Comparative Evaluation of In Situ U-Value Measurement Techniques of an External Wall in a Multi-Method Field Study

Abstract

Accurate knowledge of the thermal transmittance (U-value) of existing building envelopes is essential for reliable energy performance assessment and the planning of energy-efficient refurbishment measures. However, in practice, the material composition of existing walls is often unknown, and installing measurement devices may be restricted due to limited accessibility, the risk of structural damage, or varying on-site boundary conditions. Although several in situ methods for determining the U-value have been proposed in the literature, systematic comparisons of their performance under real environmental conditions remain limited. This lack of comparative evaluation makes it difficult to select the most appropriate method under specific practical constraints. To address this gap, this study presents a comprehensive experimental comparison of four in situ U-value measurement methods applied simultaneously to the same building element under identical real boundary conditions, providing new insights into their accuracy, uncertainty, and practical applicability. In this study, four in situ techniques commonly used to determine the thermal transmittance (U-value) were tested on a double-leaf brick wall at the University of Stuttgart: heat flow meter (HFM), infrared thermography (IRT), infrared thermometer (IRTM), and thermometric method (THM). The measurements were carried out over several days under real boundary conditions, during which air temperature, surface temperature, and heat flux were recorded at regular intervals. The results show that all four techniques can be reliably used under real boundary conditions, with the measured U-values lying within a comparable range. Differences among the methods were observed, largely due to their varying sensitivity to environmental influences and sensor placement. A comparison between the upper and lower parts of the wall indicated that its thermal response is non-uniform, and the observed deviations can be attributed to its inhomogeneous structure. By outlining the strengths and limitations of each technique and comparing their measurement outcomes, this study provides practical guidance for selecting suitable approaches for in situ U-value determination. Furthermore, the findings support future efforts to refine thermal evaluation methods and improve energy performance in existing buildings.

1. Introduction

Reducing energy consumption and improving thermal performance of buildings are key priorities in sustainable construction and climate-change mitigation, particularly in the retrofit of existing building stock and heritage structures [1]. In such contexts, reliable knowledge of the thermal transmittance (U-value) of envelope components is essential for accurate energy assessment and retrofit design. However, determining U-values in real buildings remains challenging due to unknown material compositions, construction heterogeneity, and measurement constraints under field conditions [2,3]. Although several standardized methods for estimating the U-value exist, accurately measuring it under real-world conditions remains challenging. This challenge is even more pronounced in heritage buildings, where unique materials and complex environmental conditions complicate measurements [2,3].

The U-value describes the rate of heat transfer through a building element per unit area and temperature difference between interior and exterior environments [4]. It is defined as Equation (1):

$$U={\displaystyle \frac{Q}{A\times \left(T_i-T_e\right)}}$$

(1)

where U is the thermal transmittance (W m⁻² K⁻¹), Q is the heat flow rate (W), A is the heat transfer area (m²), T_i is the interior air temperature (K), and T_e is the exterior air temperature (K) [5]. A lower U-value indicates higher thermal resistance and reduced heat loss [6]. Using heat flux, q (W m⁻²; Equation (2)), the U-value can also be expressed as Equation (3).

$$q={\displaystyle \frac{Q}{A}}$$

(2)

$$U={\displaystyle \frac{q}{\left(T_i-T_e\right)}}$$

(3)

Several in situ and analytical approaches exist for U-value determination, including the calculation (CAL) method [7], the heat flow meter (HFM) method standardised in ISO 9869-1 [8], infrared-based methods including infrared thermography (IRT) [9,10] and infrared thermometer measurements (IRTM), as well as thermometric (THM) approaches based on surface and air temperature measurements [11,12]. Among these, the HFM method is widely regarded as a reference technique due to its direct measurement of heat flux; however, it requires long monitoring periods and careful sensor installation, as improper contact can significantly affect accuracy [13,14].

Infrared thermography has been widely investigated as a non-contact alternative for in situ thermal assessment. Previous studies have demonstrated its applicability under field conditions [15,16], but they have also highlighted its sensitivity to emissivity, reflected temperature, and environmental stability. Comparative studies have reported systematic deviations between infrared-based and heat-flux-based methods, particularly under non-ideal boundary conditions [17]. Recent reviews further indicate that infrared approaches may exhibit high propagated uncertainties, sometimes approaching 70%, due to sensitivity to surface-temperature estimation and radiative exchange effects [18,19]. Nevertheless, under stable environmental conditions with sufficient thermal gradients, good agreement between infrared and reference methods has been reported, with deviations as low as approximately 4–5% [20].

The thermometric method (THM) provides a simplified approach based on surface and air temperature measurements and has been investigated as a practical alternative for in situ assessment [21,22]. However, its formulation neglects radiative heat transfer components, which may lead to systematic underestimation of heat loss. Comparative studies between HFM and THM methods further emphasize that measurement agreement strongly depends on environmental conditions and implementation details [23].

Although previous research has demonstrated the potential of these methods, discrepancies between results remain common due to differences in measurement principles, boundary conditions, and uncertainty propagation mechanisms [15,23,24]. Moreover, most existing studies focus on pairwise comparisons, while simultaneous multi-method evaluations under identical field conditions remain limited.

This study addresses this gap by providing a simultaneous in situ comparison of four U-value determination methods (HFM, IRT, IRTM, and THM) under identical boundary conditions. The measurements were conducted on a double-leaf brick wall at the University of Stuttgart, including an additional analysis of vertical thermal heterogeneity across different wall zones.

Experimental design and monitoring procedures were defined in accordance with ISO 9869-1 [8], which recommends a minimum 72 h measurement period and sufficient interior–exterior temperature difference to ensure quasi-steady-state conditions in dynamic environments. In this study, monitoring was conducted from 3 to 11 February 2025, exceeding the minimum duration requirement and maintaining a temperature difference consistently above 15 °C. Continuous methods (HFM and THM) were applied throughout the full period, while IRT and IRTM were performed during dedicated multi-day measurement phases. Time-averaging over extended periods was used to reduce the influence of transient fluctuations, consistent with the convergence behavior described in ISO 9869-1.

This systematic comparison under real boundary conditions contributes to improving the reliability of in situ thermal performance assessment and supports more robust energy retrofit strategies for existing buildings [25].

This study provides a simultaneous multi-method experimental comparison of four in situ U-value measurement techniques (HFM, IRT, IRTM, and THM) under identical real boundary conditions applied to a double-leaf brick wall. Unlike previous research focusing on pairwise comparisons, the present work integrates measurement results, uncertainty analysis based on the ISO GUM framework, and spatial thermal variability assessment. This study further introduces a structured methodological framework linking experimental design, data acquisition, and comparative evaluation, thereby providing both methodological and practical contributions for in situ thermal performance assessment of existing buildings.

2. Common Methods for Determining the U-Value

A comparative evaluation of commonly used U-value measurement methods helps practitioners select the most appropriate technique based on available equipment, field conditions, and building characteristics. Comparing in situ measurement results also provides a clearer understanding of each method’s performance and applicability. This, in turn, helps professionals choose the most suitable approach depending on the measurement context, site constraints, and building characteristics.

The calculation method (CAL) estimates the U-value by summing the thermal resistances of material layers. While widely applied in design and assessment practices, this method was not used in this study, which focuses on in situ measurement techniques [7]. However, it was used to calculate a reference U-value based on the wall construction details, providing a theoretical context for interpreting the experimental results.

Reliable results from all in situ methods depend on appropriate environmental and boundary conditions. These include maintaining an air temperature difference of 15 °C between indoor and outdoor environments [22]. To achieve the 15 °C air temperature difference more easily, the test is usually done in the cold season [15,17,26]. Moreover, as long as the required indoor–outdoor temperature difference is maintained, measurement outcomes remain largely independent of the test duration. Indoor environments usually exhibit more stable temperatures; therefore, the infrared camera, surface-temperature sensors, and heat flux sensor are positioned indoors to measure interior surface temperature and heat flux. Direct solar radiation on the measured surfaces should be avoided [9], and measurements should not be conducted under wind speeds exceeding 1 m s⁻¹. Surfaces exhibiting moisture, damage, or visible defects should be excluded from the evaluation [21].

2.1. Calculation (CAL) Method

The calculation (CAL) method is used to calculate the thermal transmittance (U-value) and thermal resistance (R-value) of building envelopes and elements. This method can be applied to elements consisting of thermally homogeneous layers. In this method, the U-value can be determined by simple calculations without experimental measurements [7].

The total thermal resistance of a plane building component made of thermally homogeneous layers is calculated from Equation (4) or (5) (

Table 1. Main principles of U-value measurement methods.

Method	Setup	Measuring Devices: Number	Formula
Calculation (CAL) method		-	$U={\displaystyle \frac{1}{{(R}_{si}+R_{se}+{\displaystyle {\sum }_{j=1}^n}{R}_{i\left(j\right)})}}$
Heat flow meter (HFM) method		Air-temperature sensor: 2 Heat flux sensor: 1 Data logger: 1	$U={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n}{q}_{\left(j\right)}}{{\displaystyle {\sum }_{j=1}^n}\left(T_{i\left(j\right)}-T_{e\left(j\right)}\right)}}$
Infrared thermography (IRT) method		Infrared camera: 1 Air-temperature sensor: 2 Data logger: 1	$U={\displaystyle \frac{4\epsilon \sigma T_{si}^3\left(T_{si}-T_{ref}\right)+h_{ci}\left(T_{si}-T_i\right)}{\left(T_i-T_e\right)}}$ $U={\displaystyle \frac{{\sum }_{j=1}^n{U}_{\left(j\right)}}{n}}$
Infrared thermometer (IRTM) method		Digital pyrometer (infrared thermometer): 1 Air-temperature sensor: 2 Data logger: 1	$U={\displaystyle \frac{4\epsilon \sigma T_{si}^3\left(T_{si}-T_{ref}\right)+h_{ci}\left(T_{si}-T_i\right)}{\left(T_i-T_e\right)}}$ $U={\displaystyle \frac{{\sum }_{j=1}^n{U}_{\left(j\right)}}{n}}$
Thermometric (THM) method		Surface-temperature sensor: 1 Air-temperature sensor: 2 Data logger: 1	$U=h_i{\displaystyle \frac{\left(T_i-T_{si}\right)}{\left(T_i-T_e\right)}}$ $U={\displaystyle \frac{{\sum }_{j=1}^n{U}_{\left(j\right)}}{n}}$

):

$$R_T=R_{si}+R_1+R_2+\cdots R_n+R_{se}$$

(4)

$$R_T=R_{si}+R_{se}+{\displaystyle \sum _{i=1}^n}{R}_i$$

(5)

where $R_T$ is the total thermal resistance (m² K W⁻¹); $R_{si}$ is the interior surface resistance; $R_1$, $R_2$… $R_n$ are the thermal resistances of each layer; $R_{se}$ is the exterior surface resistance; and ${\displaystyle {\sum }_{i=1}^n}{R}_i$ is the sum of thermal resistances of all layers. When the resistance of the component from surface to surface is required, the surface resistances are omitted in Equations (4) and (5).

The thermal resistance of each layer can be obtained from its thermal conductivity using Equation (6):

$$R={\displaystyle \frac{d}{\lambda }}$$

(6)

where R is the design thermal resistance of the layer (m² K W⁻¹), d is the thickness of the layer in the component (m), and λ is the design thermal conductivity of the material (W m⁻¹ K⁻¹), which can be calculated or obtained from tabulated values [7].

The surface thermal resistances depend on the direction of the heat flow and can be obtained from

Table 2. Surface thermal resistance [7].

Surface Thermal Resistance (m2 K W−1)	Direction of Heat Flow
	Upwards (Roof or Floor)	Horizontal (Wall)	Downwards (Roof or Floor)
${\mathit{R}}_{\mathit{s}\mathit{i}}$	0.10	0.13	0.17
${\mathit{R}}_{\mathit{s}\mathit{e}}$	0.04	0.04	0.04

. The surface thermal resistances apply to surfaces in contact with air.

To apply the calculation method, the thickness and thermal conductivity of each layer must be known. Determining layer thickness may require destructive investigation, which is not always feasible, and laboratory measurement of thermal conductivity is often difficult and expensive. In addition, building materials may not be thermally homogeneous. Consequently, the calculation method may provide only approximate values of thermal resistance (R-value) and thermal transmittance (U-value), particularly for buildings constructed with traditional materials and techniques [13,14,27].

In the calculation method, the U-value (U) can be determined from Equation (7) or (8):

$$U={\displaystyle \frac{1}{R_T}}$$

(7)

$$U={\displaystyle \frac{1}{R_{si}+R_{se}+{\displaystyle {\sum }_{i=1}^n}{R}_i}}$$

(8)

When the building component includes an air layer, Equation (9) or (10) is used, where $R_{\mathrm{a}\mathrm{i}\mathrm{r}}$ is the thermal resistance of the air layer:

$$U={\displaystyle \frac{1}{R_T+R_{\mathrm{a}\mathrm{i}\mathrm{r}}}}$$

(9)

$$U={\displaystyle \frac{1}{{(R}_{si}+R_{se}+{\displaystyle {\sum }_{i=1}^n}{R}_i)+R_{\mathrm{a}\mathrm{i}\mathrm{r}}}}$$

(10)

When the U-value of the building component from surface to surface is required, the surface thermal resistances are omitted in Equations (7)–(10).

2.2. Heat Flux Meter (HFM) Method

The heat flux meter (HFM) method, standardized in ISO 9869-1 [8], is used to determine the thermal transmittance and thermal resistance of building envelopes and elements through in situ measurements. In this method, the U-value is obtained by measuring the heat flux through a building element using a heat flow meter together with the air temperatures on both sides of the element. Therefore, a heat flux sensor, two ambient air-temperature sensors, and a data logger are required.

Since true steady-state conditions are rarely achieved under real environmental conditions [14,17,28], ISO 9869-1 [8] recommends performing measurements over a sufficiently long period so that the averaged values provide a reliable approximation of steady-state behavior. The standard further states that the method is valid when the change in stored heat within the element becomes negligible compared with the heat passing through the element during the measurement period.

According to ISO 9869-1 [8], the minimum recommended measurement duration is 72 h for lightweight building elements, while longer monitoring periods are generally required for heavy structures with high thermal inertia. In the present study, measurements were conducted continuously from 3 to 11 February 2025, exceeding the minimum duration recommended by the standard. Furthermore, during the monitoring period, the interior–exterior temperature difference should remain greater than 15 °C, which is generally considered favorable for reliable in situ U-value determination [29].

In accordance with ISO 9869-1 [8], the heat flux sensor is installed on the interior side of the wall, where the thermal conditions are more stable. The sensor is mounted in direct thermal contact with the wall surface in order to minimize contact resistance and measurement disturbance. Interior and exterior air temperatures are simultaneously recorded using ambient temperature sensors [30]. In addition, the measurement location is selected away from visible thermal bridges, cracks, and local defects, following the installation recommendations of ISO 9869-1 [8].

Heat flux (q) always flows from the warmer side to the colder side. During winter conditions, $T_i>T_e$, and heat flows from indoors to outdoors, resulting in heat loss. During summer conditions, $T_i<T_e$, heat flows from outdoors to indoors, resulting in heat gain. In this case, the absolute value $\left|T_i-T_e\right|$ is used in Equation (11) to avoid negative U-values [8].

When n measurements are taken during the test period, the U-value is obtained using the progressive average method using Equation (11) [8,12]:

$$U={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n}{q}_{\left(j\right)}}{{\displaystyle {\sum }_{j=1}^n}(T_{i\left(j\right)}-T_{e\left(j\right)})}}$$

(11)

where U is the thermal transmittance (W m⁻² K⁻¹); q_(j) is the measured heat flux at time step j (W m⁻²); T_i(j) and T_e(j) are the indoor and outdoor air temperatures at time step j (K); j = 1, …, n denotes the time index of the measurement; and n is the total number of measurements.

2.3. Infrared Thermography (IRT) Method

In the infrared thermography (IRT) method, the thermal transmittance, and thermal resistance of building envelopes, components, and elements are determined in steady state by measuring the interior surface temperature using an infrared camera and the interior and exterior air temperatures using two air-temperature sensors. All temperature data from the sensors are recorded and stored by a data logger for subsequent analysis (Table 1). Unlike other methods, heat flux is not directly measured in IRT [9].

In this method, Equation (12) is used for each measurement, in which it is assumed that the heat transfer from the building component to the thermal camera sensor is due to thermal radiation, as well as thermal convection [24]. In this method, an infrared camera connected to software running on a laptop is used to continuously measure and record temperature.

$$U={\displaystyle \frac{4\epsilon \sigma T_{si}^3\left(T_{si}-T_{ref}\right)+h_{ci}\left(T_{si}-T_i\right)}{\left(T_i-T_e\right)}}$$

(12)

where U is the thermal transmittance (W m⁻² K⁻¹); ε is the thermal emissivity (-); σ is the Stefan–Boltzmann constant (W m⁻² K⁻⁴); T_si is the interior surface temperature (K); T_ref is the reflected temperature, measured using a reflective target such as aluminum foil (K); T_i is the interior air temperature (K); T_e is the exterior air temperature (K); and h_ci is the convective heat transfer coefficient (W m⁻² K⁻¹). In Equation (12), the interior total heat-transfer coefficient is expressed as the sum of radiative and convective components. Therefore, only the convective heat transfer coefficient, $h_{ci}$, is used, while the radiative coefficient is explicitly calculated by the term $4\epsilon \sigma T_{si}^3$.

At the end of the test, the average of all measurement results should be computed to obtain the U-value using Equation (13):

$$U={\displaystyle \frac{{\sum }_{j=1}^n{U}_{\left(j\right)}}{n}}$$

(13)

where U_(j) is the U-value obtained from the j-th individual measurement, and n is the total number of measurements. The reflected temperature of the environment, $T_{ref}$, is measured using an infrared camera directed at a piece of crumpled and then unfolded aluminium foil attached to the surface [10,31]. The thermal emissivity, $\epsilon$, is the ratio of the radiance of the building surface to the radiance of a black target at the same temperature over the same spectral interval, using specialized software or equations [31,32]. The Stefan-Boltzmann constant is 5.67 × 10⁻⁸ (W m⁻² K⁻⁴), and the value of the horizontal convective heat transfer coefficient is 2.5 (W m⁻² K⁻¹) [7].

2.4. Infrared Thermometer (IRTM) Method

In the infrared thermometer (IRTM) method, the interior and exterior air temperatures, and the interior surface temperature, are measured, but heat flux is not. Using a non-contact digital laser point infrared thermometer (digital pyrometer) allows the measurement of surface temperature from a short distance without physical contact. Concurrently, the air temperatures are recorded by air-temperature sensors connected to a data logger (Table 1). In the IRTM method, as in the IRT method, measurements are carried out under steady-state conditions. For each measurement, the U-value is computed using Equation (12). Afterwards, the average of all measurements is calculated to determine the U-value using Equation (13).

2.5. Thermometric (THM) Method

In the thermometric (THM) method, the interior and exterior air temperatures, as well as the interior surface temperature, are measured, while heat flux is not directly measured. The surface-temperature sensor is installed in direct contact with the building element surface and allows for continuous temperature measurement. Along with this, the interior and exterior air temperatures are measured using two ambient temperature sensors. All sensors are connected to a data logger for continuous data acquisition and storage (Table 1). In the thermometric (THM) method, the U-value is determined by measuring the interior surface temperature using a sensor such as a thermocouple or NTC, along with the interior and exterior air temperatures using two ambient temperature sensors (Table 1). The interior total heat-transfer coefficient values presented in

Table 3. Interior total heat-transfer coefficient.

Interior Total Heat-Transfer Coefficient (W m−2 K−1)	Direction of Heat Flow
	Upwards (Roof)	Horizontal (Wall)	Downwards (Floor)
${\mathit{h}}_{\mathit{i}}$	10	7.69	5.88

can be used in Equations (14) and (15) [21,22,23,33].

Equation (14) relates the heat flux to the interior air temperature and the interior surface temperature [9,32]:

$$q=h_i\left(T_i-T_{si}\right)$$

(14)

where q is the heat flux (W m⁻²), $h_i$ is the interior total heat-transfer coefficient (W m⁻² K⁻¹), $T_i$ is the interior air temperature (K), and $T_{si}$ is the interior surface temperature (K).

By substituting Equation (14) in Equation (3), the U-value is given by Equation (15):

$$U=h_i{\displaystyle \frac{\left(T_i-T_{si}\right)}{\left(T_i-T_e\right)}}$$

(15)

where U is the thermal transmittance (W m⁻² K⁻¹), $h_i$ is the interior total heat-transfer coefficient (W m⁻² K⁻¹), $T_i$ is the interior air temperature (K), $T_{si}$ is the interior surface temperature (K), and $T_e$ is the measured exterior air temperature (K).

In contrast to Equation (12), where the radiative and convective coefficients are treated separately, Equation (15) uses the interior total heat-transfer coefficient, $h_i$, which includes both convective and radiative coefficients ($h_i=h_{ci}+h_{ri}$), as provided by ISO 6946 [7].

Taking into consideration that the interior surface thermal resistance, $R_{si}$, is the reciprocal of the interior total heat-transfer coefficient, $h_i$ (W m⁻² K⁻¹; Equation (16)), and the values of $R_{si}$ set by ISO 6946 [7] for upwards (roof), horizontal (wall), and downwards (floor) heat flow are 0.13, 0.10, and 0.17 m² K W⁻¹, respectively (Table 2), $h_i$ values are obtained as 10, 7.69 and 5.88 W m⁻² K⁻¹, as indicated in Table 3. Hence, $h_i$ = 7.69 (W m⁻² K⁻¹) is used for the wall in this study. It should be noted that the interior total heat-transfer coefficient, hi, was assumed to be spatially uniform over the investigated wall surface according to ISO 6946 [7]. Therefore, possible local variations in convective heat transfer between different wall zones were not directly captured in the THM calculations. Such effects may contribute to slight variations in the calculated U-values under real experimental conditions and are considered indirectly within the uncertainty analysis presented in the section entitled “Uncertainty Estimation”.

$$h_i={\displaystyle \frac{1}{R_{si}}}$$

(16)

The measurement of interior surface, interior, and exterior air temperatures, and the calculation of the U-value should be repeated several times. An average of calculations according to Equation (13) gives the U-value.

2.6. Research Concept and Methodological Framework

This study follows a comparative multi-method framework to evaluate the performance of four in situ U-value measurement techniques under identical real boundary conditions. The research concept is based on applying different measurement principles to the same building element and comparing their resulting U-values, uncertainties, and practical limitations.

The methodological framework consists of four main steps. First, the theoretical basis and measurement requirements of the selected methods, HFM, IRT, IRTM, and THM, are defined. Second, all methods are applied to the same external double-leaf brick wall under comparable indoor and outdoor boundary conditions. Third, the recorded air temperature, surface temperature, and heat flux data are processed to calculate the U-value for each method using the corresponding equations. Fourth, the calculated U-values are compared in terms of agreement, uncertainty, sensitivity to environmental conditions, and ability to detect spatial variations in the wall.

This framework establishes a direct link between measurement theory, experimental implementation, uncertainty evaluation, and comparative performance assessment under real field conditions.

To further formalize the research design, this study can be interpreted as a comparative experimental framework for in situ thermal characterization under controlled field conditions. The framework integrates five complementary approaches with distinct functional roles: the CAL method provides a theoretical reference baseline derived from material properties; the HFM method serves as the experimental reference benchmark based on direct heat flux measurement; the IRT and IRTM methods represent radiative inference approaches relying on surface temperature field acquisition; and the THM method provides a simplified contact-based approximation of heat transfer behavior.

Within this framework, all methods are applied in parallel to the same building element under identical boundary conditions, enabling direct comparability of outputs. The research workflow follows a structured input–process–output logic, where measured physical quantities (temperature and heat flux) are transformed into U-values using method-specific governing equations, followed by a multi-level evaluation of agreement, uncertainty propagation, and spatial variability across measurement zones. This system-based formulation ensures that the study is not only a measurement campaign, but also a structured benchmarking of methodological performance under real environmental conditions.

A schematic representation of the methodological framework is provided in Figure 1, illustrating the sequential and parallel structure of the experimental and analytical workflow.

3. In Situ Experimental Tests

Measurements were conducted from 3 to 11 February 2025 on the north-facing wall of an office building at the University of Stuttgart, Germany, as shown in Figure 2a. The wall consists of a double-leaf hollow clay brick masonry system with an overall thickness of approximately 0.30 m. As illustrated in Figure 2b,c, the wall assembly comprises an inner hollow clay brick layer with a thickness of approximately 0.115 m, followed by a 0.04 m air cavity, a 0.03 m cork insulation layer, and an outer hollow clay brick layer with a thickness of approximately 0.115 m. Due to the perforated geometry of the hollow clay bricks and the non-uniform distribution of mortar within the perforations, the wall cannot be considered a fully homogeneous or monolithic element.

During the measurement period, the indoor thermal environment was maintained under stable conditions using an electric heater positioned near the center of the room after the building’s central heating system, adjacent to the test wall, had been switched off. The indoor air temperature remained relatively constant throughout the test period, varying between approximately 21.7 °C and 23.2 °C.

The outdoor climatic conditions were representative of winter boundary conditions in Stuttgart during February, with exterior air temperatures ranging from approximately −4 °C to 9.2 °C. The measurements were conducted on a north-facing façade in order to minimize the influence of direct solar radiation. The difference between indoor and outdoor air temperatures remained above 15 °C for almost the entire measurement period, thereby ensuring suitable thermal conditions for reliable steady-state evaluation. Four in situ U-value measurement methods were applied during this period. The heat flow meter (HFM) and thermometric (THM) methods were installed and operated continuously throughout the entire testing period. In contrast, the infrared thermometer (IRTM) and infrared thermography (IRT) methods were conducted in two separate phases to avoid overlap and minimize measurement interference, particularly due to the influence of human presence on infrared readings.

From 3 February 2025, 13:35, to 7 February 2025, 10:15, the IRTM measurement campaign was conducted. Unlike the continuously logged HFM and THM methods, the IRTM measurements were performed manually at predefined time intervals using a handheld pyrometer (Testo—830 T2 (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)). Measurements were repeated several times per day under stable thermal conditions, and the recorded values were subsequently processed together with the simultaneously measured air temperatures.

From 7 February 2025, 13:35, to 11 February 2025, 10:15, the IRT method was applied using continuous thermographic image acquisition at one-minute intervals.

The HFM and THM methods were continuously active from 3 to 11 February 2025.

The IRT method was intentionally scheduled after completing the THM measurements to ensure that the infrared camera had an unobstructed view of the test surface. This scheduling strategy was designed to eliminate any visual obstructions and avoid thermal disturbances caused by human presence during IRT imaging.

An overview of the measurement schedule and method combinations is shown in Figure 3.

During the test periods, HFM sensors were installed in both the upper and lower zones of the wall to enable vertical comparison. Similarly, infrared thermography (IRT) measurements were conducted at both heights using appropriately positioned thermal cameras. This setup was intended to examine whether the wall exhibits homogeneous thermal properties along its height. In general, the instrument arrangement follows the recommendations of Bienvenido-Huertas et al. [21].

In accordance with these recommendations, the measurement positions were selected in representative central areas of the wall surface in order to minimize the influence of edges, corners, and possible thermal bridges. The sensors and heat flux plates were positioned on the brick surface rather than directly on mortar joints to avoid local thermal irregularities associated with mortar distribution. In addition, the measurement points were arranged at the same height level on the wall to reduce the influence of vertical temperature stratification and convective effects. The wall area below the window was additionally insulated during the measurements to minimize possible thermal bridge effects from the window connection region. The exterior sensors were installed on the north-facing façade and positioned to avoid direct solar radiation, in accordance with standard recommendations for reliable in situ thermal transmittance measurements under stable boundary conditions.

A total of two contact sensors for indoor surface temperature (Testo Thermocouples (Testo SE & Co. KGaA, Titisee-Neustadt, Germany) and NTC (Testo SE & Co. KGaA, Ti-tisee-Neustadt, Germany)), two sensors for air temperature (Sensirion SHT25 (Sensirion AG, Stäfa, Switzerland) and Testo NTC (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)), two heat flux plates (Ahlborn FQA018 (Ahlborn Mess- und Regelungstechnik GmbH, Holzkirchen, Germany)), one thermal-imaging camera, and one pyrometer (non-contact digital laser point infrared thermometer) were used. The three data loggers, A to C, recorded a measured value every 10 min. The base module, to which the sensors for the air temperature and the heat flux plates were connected, also sent data online every ten minutes to a master module, i.e., the wireless measurement node in Figure 4a,b. As shown in Figure 4a, the infrared camera was about 1.20 m away from the wall surface and connected to an adjacent computer that had opened a thermography program (IRBIS) produced by the manufacturer InfraTec (InfraTec GmbH, Dresden, Germany). The software saved thermal images, also known as thermograms, of the front wall surface every minute. Such a snapshot is shown in Figure 4d as an example. A crumpled and unfolded aluminum foil was placed at the upper zone and lower zone of the wall to measure the reflected temperature of the environment, T_ref. The thermal emissivity was set to 0.9 for the double-leaf brick wall surface temperature and to 1.0 for the reflected environmental temperature from the aluminum foil, for both the pyrometer and the thermal camera [15,17].

The pyrometer measurements used in the IRTM method were carried out manually at predefined measurement intervals during the monitoring period. At each measurement session, the surface temperatures of the predefined wall locations were recorded together with the corresponding interior and exterior air temperatures. Since the IRTM method relies on discrete point-based measurements rather than continuous image acquisition, only selected representative measurement periods are illustrated in the time-series figures. Consequently, the graphs shown in Figure 5 and Figure 6 represent representative datasets extracted from the full monitoring campaign rather than continuous measurements over the entire IRTM campaign period. On the exterior side of the wall, the measurement setup mirrored the arrangement on the interior to enable direct comparison. Two surface-temperature sensors (Testo thermocouples (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)) were installed at the same vertical positions as the indoor surface-temperature sensors. Additionally, two air-temperature sensors (Sensirion SHT25 (Sensirion AG, Stäfa, Switzerland), Testo NTC (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)) were placed outside to capture the ambient outdoor temperature under real environmental conditions in Figure 4c.

Data are recorded at ten-minute intervals using the data logger. Following data processing, U-values are calculated and compared across methods.

Table 4. Devices used to measure thermal data of the double-leaf brick wall.

Instrument	Model	Specifications
Thermal camera	InfraTec—VarioCAM hr (InfraTec GmbH, Dresden, Germany)	Detector: Uncooled Microbolometer FPA Image resolution: up to 1.280 × 980 pixels Spectral range: 7.5 to 14 µm Temperature measurement range: −40 to 1200 °C NETD (thermal resolution): <0.05 °C at 30 °C Measurement accuracy: ±1.5 °C or ±2% of reading (whichever is greater) Image rate: 50/60 Hz
Pyrometer (infrared thermometer)	Testo—830 T2 (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −30 to 400 °C Accuracy: ±1.5 °C or ±1.5% v. Mw. of reading (+0.1 to 400 °C) ±2 °C or ±2% of reading (−30 to 0 °C) (whichever value is greater applies) Resolution: 0.1 °C
Measurement node and base module	Custom MPA Module (Materials Testing Institute, University of Stuttgart, Stuttgart, Germany)	Channels: 4 (2× sensor SHT25 and 2× heat flux plate)
Air-temperature sensor	Sensirion—SHT25 (Sensirion AG, Stäfa, Switzerland)	Temperature measurement range: −40 to 125 °C Accuracy: ±0.2 °C
Heat flux plates	Ahlborn—FQA018 (Ahlborn Mess- und Regelungstechnik GmbH, Holzkirchen, Germany)	Dimensions (mm): 120 × 120 × 3 Temperature resistance: −40 to 80 °C Calibration value: <15 W/m2 ≈ 1 mV Accuracy: 5% at 23 °C
Data logger A	Testo—175 H1 (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −20 to 55 °C
Air-temperature sensor	Testo—NTC (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −20 to 50 °C Accuracy: ±0.5 °C Resolution: 0.1 °C
Data logger B	Testo—175 T3 (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −20 to 55 °C
Surface-temperature sensor	Testo—Thermocouples (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Materials: copper and copper–nickel Temperature measurement range: −50 to 250 °C Accuracy: ±0.5 °C (−50 to +70 °C), ±0.7% v. Mw. (+70.1 to 400 °C) Resolution: 0.1 °C
Data logger C	Testo—175 T2 (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −35 to 55 °C
Surface-temperature sensor	Testo—NTC (Testo SE & Co. KGaA, Titisee-Neustadt, Germany)	Temperature measurement range: −50 to 80 °C Accuracy: ±0.2 °C (−50 to 80 °C) Resolution: 0.1 °C

lists a detailed summary of the measurement instruments used.

Figure 5 shows the measurement results over the period from 3 to 11 February 2025. In Figure 5a, T_i is the average of interior air temperature by data logger A and measurement node and base module; T_si is the average of the interior surface temperature by the surface sensor data logger B (thermocouple) and data logger C (NTC); T_si,cam is the interior surface temperature from the thermal image; T_ref is the reflected temperature of the environment recorded by the infrared camera from aluminum foil; T_si,pyr is the interior surface temperature from the pyrometer; T_se is the average of the exterior surface temperature measured by surface sensors from data logger B (two thermocouples surface sensors); and T_e is the average exterior air temperature by data logger A and measurement node and base module. For the IRTM method using the pyrometer, an emissivity of 0.9 was applied for measuring the double-leaf brick wall surface temperature, while an emissivity of 1.0 was used for the reflected environmental temperature from the aluminum foil. For clarity, the pyrometer-based IRTM data shown in Figure 5a correspond to representative measurement sessions conducted during the overall monitoring campaign between 3 and 7 February 2025. Figure 5a illustrates the temperatures measured by different methods. The following aspects of the temperature measured values are observed:

• During the test period, the value of the interior air temperature was constant but with slight variations between 21.7 °C and 23.2 °C.
• The exterior air temperature varied between −4 °C and 9.2 °C.
• According to recommendations from the literature, the temperature difference between the interior and exterior was always greater than 15 °C throughout the experiment [21], except for 45 min on 10.02.2025 from 13:45 to 14:25 in the text, when it ranged from 13.6 °C to 14.9 °C.
• The exterior surface temperature measured by a surface sensor was always, as expected, higher than the exterior temperature, with a variation of −1.6 °C to 8.8 °C. The temperature difference is more pronounced during cooler hours of the day and night.
• For the interior surface temperature, the values measured by thermal camera (IRT method), pyrometer (IRTM method), and the sensor (THM method) are generally close to each other. At the wall’s upper zone, the measured interior surface temperatures are approximately from 20.1 °C to 22.6 °C (IRT), 20 °C (IRTM), and from 19.8 °C to 20.8 °C (THM).
• The interior surface temperature of the lower zone measured by the IRT method is about 20.7 °C, which is around 0.5 °C lower than the upper zone’s interior surface temperature, 21.2 °C, measured by the same method.
• The reflected temperature of the environment recorded by the infrared camera (IRT) in the upper zone ranges from 22.3 °C to 24.8 °C, which is slightly higher than the corresponding values in the lower zone (21.6 °C to 24.2 °C).

The heat flux (q) shown in Figure 5b demonstrates a rapid initial increase, followed by a quasi-steady state where it slowly decreases with a constant slope after approximately 24 h. This trend is consistent for both the upper and lower zones. The initial rapid increase, leading into this quasi-steady condition, indicates the thermal mass of the double-leaf brick wall reaching an evolving equilibrium with the changing temperature difference between the warm interior and the exterior. The observed constant decrease in heat flux is likely due to a continuous, gradual increase in the exterior air temperature over the experimental period, as shown in Figure 5a, which slowly reduces the overall temperature difference driving heat transfer through the wall.

The heat flux of the upper zone, ranging from 11.60 W m⁻² to 25.50 W m⁻², is always slightly lower than that of the lower zone, from 13 W m⁻² to 25.70 W m⁻².

Uncertainty Estimation

The uncertainty associated with the investigated in situ measurement methods was evaluated following the ISO/IEC Guide 98-3 (GUM) [34]. The thermal transmittance (U-value) was calculated according to the corresponding equations presented in the previous section for each method.

The first type of uncertainty was determined from the statistical analysis of the measured time series and calculated as the standard deviation divided by the square root of the number of observations (Type A evaluation). The second type of uncertainties was derived from the manufacturers’ specifications for the measuring instruments, such as the heat flux plates, temperature sensors, infrared camera, and infrared thermometer (Type B evaluation).

The combined standard uncertainty, u_c, was calculated from Equation (17) using the law of propagation of uncertainty according to the ISO/IEC Guide 98-3 (GUM) [31]:

$$u_c\left(U\right)=\sqrt{\sum {\left(c_{\mathrm{x}}u\right(x\left)\right)}^2}$$

(17)

where u(x) represents the standard uncertainty of the input quantity, x; and c_x = ∂U/∂x denotes the sensitivity coefficient.

The reported uncertainty values correspond to the combined standard uncertainty coverage factor, k = 1, representing one standard deviation of the estimated U-value.

Due to differences in the measurement principles, the dominant sources of uncertainty differ between the investigated methods. The specifications of the sensors, including their measurement accuracy, are presented in Table 4 in the previous section.

For the HFM method, the uncertainty analysis includes the heat flux (q) and the temperature measurements of interior air (T_i) and exterior air (T_e).

For the IRT and IRTM methods, the input quantities considered are the surface temperature (T_si), interior and exterior air temperatures (T_i and T_e), emissivity (ε), reflected temperature (T_ref), and the interior convective heat transfer coefficient (h_ci).

For the THM method, the uncertainty contributions are associated with surface temperature (T_si), interior and exterior air temperatures (T_i and T_e), and the interior total heat-transfer coefficient (h_i). Uncertainty budget and sensitivity coefficients for all methods are presented in

Table 5. Uncertainty budget and sensitivity coefficients for U-value estimation.

Quantity	Symbol	Standard Uncertainty u(x)	Sensitivity Coefficient cx = ∂U/∂x	Contribution cxu(x)
Heat flux	q	u(q)	∂U/∂q	cqu(q)
Interior air temperature	Ti	u(Ti)	∂U/∂Ti	cTiu(Ti)
Exterior air temperature	Te	u(Te)	∂U/∂Te	cTeu(Te)
Interior surface temperature	Tsi	u(Tsi)	∂U/∂Tsi	cTsiu(Tsi)
Convective heat transfer coefficient	hci	u(hci)	∂U/∂hci	chciu(hci)
Reflected temperature	Tref	u(Tref)	∂U/∂Tref	cTrefu(Tref)
Emissivity	ε	u(ε)	∂U/∂ε	cεu(ε)
Interior total heat-transfer coefficient	hi	u(hi)	∂U/∂hi	chiu(hi)

.

4. Results and Discussion

This section summarizes the findings from the in situ measurement campaign. First, the recorded time-series data for each method are presented, followed by a comparative evaluation of the U-values determined using HFM, IRT, IRTM, and THM.

4.1. Experimental Conditions and Instantaneous U-Value Measurements

The recorded time-series data for each method, including interior and exterior air temperatures, surface temperatures, heat flux, and the continuously calculated U-value, are shown in Figure 6. These plots illustrate both the thermal behavior of the wall and the instantaneous U-value trends obtained by each technique.

4.2. Comparative Analysis and Discussion of U-Value Measurements

This section presents a comparative analysis of the U-values obtained using four in situ methods: the heat flow meter (HFM), the infrared thermography (IRT), the infrared thermometer (IRTM), and the thermometric method (THM). All measurements were performed under steady-state conditions.

Figure 7 presents the results for the upper zone, and

Table 6. Determined U-values from HFM, IRT, IRTM, and THM methods at the upper zone.

Method	U-Value (W m−2 K−1) Upper Zone	Uncertainty (k = 1)
Heat flow meter (HFM)	0.77 ± 0.03	3.9%
Infrared thermography (IRT)	0.62 ± 0.43	69%
Infrared thermometer (IRTM)	0.69 ± 0.40	62%
Thermometric (THM)	0.67 ± 0.04	5.9%

summarizes the numerical values.

The calculated U-values for the upper zone are 0.77 ± 0.03 W m⁻² K⁻¹ (HFM), 0.62 ± 0.43 W m⁻² K⁻¹ (IRT), 0.69 ± 0.40 W m⁻² K⁻¹ (IRTM), and 0.67 ± 0.04 W m⁻² K⁻¹ (THM).

The closest agreement is between IRTM and THM, with a difference of only 0.02 W m⁻² K⁻¹, corresponding to about 3%. In contrast, HFM yields the highest U-value among all methods. HFM produces the highest U-value, exceeding IRT by 0.15 W m⁻² K⁻¹ (approximately 24%), IRTM by 0.08 W m⁻² K⁻¹ (approximately 11%), and THM by 0.10 W m⁻² K⁻¹ (approximately 15%). The lowest U-value is given by IRT, which is 0.15 W/(m²·K) lower than HFM (−24%) and 0.07 W m⁻² K⁻¹ lower than IRTM (−10%). The overall spread between the highest (HFM) and lowest (IRT) results is 0.15 W m⁻² K⁻¹, representing a relative variation of around 24%.

As shown in Figure 6, the surface temperature measured by the pyrometer in the IRTM method is approximately 0.5–1.0 °C lower than the corresponding values obtained from the other methods.

This deviation can be explained by the spot-based measurement principle of pyrometers, which capture temperature from a highly localized surface area and are therefore more sensitive to localized anomalies, such as areas with reduced thermal mass, darker finishes, or slight internal air movements.

Consequently, even a slight underestimation of surface temperature increases the (Ti − Tsi) temperature gradient, leading to an overestimation of the calculated U-value.

Furthermore, the IRTM method is inherently sensitive to the accuracy limits of the measurement equipment. With an average surface temperature of approximately 20 °C and an instrument accuracy of ±1.5 °C for the temperature sensors, the resulting uncertainty in the calculated U-value is approximately 62% in this case.

This level of sensitivity is typical for non-contact, point-based measurement devices.

Additional factors, such as emissivity variations or reflections from nearby heat sources, can further affect the results, particularly in indoor environments, where radiative heat exchange contributes significantly to the overall heat transfer.

However, despite a similar instrumental accuracy limit, the IRT method employs thermal imaging to derive a spatially averaged surface temperature over a wider area, thereby reducing the influence of localized anomalies and yielding more stable measurements. This averaging effect may explain the lower U-value observed in this study, particularly when small high-loss regions are diluted within a broader surface pattern.

However, despite a similar instrumental accuracy limit (±1.5 °C), the IRT method exhibits a higher relative uncertainty in the calculated U-value (approximately 69%) compared to the IRTM method (approximately 62%).

This difference can be attributed to the greater variability in the spatially distributed temperature data that is reflected in the Type A component of the uncertainty analysis within the GUM framework. The use of thermographic measurements inherently introduces additional scatter due to the larger number of recorded data points, capturing increased temporal and spatial variability in surface temperature.

In contrast, the IRTM method relies on discrete manual point-based measurements obtained at predefined time intervals rather than continuous spatial monitoring, resulting in lower statistical dispersion and a reduced Type A uncertainty component.

The THM method, which combines surface contact thermocouples with air temperature measurements, produced results close to those obtained with the IRTM method, but slightly lower. This behavior is consistent with the neglect of radiative heat transfer effects in the THM formulation, which may lead to an underestimation of the total heat loss, particularly under conditions where radiative exchange is significant.

Furthermore, the THM method exhibits a significantly lower uncertainty (±0.04 W m⁻² K⁻¹, approximately 5.9%) compared to the IRTM method (±0.40 W m⁻² K⁻¹, approximately 62%). This can be attributed to the simplified formulation of the THM approach, which relies primarily on a limited number of input parameters, such as interior air and surface temperatures, both of which can be measured with relatively high accuracy (±0.5 °C).

In contrast, the instrumental error margin of the IRTM method (±1.5 °C) is approximately three times greater, which, within the GUM-based uncertainty propagation framework, significantly amplifies the resulting uncertainty in the calculated U-value. Since temperature differences directly influence the heat transfer gradient, even small increases in the input temperature error lead to a substantial amplification of the propagated uncertainty.

Similar discrepancies between HFM and IRT have been reported in the literature [16,29]. In the present study, the closest agreement is between IRTM and THM, while both show larger deviations from HFM and IRT. This highlights that although different methods can yield comparable results under certain conditions, their agreement depends strongly on the specific measurement principles and environmental factors involved.

This interpretation is further supported by recent experimental studies on aerial infrared thermography which demonstrate that infrared-based methods can achieve close agreement with thermometric approaches when measurements are conducted under controlled conditions with sufficient thermal gradients. In particular, deviations as low as approximately 4–5% have been reported during stable winter nighttime conditions, while significantly larger deviations occur under low thermal gradients or transient daytime conditions [20]. This confirms that the discrepancies observed in the present study are consistent with the known sensitivity of infrared methods to environmental stability rather than indicating a fundamental limitation of the technique.

To compare the determined U-values at the upper and lower zones of the wall, the results obtained from the heat flow meter (HFM) and infrared thermography (IRT) methods are analyzed. The comparison is illustrated in Figure 8, and the corresponding numerical values are summarized in

Table 7. Determined U-values from the HFM and IRT methods at the upper zone and lower zone.

Method	U-Value (W m−2 K−1)
	Upper Zone	Lower Zone
Heat flow meter (HFM)	0.77 ± 0.03	0.83 ± 0.04
Infrared thermography (IRT)	0.62 ± 0.43	0. 68 ± 0.43

.

For the HFM method, the U-values are 0.77 W m⁻² K⁻¹ in the upper zone and 0.83 W m⁻² K⁻¹ in the lower zone. Similarly, the IRT method yields values of 0.62 W m⁻² K⁻¹ and 0.68 W m⁻² K⁻¹ for the upper and lower zones, respectively. Both methods consistently indicate that the upper zone exhibits a lower U-value than the lower zone.

For both methods, the U-value increases by 0.06 W m⁻² K⁻¹ from the upper to the lower zone, corresponding to approximately 7.8% for HFM and 9.7% for IRT. This consistent trend across independent measurement techniques suggests that the observed difference is not method-dependent, but rather reflects a genuine thermal inhomogeneity within the wall structure.

This observation is further supported by the heat flux measurements, as shown in Figure 5b, where the upper zone exhibits lower heat flux values than the lower zone. According to Equation (3), this directly leads to a lower calculated U-value in the upper zone, which is consistent with the measured data.

When comparing the two methods, the IRT method systematically yields lower U-values than the HFM method in both zones. In the upper zone, the IRT result is lower by 0.15 W m⁻² K⁻¹ (approximately 19.5%), while in the lower zone, the difference is also 0.15 W m⁻² K⁻¹ (about 18.1%).

This consistent offset suggests a systematic difference between the two measurement approaches rather than random measurement error. Despite this discrepancy, both methods exhibit a consistent trend in identifying the relative differences between zones, reinforcing the reliability of the observed spatial variation. Such discrepancies between infrared and contact-based measurements have also been reported in previous studies [17,23,32].

Furthermore, the relatively high uncertainty associated with the IRT method in this study (approximately 69%) should be interpreted in the context of uncertainty propagation rather than measurement inconsistency. Infrared-based methods are highly sensitive to input parameters such as emissivity, reflected temperature, and surface temperature gradients, which can significantly amplify the propagated uncertainty in the calculated U-value. However, as demonstrated in recent research, when environmental conditions are stable—particularly under high interior–exterior temperature differences—infrared thermography can still yield consistent and reliable U-value estimates despite large theoretical uncertainty bounds [18]. This behavior is consistent with the present results, where the IRT method shows reasonable agreement with the HFM method while exhibiting a wider uncertainty range, indicating that high propagated uncertainty does not necessarily imply poor measurement reliability.

For reference, the analytically calculated U-value based on the wall construction details is approximately 0.71 W m⁻² K⁻¹. This value serves as a theoretical benchmark and shows good agreement with the experimental results, particularly those obtained using contact-based methods. The comparison of the four in situ techniques reveals consistent trends in U-value estimation, despite differences in measurement principles and associated uncertainties. Methods based on direct measurements or simplified formulations tend to yield values closer to the analytical reference, whereas infrared-based approaches exhibit larger deviations due to their higher sensitivity to surface-temperature estimation and radiative effects.

It should be noted that the analytical calculation assumes thermally homogeneous material properties and idealized boundary conditions, and therefore, it cannot capture local variations within the wall. In contrast, in situ measurements provide additional insight into the actual thermal behavior by revealing spatial differences between wall zones.

These findings highlight the importance of selecting an appropriate measurement technique based on the required balance between accuracy, uncertainty, and practical applicability, rather than relying on a single method.

5. Conclusions and Further Research

Accurate determination of the thermal transmittance (U-value) of existing building envelopes is essential for reliable energy performance assessment and the planning of energy-efficient refurbishment measures. However, practical limitations, including unknown material properties and on-site measurement constraints, complicate the application of in situ methods.

Although several techniques have been proposed, systematic experimental comparisons under real operating conditions remain limited.

This study presents a comprehensive experimental comparison of four in situ U-value measurement methods, namely heat flow meter (HFM), infrared thermography (IRT), infrared thermometer (IRTM), and thermometric method (THM), applied to a double-leaf brick wall at the University of Stuttgart. Measurements were conducted under real boundary conditions over several days, including continuous monitoring of air temperature, surface temperature, and heat flux.

The results demonstrate that, while all methods yield U-values within a comparable range, significant differences arise in terms of uncertainty and sensitivity to environmental conditions. Contact-based methods (HFM and THM) provide more stable and reliable estimates, while infrared-based methods (IRT and IRTM) exhibit higher propagated uncertainty due to their sensitivity to emissivity, reflected temperature, and sensor accuracy. Furthermore, the analysis revealed consistent spatial variations between different wall zones, indicating thermal inhomogeneity that cannot be captured by analytical calculations, thereby emphasizing the importance of in situ measurements for realistic thermal performance assessment.

The findings highlight that no single method is universally optimal, and that method selection should consider the trade-off between accuracy, measurement practicality, and environmental sensitivity. This study provides practical guidance for selecting appropriate in situ techniques and contributes to improving the reliability of thermal performance assessment in existing buildings.

In addition, some assumptions adopted in standardized in situ U-value measurement approaches may not fully capture local thermal variations under real boundary conditions. Therefore, local environmental effects and spatial variations in heat transfer may influence the calculated U-values and contribute to the overall measurement uncertainty.

The results of this study provide direct implications for the application of in situ U-value measurement techniques in building renovation projects. The calculation-based method (CAL) is suitable for preliminary thermal assessments when reliable construction details and material properties are available. However, since CAL does not account for construction defects, material aging, moisture effects, or in situ thermal heterogeneity, its applicability for existing buildings is limited and should preferably be complemented by in situ measurements.

Among the investigated in situ methods, the heat flux meter (HFM) method remains the most reliable and accurate approach due to the direct measurement of heat flux through physical contact with the wall surface and its compliance with ISO 9869-1. Although the HFM method requires more instrumentation and higher sensor cost, it is particularly recommended for post-renovation verification and detailed thermal performance assessment where high accuracy is required.

The thermometric method (THM) also showed acceptable agreement with the reference measurements while requiring simpler and less expensive instrumentation. Since THM is also based on direct contact measurements, it can provide reliable results for practical retrofit applications when the use of heat flux sensors is not feasible.

In situations where direct contact with the wall surface is not permitted, such as historical buildings or protected surfaces, infrared thermography (IRT) represents a suitable non-contact alternative. In addition to U-value estimation, IRT enables rapid qualitative assessment of thermal bridges, façade heterogeneity, insulation defects, and moisture-related anomalies prior to renovation planning.

Finally, when access to thermal-imaging cameras is limited, the infrared thermometer method (IRTM) can be considered a simplified and low-cost alternative. Although its accuracy is lower compared to HFM and THM, the present study demonstrated that the method can still provide acceptable estimations under controlled environmental conditions.

For practical field application, it is recommended that measurements be conducted under winter conditions, with a minimum indoor–outdoor temperature difference of 15 °C, a monitoring duration of at least 72 h (preferably 5–10 days for high-thermal-mass walls), and strict avoidance of solar radiation and high-wind conditions. A multi-zone measurement strategy (upper and lower façade regions) is also recommended to capture potential thermal inhomogeneities. Finally, uncertainty quantification using ISO GUM should be systematically applied when comparing different in situ methods to ensure reliability of retrofit decisions.

The present study is subject to several limitations that should be considered when interpreting the results. The experimental campaign was conducted on a single north-facing double-leaf brick wall under winter climatic conditions with relatively stable interior temperatures and a sufficiently high interior–exterior temperature difference. Consequently, the reported U-values and the observed agreement between methods may be influenced by the specific wall typology, thermal properties, façade orientation, and environmental boundary conditions investigated in this study. In addition, the measurement configuration, including sensor placement, local surface characteristics, and the duration and timing of infrared measurements, may affect the obtained results. Therefore, while the comparative trends between the investigated methods provide valuable insight into their relative behavior and uncertainty characteristics, the quantitative findings should not be directly generalized to all wall systems, climatic conditions, or operational scenarios without further validation.

Future research should focus on reducing uncertainty in infrared-based methods, improving measurement protocols under real boundary conditions, and developing hybrid approaches that combine the advantages of different techniques.

Further studies on different wall constructions, climatic regions, façade orientations, and transient environmental conditions would also help evaluate the broader applicability and robustness of the investigated methods.