Thermal Performance Evaluation of Phase Change Material-Integrated Triple-Glazed Windows Under Korean Climatic Conditions

Abstract

Passive design strategies incorporating phase change materials (PCM) provide effective thermal energy storage, improve indoor comfort, and reduce building energy demand. This study aimed to evaluate the effectiveness of partially filled PCM glazing systems in stabilizing indoor thermal comfort under Korean climate conditions, testing the hypothesis that partial integration can provide meaningful diurnal temperature regulation without compromising daylight access. Indoor air, interior and exterior glazing surfaces, and the PCM layer were monitored to evaluate heat transfer, while EnergyPlus simulations extended the analysis to seasonal conditions. The PCM model was developed using the Conduction Finite Difference (CondFD) algorithm and validated against experimental data, reliably reproducing dynamic phase change behavior. Field tests with a 28 °C PCM showed reductions in indoor peak temperatures of about 2.0 °C during daytime and increases of 1.5 °C at night. Under broader climatic simulations, the same PCM achieved up to 3.7 °C daytime reductions and 2.0 °C nighttime increases, depending on outdoor conditions. These findings highlight the potential of PCM-integrated glazing systems for adaptive thermal regulation in Korean climates and suggest broader applicability for passive cooling and heating strategies in buildings facing increasingly variable weather conditions.

1. Introduction

Global temperatures continue to rise due to global warming, and the increase is expected to reach 1.5 °C between 2030 and 2052 [1]. Building energy consumption is reported as the main cause of global warming, as it represents approximately 28% of global energy-related CO₂ emissions [2]. Accordingly, international responses to reduce greenhouse gas emissions from buildings have been accelerated [3]. South Korea also submitted its 2030 national greenhouse gas reduction target and 2050 long-term low-carbon development strategy to the Secretariat of the United Nations Framework Convention on Climate Change, which reflected that buildings accounted for 20.3% of total energy consumption as of 2023 [4,5].

In tandem with this policy trend, various strategies have been proposed to reduce energy consumption in the building sector. Among them, passive design that can effectively utilize natural energy sources has become essential. It is defined as technology that improves the thermal performance of buildings by enhancing insulation performance, introducing high-performance windows, and installing external shading devices, thereby reducing the heating and cooling energy demand in buildings [6]. In particular, since passive design is applicable to old buildings, it is expected to exhibit higher applicability and energy-saving effects in Korea, where buildings over 15 years old represent approximately 75% (5.4 million buildings) of all buildings [7,8].

Among passive technology strategies, thermal energy storage (TES) is a technology that can store atmospheric heat and solar radiant energy and then resupply the energy at night or during periods with high thermal demands. It is evaluated as an effective method in mitigating indoor overheating and reducing heat loss at night. In general, the heat storage materials used in TES are substances with high specific heat capacity, such as water and molten salt. Among them, phase change materials (PCMs) have been actively applied in TES because they can store a large amount of latent heat through phase changes within a certain temperature range and have high heat storage efficiency compared to the same volume [9]. Due to these benefits, PCMs have been utilized to reduce heating and cooling energy consumption in buildings. In particular, studies have been actively conducted to apply PCM to windows that experience overheating and high heat loss. Several studies have demonstrated significant thermal buffering, e.g., up to ~9 °C indoor peak reduction and ~28% transmitted heat reduction in PCM-integrated windows [10,11].

These research results indicate that window systems integrated with PCM are effective in mitigating rapid changes in the indoor thermal environment and reducing building energy consumption. The heat storage performance of PCMs, however, is affected by solar radiation conditions, outdoor temperature, and the melting temperature of the PCM, and design optimization based on climate characteristics and the location of application is essentially required [12,13]. Lisa et al. [14] aimed to propose the optimal application form of PCMs based on climate through a literature survey. Through the studies conducted in temperate, dry, cold, and tropical climates, they reported that the most difficult part in PCM design is temperature variation, and that the melting temperature and application method of PCMs vary depending on the climate data of the site. Gowreesunker et al. [15] found that PCMs filled in double-glazed windows have a transmittance of 40% in the solid state through spectrophotometry and simulation, and verified that the transmission spectrum is unstable during phase change. They emphasized that PCMs should be allowed or hidden depending on the application method when PCM windows are designed, considering visual comfort.

Therefore, system design that comprehensively considers heat transfer characteristics, application thickness and location, and interactions with light transmittance, which vary depending on weather conditions, is required to effectively utilize the latent heat characteristics of PCM-integrated windows. In particular, a quantitative analysis that reflects seasonal characteristics is required in temperate climate regions that exhibits the weather change during four seasons, such as Korea [16]. Recent studies have also highlighted the potential of combining PCM with other passive strategies, such as natural ventilation, in diverse climatic contexts [17]. These works support the broader relevance of PCM applications and underscore the need for climate-adapted design approaches.

While previous PCM window studies have demonstrated large thermal buffering effects under specific conditions, they often relied on nearly full PCM filling (80–90% of glazing area) and did not address practical limitations such as daylight provision and seasonal adaptability. This study addresses these gaps by testing a partial PCM integration strategy and quantifying its thermal impacts under realistic Korean climate conditions.

Introduced a limited design method that fills only 50% of the inside of the triple-glazed window, which corresponds to the lower section, with a PCM to utilize its transmission and non-transmission characteristics and latent heat storage effect while maintaining the daylight access of the upper section of the window. This is a new application method that considers the balance between securing natural light indoors and thermal performance beyond previous studies. For the designed PCM window, the thermal behavior of PCM and the temperature gradient of the window were quantitatively analyzed through a small-scale mock-up experiment. Based on this, the impact of the PCM triple-glazed window on indoor thermal comfort by season was evaluated under annual Korean climate conditions through the EnergyPlus simulation model. This study aimed to verify the effectiveness of the partial PCM triple-glazed window system under various weather conditions and to propose an application method suitable for the domestic climate.

2. Methodology

2.1. Field Measurement Setup

2.1.1. Description of Mock-Up Test

To experimentally evaluate the thermal behavior of the PCM window under the inflow of solar radiation and the resulting changes in the indoor thermal environment, mock-ups were fabricated as shown in Figure 1. The mock-ups were scaled models of an office building and had a size of 736 mm (W) × 1163 mm (L) × 736 mm (H). The depth of the interior space was set to approximately twice the height of the window by reflecting the actual transmission of solar radiation in the office space [18].

The outer walls of the mock-ups were constructed with extruded polystyrene (XPS) insulation to minimize heat transmission. The physical properties of the triple-glazed window used in the mock-ups and the dimensions of the internal experimental space are summarized in

Table 1. Mock-up information.

Type		Details
Mock-up Size	Indoor	Width	536 mm
		Depth	980 mm
		Height	536 mm
Material	Triple Glazing (Polycarbonate)	Thickness	33 mm (3PC + 12A + 3PC + 12A + 3PC)
		Visible Light	Transmittance (91%)
			Reflectance (7%)

.

The PCM was placed between the middle and rear glazing to effectively store solar heat transmitted through the outer glazing and to minimize direct exposure to outdoor temperature fluctuations, thereby enhancing its heat-storage efficiency and the nighttime release of stored heat into the indoor space.

Structural safety is essential for PCM windows because expansion or leakage during phase change can damage the glass and lower performance. To address these issues, high-strength transparent polycarbonate with excellent impact resistance was employed. The PCM was injected into a cylindrical polycarbonate container, with the lower section sealed using rubber to prevent leakage. This design minimized airtightness concerns and ensured stable, long-term indoor thermal measurements.

The transmittance and reflectance listed in Table 1 were measured using a Window Energy Profiler for the triple-glazed assembly without PCM to establish baseline envelope performance. However, after the PCM was inserted between glazing layers, leakage during sealing prevented reliable measurement with the same device. Therefore, the optical and thermal parameters in Table 1 represent the glazing-only configuration. In the EnergyPlus simulation, the optical behavior of PCM was not modeled because the software does not support phase-dependent transmittance; instead, only the thermal properties of PCM were considered. Previous research [15,19] has shown that paraffin-based PCM glazing typically exhibits a 15–25% increase in transmittance when melted, which aligns with the qualitative observations of this study.

2.1.2. PCM-Integrated Window Design

The windows of the two mock-ups used in the experiment were fabricated in a triple-glazed window structure by placing 520 mm (W) × 520 mm (H) × 3 mm (T) glazing at 12 mm intervals. The PCM-integrated window was applied to one of the mock-ups.

In this study, an organic paraffin-based solid PCM was applied, which is easily applicable to building materials and exhibits stable crystallization even under supercooled conditions [20].

For the PCM-integrated window, the solar radiation transmittance varied during the phase change, as shown in Figure 2. In the solid state, most incident solar radiation was reflected or absorbed, whereas transmittance increased as the PCM transitioned to the liquid state through melting. However, if the melting temperature of the PCM is not appropriate, delayed or inefficient phase change reactions may lead to issues such as indoor overheating or obstruction of external views [19,21].

Therefore, a preliminary experiment was conducted before the field test to select a PCM with a suitable melting temperature. Two types of paraffin-based PCMs with melting temperatures of 28 °C and 35 °C were injected into the cavity of triple-glazed windows and installed facing south to observe phase change behavior under outdoor conditions from 08:00 to 18:00. During the experiment, the outdoor temperature ranged from 22 °C to 25.8 °C. The 28 °C PCM fully melted around 14:00 as the temperature peaked at 25.8 °C, whereas the 35 °C PCM underwent only partial melting, showing limited latent heat storage and insufficient indoor thermal regulation. Figure 3 shows the phase changes for the PCMs with different melting temperatures. Based on the results of the preliminary experiment, the 28 °C PCM, which could induce a stable phase change at a relatively low temperature, was selected as the one to be applied to field experiments.

Table 2. Phase change material properties.

Type	Details
PCM	Formula	C18H38
	Melting point	28 °C
	Heat of fusion	241 kJ/kg
	Flash point	165 °C
	Specific heat capacity	2 kJ/kg °C
	Volume expansion	12.5%

summarizes the key material properties of the PCM used in the field experiments.

2.1.3. Data Collection and Analysis Process

To quantitatively evaluate the thermal behavior of the PCM-integrated triple-glazed window and its impact on the indoor thermal environment, a total of 18 T-type thermocouples with verified initial performance were used in this study. They were attached to the nine points selected for each of the two mock-ups as shown in Figure 4. As summarized in

Table 3. Names for each sensor measurement point.

Measure Points
P1	Indoor Air
P2	Indoor Glazing Inner Surface
P3	Indoor Glazing Outer Surface
P4	Phase Change Materials
P5	Middle Glazing Inner Surface
P6	Middle Glazing Outer Surface
P7	Window Air Layer
P8	Outdoor Glazing Inner Surface
P9	Outdoor Glazing Outer Surface

, P1 was set as the location for measuring indoor air temperature, and the measurements were used as reference data to identify changes in the indoor thermal environment caused by the phase change of the PCM. The sensors from P2 to P9 were attached to each layer of the triple-glazed window (glass layers, air layers, and the layer including PCM). This arrangement was adopted to quantitatively capture the heat transfer paths and thermal gradients across the glazing layers, thereby enabling a more reliable evaluation of the thermal behavior of the PCM-integrated window.

For solid PCMs, the melting and cooling rates of the exterior, center, top, and bottom sections are different depending on the external conditions. Thus, precise setting of temperature measurement points is required for accurate thermal behavior evaluations. Preliminary tests showed that sensors installed in the middle of the PCM recorded higher temperatures once the upper layers began to melt, even though the bulk PCM was not fully liquefied. Therefore, the sensor P4, measuring the PCM temperature, was installed at the bottom of the PCM, which can most sensitively reflect the progress of the phase change. Field experiments were conducted in Daejeon, Korea, from March to June 2023. Among the data collected, that of 31 May, when weather and shading effects were minimal, was selected as the representative case. Temperature data were recorded every 30 s and averaged every five minutes. The mock-ups were installed on a rooftop to reduce external interference, maintaining a 10 cm gap to prevent thermal interaction. The analysis proceeded in three stages: (1) evaluation of the PCM’s thermal behavior within the triple-glazed window, (2) examination of the temperature gradient across window layers, and (3) assessment of the resulting indoor temperature variation.

2.2. Simulation Setup

In this study, the simulation analysis was conducted using EnergyPlus Version 9.6.0 (U.S. Department of Energy, Washington, DC, USA). EnergyPlus is a high-precision building energy simulation software developed by the US Department of Energy (DOE), which can implement thermal and material equilibrium models by integrating the functions of DOE-2 and BLAST [22,23]. In particular, it facilitates simulation in consideration of various climate conditions and the physical environments of buildings, making it suitable for the analysis of materials with unsteady heat transfer characteristics, such as PCM. Simulation configuration was focused on the process of implementing mock-up models in EnergyPlus based on the physical conditions and geometry of the field experiments performed earlier. Based on this, an attempt was made to numerically verify the thermal behavior of the PCM-integrated triple-glazed window model and changes in the indoor thermal environment under a thermal environment similar to reality. The following section describes the process of setting energy simulation for the model integrated with the PCM-integrated triple-glazed window, including the setting of thermal property parameters to mathematically model the physical properties of the PCM and the method of implementing heat transmission and heat storage analysis models.

2.2.1. Simulation Model Development

Two spatial regions were set inside one mock-up to implement the triple-glazed window structure with PCM in the EnergyPlus environment. The front section was a space with a conventional double-glazed window, while the rear section was a space with a single window. Figure 5 shows each component of the model.

The thermal behavior of the PCM was implemented by referring to 1ZoneUncontrolledWithHysteresisPCM among the EnergyPlus example files. This model is based on the Phase Change Hysteresis (PCH) characteristics, and is suitable for simulating supercooling and delayed coagulation that may occur in the actual building environment because it can reflect nonlinear behavior with different critical temperatures during the melting and solidification of PCMs [24]. In this study, three PCMs with different melting temperatures (28 °C, 35 °C, and 44 °C) were applied to evaluate their impacts on the indoor thermal environment. Polycarbonate glazing and air layers were modeled with thermal conductivities of 0.20 W/m·K and 0.024 W/m·K, and densities of 1200 and 1.2 kg/m³, respectively. Internal and external surface heat transfer coefficients were set to 8.3 and 23.9 W/m²·K following EnergyPlus defaults. Their physical properties are summarized in

Table 4. Properties of the PCMs used in simulation.

Type	Details
PCM	Formula	C18H38	C20H42	C22H46
	Melting point	28 °C	35 °C	44 °C
	Heat of fusion	241 kJ/kg	208 kJ/kg	230 kJ/kg
	Flash point	165 °C	177 °C	186 °C
	Specific heat capacity	2 kJ/kg °C	2 kJ/kg °C	2 kJ/kg °C
	Volume expansion	12.5%	12.0%	12.5%
	Thermal Conductivity	0.2 W/mK	0.2 W/mK	0.2 W/mK
	Solid State Density (Liquid State Density)	880 kg/m3 (770 kg/m3)	880 kg/m3 (770 kg/m3)	880 kg/m3 (760 kg/m3)

.

Typical Meteorological Year (TMY) data of Daejeon provided by ClimateOneBuilding were utilized as weather data for simulation. Due to Korea’s climate characteristics, there are four distinct seasons, and weather conditions significantly vary depending on the time of the year. Simulation was performed by reflecting annual weather conditions under each melting temperature condition. In this instance, the actual measurements collected from an outdoor experimental facility were applied for solar irradiation, temperature, and humidity to verify the reliability of the simulation model.

2.2.2. Calculation Method for the PCM-Integrated Window

The thermal behavior of the PCM-integrated triple-glazed window was calculated through the Conduction Finite Difference Solution (CondFD) algorithm of the EnergyPlus software. CondFD is a method of numerically analyzing unsteady heat transfer that varies over time, and it can reflect changes in latent heat storage and heat capacity during the phase change process of PCM in detail [25]. The algorithm provides two numerical analysis methods. One is the semi-implicit (Crank-Nicolson) method and the other is the fully implicit (Adams-Moulton) method. In this study, the fully implicit method was adopted for the stable convergence of rapid thermal behavior. This method calculates the temporally continuous node temperature $T_i^{j+1}$ using Equation (1).

$$C_p\rho ∆x{\displaystyle \frac{T_i^{j+1}-T_i^j}{∆t}}=\left(k_W{\displaystyle \frac{\left(T_{i+1}^{j+1}-T_i^{j+1}\right)}{∆x}}+k_E{\displaystyle \frac{\left(T_{i-1}^{j+1}-T_i^{j+1}\right)}{∆x}}\right)$$

(1)

where $C_p$ is the specific heat capacity (J/kg) and $\rho$ is the density (kg/m³). $∆x$ is the finite difference layer thickness (m) and $T$ is the node temperature (°C). $j+1$ is the new time step, $j$ is the previous time step, and $∆t$ is the calculation time step. $i$ is the node being modeled, $i+1$ is the adjacent node to the interior of the construction, and $i-1$ is the adjacent node to the exterior of the construction.

$$k_W{\displaystyle \frac{\left(T_{i+1}^{j+1}-T_i^{j+1}\right)}{∆x}}$$

(2)

$$k_E{\displaystyle \frac{\left(T_{i-1}^{j+1}-T_i^{j+1}\right)}{∆x}}$$

(3)

Equations (2) and (3) represent the thermal conductivities of the following interfaces. $k_W\left(W/mK\right)$ is the thermal conductivity for the interface between the nodes $i$ and $i+1$, while $k_E\left(W/mK\right)$ is the thermal conductivity for the interface between the nodes $i$ and $i-1$.

In CondFD, the distance between nodes is determined by the thermal diffusivity of the material $\left(\alpha \right)$ and the time interval, and follows the discretization conditions below (Equation (4)).

$$∆x=\sqrt{c\alpha ∆t}=\sqrt{{\displaystyle \frac{\alpha ∆t}{F_o}}}$$

(4)

where $c$ is the space discretization constant, which is the reciprocal of the Fourier number ($F_o=\alpha ∆t/∆x^2$). Since stability requirements are mitigated for the fully implicit method compared to the semi-implicit method, the default value of $c=3$ was applied in this study. Adopted settings (space discretization constant = 3, time step = 1 min) followed EnergyPlus recommendations and were confirmed to yield stable solutions without numerical oscillation. Similar configurations have been validated in previous PCM modeling research [26].

$$C_p={\displaystyle \frac{h_{i,new}-h_{i,old}}{T_{i,new}-T_{i,old}}}$$

(5)

Since PCMs could be defined previously only through the MaterialProperty:PhaseChange object that calculates the specific heat from the input data prepared in a table of enthalpy-temperature pairs, most studies were conducted using the object. This, however, decreases accuracy when PCMs with clear hysteresis are simulated, because only a single enthalpy-temperature curve is used [27]. To address this problem, a simulation was performed using the new MaterialProperty:PhaseChangeHysteresis object. This object captures hysteresis that exists between solidification and melting processes, considering the previous state as well as the current state to determine the specific heat of the PCM, and increases the accuracy for the change in state (Equation (5)). In addition, the hysteresis model also has inputs for solid- and liquid-state values with respect to thermal conductivity and density, and the average of the two values is used in the transition area of the hysteresis model.

3. Results and Discussion

3.1. Field Test-Based Results

3.1.1. Thermal Behavior and Phase Change Cycle of the PCM

Figure 6 shows the results of comparing thermal behavior when air and PCM are present inside the triple-glazed window, respectively. As the outdoor temperature increased, the air temperature inside the triple-glazed window increased to 45.9 °C, while PCM went through phase transition for approximately four hours from 08:30 to 12:30 after its temperature reached a melting point of 28 °C. This section is a “Mushy” zone where solid and liquid states coexist, indicating that the PCM delayed heat inflow into the indoor space by absorbing latent heat.

The PCM temperature then increased to a maximum of 32.7 °C and gradually decreased after 16:00. From 19:00, the phase transition process was entered again and stored thermal energy was released indoors and outdoors while continuously maintaining a temperature of 28 °C.

While the solid-to-liquid phase transition occurred for approximately four hours under the combined influence of temperature and solar radiation, the liquid-to-solid phase transition lasted for over eight hours as it was mainly dominated by temperature. This reflects the asymmetry of the phase change cycle, showing that the PCM releases heat longer during cooling.

Figure 7 shows the indoor temperature and the temperature gradient from the inner surface to the outer surface of the triple-glazed window, analyzed every two hours on 31 May. Case 1 represents the conventional triple-glazed window, and Case 2 represents the PCM-integrated triple-glazed window. On this day, the highest temperature was 27.3 °C and the lowest temperature was 18.9 °C.

During the night, the PCM existed in the solid state at 24.5 °C. By releasing latent heat through liquid-to-solid phase transition, the indoor temperature of the mock-up with the PCM triple-glazed window remained at 7.4 °C higher than the mock-up with the conventional triple-glazed window.

Between 07:00 and 11:00, when the inflow of solar radiation began, the indoor temperature of Case 2 increased from 27.3 °C to 37.0 °C while that of Case 1 rose from 19.9 °C to 35.2 °C. The indoor temperature of Case 2 was 1.8 °C higher, but this was affected by the transmission of solar radiation through the upper glass. The overall temperature rise was 15.3 °C for Case 1, but it was 9.7 °C (36% lower) for Case 2. This indicates that the PCM-integrated triple-glazed window effectively inhibited the indoor temperature rise through latent heat storage.

After 13:00, the indoor temperatures of Case 1 and Case 2 were reversed due to the latent heat storage of the PCM caused by solar radiation coming from the south, but the maximum difference remained less than 1.5 °C. These results show the influence of solar radiation that penetrates liquid PCM after phase transition.

After 15:00, solar irradiance decreased and the heat flow at the window changed from indoors to outdoors. From this time point, indoor temperature rapidly decreased for Case 1, but indoor temperature remained stable for Case 2 as the PCM continued releasing heat over an extended period of time while maintaining the state of phase transition until 05:00.

3.1.2. Indoor Thermal Performance of the PCM-Integrated Mock-Up Window

As analyzed earlier, PCM exhibited solar radiation absorption and heat release effects through solid-to-liquid phase transition and nonlinear thermal behavior under prevalent weather conditions. To evaluate the effects of these thermal characteristics of the PCM on the actual indoor thermal environment, indoor temperature was compared between two mock-ups in which the conventional triple-glazed window and the triple-glazed window with PCM were installed, respectively.

Figure 8 compares the indoor temperature changes of the two mock-ups. The indoor peak temperature of the mock-up without the PCM reached 40.9 °C at 13:00, while that of the mock-up with the PCM was approximately 2.0 °C lower (38.9 °C). In particular, the indoor temperature of the mock-up with the PCM remained 1.3 °C lower on average while exhibiting a gentler temperature increase curve compared to the indoor temperature of the mock-up without the PCM in the solar radiation acquisition section. At night, when the influence of solar radiation decreased and the outdoor temperature dropped, the indoor temperature of the mock-up with the PCM remained 1.5 °C higher on average while showing a gentler temperature decrease curve due to the thermal energy released during the PCM solidification process.

These results show that PCM applied in the window is effective in reducing changes in the indoor thermal environment by reducing heat gain during daytime and delaying the temperature drop at night. In particular, it serves as a thermal buffer that stabilizes indoor temperatures through the phase change process within the melting temperature range.

3.2. Simulation-Based Results

3.2.1. Validation Through Measurement Comparisons

Before conducting simulation analysis, the measurement results from the small mock-ups were compared with the simulation prediction results to verify the reliability of the performance analysis software-based model used in this study. The indoor temperature was measured every minute, and the analysis was conducted based on the data collected for three days. Especially, Validation employed the measured solar irradiance data recorded on-site to ensure consistent boundary conditions with the experimental setup.

The left figure of Figure 9 shows the indoor temperature measured through the small mock-up experiment and that predicted through simulation according to the outdoor temperature, as well as the difference between them. Overall, the simulation results showed a tendency similar to that of the measurement results, and the temperature difference was less than 6 °C.

The right figure of Figure 9 is the scatter plot that quantitatively analyzed the correlation between the measured values and simulated values. The linear relationship between the two values can be identified through the regression line. The coefficient of determination (R²) was 0.94, showing high predictive accuracy. In addition, the coefficient of variance of the root mean square error (CV(RMSE)) was 7.6%, which sufficiently met the predictive accuracy criterion (less than 30%) for time-based results presented in the ASHRAE Guideline 14 [27]. In addition to R² and CV(RMSE), the Normalized Mean Bias Error (NMBE) was calculated following ASHRAE Guideline 14, resulting in an NMBE of +2.4%, which falls within the ±10% acceptance range for hourly data.

Therefore, the simulation model has precision at a level high enough to reasonably reproduce dynamic changes in the indoor thermal environment. It is judged to be suitable for PCM behavior analysis and thermal environment evaluation in subsequent stages.

3.2.2. Annual Thermal Behavior from the Model-Based Analysis

Figure 10 shows seasonal indoor air temperature changes according to the melting temperature of the PCM. The results were analyzed based on the representative date of each month (21st of each month).

Spring and autumn (Mar–May and Sep–Nov) in Figure 10a,c showed an indoor temperature tendency similar to that of the winter PCM window model, but a higher indoor temperature maintenance effect could be seen due to the longer thermal energy release time at night under the influence of relatively high outdoor temperatures. As can be seen from Figure 10b, in summer (Jun–Aug), overall indoor temperature during daytime could be reduced, regardless of the melting temperature, by applying PCM in the window alone. In detail, the PCM with a melting temperature of 28 °C inhibited daytime overheating by effectively absorbing heat inflow into the indoor space through phase change, while PCMs with melting temperatures of 35 °C and 44 °C showed relatively low levels in terms of the daytime heat reduction effect. As shown in Figure 10d, during winter, the daytime outdoor temperature often remained below the 28 °C melting point, the PCM did not fully melt; however, even partial melting under limited solar radiation was sufficient to store residual heat that was gradually released at night, providing a modest but measurable moderation of indoor temperature. The 28 °C PCM exhibited the highest nighttime heat release effect through phase change, followed by 35 °C PCM and 44 °C PCM. This indicates that, although the heat storage capacity is reduced in winter, PCM glazing still offers potential benefits by mitigating nighttime cooling and improving indoor comfort, rather than hindering energy conservation.

Despite the application of the PCM window model, the overheating phenomenon when the indoor temperature exceeds the comfort zone, occurred in all seasons. This appeared to be due to the combination of the low radiant heat release caused by the high airtightness performance of the mock-up model and the introduction of direct solar radiation through the upper opening. In particular, it appears that the daytime indoor temperature in winter was higher than that in summer, also due to the excessive introduction of direct solar radiation caused by the low solar altitude in winter.

Figure 11 shows the indoor temperature difference between the mock-up without the PCM and the mock-up with the 28 °C PCM, which exhibited the highest performance throughout the year, according to the global horizontal irradiance and the outdoor air temperature. As the irradiance increased, the indoor temperature of the model with the PCM was relatively lower. As the outdoor air temperature increased, the temperature difference between the two models decreased due to the limited heat storage effect of the PCM.

Figure 12 compares the indoor temperature of each mock-up and the corresponding temperature of the PCM on 21 October, when the highest PCM window performance of the year was observed. For all three PCMs, the temperature continuously increased after the phase change caused by solar radiation, and the indoor temperature also increased after the complete liquid state was reached. The 28 °C PCM with the lowest melting temperature exhibited the largest increase in indoor temperature. This appeared to be because indoor overheating was caused by the continuous temperature rise after rapid melting. On the other hand, the PCMs with higher melting temperatures exhibited relatively slow phase changes. However, they exhibited unstable thermal behavior during the melting and solidification processes under the influence of low outdoor air temperatures, thereby showing lower nighttime heat release effects than the 28 °C PCM.

Based on this finding, the integration of PCM into the window is expected to reduce the indoor temperature during daytime by up to 10 °C and increase the nighttime indoor temperature by up to 8 °C. This performance was found to vary depending on the melting temperature of the PCM, outdoor air conditions, and indoor and outdoor heat flow conditions. The discrepancy of approximately 6 °C between field and simulation results can be attributed to differences in testing periods, as the mock-up experiment was conducted in May–June while the simulation results were derived from representative days throughout the year, including November, yet the relative trends of daytime peak reduction and nighttime elevation remained consistent.

These results are attributed to the downscaled mock-up conditions. In this setup, the high window-to-wall ratio and lack of air exchange caused heat transfer to be predominantly governed by the glazing system. Therefore, the findings should be interpreted as relative thermal behavior rather than absolute energy savings. Previous studies have reported that PCM-integrated glazing can reduce indoor peak temperature by about 9 °C and transmitted solar heat by approximately 28%, resulting in 3–5% annual cooling and heating energy savings [28]. Although PCM applied to opaque components such as walls and floors has shown higher cumulative energy reductions of 13–50% annually due to its greater heat-storage capacity, its slow thermal response limits short-term comfort control [29]. In contrast, PCM glazing responds rapidly to outdoor temperature fluctuations and solar gains. This makes it more suitable for climates with strong diurnal variations, such as Korea, where maintaining thermal comfort during rapid day–night temperature changes is critical.

4. Conclusions

This study proposed a window model in the triple-glazed window structure integrated with phase change materials (PCM) to improve indoor thermal environments, and evaluated its applicability under the Korean climate conditions. In particular, field experiments were performed using small mock-ups along with building energy simulations to analyze the indoor temperature control effect of a PCM according to its melting temperature and weather conditions. The main results of this study are as follows.

• For the n-Octadecane-based solid PCM, the solid-to-liquid phase change cycle was four hours on average and the liquid-to-solid phase change cycle was eight hours or more based on May and June. In other words, the duration of heat release was approximately twice as long as the heat storage time. This confirms that PCM plays a crucial role in maintaining indoor temperature at night during intermediate seasons with large daily temperature fluctuations. These asymmetric charge–discharge cycles suggest that, over extended periods, PCM glazing can consistently mitigate daytime overheating while sustaining nighttime comfort, thereby reducing cooling and heating energy demand in real building operation.
• The indoor temperature control effect of the PCM window model showed the following seasonal tendencies. In summer, the rise in indoor temperature during daytime was inhibited under all melting temperature conditions. In particular, the 28 °C PCM exhibited the highest overheating inhibition effect during daytime. During winter and intermediate seasons, daytime indoor temperatures showed similar levels due to the limited phase change of the PCM caused by low outdoor air temperatures, while the 28 °C PCM exhibited the highest nighttime heat release effect followed by the 35 °C PCM and 44 °C PCM. This confirms that PCM inhibits overheating in summer and exhibits the nighttime heat release effect in winter.
• During the experiment, however, the temperature of the PCM inside the triple-glazed window increased to a maximum of 32.7 °C after phase change. Consequently, the indoor temperature also increased to 40.9 °C. This appears to be due to the high airtightness of the mock-up and the insufficient installation capacity of the PCM. Therefore, it is necessary to properly design the melting temperature and installation capacity of PCM according to indoor airtightness and solar radiation conditions.

This study empirically analyzed the impact of the PCM-integrated triple-glazed window system on indoor thermal environment and its applicability through a combination of measurements and simulation, and presented a direction for its future application. Unlike PCM applied in opaque elements such as walls and floors, which mainly contribute to long-term energy savings, our results demonstrate that PCM-integrated glazing directly enhances indoor thermal comfort by moderating daytime overheating and nighttime cooling. This distinctive role also provides a foundation for future studies on the visual comfort impacts of PCM windows. In this study, however, the analysis of the optical characteristics of PCMs, i.e., solar radiation transmission and visible light shielding performance, was not considered. In the future, follow-up studies need to be conducted considering indoor visual comfort and lighting energy loads.

Furthermore, the small-scale mock-up was intentionally designed as a sealed system to isolate the effect of PCM heat storage. However, this condition results in higher peak temperatures than would occur in occupied rooms with ventilation and thermal mass. The experimental validation was limited to three days, which cannot fully reflect nonlinear PCM behavior under seasonal extremes. Therefore, the present results should be interpreted as relative trends rather than absolute values. In full-scale buildings, indoor temperature variation would be mitigated by air exchange and internal heat capacity. Future research will extend the analysis to incorporate ventilation/infiltration models, internal loads, and adaptive control strategies to evaluate PCM glazing under realistic operational conditions.