Advancing Energy Efficiency in Educational Buildings: A Case Study on Sustainable Retrofitting and Management Strategies

Abstract

Public educational buildings, particularly schools, are often overlooked in energy efficiency initiatives, despite their potential for substantial energy and cost savings. This study presents an integrative, measurement-informed, calibrated model-based approach for assessing and enhancing energy performance in elementary schools located in Israel’s hot-arid climate. By combining multiscale environmental monitoring with a rigorously calibrated Energy Plus simulation model, the study evaluates the impact of three demand-side management (DSM) strategies: night ventilation, external envelope insulation, and a combination of the two. Quantitative results show that night ventilation reduced average indoor temperatures by up to 3.3 °C during peak occupancy hours and led to daily energy savings of 10–15%, equating to approximately 1500–2200 kWh annually per classroom. Envelope insulation further reduced diurnal temperature fluctuations from 7.75 °C to 1.0 °C and achieved an additional 9% energy savings. When combined, the two strategies yielded up to 20% energy savings and improved thermal comfort. The findings provide a transferable framework for evaluating retrofitting options in public buildings, offering actionable insights for policymakers and facility managers aiming to implement scalable, cost-effective energy interventions in educational environments.

1. Introduction

School buildings account for a significant share of public-sector energy consumption and are crucial to ensuring healthy indoor environments for children. In hot-arid and Mediterranean climates, classrooms face the dual challenge of maintaining acceptable thermal comfort and indoor air quality (IAQ) while limiting cooling demand. Rising temperatures and growing concern over students’ cognitive performance under poor IAQ conditions have intensified the need for reliable, cost-effective retrofit strategies tailored to this building typology.

Recent studies confirm that school buildings in the Mediterranean and Middle East exhibit high energy intensities and suboptimal IEQ. Campagna and Fiorito [1] analyzed the Apulia Region school stock, highlighting the sensitivity of Mediterranean schools to envelop performance. Belpoliti et al. [2] mapped UAE public schools, showing sector heterogeneity and high baseline consumption. In Saudi Arabia, Aloshan and Aldali [3] and Alfaoyzan and Almasri [4] documented facade retrofits and benchmarking approaches, while Zamani and Amiri [5] introduced renewable energy integration for “smart schools.” In Israel, Schwartz et al. [6] developed a data-driven stock model integrating energy and IEQ factors, but practical retrofit pathways at the classroom scale remain underexplored. The European PEDIA project [7] and Congedo et al. [8] further confirm the relevance of resilient, ventilation-based cooling strategies in Mediterranean schools, yet highlight feasibility constraints.

Recent works emphasize the effectiveness of low-cost envelope and ventilation measures. Randelovic et al. [9] evaluated insulation and window replacement via cost–benefit analysis, while Maiques et al. [10] quantified the energy and IAQ implications of natural versus mechanical ventilation strategies in Mediterranean schools. Pollozhani et al. [11] compared ventilation strategies across health, environmental, and energy criteria, and Aloshan and Aldali [3] showed how targeted envelope retrofits substantially reduced cooling demand. Llanos-Jimenez et al. [12] introduced a Retrofit Potential Index to prioritize Mediterranean schools, while Sanchez et al. [13] developed cost-optimal deep retrofit strategies for public high schools. Together, these studies demonstrate the scalability and replicability of relatively simple interventions, consistent with the envelope insulation and ventilation measures investigated in this paper.

Building performance depends strongly on user interactions. Kharvari and Rostami-Moez [14] documented adaptive behavior impacts in educational spaces, while Li et al. [15] applied the Theory of Planned Behavior to student energy practices in China. Feng et al. [16] showed how occupants adapt post-retrofit, underscoring the persistence of performance gaps. Song et al. [17] conducted a systematic review of occupant behavior impacts on energy efficiency, confirming the centrality of accurate occupant modeling. These works extend the broader review by Zhang et al. [18] and demonstrate that occupant behavior must be explicitly considered in retrofit assessment.

CO₂ buildup is a persistent problem in classrooms. Rey-Hernandez et al. [19] assessed natural ventilation to mitigate infection risks, while Maiques et al. [10] evaluated natural ventilation and mechanical heat-Kohl recovery strategies, respectively, in Mediterranean classrooms. Kohl et al. [20] introduced the “comfort performance gap” concept in new educational buildings, highlighting deviations between simulated and real comfort outcomes. These benchmarks are crucial for validating the monitoring results presented here.

While broader links between sustainability and pedagogy are recognized (Onyeaka and Akinsemolu [21]; Craig and Allen [22]), the practical focus remains on quantitative performance. Abbas et al. [23] assessed resilient cooling strategies in Mediterranean offices under climate change, offering transferable insights for school retrofits. Such works highlight the importance of contextualized resilience alongside traditional efficiency measures.

This study aims to answer the following research question: How can data-driven DSM strategies improve energy performance in public elementary schools operating in hot-arid climates? To address this, authors: (1) develop a multiscale environmental monitoring system, (2) calibrate and validate an energy model, (3) assess the effects of night ventilation and envelope insulation, and (4) evaluate the replicability of these strategies across different climatic contexts. The hypothesis of this research is that combining these DSM measures will yield significant improvements in thermal comfort and energy savings.

The novelty of this study lies in integrating multi-scale empirical monitoring with calibrated simulation to evaluate climate-specific DSM strategies, providing a combined framework for both scientific validation and practical application.

2. Methodology

This study aims to assess the energy performance of elementary school buildings in Israel and to develop measurement-informed, calibrated model-based strategies for improving energy efficiency through multiscale monitoring and numerical modeling. The approach integrates empirical measurements, simulation calibration, and Demand-Side Management (DSM) strategy evaluation.

The methodological process flowchart in Figure 1 illustrates the sequential integration of data collection from test case sites, multiscale monitoring, numerical model calibration, and DSM strategy evaluation in educational buildings, clarifying the rationale for each step and showing how the combined process yields validated simulation outcomes for both thermal comfort and energy savings.

2.1. Case Study Selection

The selected schools represent a cross-section of building typologies in Israel, including one constructed in 1965 and another in 2020. The two test cases were selected because they represent typical construction archetypes in Israel, an older legacy building requiring retrofits and a newer facility built under modern standards.

Figure 2 depicts the location of each school. Both are located in southern Israel with approx. 40 km separates them.

Table 1. Test case parameters.

School ID	Location	Construction Year	HVAC Type	Number of Students	Walls/Roof Insulation U-Value (W/m2·K)	On-Site RES
A	Eilat	1965	Centralized Chiller Fan coils in the classes	300	1 Wall 1.7–2.0 Roof 1.5–1.8	Solar Generation Battery Storage
B	Eilot	2020	Split systems, inverter A/C in each class	450	2 Wall 0.8–1.0 Roof 0.4–0.6	Solar Generation Battery Storage

¹ No mandatory thermal insulation standard existed at the time. Typical construction included single-layer concrete or block walls without insulation. Roof: Often reinforced concrete slab with bitumen, minimal or no insulation. ² Complies with SI 1045 [24], SI 5282 [25] for building envelope energy performance. Post-2014 regulations require minimum thermal resistance for educational buildings. According to the SI 1045, Israel is mapped to 4 zones of insulation requirements. The Eilot regional council and Eilat city refer to zone D, where thermal insulation is mandatory.

summarizes their HVAC types, insulation levels, and renewable energy systems. This variation allows the study to address energy performance across different architectural eras and system configurations. A representative classroom in each school was selected based on typical occupancy, geometric orientation, and envelope characteristics to serve as the focal point for microscale analysis.

2.2. Pilot Classroom Characteristics and Monitoring Setup

The selected pilot classroom represents the typical architectural and operational conditions of classrooms in the investigated schools. It reflects common occupancy patterns, window-to-wall ratios, equipment usage, and HVAC configuration, making it a suitable reference for broader applicability.

Measuring approximately 60 square meters, the classroom accommodates around 25 students and one teacher during school hours. It is equipped with standard educational technology, including a smartboard, and has an independent HVAC unit set at a cooling target of 22–23 °C. The cooling target was chosen to align with the lower bound of the comfort zones defined by ASHRAE 55 [26] and EN 16798-1:2019 [27]. Preliminary monitoring revealed distinct energy use patterns, with peaks at the beginning of the school day and intermittent spikes related to equipment use, breaks, and thermal adjustments.

Focusing on a single classroom enables high-resolution analysis of energy dynamics in a controlled yet representative setting. This approach supports the development of scalable strategies that may inform energy efficiency improvements across similar educational environments.

Figure 3 describes the pilot classroom configuration. Equipment includes standard educational technology and an HVAC unit, which is representative of typical classrooms in study schools. The windows are marked on Figure 3 with red lines.

The data collection segment of the methodology aims to garner a comprehensive set of information that will enable a nuanced analysis of energy performance. Several types of data will be collected to fulfill the objectives of this study. The data regarding electricity use by lighting, HVAC, and other electronic equipment will be captured to provide a complete picture of energy utilization within the classroom and the broader school.

The selected classroom is located on the first floor of the southern wing of School A and has a full-height south-facing window that spans approximately 4.2 m in width and 1.5 m in height, representing roughly 30% of the façade area (see Figure 4). The glazing is single-pane clear glass with no thermal coatings. Fabric curtains are used for limited internal shading. External shading is limited, the window is partially shaded during morning hours by a nearby tree located approximately 3 m from the facade, but remains fully exposed to solar radiation during peak afternoon hours. The absence of overhangs or external louvers results in significant direct solar gain during mid-day periods, contributing to notable indoor temperature fluctuations. These characteristics were factored into the simulation setup and are representative of common construction practices in regional public schools built prior to the implementation of national energy standards.

2.3. Multiscale Monitoring Approach

Monitoring equipment placement was carefully selected based on preliminary site surveys and consultations with school maintenance personnel. Sensors for indoor temperature, humidity, and CO₂ were positioned at standard occupant heights (approximately 1.5 m from the floor), away from direct airflow from HVAC units and windows, to capture representative indoor environmental conditions. The spatial variation in IAQ measurements across different locations within the classroom is analyzed and presented in the Results section. Outdoor meteorological stations were installed on unobstructed rooftop areas to ensure accurate recording of external conditions such as solar radiation, wind speed, and ambient temperature. Data were collected at consistent intervals (every 5 min), allowing comprehensive time-series analysis and facilitating accurate numerical model calibration.

A customized monitoring system was installed, combining third-party PV production and grid consumption meters with in-class smart plugs and air quality sensors. Outdoor environmental conditions were captured using roof-mounted meteorological stations. Surface temperatures of walls and windows were monitored using custom-adapted iButton devices to assess envelope performance and inform model calibration.

Monitoring during the period of 12 months was conducted at three scales:

• Macro (school-wide): capturing total energy use, HVAC performance, and lighting.
• Meso (classroom-level): evaluating localized loads and indoor environmental quality.
• Micro (device-level): measuring real-time consumption of individual appliances and indoor conditions such as temperature, humidity, CO₂, and PM2.5.

Sensor deployment and data logging, including a summary of all monitoring equipment, measured parameters, and uncertainty ranges, are detailed in clause 3.

Calibration and integration across scales enable a holistic understanding of system performance.

2.4. Numerical Modeling

Numerical modeling was conducted using Energy Plus (v22.2.0) to simulate classroom and whole school performance. Calibration against monitored energy data ensured accuracy (NMBE, CvRMSE within accepted thresholds). This approach enabled testing of retrofit and DSM scenarios under realistic operating conditions. This clear delineation between empirical and simulated datasets ensures the robustness of our results and provides confidence in the theoretical assessments of energy-saving interventions.

The numerical modeling component is a cornerstone for simulating and understanding the energy and environmental performance dynamics within the selected educational setting. Numerical modeling is required for predicting energy performance in addition to allowing systematic investigation of numerous hypothetical scenarios. The DSM analysis focused on strategies relevant to hot-arid schools: optimized HVAC schedules, lighting retrofits, night ventilation, and envelope insulation. The latter two management tools were selected as the primary DSM strategies because prior research highlights their high impact and low-cost feasibility. These were evaluated comparatively through calibrated simulations to assess energy savings and comfort impacts. Wall insulation and controlled night ventilation are prioritized because they are high-impact yet low-cost measures in hot arid schools, require minimal occupant behavior change, and can be implemented with limited disruption during school holidays, with documented payback periods typically within two to five years in comparable public sector retrofits [28,29]. Both measures have low design risk and limited maintenance burden compared with active system retrofits, which aligns with budget-constrained school districts and with procurement practices that favor simple envelopes and schedule-based controls [30].

Calibration ensures the accuracy and reliability of the numerical models and follows the process shown in Figure 1. First, prior to predictive runs, the model is calibrated to historical energy data from the selected school and the pilot classroom, aligning simulated and monitored performance. Parameters being adjusted include U-values, internal gains, and infiltration rates, and fit quality is reported with RMSE and CvRMSE, as presented in Section 5.1. The calibrated model is then used to evaluate DSM scenarios at classroom and whole school scales, focusing on night ventilation and envelope insulation, as depicted in Section 6. This simulation-based approach enables a comparative evaluation of interventions, supporting informed decision-making for future implementation. The findings from this tier contribute to the overall objective of developing scalable, evidence-based DSM recommendations for public educational buildings.

The monitoring, calibration, and retrofit simulations were structured to test the study’s hypothesis on comfort and energy savings.

3. Mathematical Formulation

The presented methodology is designed to offer a comprehensive assessment of energy performance in selected elementary schools in Israel, specifically focusing on numerical modeling for precise evaluation. The thorough approach addresses the building design, usage, and energy consumption patterns. Numerical modeling forms the cornerstone of this methodology, as it allows for a nuanced understanding of how different variables interact within the school environment to affect energy performance.

The model incorporates detailed information on construction materials, thermal properties, and configurations, starting with the building envelope. U-values, representing the thermal transmittance of building materials, are thoroughly modeled for structural elements like walls, roofs, and windows. The envelope build-ups are modeled layer by layer, considering each material’s thickness and thermal conductivity.

The envelope conductive heat flow is modeled with Fourier’s law [26]. Envelope properties, including U-values, solar transmittance parameters, and air infiltration rates, were derived from audits and from Israeli standards [24] and international guidance [31], then calibrated against measured indoor and outdoor temperatures to ensure realism of inputs. Solar gains are represented as $Q_{solargain}=A\cdot GHI\cdot (1-\mu )\cdot \kappa$, where $\kappa$ is an effective transmittance that depends on wall thermal properties and $\mu$ is surface reflectance [32]. Internal gains are modeled as the sum of equipment, lighting, and occupants following the school timetable, with the occupant term written as $Q_{internal}=N·Q_{person}$, based on standard metabolic rates.

Indoor air quality monitoring often involves using sensors to measure parameters like temperature, humidity, and CO₂ levels, estimating the heat emitted per person by correlating these parameters with occupancy levels. For example, increased CO₂ levels indicate higher occupancy and, consequently, higher heat emissions. This data is integrated into a building energy model to estimate the heat gain due to occupants.

In numerical modeling, the heat emitted per person is considered a constant value or variable depending on activities and occupancy schedules. Commercial software can simulate the heat gains from occupants and other heat sources like lighting and equipment. These models often use predefined values for metabolic rates associated with different activities (e.g., sitting, walking) to estimate the heat emitted per person.

Another approach is to use real-time occupancy data, possibly obtained through motion sensors or cameras, and feed this into the numerical model, allowing for a more dynamic estimation of internal heat gains.

It is also possible to combine both methods for a more accurate estimation. Real-time indoor air quality data can be used to validate and calibrate the numerical model, ensuring that the estimated heat gains closely match the actual conditions.

The following equations serve as a foundation for quantifying the benefits of DSM strategies, particularly night ventilation. By clearly relating temperature differentials (indoor vs. outdoor) and airflow rates, these mathematical relationships guide precise adjustments in ventilation scheduling and operational parameters, optimizing thermal comfort and energy consumption. The overall indoor heat to handle $Q_{total}$ can be represented as a sum of $Q_{transfer}$ and $Q_{internal}$.

Infiltration rates, which refer to the air leakage into and out of the building, need to be carefully handled since they significantly impact the indoor heat load and, thus, the energy consumption of the building. The model considers factors like wind speed, temperature differential, and building geometry to evaluate the infiltration rate that different impact factors can cause. Here, the authors address the impact of night ventilation, the occupant-controlled approach for utilizing the DSM strategy to initiate artificial infiltration.

The efficiency of night ventilation is modeled [33] as:

$$\eta ={\displaystyle \frac{T_{indoor}^{before}-T_{indoor}^{after}}{T_{indoor}^{before}-T_{outdoor}}}$$

(1)

where $T_{indoor}^{before}$ and $T_{indoor}^{after}$ are the indoor temperatures before and after the deployment of night ventilation and $T_{outdoor}$ refers to the average ambient outdoor temperature during the night ventilation period, i.e., the time window between the initiation and conclusion of night ventilation (from 19:00 to 06:00).

The suggested energy model also considers HVAC systems, schedules, and efficiency coefficients, allowing for an all-encompassing energy performance assessment that includes heating, cooling, and ventilation loads.

The cooling load reduction due to improved wall insulation and night ventilation [34] can be modeled as:

$${\Delta Q}_{cooling}=Q_{total}^{before}-\left[Q_{transfer}·{\displaystyle \raisebox{1ex}{$U_{before}$}\!\left/ \!\raisebox{-1ex}{$U_{after}$}\right.}+\eta Q_{internal}\right]$$

(2)

where ${\displaystyle \raisebox{1ex}{$U_{before}$}\!\left/ \!\raisebox{-1ex}{$U_{after}$}\right.}$ is the improvement factor for wall insulation, considered here as the selected approach for structure-related DSM. It is defined as the fractional reduction in conductive heat transfer through the building envelope resulting from improved wall insulation, based on the ratio of thermal transmittance before and after the retrofit, where $U_{before}$ is the original U-value of the wall before adding insulation, and $U_{after}$ is the improved U-value after insulation enhancement.

Quantifying the benefits of reducing energy demand through mathematical equations provides a more explicit understanding of the impact, such as:

$$E_{before}=Q_{total}^{before}t$$

(3)

where ${\mathrm{E}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}}$ is the energy consumption before optimization, ${\mathrm{Q}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}^{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}}$ is the initial total heat load, and $\mathrm{t}$ is the considered time duration. In this paper, the authors focus on relatively short-term daily improvement. This paper does not analyze annual optimization, which is also important for evaluating the seasonal impacts, given the minor seasonal effects in the selected case study buildings. A more robust formulation for scaling the suggested approach to other climate zones will be investigated in future work.

Energy Consumption After Optimization:

$$E_{after}=\left[Q_{total}^{before}-{\Delta Q}_{cooling}\right]t$$

(4)

Energy Savings:

$$E_{savings}=E_{before}-E_{after}$$

(5)

Figure 5 below illustrates the summary of the mathematical methodology.

Below is a summarizing

Table 2. Summary of Key Variables in the Modeling and Calibration Process.

Symbol	Parameter Description	Unit	Source	Role in Model
A	Surface area	m2	Building plans, field survey	Determines heat transfer area (walls, windows)
k	Thermal conductivity	W/m·°K	Material databases, the literature	Input to Fourier’s Law for conduction
ρ	Density of wall materials	kg/m3	Literature, material specs	Used in thermal mass and heat storage calculations
cp	Specific heat capacity	J/kg·°K	Literature, material specs	Affects the thermal inertia of walls
U	Thermal transmittance	W/m2·°K	Calibration, standards	Overall heat loss/gain through the envelope
G	Solar heat gain coefficient (glazing)	–	Window manufacturer data	Quantifies solar radiation admitted indoors
Qtransfer	Heat transfer through the envelope	W	Derived from model equations	A major component of building a heat balance
Qinternal	Internal gains from occupants/equipment	W	Monitoring, audits	Adds to the indoor thermal load
η	Ventilation effectiveness	–	Modeled, validated via temperature drop	Quantifies the impact of the night ventilation strategy
N	Number of occupants	persons	Occupancy schedules, audits	Used to compute metabolic/internal gains
t	Duration of operation or simulation step	h	Simulation setting	Time basis for cumulative energy calculations
E	Energy consumption	kWh	Monitoring, simulation output	Used for calibration and DSM scenario comparison
Tin, Tout	Indoor and outdoor temperatures	°C	Sensor network, weather station	Validation and boundary condition inputs
ACH	Air changes per hour	1/h	On-site estimation, calibration	Impacts infiltration and ventilation loads

of the key variables used in the modeling and calibration, including their symbols, physical meaning, units, data sources, and roles in the model.

4. Monitoring System

A multiscale monitoring system was developed and deployed to support numerical model calibration and evaluate energy efficiency interventions within the school environment. This section outlines the architecture, instrumentation, and data acquisition methods used across macro (school-wide), meso (classroom), and micro (device-level) scales.

4.1. Multiscale Energy Monitoring Architecture

The energy monitoring system integrates municipal and third-party hardware data with customized classroom instrumentation. Figure 6 illustrates the schematic diagram of the multiscale energy monitoring architecture, integrating macro-level grid and PV data with classroom-level electrical monitoring and individual appliance tracking.

At the macro scale, data on photovoltaic (PV) production and overall grid consumption are collected using devices installed by third-party contractors. These systems are connected via a flexible software platform capable of real-time data acquisition and modular expansion.

At the meso and micro scales, energy consumption within the pilot classroom is captured through:

• Clamp gauges at the classroom’s distribution board to monitor total load,
• Smart plugs installed at selected sockets to record real-time energy use of individual appliances.

Smart plugs are well-suited for low-power equipment (e.g., lights, computers). Still, they are limited in handling high inrush currents from HVAC systems, which are monitored via more robust electrical meters.

To ensure the robustness of this study, all monitoring equipment was deployed following a standardized protocol. Indoor environmental sensors were placed at occupant height (approximately 1.5 m above the floor), positioned to avoid localized effects from windows or HVAC vents. Outdoor conditions were monitored using rooftop stations in unobstructed locations. Data were logged at 5 min intervals, capturing diurnal variations in temperature, humidity, and air quality. These high-resolution measurements formed the empirical basis for evaluating baseline building performance and calibrating the numerical simulations.

4.2. Indoor Environmental Quality Monitoring

IAQ significantly impacts occupant comfort and energy demand. Classrooms, libraries, and laboratories were equipped with multiple sensors that measure:

• Temperature and relative humidity
• Carbon dioxide (CO₂)
• Particulate matter (PM2.5)
• Total volatile organic compounds (TVOC)

Figure 7 presents the four types of IAQ sensors used. Sensors were selected to enable a cross-validation procedure, presented in the Results section, and flexibility in deployment. Types (a) and (c) support multi-parameter measurement and were installed in the pilot classroom and laboratory. Types (b) and (d), which measure only temperature and humidity, differ mainly in power source and communication interface.

Table 3. Indoor monitoring devices.

Measuring Device	Measured Entities	Uncertainty	Manufacturer
a	Temperature, humidity, PM2.5, CO2, TVOC	Up to 3% or 1PPM	Tuya DM72B
b	Temperature, humidity	Up to 5%	Tuya TH
c	Temperature, humidity, PM2.5, TVOC, Radon	Up to 3% or 1PPM	Airthings View Plus
d	Temperature, humidity	Up to 1%	Lascar EL-WIFI-TH

below provides a more detailed information about monitoring devices, Type (b) devices are battery-powered with intermittent data upload, which is suitable for retrofit scenarios. Type (d) devices require external power and maintain continuous data communication.

4.3. Outdoor Weather Station

To evaluate environmental influences on building performance, two types of rooftop meteorological stations were deployed (Figure 8) to measure:

• Outdoor temperature and humidity
• Wind speed and direction
• Global Solar irradiance (GHI)
• UV index

Unlike municipal weather services, local data acquisition provides greater spatial and temporal precision (see

Table 4. IAQ measured parameters.

Measured Entities	Resolution/Uncertainty
Outdoor Temperature	0.1 °C/0.6 °C
Outdoor humidity	1%/1%
Solar irradiance	1 W/m2/5%
UV index	N/A
Wind speed and direction	1 m/s/1%

), which is crucial for validating envelope behavior and passive strategies (e.g., night ventilation).

Both stations include independent cellular data modems, avoiding dependency on the school’s internal communication infrastructure, a key consideration in older buildings.

To enable compatibility with Energy Plus simulation requirements, the measured global horizontal irradiance (GHI) obtained from the local weather station is decomposed into its two essential components: direct normal irradiance (DNI) and diffuse horizontal irradiance (DHI). This decomposition is necessary because Energy Plus requires beam and diffuse radiation inputs to simulate surface solar gains accurately. The Erbs model [35] is applied to estimate these components based on the clearness index $K_t$. To improve the accuracy of the Erbs decomposition, the clearance index from site-specific solar geometry is computed for each timestamp, and standard data quality control actions are applied to the GHI series, such as range checks, despiking, time alignment, etc. The clearness index is calculated as:

$$K_t={\displaystyle \raisebox{1ex}{$GHI$}\!\left/ \!\raisebox{-1ex}{$H_0$}\right.}\#\left(7\right)$$

(6)

where $H_0$ is the extraterrestrial horizontal irradiance, which represents the solar radiation incident on a horizontal surface at the top of the Earth’s atmosphere, for a given location, day, and timestamp. DHI and DNI are computed as

$$\begin{array}{c}DHI=D_f·GHI\\ DNI={\displaystyle \frac{GHI-DHI}{\mathrm{c}\mathrm{o}\mathrm{s}\left({\theta }_z\right)}}\end{array}$$

(7)

where $D_f$ is the diffuse fraction derived from empirical correlations proposed by Erbs, and ${\theta }_z$ is the solar zenith angle. This method allows accurate and physically consistent estimation of solar radiation components needed for dynamic thermal simulation.

4.4. Surface Temperature and Thermal Envelope Monitoring

Understanding thermal envelope performance is vital for energy modeling. Surface temperatures of internal and external walls were recorded using Thermochron iButton devices (Figure 9), which were retrofitted with a custom interface to enable continuous online monitoring.

Data were logged every 5 min and uploaded to a remote FTP server hourly. This configuration enabled high-resolution tracking of thermal dynamics about HVAC operation, insulation characteristics, and solar exposure.

The surface temperature was measured using a thermocouple sensor placed in direct contact with the wall surface. To minimize radiation effects, the exterior sensor was covered with a radiation shield, ensuring that the measurement reflected surface-air interaction rather than direct solar heating. The thermocouple was fixed to the wall using a highly conductive adhesive paste, which ensured stable positioning and reliable thermal contact with the substrate while avoiding the thermal bridging and material damage that can occur with mechanical fasteners. This approach is consistent with ISO 7726 [35] and ASHRAE Fundamentals Handbook [36] recommendations for surface temperature measurements.

Collected data support model calibration and validation, adjusting U-values, G-values, and internal gains to match observations, and RMSE to quantify accuracy. Monitoring over 12 months provides comprehensive insights for improving energy efficiency in schools.

However, the encountered challenges include occasional sensor calibration drift, temporary signal loss due to structural interference, and limited power availability in older classroom setups. These issues required periodic maintenance and recalibration procedures, emphasizing the necessity of robust operational practices for long-term reliability and accuracy.

Periodic maintenance, though necessary, introduces additional operational costs and logistical considerations. Future implementations should include budget allowances and personnel training to minimize downtime, which could affect the monitoring system’s long-term effectiveness and scalability.

5. Modeling and Calibration

One of the main challenges of this research was the simulation of an aged structure with limited documentation regarding construction methods and material properties. A multi-stage calibration strategy was adopted, combining building geometry data, monitored environmental conditions, and literature-based material properties to refine the model’s accuracy.

The first calibration stage focused on aligning indoor ambient temperatures in the test classroom with monitored data, considering ventilation habits and occupant schedules. Key adjustable parameters included air exchange rates and material properties, such as density, conductivity, and specific heat capacity of masonry blocks. Where values were unavailable (e.g., solar absorptance), parameters were treated as free variables during calibration.

Geometrical inputs (e.g., floor area, wall height, window orientation) were extracted from architectural drawings and field surveys, forming the basis for thermal envelope modeling. Material transmittance (U-values) and window G-values were included for conductive and solar heat gains.

To capture internal gains, occupancy profiles and equipment loads were estimated from audits and monitoring. At the same time, HVAC performance, such as nominal efficiency and control behavior, was based on technical documentation and refined through empirical comparison. Given the age of the systems, degradation effects were acknowledged, for example, studies have shown HVAC performance may decline by up to 40% annually without maintenance (e.g., Florida Solar Energy Center [37]).

Weather data (temperature, humidity, wind, solar radiation) were taken from local on-site meteorological stations, enhancing the reliability of passive strategy simulations such as night ventilation.

All parameter selections were based on their established role in the fundamental building physics equations governing energy performance, particularly heat transfer and mass flow. The parameters were sourced from building plans, on-site measurements, technical manuals, or validated defaults from standards and guidelines. This comprehensive approach ensures the calibrated model is a robust and credible tool for evaluating retrofit strategies and DSM measures.

Figure 10 presents the models of School A, such that (a) depicts the SketchUp model and (b) refers to the model of Energy Plus simulation. The red dotted line on Figure 10b marks the location of the pilot classroom.

5.1. Calibration Outcomes

Considering the abovementioned challenges, this research developed a methodology to calibrate the software results for an existing structure. The calibration process involved selecting the construction materials used in the building, which were obtained through discussions with the maintenance personnel on-site and assessments based on the construction practices prevalent during the building’s construction period.

The most crucial parameter for user comfort is the temperature of the indoor area where the residents spend most of their time. Therefore, the first parameter calibrated in this study was the indoor ambient temperature. The ambient temperature was tested and monitored continuously over a week to ensure accuracy and minimize deviations.

The temperature was calibrated using information on the room and wall temperatures. The external and internal wall sensors measured the temperature of the surfaces. Figure 11 below compares the daily graphic results of measurements with software results produced during the model calibration.

The calibration process, which involved comparing simulated and measured surface temperatures, has yielded promising results. The close match between the simulation and measurement data in terms of both the temperature trends and the low error percentage suggests that the material properties used in the simulation have been effectively calibrated.

Following the accepted standard [38], several error estimation methods were employed here to match the model parameters with the actual field measurements, such as Normalized Mean Bias Error (NMBE) and Coefficient of Variation of Root Mean Square Error (CvRMSE). NMBE assesses the global over- or under-compliance, compared with the CvRMSE score, which reflects how close the individual data points are.

$$NMB{E}_{IndoorAmbientTemp}={\displaystyle \frac{1}{N}}\sum _{i=1}^N{\displaystyle \frac{T_{meas,i}-T_{sim,i}}{T_{meas,i}}}$$

(8)

$$CvRMS{E}_{IndoorAmbientTemp}={\displaystyle \frac{1}{T_{meas,avg}}}\sqrt{\sum _{i=1}^N{\displaystyle \frac{1{\left(T_{meas,i}-T_{sim,i}\right)}^2}{N}}}$$

(9)

The graphs in Figure 12 below compare the simulated surface temperature with the measured surface temperature throughout the day.

In this manuscript, NMBE introduces a 5% mismatch, whereas CvRMSE presents only a minor error of 0.5%.

Figure 13 below illustrates the calibration procedure flowchart, as described above.

Surface Temperature Calibration is essential for reliable insights regarding the impact of the evaluated DSM measures, such as night ventilation or envelope insulation. Indoor and outdoor sensors were installed in several locations to provide a comprehensive calibration procedure. Following that, discrepancy evaluation is performed based on the CvRMSE error criterion:

The discrepancy values presented in

Table 5. CvRMSE for Surface Temperature.

Sensor Location	CvRMSE
Indoor Surface Temperature–Northern Wall	2.78%
Indoor Surface Temperature–Southern Wall	2.43%
Outdoor Surface Temperature–Southern Wall	4.4%

fit well inside the limitations of 15% discrepancy for monthly calibrations, specified by ASHRAE Guideline 14 ([38]).

An Energy Plus model incorporates construction materials’ U-values, HVAC schedules, and occupancy data. Calibration was conducted using historical data from the monitored schools.

Table 6. School-specific calibration values.

Parameter	Source	School A (Older Building)	School B (Newer Building)
U-value (walls)	Building codes	1.7–2.0 W/m2K	0.8–1.0 W/m2K
U-value (roof)	Building codes	1.5–1.8 W/m2K	0.4–0.6 W/m2K
Occupancy-normalized heat gain	Monitoring data + ASHRAE Handbook (2021) + EN 16798-1 (2019)	5.8 W/person	5.2 W/person
HVAC efficiency	Manufacturer manuals	78%	90%

summarizes key model parameters for each school:

School A, an older building, exhibited higher U-values, indicating poorer insulation, while the results for School B indicate better thermal performance. Occupancy heat gain varied slightly due to differing classroom densities, and HVAC efficiency was lower in School A due to outdated systems.

The sensible heat gain of 5.8 W/person was adopted for aged School A, while in Case B of modern school, the latent heat gain of 5.2 W/person was considered. These values represent occupancy-normalized ASHRAE handbook [36] values for classroom activity, assuming a typical classroom occupancy of 30 students and a teacher, operating for approximately 7–8 h per school day, 5 days per week, with no occupancy during weekends.

Calibration achieved NMBE of 4.7% and CvRMSE of 2.9% across both schools, ensuring a reliable model representation of energy performance in different building conditions.

5.2. Modeling Assumptions and Uncertainty Analysis

While the numerical modeling framework employed in this study is rigorously calibrated, several assumptions and potential uncertainties may influence the accuracy of the results. Identifying these factors enhances transparency and provides a clearer understanding of the model’s reliability.

Assumptions in the Energy Simulation Model

• The thermal properties of the school buildings were estimated based on historical construction data and field observations, as detailed material specifications were unavailable. Assumptions regarding insulation thickness, U-values, and thermal conductivity may introduce variability in heat transfer calculations.
• All simulations were performed at the classroom scale, representing the monitored room as a single Energy Plus thermal zone with internal partitions treated as adiabatic, external wall, roof, and windows exposed to weather and solar loads, and the floor modeled as ground contact. The model was calibrated to the measured indoor temperature, ensuring the single-zone representation reproduces the observed dynamics.
• Energy consumption models assume standardized occupancy schedules, as obtained during the initial audit, though real-world variations (e.g., after-school programs and maintenance activities) may lead to deviations.
• The model assumes that HVAC systems operate at consistent efficiency levels, but real-world performance can vary due to aging equipment, irregular maintenance, or user behavior.
• The infiltration rates and night ventilation efficiency are estimated based on measured wind speeds and external temperature differentials. However, occupant-driven variations in window usage may alter airflow patterns.

Sensitivity and Uncertainty Analysis

The following validation steps were conducted to quantify model sensitivity and uncertainty:

• Key parameters, such as wall U-values, HVAC efficiency, and infiltration rates, were varied by ±10% to assess their impact on predicted energy performance. Results indicate that HVAC efficiency variations contributed the highest uncertainty (±4% in total energy consumption).
• Calibration of the model using real-world energy consumption data yielded an NMBE of 4.7% and CvRMSE of 2.9%, indicating a high level of accuracy within industry standards.
• A comparison of simulated vs. measured indoor temperatures showed deviations of ≤0.5 °C, validating the model’s predictive capabilities.

Potential Impacts of Uncertainties on Results

• Underestimation of heat transfer losses could lead to overstated energy savings from insulation retrofits.
• Uncertainty in air exchange rates could impact the predicted effectiveness of night ventilation, particularly in varying wind conditions.
• Future work will incorporate probabilistic uncertainty modeling (e.g., Monte Carlo simulations) to refine predictions under varying climatic and operational conditions.
• By recognizing these potential uncertainties, this study aims to enhance the credibility of its findings while providing a transparent basis for future refinement of energy modeling in educational buildings.

While the calibration process significantly improved the model’s accuracy, residual uncertainties remain due to limitations in directly measured material properties, HVAC system degradation, and occupant behavior variability. Sensitivity analysis indicated that uncertainties in thermal transmittance (U-values), infiltration rates (ACH), and HVAC efficiency had the most significant influence on model outputs, with deviations in indoor temperature predictions remaining within ±0.5 °C and energy consumption estimates varying by up to 8%. These results confirm that, despite partial data limitations, the calibrated model provides a robust and reliable basis for scenario evaluation and DSM strategy comparison, especially when uncertainty margins are acknowledged in interpreting the results.

6. Results and Discussion

During the initial calibration phase, all IAQ sensors demonstrated good alignment, with discrepancies of less than 5% across devices. These variations are within the accuracy range specified by the manufacturers and confirm the reliability of the measured data.

As mentioned, achieving thermal comfort in space involves considering various building parameters and how the building is used. The current study focuses on influencing factors broadly categorized into two main types: the type of structural materials and user behavior.

Regarding structural materials, the possibility of adding insulating materials to the exterior of the building with a thickness of 5 cm is explored. This change aims to improve the building’s thermal performance by enhancing its insulation properties.

On the other hand, regarding user behavior, the option of implementing night ventilation is examined and analyzed in terms of its associated effects. Night ventilation involves opening windows during colder nights to facilitate natural cooling while keeping the windows closed during hotter daytime hours to minimize heat gain. This behavioral change can significantly impact the indoor temperature dynamics.

Importantly, the study does not treat these two parameters in isolation. A central focus of the research is to investigate how the combination and correlation between structural improvements and user-driven strategies interact to influence thermal comfort and energy efficiency. By exploring both independent and integrated effects, the study aims to uncover synergistic relationships that may offer greater benefits than applying each measure separately.

6.1. Night Ventilation

Night ventilation is pivotal in energy-efficient building design, particularly in reducing cooling loads. Night ventilation utilizes the cooler outdoor air during nighttime to decrease the indoor air temperature, reducing the need for mechanical cooling during the daytime. Factors like the ventilation rate, schedule, and external wind conditions influence the strategy’s efficiency.

The amount of heat removed by night ventilation is evaluated by the following equation [39]:

$${\mathrm{Q}}_{\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{d}}=\mathsf{\rho }\mathrm{V}{\mathrm{C}}_{\mathrm{p}}\mathsf{\Delta }\mathrm{T}$$

(10)

where $\mathsf{\rho }$ stands for air density, $\mathrm{V}$ is the volume flow rate of air, ${\mathrm{C}}_{\mathrm{p}}$ depicts the specific heat capacity of air and $\mathsf{\Delta }\mathrm{T}$ is the temperature difference between indoor and outdoor air.

A higher airflow rate $\mathrm{V}$ will lead to a more significant reduction in indoor temperature $\mathsf{\Delta }{\mathrm{T}}_{\mathrm{v}\mathrm{e}\mathrm{n}\mathrm{t}}$, thereby decreasing the cooling load the following day. However, a rate that is too high could compromise indoor air quality. A schedule synchronized with the coldest outdoor temperatures maximizes the benefits of night ventilation. An incorrectly timed schedule could draw in warmer air, negating the intended benefits. The efficiency of night ventilation depends on the specific climatic conditions of the building. Given the desert location of the case study buildings and a significant drop in night temperature, night ventilation introduces an excellent potential for reducing the cooling load and, thus, the energy demand. The local weather station measures real-time wind speed and direction, whereas temperature loggers inside the building provide data on internal temperature variations due to night ventilation.

The impact of variations in night ventilation rate and schedule, particularly in conjunction with local wind conditions, can be substantial and multifaceted. The interplay of these factors requires a holistic approach for accurate analysis, making the combined use of empirical measurements and calibrated numerical simulations essential. Through a well-structured methodology, it is possible to optimize night ventilation strategies to maximize energy efficiency.

The simulation of night ventilation strategies was based on established airflow rate assumptions commonly used in building energy performance studies. In this research, an air exchange rate of 3–5 ACH was applied, corresponding to an estimated airflow of 20–40 ${\mathrm{m}}^3/\mathrm{h}$ per occupant, depending on the classroom’s occupancy density and window opening areas. These values were selected based on prior experimental studies on passive ventilation effectiveness in similar climatic conditions. The selected rates ensure sufficient cooling without excessive infiltration, which could introduce additional latent cooling loads. Future studies could refine these assumptions by incorporating real-time airflow monitoring and occupant-controlled ventilation schedules to optimize energy savings.

The following graph illustrates the impact of night ventilation between 19:00 and 6:00 through a typical day’s comparison of the indoor temperatures obtained both with and without night ventilation (blue and green curves, respectively). For validation purposes, the measurement results are also presented (red line), showing a good agreement with the base simulation (without ventilation). An outdoor temperature (black curve) is added for reference.

Figure 14 presents an indoor temperature comparison on a typical summer day with the following test conditions: external temperature peak: 39 °C, occupancy: 26 persons, illustrating night ventilation effects. It can be observed that without night ventilation, the indoor temperature gradually rises throughout the day, reaching its peak in the afternoon, according to measurement results on the same day. However, when night ventilation is implemented, there is a noticeable improvement in maintaining a lower and more comfortable indoor temperature. The temperature drops significantly during the night, and with proper ventilation, the subsequent day starts with a much cooler indoor environment.

This comparison demonstrates the overall positive impact of night ventilation on maintaining thermal comfort by effectively reducing indoor temperature during hot periods.

In addition to the visual analysis of the night ventilation effect, an accurate statistical analysis is performed to quantify the impact of the suggested activity. The T-test statistical tool is utilized here to compare the means before and after applying the night ventilation. T-test analysis results introduce a superscript probability of insignificant impact, meaning that the result is obtained with almost absolute certainty.

Due to night ventilation, lower temperatures are observed throughout the day during the daytime period. With an increase in outside temperature, the indoor benefit decreases, dropping from 4 °C in the morning to 2 °C during noon hours.

In addition, the maximum fall occurs only after 13:00. At the same time, the training session ends before 13:30. Thus, during the most used hours, the class is at a lower temperature (by an average of 2.7 °C), i.e., about a 10% reduction. Such a reduction introduces a significant benefit in cooling load. Given the A/C system installed in the pilot classroom, an independent 6 kW HVAC operating for 8 h a day between 8:00 and 16:00, an expected reduction in energy consumption might exceed 15%. Note that such a significant effect depends strongly on climate conditions. In this case study, high temperatures during the day combined with desert cold during the night introduce a high potential for a major reduction in energy consumption following the night ventilation of indoor subspaces.

A delay can be observed in Figure 14 between the “no ventilation” and “with night ventilation” curves. The indoor temperature peak in the “night-ventilation” case reflects the time-lag introduced by thermally massive, opaque envelope components under daily excitation. This dynamic response, together with the associated decrement factor, is formally defined in ISO 13786 [40] and widely documented for summer conditions [41,42]. Consistent with this theory, once night ventilation is initiated, the building material approaches a periodic steady state over successive nights, similarly to the warm-up convergence procedure used in Energy Plus, which repeats initial days until zone temperatures and loads change negligibly between cycles. In the test runs, stabilization occurred within 2 nights under comparable weather. A parametric study of ACH is outside the scope of the present work because the focus here is natural wind-driven ventilation. Instead, authors applied a representative ACH derived from on-site wind conditions and consistent with classroom ranges reported in the literature for naturally ventilated spaces. With partial windows opening on both north and south facades and a classroom volume of approximately. 200 m³, and typical winds in this geographic area of 3–5 m/s, ACH of 3 h⁻¹ as a conservative representative value is adopted.

Figure 15 below depicts the daily effect of night ventilation on reducing indoor ambient temperature.

The reduction peaks at 4:00 in the morning, correlating with the lowest outdoor temperature, and drops to just below 2 °C at the hottest noon hours.

Throughout the results, a clear distinction is made between measured data and simulated outcomes. Empirical measurements reflect actual classroom conditions under typical operation, including irregular HVAC usage and occupancy-driven fluctuations. Simulated scenarios, by contrast, were constructed under controlled boundary conditions to assess the isolated and combined impact of envelope insulation and night ventilation. These simulations, calibrated using real data, allow for predictive comparisons and scenario testing while maintaining fidelity to observed behavior.

In this research, a coupled analysis of indoor surface temperatures and simulated ventilation scenarios is performed to evaluate the impact of night ventilation as part of a holistic DSM strategy. Figure 15 compares the measured surface temperatures on the southern and northern walls of the pilot classroom with simulated results under night ventilation conditions.

The data show a pronounced asymmetry between the walls. The southern wall, heavily exposed to solar radiation during the day, exhibits significantly higher surface temperatures compared to the northern wall. This thermal disparity complicates maintaining a uniform indoor temperature, as heat accumulation varies by orientation.

Figure 16 also quantifies the temperature reduction (ΔT) attributable to night ventilation. While the cooling effect is present, approximately 1.0–1.5 °C on average, it is more notable on the northern wall. Due to prolonged sun exposure, the southern wall shows limited response to nighttime cooling, underscoring the limits of passive ventilation in sun-exposed zones.

The analysis highlights that although night ventilation modestly reduces surface temperatures, particularly on the shaded northern wall, its direct effect on overall thermal comfort is constrained. Nevertheless, when interpreted at the whole-room scale and integrated with occupancy patterns, night ventilation contributes its portion to the estimated daily energy savings of 10–15%. This translates to annual savings of approximately 1500–2200 kWh per classroom under typical usage conditions.

6.2. Envelope Insulation

The impact of variations in envelope insulation on energy consumption can be significant, affecting both heating and cooling loads in a building. Empirical measurements and numerical simulations are commonly employed to analyze this impact, offering unique insights into the building’s thermal performance.

Increasing the insulation level decreases the U-value, reducing the heat flux. ${\mathrm{Q}}_{\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{s}\mathrm{f}\mathrm{e}\mathrm{r}}$ through the building envelope, resulting in lower energy consumption for winter heating and summer cooling. Changing the insulation material to one with lower thermal conductivity will also lower the U-value, reducing the energy loads.

Conversely, a thinner insulation layer increases the U-value and heat flux, dramatically raising energy costs, especially in extreme climatic conditions, such as the case study example in hot, dry desert-like conditions.

Computational software packages can model the energy performance of a building with different insulation levels. The impact on energy consumption is estimated by inputting the various U-values based on different insulation scenarios. Models can also be run to perform a sensitivity analysis, demonstrating how sensitive energy consumption is to variations in insulation levels, providing quantitative metrics that can guide decision-making. Combining measurements and simulations can accurately quantify this impact, providing valuable insights for optimizing building energy performance.

Adding or upgrading envelope insulation is a viable strategy for new and existing buildings seeking higher energy efficiency. Even for existing structures, retrofitting insulation can be relatively cost-effective and provide substantial benefits. While insulated materials may have higher upfront costs than traditional building materials, long-term energy savings and improved comfort levels justify their use.

By implementing thermal insulation, the internal temperature of the building can remain relatively stable, reducing the need for excessive heating or cooling, enhancing user comfort, reducing energy consumption, and lowering utility costs. While adding insulation may necessitate design and maintenance considerations, its contribution to improved thermal comfort and energy efficiency can substantially enhance indoor environmental quality.

A model incorporating external insulation in the wall introduces more moderate internal surface temperature fluctuations. On average, the temperature on the south wall remains relatively stable. The insulation helps to mitigate extreme temperature variations, leading to a more comfortable and consistent indoor environment.

Figure 17 below illustrates the daily profiles of outdoor ambient temperature and exterior surface temperature, which illustrate facade solar absorption: during sunlit hours, the surface temperature consistently exceeds ambient, evidencing absorbed solar gains. As this surface temperature load increases, the driving potential for conduction into the envelope also increases, raising cooling demand. Accordingly, higher exterior surface temperatures imply a greater need for improved insulation with lower U-values to limit heat transfer to the indoor zone and moderate peak loads.

Figure 18 shows the indoor surface temperature, the ambient, and the operative temperatures for both test cases: baseline with no insulation and after the insulation was applied.

In the baseline test case, the operative temperature follows a pronounced daily cycle, rising with solar loading and internal gains: it varies from about 23.0 °C in the early morning to nearly 26 °C at the midday peak, a range of 2.8 °C. During school hours (08:00–15:00), the mean operative temperature is 24.97 °C.

After insulation is applied, the indoor response is markedly flattened, as operative temperature remains within 22–26 °C over the entire day (range of 4 °C), with a midday ambient peak temperature lower by 1.4 °C compared to baseline and an occupied-hours mean ≈ 24.29 °C (≈0.7 °C lower than baseline).

This demonstrates that improved insulation partially decouples the classroom from outdoor solar fluctuations, suppressing the amplitude of indoor temperature swings and improving comfort stability under the same HVAC settings. The slightly higher early-morning level in the insulated case reflects reduced night heat loss and the thermal-mass effect, but the overall occupied-period conditions are cooler and more stable.

Figure 18 outcomes support a 0.5–1 °C increase in the occupied cooling setpoint, keeping the occupied hours mean operative temperature equal to or below the baseline, corresponding to up to 5% operational savings.

6.3. Combined DSM Analysis

Envelope insulation and night ventilation represent two distinct yet complementary strategies to improve indoor thermal comfort and manage energy performance. Each aims to reduce both the mean indoor temperature and the amplitude of daily fluctuations, thereby moderating indoor thermal conditions. Their effectiveness is context-dependent, influenced by building design, climatic conditions, insulation quality, and operational constraints.

Figure 19 presents measured indoor temperatures compared with simulated scenarios: (i) no intervention, (ii) insulation only, and (iii) the combined application of night ventilation and insulation.

Both insulation and night ventilation reduced indoor temperatures relative to baseline conditions, with average reductions of 2.7 °C and 2.0 °C, respectively. Their combined application yielded the largest benefit (above 3.3 °C reduction), confirmed by statistical tests indicating up to 20% energy savings compared to baseline. This demonstrates the additive value of integrating envelope and operational strategies. These findings highlight that insulation and night ventilation are comparably effective in reducing average higher temperatures.

As expected, envelope insulation and night ventilation yield the most significant improvement in indoor thermal conditions. However, it is important to emphasize that even in cases where insulation cannot be implemented due to structural or budgetary constraints, night ventilation alone still offers a simple, effective, and accessible strategy for enhancing thermal comfort and reducing energy loads.

Their combination yields the most substantial improvement. Contrary to earlier assumptions, insulation has a significant standalone impact in this context.

A rigorous statistical analysis, including paired-sample t-tests, compared scenarios with combined DSM strategies (ventilation and insulation) versus individual implementations. Results indicated a statistically significant improvement in indoor temperature stabilization and cumulative energy savings of up to 20% compared to baseline conditions. This analysis quantitatively confirms the superior effectiveness of integrating multiple DSM measures.

The following summary

Table 7. School A. DSM Efficiency Comparative Table.

DSM Strategy	Average Daily Temp. Reduction (°C)	Average Daily Energy Savings (%)	Annual Energy Savings per Classroom (kWh)
Night Ventilation	2.7–3.0	10–15%	1500–2200
Envelope Insulation	0.8–1.0	Approx. 5%	900–1300
Combined DSM	3.5–4.0	Up to 20%	2200–3000

explicitly compares individual and combined DSM strategies, clarifying their relative benefits to facilitate informed decision-making for practitioners.

Comparative testing confirmed that combined DSM strategies outperformed individual ones, producing statistically significant improvements in comfort stability and cumulative energy savings of up to 20%.

6.4. Discussion

The outcomes of this study align with trends reported in comparable research, while extending their applicability to hot-arid educational settings. For example, Congedo et al. [8] found that night ventilation in Mediterranean schools reduced operative temperatures by 1.5–2.5 °C; the simulations showed a similar order of magnitude, confirming its effectiveness under different climatic conditions. Ranđelović et al. [9] reported cooling load reductions of 18–22% from insulation retrofits, which is comparable to the approx. 20% savings were achieved here when insulation was combined with night ventilation. Importantly, this research’s integrated monitoring and simulation approach demonstrates how these measures perform specifically in Israeli schools, thereby adding empirical validation in a local context. This strengthens both the robustness and the transferability of the findings beyond the case study.

Table 8. DSM Efficiency Comparative Table.

Study	Climate	Building Type and Vintage	Intervention	Baseline Description	Calibration Method	Primary Metric	Reported Magnitude	Notes on Comparability
Rizzo et al., 2024 [43]	Mediterranean	Primary school, contemporary	Demand-controlled ventilation with summer night purging	Predominantly naturally ventilated baseline with limited AC	Model validated against on-site measurements, EN 16798 comfort compliance	Annual cooling energy and IAQ compliance	Qualitative energy reduction with night purging and improved comfort while meeting IAQ targets	Same climate class as our site, with explicit night purge control window and comparable school occupancy
Chiesa and Vigliotti, 2024 [44]	Northern Italy	Junior high school, 2000s	Optimized mechanical ventilation control strategies with schedules and setpoints	Mechanically ventilated with heat recovery baseline	Monitoring and simulation with control strategy validation	Heating and fan energy with comfort compliance	Energy reduction under optimized control with maintained IAQ	Relevant as schedule-based controls and setpoints drive DSM with limited occupant burden
Aloshan and Aldali, 2024 [3]	Hot arid Saudi Arabia	Primary and secondary schools, 1990s, 2000s	Facade insulation and glazing retrofit scenarios	Split AC baseline	Simulation calibrated to utility data at the building level	Annual cooling electricity and cooling load	Up to 17% reduction in cooling load for optimized facade packages	Hot arid public schools with envelope-driven savings comparable to our envelope strategy
Campagna and Fiorito, 2023 [1]	Apulia, Italy	Preschool, primary, lower secondary, mixed vintages	Stock level envelope and systems characterization with upgrade scenarios	Varied baselines across vintages	Statistical and simulation-based assessment across typologies	End-use intensities and retrofit potential ranges	Identifies envelope upgrades as the primary lever with climate-sensitive performance	Establishes a regional baseline and comparability of typologies to our Mediterranean school context

compares the current results with representative studies spanning Mediterranean and hot dry climates, focusing on school buildings, envelope-centric DSM, and night ventilation controls. Metrics to annual cooling energy and peak load are harmonized, and when needed, are converted to reported temperatures to occupied period maxima for comparability (Rizzo et al. [43], Chiesa and Vigliotti [44]; Aloshan and Aldali [3]; Campagna and Fiorito [1]).

Across climates and vintages, the most consistent gains arise when night ventilation or optimized ventilation control is paired with improved roof facade thermal resistance, which is reproduced in the authors’ findings.

7. Scalability and Replicability of Findings

While this study focuses on elementary schools in Israel, the methodology and findings might have broader implications for different climates, school types, and geographical regions. This section explores the generalizability of the results and potential adaptations for diverse contexts.

Climate Considerations

The study was conducted in hot, arid regions, where cooling loads dominate energy consumption. However, the proposed DSM strategies can be adapted for other climates:

Temperate and Humid Climates:

• Night Ventilation Effectiveness: Night ventilation may increase latent cooling loads due to moisture infiltration in higher-humidity climates. Dehumidification strategies or hybrid ventilation (natural + mechanical) should be considered.
• Insulation Strategies: While external insulation improved thermal stability in the Israeli case study, additional glazing upgrades and airtightness improvements may be required in colder climates to minimize heating losses.

Cold Climates:

• Heat Recovery Ventilation (HRV): Night ventilation might not be feasible in regions where heating dominates in winter. Instead, HRV systems can recover heat from exhaust air while maintaining indoor air quality.
• Solar Gains Utilization: Schools in high-latitude regions may benefit from passive solar heating, incorporating thermal mass materials to store solar energy during daylight hours.

Adaptation to Different Building Types

The research focused on elementary schools with moderate occupancy density and relatively small HVAC loads per unit area. However, the findings can be scaled to other types of buildings with key adaptations:

Different Building Occupancy and Facilities:

• Increased occupant density and electronic equipment usage (e.g., computers, laboratory equipment) may require more dynamic load management strategies.
• Zoned HVAC control and real-time occupancy sensors would be more effective than whole-building ventilation strategies.

Older vs. Newer Buildings:

The effectiveness of retrofit interventions (e.g., insulation, smart lighting, envelope upgrades) depends on the baseline condition of the building.

Building envelope retrofits should be prioritized for aging buildings, whereas newer buildings may benefit more from advanced DSM strategies like automated energy management systems.

Policy and Implementation Considerations

To replicate and scale these strategies, policymakers and facility managers should consider the following:

• Customization Based on Climate and Energy Costs: The cost-effectiveness of DSM strategies varies by region. Incentive programs and regulatory frameworks should align with local energy prices, carbon reduction targets, and infrastructure constraints.
• Integration with Smart Grid Technologies: Buildings with on-site renewable energy (e.g., solar panels + battery storage) can optimize energy use by selling excess electricity to the grid or participating in demand-response programs.

Integration of on-site energy sources

Integrating renewable energy sources, such as PV panels with battery storage, can significantly enhance school buildings’ energy efficiency and sustainability. By leveraging on-site solar generation, schools can reduce reliance on grid electricity, particularly during peak demand periods, and improve resilience against power disruptions. In our case study, schools with PV-battery systems achieved high energy self-sufficiency, demonstrating the potential for widespread implementation. However, optimal energy management strategies, such as load shifting and demand-response participation, are needed to maximize savings. Policy recommendations to support these efforts include government incentives for retrofitting school buildings with renewable energy technologies, the introduction of energy performance benchmarks for public institutions, and funding mechanisms for demand-side management initiatives. Encouraging schools to participate in smart grid programs and providing subsidies for energy-efficient retrofits could further enhance adoption rates, aligning with national energy sustainability goals.

Future Research Directions on Scalability

To further generalize the findings, future research should:

• Conduct longitudinal studies across multiple climatic zones to refine the effectiveness of DSM strategies under varying seasonal conditions.
• Explore machine learning-based predictive models for real-time adaptive control of HVAC and lighting.
• Expand field studies to buildings with different construction typologies (e.g., prefabricated vs. masonry buildings).

By demonstrating the scalability and adaptability of these findings, this research provides a foundation for energy efficiency improvements in diverse educational settings worldwide.

The summary in

Table 9. Scalability summary table.

Climate Zone	Recommended Methodological Adjustments
Hot, Desert (Current Case Study)	Night ventilation prioritized, envelope insulation beneficial.
Temperate, Humid	Hybrid ventilation strategies: enhanced moisture control systems required, air-tightness upgrades recommended.
Cold Climates	Prioritize thermal mass and passive solar gains, adopt heat recovery ventilation systems (HRV), significantly upgrade insulation thickness.
High Occupancy Buildings	Implement zoned HVAC systems, occupancy sensors, and adaptive ventilation controls.

below explicitly clarifies the necessary methodological adjustments for replicating the presented DSM strategies across diverse climates and building typologies, significantly enhancing this research’s practical applicability.

Specifically, school administrators and policymakers should prioritize the following actionable recommendations based on this study’s findings:

• Short-term (within 1 year): Implement structured night ventilation schedules, particularly effective in climates with significant diurnal temperature variation, to achieve immediate energy savings.
• Mid-term (1–3 years): Conduct targeted external insulation retrofits on older buildings, as the study demonstrates clear benefits in thermal stability and reduced cooling loads.
• Long-term (3–5 years): Integrate real-time monitoring systems to adaptively manage IAQ and HVAC systems, thus ensuring ongoing energy optimization and occupant comfort.

8. Conclusions

The study developed an energy simulation model deployed in an elementary school in southern Israel, employing rigorous methodology and numerical simulations. The primary focus was on understanding the impact of structural materials on thermal comfort and energy usage within the building. Various parameters, including conductivity, specific heat, and density of materials, were examined to understand their implications on the temperature results obtained in the simulation. The study’s novelty lies in combining multi-scale monitoring with calibrated simulation to evaluate DSM strategies in a climate-specific context.

The outcomes of the study serve multiple purposes. Firstly, they offer a quantifiable impact of building features like better insulation and night ventilation strategies, which are suggested to be used by policy decision-makers to advocate for legislation that mandates or incentivizes higher energy-efficient standards in schools. The findings could also inform policy instruments like subsidies for schools that adopt such energy-efficient technologies or practices.

Secondly, the research can serve as a guide for architects, engineers, and building designers. The meticulous data collection, analysis, and validation offer a robust foundation for understanding the energy dynamics at play in school buildings, enabling the integration of best practices into the design process, which is particularly beneficial for new constructions. The study identifies areas where retrofits could be most impactful for existing buildings, especially older ones that may be less energy-efficient, such as enhancing insulation or implementing night ventilation systems.

This study demonstrates that combining multiscale monitoring with calibrated simulations provides a robust framework for evaluating DSM in schools. Integrating night ventilation with insulation yielded measurable comfort and energy benefits (up to 20% savings). Beyond this case, the framework offers a transferable method for prioritizing retrofits across educational contexts.

Based on the findings, several actionable recommendations can be drawn:

(1) Implement real-time indoor environmental monitoring systems to support data-driven decision-making in facility management and improve response to thermal discomfort.

(2) Prioritize cost-effective retrofitting measures, such as controlled night ventilation and selective insulation upgrades, which offer high returns under limited budgets.

(3) Develop national or municipal support programs that facilitate the scaling of validated DSM strategies in public schools, particularly in hot climatic regions, by linking performance data to targeted funding mechanisms.

These measures can significantly advance the transition toward sustainable and resilient school infrastructure.

Given the modular nature of the monitoring and modeling framework, the approach is readily adaptable to other public buildings beyond schools, offering potential for widespread replication across the Mediterranean region and beyond.