	2071	30.9	29.7	34	32.4	31.9	34

clearly reveals an upward trend in maximum temperatures over the years across all locations. Additionally, the assessment of seasonal patterns indicates that the highest temperatures tend to occur in either June, July, August, or September between 12:00 and 21:00 across all years and locations.

The previous study [65] underscores the necessity of employing climate files that forecast future conditions in building performance simulations. It emphasises the impact of the chosen projection method for generating future climatic files on thermal comfort and energy demand analysis. In this study, assessing future temperature variations across both years and locations facilitates a robust understanding of the future climate. This study focused on three time periods, including “near-term” (2021–2050), “mid-century” (2041–2070), and “end-century” (2071–2100) conditions.

This study selected various locations in Ireland, including coastal and inland cities with diverse climates, as shown in Figure 8. By conducting research in these locations, the findings contribute to a comprehensive understanding of how buildings perform in different environments in Ireland, as detailed in the references [64,66,67]. The comparison of weather files in all six locations showed that the air temperature in Belmullet is similar to that of Cork, Birr is similar to Limerick, and Clones is similar to Dublin. As a result, this study has selected three main cities in Ireland, namely, Dublin, Cork, and Limerick, for further investigation. Although Ireland has a relatively uniform temperate maritime climate, subtle regional and urban differences can influence building performance. All three cities, Cork, Dublin, and Limerick, vary in terms of urban density and microclimatic conditions. These distinctions can affect the effectiveness of natural ventilation and the risk of overheating.

This study assesses the vulnerability of classrooms, identifying and addressing risks based on the overheating criteria in the CIBSE TM52 [68] and BB101 [69] standards. There are three criteria to assess when overheating is likely to happen during non-heating season, which covers the period 1 May to 30 September: (1) total hours of exceedance, (2) daily weighted exceedance, and (3) upper limit temperature. Based on the adaptive thermal comfort descriptions used in the new BB101 [69], specifically target schools, and CIBSE TM52 [68], as shown in Equation (1), Criterion 1 sets a limit on the number of hours the operative temperature (T_op) can exceed the maximum comfort temperature (T_max); the indoor comfort temperatures are influenced by the outdoor running mean temperature. In this study Top was obtained directly from IESVE/ApacheSim outputs (hourly-0.6 m) [47].

CIBSE TM52 sets a percentage-based compliance criterion (3% of occupied hours), which allows for some flexibility in managing overheating, while BB101 sets an absolute limit (40 h), providing a stricter threshold. Criterion 2 sets a daily limit for acceptability, this amount should not be greater than six degree-hours. Criterion 3 is expressed as T_o − T_max = 4 °C, which the temperature of a building cannot exceed. Unlike BB101, where only Criterion 1 is mandatory, TM52 requires failure in at least two criteria to classify a space as overheated. This approach provides some flexibility by allowing short periods of exceedance, as long as they do not significantly impact occupant comfort.

$$T_c{>0.33\times (T}_{rm})+18.8 \pm T_{lim}$$

(1)

where $T_c$ is comfort temperature, $T_{rm}$ is running mean temperature and $T_{lim}$ is the category range limit of comfort. Categories of comfort are typically determined with varying category range limits for those with an average level of expectation (i.e., +3–4, Category II) and exceptional cases or those with a high level of expectation (±2, Category I) recommended for young children [41,70].

Based on the literature review [71,72], the resilience assessment method in this study consists of three metrics, called Indoor Overheating Degree (IOD), Ambient Warmness Degree (AWD_18°C), and Overheating Escalation Factor (α_IOD/AWD) [24]. IOD, as shown in Equation (2), quantifies the severity and frequency of indoor overheating risk.

$$IOD={\displaystyle \frac{\sum _{z=1}^z{\sum }_{i=1}^{N_{{occ }^{\left(z\right)}}}\left[({T_{fr,i,z}{-T}_{op,i,comf,z})}^{+}\times t_{i,z}\right]}{\sum _{z=1}^z{\sum }_{i=1}^{N_{{occ }^{\left(z\right)}}}{t}_{i,z}}}$$

(2)

where t is the time step, i is the occupied hour counter, Z is the total building zones, N_occ is the total number of occupied hours, T_fr,i,z is the free-running indoor operative temperature in zone z at time step i [°C], and T_fop,i,comf,z is the comfort temperature in zone z at time step i [°C]. Only positive values of (T_fr,i,z − T_fop,i,comf,z) + are considered. AWD, as shown in Equation (3), is used to assess the severity of outdoor air temperature compared to a reference temperature T_b. The reference temperature in this study was set at 18 °C.

$${AWD}_{18 °\mathrm{C}}={\displaystyle \frac{{\sum }_{i=1}^N\left[{(T_{a,i}-T_b)}^{+}\times t_i\right]}{\sum _{i=1}^N{t}_i}}$$

(3)

where T_a,i is the outdoor dry-bulb air temperature at time step i [°C], T_b is base temperature set at 18 °C, and N is the number of occupied hours. Overheating Escalator Factor (α_IOD/AWD), as shown in Equation (4), is the slope of the regression line between IOD and AWD. It shows the resistance of the building toward global warming.

$${\alpha }_{IOD/AWD}={\displaystyle \frac{IOD}{{AWD}_{18 °\mathrm{C}}}}$$

(4)

An α_IOD/AWD greater than the unit (α_IOD/AWD > 1) means that the building is not resilient to overheating and indoor thermal conditions get worse when compared to outdoor thermal stress. On the other hand, an α_IOD/AWD lower than the unit (α_IOD/AWD < 1) means that the building is resilient to overheating and can resist some outdoor thermal stress. Values between 0 and 1 represent partial resilience, where the building still moderates indoor temperatures but with reduced effectiveness under extreme outdoor conditions.

3. Ventilation Strategies

This study outlines a set of ventilation strategies to enhance the resilience of the studied NV low-energy school building against overheating. According to previous studies [24,28,73], these strategies sought to evaluate the thermal resilience of the building based on four criteria: vulnerability, resistance, robustness, and recoverability. A previous review [28] indicated that the majority of studies focused on two resilience factors: vulnerability (the likelihood of facing overheating risks) and resistance (the ability to employ passive strategies to enhance thermal comfort and mitigate overheating risks). Although the term “resilience” is frequently used in scholarly articles, specific resilience criteria are rarely specified. Moreover, only a single study considered robustness (backup plans in case the ventilation control system fails) and none evaluated recoverability (strategies for restoring system functionality and enhancing the adaptability of systems and occupants to overheating conditions).

This study used a multi-criteria approach for evaluating resilience involving three criteria.

Table 4. Control strategies of window opening during occupied hours.

No	NV Strategy	Season	Time/Date	Condition	Air Flow (ACH)				Variables	Function
					Min	Mean	Max	Std
1	NoV (Base)	Oct to Apr	09:00–14:30	0% POF	0	0.01	0.03	0.02	-	All windows/Closed
		May, June, Sep	09:00–14:30	0% POF	0	0.01	0.02	0.01
2	LoV	Oct to Apr	09:00–14:30	1.5% POF (1/2 Lower Windows) If To > 12 °C and If To < Ta	0	0.04	0.04	0.1	Air Temperature	Ramp (Ta,21,0,24,1)
		May, June, Sep	09:00–14:30	1.5% POF (1/2 Lower Windows)	0	0.3	0.9	0.2	Time	09:00: 1 14:30: 0
3	StdV	Oct to Apr	09:00–14:30	4.8% POF (1/2 All Windows) If To > 12 °C and If To < Ta	0	0.08	0.05	0.3	Air Temperature	Ramp (Ta,21,0,24,1)
		May, June, Sep	09:00–14:30	4.8% POF (1/2 All Windows)	0	1.5	4	0.9	Time	09:00: 1 14:30: 0
4	MaxV	Oct to Apr	09:00–14:30	9.6% POF (All Windows)	0	0.1	0.05	0.4	Air Temperature	Ramp (Ta,21,0,24,1)
		May, June, Sep	09:00–14:30	9.6% POF (All Windows)	0	3	7	1.8	Time	09:00: 1 14:30: 0

shows the control strategies for four NV strategies that were evaluated in the school building studied: no ventilation (NoV), low ventilation (LoV), standard ventilation (StdV) and maximum ventilation (MaxV). The four control strategies are designed based on the POF [34] in May, June, and September.

The selected POF values (1.5%, 4.8%, and 9.6%) were derived directly from the geometry and operability of the studied classroom windows. The maximum value of 9.6% represents the actual POF ratio when all six operable windows are fully opened, reflecting the upper limit of achievable NV in the as-built design. The intermediate value of 4.8% corresponds to a half-opening scenario, representing typical or realistic user behaviour where windows are not fully opened due to comfort, noise, or safety considerations. The lowest value of 1.5% reflects a restricted-opening condition, capturing situations where only a small portion of the available openings are used, which is common in classrooms with limited occupant interaction or unfavourable weather conditions. These three POF levels, therefore, represent a realistic and physically grounded range of ventilation performance scenarios for the studied building.

This analysis was performed across three locations (Dublin, Cork, and Limerick), under three DSYs and three time periods: 2021, 2041, and 2071. The designed ventilation strategies apply to during the day as a previous study shows daytime ventilation is significantly affected when the outdoor temperature is lower than indoors and remains within the comfort range [74]. The simulations use an occupied time of 9:00 to 14:30. The window operation from October to April is automated based on the comfort temperature during occupied times. In this study, the NoV strategy, as presented in Table 4, is the base model to evaluate the effect of building physics without any ventilation on indoor air temperature.

Table 4 presents the minimum, mean, maximum, and standard deviation of the air change rate (ACH), which was computed from IESVE MacroFlo’s dynamic airflow outputs over each operating period respective to each NV control strategy. Previous studies show the effectiveness of NV depends on wind, pressure forces, and occupant window-opening patterns, which fluctuate with weather and architectural characteristics [75,76]. The maximum air change rate for the MaxV strategy (7 ACH) aligns with the upper range reported by Ferrari et al. (2023) [77], who documented values up to 12 ACH in naturally ventilated classrooms under fully open window conditions.

4. Results

4.1. Overheating Vulnerability Assessment

The calibrated model presented in Section 2.2 has yielded interesting results regarding the simulation of NoV, LoV, StdV, and MaxV strategies under future weather files (DSY1, DSY2, and DSY3) across three specific time periods (2021, 2041, and 2071).

Table 5. Overheating vulnerability results in the case study classrooms during May, June, and September, utilising four different ventilation strategies over future weather files (DSY2) (extreme national overheating hours) (2021, 2041, 2071) for three locations (Cork, Dublin, Limerick) in Ireland.

Number of Occupied Hours from May to September When Value Exceeds the Acceptable Threshold
Ventilation Strategies	Years	Cork								Dublin								Limerick
		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101
	Criteria (CRI)	CRI 1		CRI 2		CRI 3		CRI 1		CRI 1		CRI 2		CRI 3		CRI 1		CRI 1		CRI 2		CRI 3		CRI 1
	Threshold	(Top > Tcomf) ≤ 3% Occ HRS (~11)		(Top > Tcomf) ≤ 6 h on any Day		(Top − Tcomf) ≤ 4		(Top > Tcomf) ≤ 40 h		(Top > Tcomf) ≤ 3% Occ HRS (~11)		(Top > Tcomf) ≤ 6 h on any Day		(Top − Tcomf) ≤ 4		(Top > Tcomf) ≤ 40 h		(Top > Tcomf) ≤ 3% Occ HRS (~11)		(Top > Tcomf) ≤ 6 h on any Day		(Top − Tcomf) ≤ 4		(Top > Tcomf) ≤ 40 h
	Class	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF	East-GF	South-FF
NoV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	14	1	0	0	0	0	14	1
	2041	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	22	6	2	0	0	0	22	6
	2071	85	15	2	0	0	0	85	15	0	0	0	0	0	0	0	0	102	15	3	0	0	0	102	15
LoV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	9	0	0	0	0	0	9	0
	2041	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	20	6	0	0	0	0	20	6
	2071	32	2	0	0	0	0	32	2	0	0	0	0	0	0	0	0	53	10	1	0	0	0	53	10
StdV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
	2041	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	10	5	0	0	0	0	10	5
	2071	5	0	0	0	0	0	5	0	0	0	0	0	0	0	0	0	21	9	1	0	0	0	21	9
MaxV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
	2041	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	5	5	0	0	0	0	5	5
	2071	1	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	17	9	0	0	0	0	17	9

Red: Fail, Green: Pass.

shows overheating occurrences in Cork (south coastal), Dublin (east coastal), and Limerick (west coastal and more inland) under future DSY2 during May, June, and September. An identical version of the table is presented in the Appendix A as

Table A1. Overheating vulnerability results in the low-energy school during May, June, and September, utilising three different ventilation strategies over future weather files (DSY1) (2021, 2041, 2071) for three locations (Cork, Dublin, Limerick) in Ireland.

Number of Occupied Hours from May to September When Value Exceeds the Acceptable Threshold
Ventilation Strategies	Years	Cork								Dublin								Limerick
		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101		CIBSE TM52		CIBSE TM52 BB101		CIBSE TM52 BB101		BB101
	Criteria (CRI)	CRI 1		CRI 2		CRI 3		CRI 1		CRI 1		CRI 2		CRI 3		CRI 1		CRI 1		CRI 2		CRI 3		CRI 1
	Threshold	(To > Tcomf) ≤ 3% Occ HRS (~11)		(To > Tcomf) ≤ 6 h on any Day		(To − Tcomf) ≤ 4		(To > Tcomf) ≤ 40 h		(To > Tcomf) ≤ 3% Occ HRS (~11)		(To > Tcomf) ≤ 6 h on any Day		(To − Tcomf) ≤ 4		(To > Tcomf) ≤ 40 h		(To > Tcomf) ≤ 3% Occ HRS (~11)		(To > Tcomf) ≤ 6 h on any Day		(To − Tcomf) ≤ 4		(To > Tcomf) ≤ 40 h
	Class	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)	C7 (GF)	C11 (FF)
NoV	2021	0	0	0	0	0	0	0	0	7	2	0	0	0	0	7	2	6	1	0	0	0	0	6	1
	2041	6	10	0	0	0	0	6	10	13	0	0	0	0	0	13	0	0	0	0	0	0	0	0	0
	2071	54	51	4	5	0	0	54	51	18	2	1	0	0	0	18	2	5	0	0	0	0	0	5	0
LoV	2021	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	2	5	1	0	0	0	0	5	1
	2041	0	7	0	0	0	0	0	7	10	0	0	0	0	0	10	0	0	0	0	0	0	0	0	0
	2071	42	37	4	5	0	0	42	37	10	0	1	0	0	0	10	0	2	0	0	0	0	0	2	0
StdV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	3	1	0	0	0	0	3	1
	2041	0	1	0	0	0	0	0	1	7	0	0	0	0	0	7	0	0	0	0	0	0	0	0	0
	2071	23	23	0	4	0	0	23	23	4	0	0	0	0	0	4	0	2	1	0	0	0	0	2	1
MaxV	2021	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	1	1
	2041	0	1	0	0	0	0	0	1	1	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0
	2071	10	20	0	2	0	0	10	20	3	0	0	0	0	0	3	0	2	1	0	0	0	0	2	1

Red: Fail, Green: Pass.

under DSY1 and Table 2 under DSY3. A comparison of results across the future weather files shows that extreme overheating risk occurrences are observed in Limerick under the DSY2 weather file, even with MaxV. In Dublin under NoV and LoV and in Cork under all NV strategies, these overheating occurrences arise under DSY1, whereas DSY3 shows fewer exceedances, with limited failures mainly occurring in Dublin.

Under short- (2021) and medium-term (2041) weather files, the case study classrooms seem to perform well in Cork (south coastal) and Dublin (east coastal) without NV during May, June, and September. Although in Limerick (west coastal and more inland), there are some overheating occurrences under the CIBSE TM52 and BB101 criteria, with overheating hours falling within the acceptable range as indicated by the green formatted results in Table 5. However, the base model results under the long-term (2071) weather files show that, without NV, the classrooms experience overheating hours that exceed the acceptable threshold, which is denoted by a red colour in Table 5 and could represent a potential risk.

Interestingly, the results also show that when the studied low-energy school uses a LoV strategy during occupied times in the non-heating period, which is equal to 1.5% POF in this study, the classrooms could still pose an overheating risk. The maximum overheating hours with the LoV strategy were observed during DSY2 (2071) in Cork (south coastal) and Limerick (west coastal and more inland). The results in Table 5 show that increasing the level of ventilation with the StdV strategy (4.8% POF) and MaxV strategy (9.6% POF), can effectively reduce the overheating hours to acceptable thresholds in Cork (south coastal). However, these strategies alone are unable to cope with future overheating predictions in Limerick (west coastal and more inland). Natural NV alone can achieve a reduction in the number of hours where the operative temperature exceeds the CIBSE TM52 thermal comfort threshold.

The differences in overheating risk across the three cities are evident: while Cork (south coastal) and Limerick (west coastal and more inland) show notable overheating under DSY2 (2071), Dublin (east coastal) does not exhibit overheating risk in the same scenarios. This may be due to subtle climatic variations, as urban characteristics were not explicitly considered in this study. These findings suggest that even within a small geographic region, microclimatic factors can significantly influence passive cooling performance.

Figure 9 compares the projected indoor and outdoor air temperatures (primary y-axis) of two classrooms (East-GF and South-FF) under the different NV strategies along with their window opening patterns (secondary y-axis) and air change rate per hour (ACH) (third y-axis) as line graphs facing the extreme week (left) and extreme day (right) in Dublin during May, June, and September. The lighter coloured lines belong to Classroom East-GF, the darker coloured lines refer to Classroom South-FF, and the dashed lines to specific window opening patterns for each NV strategy.

Figure 9 shows that, in Dublin, the outdoor temperature peaks in unheated academic months occur in the last week of June around 26 °C from 14:30 to 16:30. The modelled low-energy school without ventilation experienced a peak indoor air temperature around 29 °C from 10:30 to 12:30 in Classroom East-GF and around 27.5 °C from 13:00 to 15:00 in Classroom South-FF. The outdoor temperature fluctuates between 15 °C and 29 °C, with peaks in the afternoon; increasing the level of NV during the occupied time could decrease the indoor air temperature by up to 2.5 °C over the day and night. Classrooms with NoV as the base model show the highest indoor temperatures, while MaxV provides the most effective cooling, especially during the night. The MaxV strategy could decrease the peak indoor air temperature to 26.5 °C in Classroom East-GF and around 26 °C in Classroom South-FF. The higher indoor air temperature in Classroom East-GF can also be attributed to sunlight entering through the east-facing windows.

The StdV and LoV strategies offer a moderate reduction in indoor temperatures, but they are insufficient to mitigate the impact of peak temperature periods. On the extreme day (right graph), window opening schedules (dashed lines) are deliberately programmed with a pre-cooling strategy, set to occur during cooler night and morning hours to mitigate overheating. These results emphasise the need for optimised ventilation timing and enhanced airflow strategies to maintain indoor comfort under future extreme weather conditions. Although airflow rates were not fixed, the simulations reflect passive ventilation performance under realistic weather conditions, rather than mechanical ventilation without air treatment. An identical version of this figure is presented in the Appendix A as Figure A1 for Limerick and Figure 4 for Cork. The comparison of figures shows that peak outdoor temperatures during unheated academic months in Limerick occur in the last week of June, reaching around 31 °C. Under these conditions, the MaxV strategy maintains indoor air temperatures within the adaptive comfort range from 09:00 to 12:00; however, between 12:00 and 16:00, indoor temperatures exceed the adaptive comfort threshold, with higher values observed in East-FF. In Cork, the peak outdoor temperature during unheated academic months occurs in the first week of September, reaching approximately 29 °C. Here, MaxV keeps indoor air temperatures within the adaptive comfort range until 12:00, but, after this period, temperatures exceed the threshold, up to 15:00 in East-GF and up to 20:00 in South-FF.

4.2. Resilience Assessment

This section aims to assess indoor thermal comfort and overheating risk under various ventilation strategies (NoV, LoV, StdV, and MaxV), by utilising the overheating resilience assessment metrics outlined in Section 2.3. The primary objective is to understand how different strategies perform in mitigating overheating and maintaining thermal comfort as the climate changes.

The IOD metric captures the severity of indoor overheating by measuring the extent to which temperatures surpass adaptive comfort thresholds, while AWD provides the severity of outdoor thermal conditions. The α_IOD/AWD analyses the regression lines between the IOD and AWD equations, showing the sensitivity of a building to outdoor thermal stress. α_IOD/AWD > 1 means that the building is unable to suppress outdoor thermal stress. α_IOD/AWD < 1 means that the building can suppress some of the outdoor thermal stress. These metrics offer a comprehensive framework for comparing and interpreting the resilience of ventilation strategies against future overheating risk. This assessment also aims to encourage the reader to focus on the variations across strategies, as well as the distinct outcomes associated with each future weather file, to appreciate the broader implications for resilience building design [78,79].

The analysis in Figure 10 demonstrates the resilience of the studied NV low-energy schools in Cork, Dublin, and Limerick with ventilation systems incorporating LoV, StdV, and MaxV strategies. This indicates that the studied school can withstand the conditions captured in the modelled future weather files used, with an overheating escalation factor of less than one.

Figure 10, based on extreme future weather files for three locations (DSY1 2071 in Cork, DSY1 2071 in Dublin, and DSY2 2071 in Limerick), highlights the differences in overheating risk between Classroom East-GF and Classroom South-FF. In Cork, Classroom South-FF consistently shows a higher indoor overheating severity (IOD), primarily due to its elevated position and greater solar exposure. This disparity becomes more pronounced with increased natural ventilation, particularly under the MaxV strategy. In contrast, Classroom East-GF benefits from morning sun and reduced afternoon gains, resulting in lower thermal stress. In Dublin and Limerick, however, Classroom East-GF shows higher IOD, especially in Limerick, though this difference diminishes with increased ventilation. These findings demonstrate that orientation and floor level influence the effectiveness of natural ventilation strategies and their interaction with local climate conditions must be considered in passive design strategies. This research examines strategies representing different levels of low-cost behavioural interaction with existing ventilation systems, aiming to evaluate their resilience under varying climate conditions. Figure 10 highlights a higher frequency and severity of indoor (IOD) and outdoor (AWD) overheating risks in Limerick compared to Cork and Dublin, underlining the necessity for further analysis of future overheating risks, particularly in Dublin. As the capital of the Republic of Ireland with a significant population, Dublin presents unique challenges, particularly the UHI effect, which is currently not accounted for in the Met Éireann weather file used in this study.

4.3. Cognition and Learner Performance

This section highlights the impact of higher indoor temperature, resulting from climate change effects, on childrens’ cognitive performance. A previous study [80] presented Equation (5) to assess the Cognitive Performance Loss (CPL) of students in school due to the indoor air temperature.

$${RP}_t={0.2269t}^2-13.441t+277.84$$

(5)

where t represents the indoor temperature and RP_t represents relative performance at a specific temperature.

CPL serves to assess the negative impact of the warming climate on cognitive performance and assess the difference between the reference performance level (100%) and the cognitive performance level (RPt) which is calculated by Equation (6). As the CPL function was developed in a different educational and climatic context, its application to Irish classrooms may introduce minor uncertainties related to age groups, acclimatisation, and occupant behaviour.

$${CPL}_t=100\%-{RP}_t$$

(6)

Figure 11 presented cognition results as boxplots to show the distribution of hourly CPL for children in the case study classrooms using four suggested NV strategies under extreme future DSY weather files in Dublin (DSY1) for short-(2021), mid-(2041) and long-term (2071) scenarios. The comparison of the boxplots shows that higher ventilation rates result in lower CPL values. The NoV strategy as a base model of this study consistently shows the highest CPL, around 20% in 2021 and 22% in 2071, indicating that a higher indoor temperature due to lack of NV has a strong negative impact on cognitive performance. Generally, the boxplots show the extent of CPL in Classroom East-GF is greater than Classroom South-FF. The LoV strategy shows a slight improvement but remains relatively ineffective in reducing CPL. The StdV strategy provides noticeable benefits, leading to a more stable and lower CPL range to less than 15% in 2021 and 2041. The MaxV strategy emerges as the most effective strategy, maintaining the lowest CPL values, particularly in 2021 and 2041, that is almost less than 10%, though even in 2071, when the maximum value of CPL is generally higher, this strategy still provides the best conditions. The trend suggests that by utilising the regulatory amount of ventilation opening area in the NV case study classrooms, there is the potential to reduce by half the CPL experienced when no ventilation is present.

To assess the cooling requirements necessary to maintain indoor temperatures that support optimal cognitive performance, peak cooling loads were calculated using the IESVE model [47], based on the admittance method described in the CIBSE Guide A-Appendix 5.A10 [81]. The fixed cooling set point for this analysis is 21 °C, reflecting the temperature threshold associated with 100% cognitive performance. The cooling load results for Classroom East-GF and South-FF in Dublin (DSY1 2071) during May, June, and September are presented in

Table 6. Cooling load result by means of admittance method for Classroom East-GF and South-FF during May, June, and September, in Dublin (DSY12071).

Room Name	Air Temperature (°C)		Peak Space Sensible (kW)	Airflow (L/s)	Engineering Checks
	Supply	Return			W/m2	(L/s·m2)
East-GF	21.72	32.83	2.69	199	36.86	2.71
South-FF	17.13	28.24	2.15	159	29.64	2.18

. These values represent the total mechanical cooling demand, as NV was not included in the calculation. Classroom East-GF experiences higher peak cooling demand (2.69 kW) and greater return air temperatures (32.83 °C), likely due to increased morning solar exposure and the absence of NV. In contrast, Classroom South-FF has a lower cooling load (2.15 kW) and return temperature (28.24 °C), benefiting from different solar gain patterns during occupied hours. Classroom East-GF also requires higher airflow (199 L/s vs. 159 L/s) and has a greater cooling load per unit area (36.86 W/m² vs. 29.64 W/m²). These findings highlight the importance of integrated shading and ventilation strategies to ensure thermally and cognitively supportive learning environments under future climate conditions.

4.4. Schools as Community Climate Shelters

The community climate shelter is an innovative initiative resulting from climate change policies, aimed at reducing urban heat through strategies such as integrating shaded areas, increasing greenery, and improving ventilation in adapted school buildings. The Climate Shelters project, led by Barcelona, aims to address overheating due to the UHI effect and prepare the city for increasingly high summer temperatures by transforming 11 pilot schools into climate shelters. This project highlights the relevance of schools as adaptive spaces due to their widespread distribution and integral role within communities to protect citizens, particularly those in vulnerable communities, from climate change. The study emphasises engagement with schools and public awareness to inspire global efforts towards sustainable building, adaptive cities and addressing the complex challenges of climate change [82].

López Plazas et al. [8] focused on comprehensive environmental interventions to enhance ventilation within climate shelter school buildings. This study emphasised the importance of both passive and active ventilation strategies to create a conducive indoor climate during extreme weather conditions. The proposed technical solutions included enhancing existing NV by optimising window designs and adding new openings to facilitate better natural cross ventilation [83]. The study also suggested the implementation of advanced mechanical devices such as energy-efficient fans, air extractors, and modern HVAC systems to supplement NV. These mechanical systems are particularly crucial in densely populated urban areas where natural airflow might be insufficient. Furthermore, the study highlighted the significance of integrating green and blue measures, such as green walls and water features, to improve the overall microclimate and air quality within and around the school buildings.

This study evaluated the resilience of the low-energy case study classrooms using the optimum NV strategy (MaxV) and identified their potential to be used as climate shelters during the summer. The findings, as shown in Figure 12 (left), highlighted non-negligible overheating risk hours (954 to 967 h) in Dublin (DSY1 2071) during summer. A HybV strategy was adopted to transform the studied NV low-energy school to act as a more effective climate shelter against overheating.

Table 7. Control strategies of natural and HybV for case study classrooms during Summer (July and August).

No	NV Strategies	Season	Time/Date	Condition	Air Flow (ACH)				Variables	Function
					Min	Mean	Max	Std
5	MaxV + 24 h	July and August	24 h	Optimum POF (max)	0	2.3	7	1.9	Time	00:00: 1 24:00: 1
6	HybV	July and August	24 h	If To > 15 °C and If To < Ta and If Ta > 18	0	3.2	10	2.7	Air Temperature	gt * (Ta,23,3)

* Fuzzy greater than (gt) function [39]: gt(x, xs, b)—a proportional band rising linearly from 0 to 1 as x increases from xs − b/2 to xs + b/2.

presents control strategies of natural and HybV use for the studied school during July and August, with representative minimum, mean, maximum, and standard deviations of ACH for each scenario. Using schools as climate shelters would require different occupancy patterns and considerations of security, accessibility, and operational logistics, which lie beyond the scope of this study. In this study, the maximum ACH under HybV scenarios was 10 ACH, which aligns with the boundary of a previous study [84], where 10 ACH yielded 28% availability in Taipei and 13% in Kaohsiung under typical classroom heat gains.

HybV systems are designed to recover and develop the building’s ability to return to the pre-risk state under unforeseen conditions to deliver the best possible performance while maintaining low energy consumption [85,86,87,88]. The proposed HybV system is designed to be energy-efficient, with a low specific fan power of 0.158 W/L/s to boost ventilation levels during summer.

Figure 12 compares the number of safe and risk hours using the max (optimum) NV (left) and HybV strategy (right). The filled boxes report the number of hours the case study classrooms need a mechanical system to be a safe climate shelter. The result shows the use of a HybV system in the case study classrooms can reduce 42–48% of the overheating risk during summer rather than optimum NV strategy.

Figure 13 presents the indoor and outdoor air temperatures (primary y-axis) under the optimum NV strategy (MaxV) and HybV for 24 h along with their opening equivalent area (Eqv) (secondary y-axis) and ACH (third y-axis) as line graphs during the extreme week (left) and extreme day (right) for Dublin (DSY1 2071) during July and August. The left graph shows outdoor temperatures peaking above 28 °C during the week, while HybV consistently maintained lower indoor temperatures than MaxV, especially in Classroom South-FF. Window opening profiles (dashed lines) indicate night ventilation contributed to pre-cooling of the spaces. The right graph focuses on 26 August, when outdoor temperatures reach nearly 30 °C; during this period, HybV reduced indoor peaks to about 28 °C in South-FF, whereas MaxV achieved only minor reductions (~0.5 °C). These results illustrate that although the adaptive comfort model technically allows indoor temperatures approaching 28 °C, such conditions are borderline for comfort and could present risks for vulnerable occupants, especially children. An identical version of this figure is presented in the Appendix A as Figure A2 for Limerick and Figure 5 for Cork. In Limerick, outdoor temperatures in the last week of August peak at around 25 °C, while the MaxV strategy maintains maximum indoor temperatures at approximately 26 °C. The HybV strategy further reduces indoor peaks to about 25 °C and lowers indoor temperatures by up to 2 °C throughout the day. In Cork, outdoor temperatures peak at 28 °C; under these conditions, MaxV keeps indoor temperatures close to 28 °C, with higher values observed in East-GF. The HybV strategy again reduces indoor temperatures by up to 2 °C compared with MaxV.

However, the duration of these peaks was limited, approximately four hours (15:00–19:00), and aligned with outdoor maximum temperature. Classroom South-FF performs better than Classroom East-GF, likely due to orientation and airflow differences. The analysis underscores the effectiveness of HybV in lowering peak temperatures and enhancing thermal resilience under extreme conditions, while also highlighting the potential need for complementary passive measures (e.g., external shading, increased thermal mass, or limited mechanical cooling) to ensure acceptable comfort during future summers. Based on hourly analysis, the HybV system’s fan energy consumption ranges from 0.7 to 1.1 kW, while its total electricity usage varies between 0.9 and 1.4 kW under the extreme weather conditions of Dublin (DSY1-2071) during July and August.

5. Discussion

To address existing gaps in the literature, this study developed a calibrated model with Class I [27] accuracy, offering a reliable method to evaluate the resilience of NV in the school building studied [27]. The findings indicate that studied NV low-energy school buildings in certain locations across Ireland could resist overheating under future weather conditions. The resilience escalation factor, as presented in Annex 80, confirmed that classrooms in Cork and Dublin demonstrate a high level of robustness against future overheating risk. However, the classrooms in Limerick, characterised by an Atlantic climate with more frequent humidity and variable outdoor conditions, are more pronounced to overheating. The vulnerability assessment highlighted the potential challenges climate change poses, emphasising the need for robust and context-specific ventilation strategies in building design. There was a consistent upward trend in maximum temperatures, highlighting the warming trend projected. Limerick is projected to experience the most significant future overheating risk under DSY2 scenarios.

The efficacy of various ventilation strategies, NoV, LoV, StdV, and MaxV was evaluated cross a range of future climate scenarios. While instances of overheating persist under extreme conditions (2071), particularly in Limerick under DSY2, the results clearly show that increasing the level of ventilation plays a critical role in mitigating thermal discomfort. The StdV (4.8% POF) and MaxV (9.6% POF) strategies demonstrate a consistent reduction in overheating hours in Cork, but still fails in Limerick, indicating that NV alone may not be adequate in all future scenarios. These findings are consistent with previous research, which found that increased NV rates can significantly reduce overheating hours [24], while they may be unable to cope with overheating risk [89].

This study applied CPL value, derived from the existing literature [80], to a real-world context, offering novel insights into how ventilation strategies affect cognitive performance in the case study classrooms. The findings link CPL outcomes to the building’s features, such as orientation, floor level, and ventilation effectiveness; the study reveals how internal layouts and opening interactions impact students’ cognitive performance. The results highlight the role of NV strategies in mitigating CPL among students; higher ventilation rates correlate with lower CPL values, which aligns with previous studies’ results [16,80,90]. The NoV strategy, as the base model of this study, consistently exhibited the highest CPL, reaching up to 23% in 2071. In contrast, the MaxV strategy proved to be the most effective, maintaining CPL levels below 10% in 2021 and 2041, with a notable increase in effectiveness compared to the LoV and StdV strategies. The MaxV strategy, which consistently yielded the lowest CPL values across all time periods, underscores the importance of optimising occupant-driven interventions, while the limited impact of the LoV strategy reflects the challenges of relying on minimal airflow under extreme conditions.

Furthermore, the distinction between Classroom East-GF and Classroom South-FF underscores the impact of orientation, ventilation performance, and solar exposure on both thermal comfort, similar to previous observations [91], and cognitive performance. Classroom East-GF consistently experienced higher peak indoor temperatures and more overheated hours than Classroom South-FF (west-facing, first floor), largely due to greater morning solar exposure and a lower elevation. These spatial differences directly impacted cooling demand, with Classroom East-GF requiring a peak cooling load of 2.69 kW and a return air temperature of 32.83 °C, compared to 2.15 kW and 28.24 °C in Classroom South-FF. CPL values were correspondingly higher in Classroom East-GF by up to 3–5% across the strategies. Similar observations were made by previous research [92], of an inverse relationship between ventilation rate and temperature, particularly in upper floors; more airflow led to lower temperatures (up to 3 °C cooler), highlighting the need for climate-responsive and floor-specific ventilation strategies. Another study [93] concludes building orientation and shape strongly influence NV performance, especially in dense urban environments, and should be carefully considered in the design stage.

The findings underscore that, while the MaxV + 24 h strategy reduces overheating during the summer and enhances the studied school’s potential as a climate shelter, a substantial number of risk hours remain when evaluated against a static threshold, such as 23 °C based on the HSE’s recommendation as a healthy indoor temperature for vulnerable people. However, results were different when using adaptive thermal comfort thresholds. The HybV system in this study aligned with previous research findings [88] and significantly reduced overheating risk, by 42–48% compared to NV, while maintaining low fan-power energy consumption. Further analysis indicates that HybV more effectively stabilises indoor temperatures, particularly during peak heat events, with Classroom South-FF performing better than Classroom East-GF due to differences in orientation and ventilation performance. Additionally, NpV was found to be crucial in pre-cooling indoor spaces, contributing to lower indoor temperatures during daytime peak hours. These findings align with previous studies [94,95] advocating for integrated, adaptive strategies and reinforcing that a combination of natural, hybrid, and passive interventions is needed to implement schools as climate shelters.

This study also highlights two key challenges, as shown in

Table 8. Control strategies of NpV and AutV for the studied school building during May, June, and September.

No	NV Strategies	Season	Time/Date	Condition	Air Flow (ACH)				Variables	Function
					Min	Mean	Max	Std
7	NpV	Oct to Apr	09:00–14:30	If To > 15 °C and If To < Ta	0	0.2	0.06	0.5	Air Temperature	Ramp (Ta,21,0,24,1)
		May, June, Sep	22:00–05:00	Upper Windows Open	0	1.3	4.6	2.5	Time	22:00: 1 05:00: 0
8	AutV	Oct to Apr	09:00–14:30	If To > 15 °C and If To < Ta	0	0.2	0.06	0.5	Air Temperature	Ramp (Ta,21,0,24,1)
		May, June, Sep	09:00–14:30 22:00–05:00	If To > 15 °C and If To < Ta	0	0.5	2	0.7	Air Temperature	Ramp (Ta,23,0,26,1)

, in achieving resilient NV cooling: the effectiveness of NpV in addressing scenarios where daytime NV is limited by external factors such as noise or air pollution and the role of automated control of window operations to enhance NV accuracy and eliminate any need for interaction from teachers, which is aligned by previous studies [35,96]. Figure 14 indicates that NpV plays a crucial role in pre-cooling classrooms in Dublin. Night-time operation helps reduce indoor temperatures, aligning with previous studies [94,95,97], allowing for more stable thermal conditions during peak daytime hours, even when outdoor temperatures exceed 31 °C in Limerick (Figure A3 in Appendix A) and exceed 28 °C in Cork (Figure A6 in Appendix A). Additionally, incorporating automated window controls as modelled in the HybV strategy proved beneficial for maintaining indoor temperatures within acceptable ranges. These findings emphasise the importance of integrating intelligent automation [98] and NpV with NV strategies to overcome practical limitations posed by urban environments and user variability.

Although this study evaluates the impact of natural and HybV strategies on the studied school building’s resilience and ability to prevent cognitive performance in students, it does not account for the broader implications of UHI effects and population growth in Dublin’s future climate scenarios [67,99]. Given that Dublin’s population is projected to reach 2.5 million by 2050, with increasing urbanisation intensifying heat stress, future research should explore the combined effects of UHI and climate change on overheating risks in schools [11,100,101,102]. Currently, there are no weather files available that integrate these combined effects, highlighting the need for further development in this area. Furthermore, the results of this study are specific to the studied low-energy school building.

These challenges present opportunities for future studies to incorporate regional climate modelling and urban microclimate analysis to assess how urban heat exposure influences ventilation performance and indoor thermal comfort under projected climate conditions. Exploring integrated mitigation strategies, such as green infrastructure, shading devices, and passive cooling techniques, alongside HybV systems, could enhance indoor thermal comfort while reducing reliance on mechanical cooling. More precise automated monitoring systems should be explored in future research to improve data accuracy. Additionally, the vulnerability and resilience assessment of the studied school building was based on current standard assumptions, which may not fully reflect future climate uncertainties or changing urban microclimates. By addressing these gaps, future studies can provide data-driven recommendations for sustainable school design and resilient climate shelter strategies.

6. Conclusions, Limitations, and Future Work

This study conducted a thorough analysis of the overheating vulnerability of the studied naturally ventilated low-energy school building across three different main cities in Ireland. The research mainly focused on instances where indoor temperatures exceeded the comfort level and provided tailored mitigation strategies to reduce overheating risks. These strategies included low, standard, and maximum levels of natural ventilation. The study aimed to assess thermal resilience to overheating using existing overheating criteria and resilience assessment indices [78,79] for the most extreme future Design Summer Years climate scenarios in each location at short-, mid-, and long-term time periods. The findings demonstrated that increasing ventilation levels significantly reduces overheating risks, particularly in Cork and Dublin, though challenges remain in Limerick under long-term projections.

The maximum ventilation strategy was identified as the optimum natural ventilated strategy during unheated academic months and when applied 24/7, it showed strong potential for supporting climate shelter functions during summer. The findings support hybrid ventilation systems that are designed to recover and develop the buildings’ ability to return to the pre-risk state and to deliver the best possible performance while maintaining low energy consumption, particularly when passive strategies like night purging and automated ventilation are included. This study also examined the impact of overheating on students’ cognitive performance. Results showed that higher indoor temperatures correlated with higher cognitive performance loss values, with the no ventilation strategy resulting in the worst performance, reaching a cognitive performance loss of up to 23% in 2071. In contrast, the maximum ventilation strategy proved most effective, keeping cognitive performance loss below 10% in 2021 and 2041. These outcomes highlight the critical role of adequate ventilation not just in reducing overheating, but in supporting cognitive health and learning outcomes in school environments. The classrooms’ location and orientation differences further emphasised the influence of wind direction and floor level, with east-facing, ground-floor classrooms showing more overheated hours, cognitive performance loss, and cooling demands compared to upper-floor spaces.

The results highlight the urban heat island phenomenon in Dublin, which poses a significant risk of overheating for vulnerable populations. According to the findings, it is recommended that schools take measures during summer to become climate shelters for vulnerable people. The study advocates using passive cooling strategies, such as maximising building thermal mass and employing shading devices, to enhance the schools’ effectiveness as efficient climate shelters while using hybrid ventilation systems.

This study acknowledges several limitations that future research should address:

• The model was calibrated for both occupied and unoccupied conditions to evaluate the impact of physical construction and occupant behaviour separately, but may not fully capture the range of occupant behaviours during the non-heating season.
• Occupancy patterns and window opening schedules were derived from teacher surveys, introducing potential inaccuracies due to manual reporting.
• A mismatch between wind speed and direction data from the local and Cork weather stations led to an overestimation of air change rates.
• Ventilation effectiveness in natural ventilation scenarios depends heavily on weather conditions. While dynamic modelling was used, airflow rates were not directly measured and real-time variability may affect performance. This represents a key limitation of the study.
• The resilience assessment focused on three criteria, vulnerability, resistance, and robustness, excluding the fourth criterion of recoverability after thermal failure.
• The study does not explicitly model urban heat island effects or microclimatic variations, which may influence overheating risk, particularly in more urbanised areas like Dublin. Future work should consider these factors to refine the accuracy of climate-based performance assessments.
• Future research should incorporate automated monitoring, airflow quantification, expanded building typologies, and refined climate projections to improve reliability and applicability.

Overall, the contributions of this study provide valuable insights for architects, engineers, and policymakers to enhance the resilience of low-energy schools and improve the cognitive performance of students in future climate conditions. The results underscore the significance of accurate weather data, calibration processes, and hybrid ventilation strategies for ensuring occupant comfort and energy efficiency in the face of climate change. As we move towards a future overheating risk, this research contributes essential knowledge for sustainable and climate-resilient building design and operation. Based on the findings, several recommendations are proposed to enhance the resilience of educational buildings against overheating conditions:

• The accuracy of simulations strongly relies on weather data inputs. It is recommended that local weather data be prioritised over airport weather data for building performance simulations. This will ensure a more accurate representation of the microclimates and urban heat islands directly influencing school buildings.
• Occupancy modelling in building simulations often involves simplifications due to the complexity of human behaviour. Future work could benefit from more sophisticated occupancy models that capture occupant activities, preferences, and interactions with the built environment for more accurate predictions. This may include the integration of machine learning approaches for more accurate predictions.
• The findings of this study are based on specific geographical and climatic conditions. The proposed recommendations in this paper may not universally apply to all regions. Future research should explore the development of guidelines and recommendations to ensure the resilience of low-energy school designs across diverse climates and future projections.