The Implications of Non-Constant Hygrothermal Parameters on Heat and Moisture Transfer in Rammed Earth Walls Across Diverse Climate Zones

Abstract

As an eco-friendly natural building material, rammed earth possesses outstanding hygrothermal performance, which plays a vital role in achieving the goals of sustainable architecture. However, most existing simulations assume constant hygrothermal parameters, resulting in considerable discrepancies between predicted and actual energy performance and consequently underestimating the true passive regulatory potential of rammed earth. To enhance the accuracy of energy consumption predictions in rammed earth buildings, this study integrates experimental measurements with dynamic simulations and experimentally determines both the constant and non-constant hygrothermal parameters of rammed earth. By integrating experimental and simulation approaches, this study reveals a strong positive linear correlation between the thermal conductivity of rammed earth and its moisture content (R² = 0.9919), increasing from 0.77 W/(m·K) to 1.38 W/(m·K) as moisture content rises from 0% to 14%, whereas the moisture resistance factor decreases exponentially with increasing relative humidity (RH). Subsequently, the two sets of hygrothermal parameters were implemented in the WUFI-Plus simulation platform to conduct annual dynamic simulations across five representative Chinese climate zones (Harbin, Beijing, Nanjing, Guangzhou, and Dali), systematically comparing the performance differences between the “non-constant” and “constant” parameter models. The results show that the non-constant parameter model effectively captures the dynamic hygrothermal regulation of rammed earth, exhibiting superior passive performance. It predicts substantially lower building energy loads, with heating energy reductions most pronounced in Harbin and Beijing (16.9% and 15.5%) and cooling energy reductions most significant in Guangzhou and Nanjing (15.8% and 15.2%). This study confirms that accurately accounting for the dynamic hygrothermal coupling process is fundamental to reliably evaluating the performance of hygroscopic materials such as rammed earth, providing a robust scientific basis for promoting energy-efficient, low-carbon, and climate-responsive sustainable building design.

1. Introduction

The building sector is one of the primary contributors to global energy consumption and environmental impact. According to a report by the International Energy Agency (IEA), this sector accounts for approximately 40% of global final energy usage and 36% of energy-related CO₂ emissions [1]. In China, building energy consumption accounts for 45.5% of the nation’s total energy consumption, with the building materials production and building operational phases contributing 22.3% and 21.7%, respectively [2]. In the pursuit of China’s “Dual Carbon” goals (peaking carbon emissions and achieving carbon neutrality), the building industry confronts significant challenges, including high energy intensity, substantial carbon emissions, and a heavy reliance on carbon-intensive materials [3]. Consequently, a supply-chain perspective is imperative for mitigating the environmental impact of building materials. The adoption of natural, low-embodied-energy materials is widely regarded as a crucial pathway toward sustainable construction.

Within this context, the earth, as an ancient natural building material, has garnered renewed attention [4,5,6]. Compared to fired bricks and concrete, earth-based materials demonstrate a notable reduction in energy consumption and carbon emissions during production, by 2–15% and 4–18%, respectively [6,7]. They exhibit superior thermal performance, the practicality of using locally sourced materials, low maintenance costs, and high energy efficiency [8,9,10,11]. More significantly, earthen materials possess the inherent capacity to passively regulate indoor ambient conditions through moisture buffering and thermal energy storage/release processes. This passive hygrothermal regulation enhances occupant comfort and can substantially reduce the operational energy demands of buildings [12,13,14]. From a materials science perspective, raw earth is a typical porous medium characterized by an internal structure of pores spanning multiple scales, which constitutes a complex multiscale pore network [15,16,17]. This structural characteristic not only endows the material with significant moisture adsorption and desorption capabilities but also couples the processes of moisture regulation and heat transfer, resulting in a complex hygrothermal coupling mechanism [18,19]. This mechanism enables earthen materials to maintain a relatively stable indoor hygrothermal environment amidst fluctuating external conditions, thereby demonstrating remarkable climate adaptability [20,21]. However, prevailing studies concerning the calculation and simulation of the thermal performance of earthen materials often simplify their hygrothermal properties as constant values. This oversimplification overlooks the material’s dynamic hygrothermal response to environmental changes, resulting in significant discrepancies between simulated predictions and actual performance.

In recent years, the hygrothermal properties of raw earth have become a prominent subject of scientific inquiry, with significant advancements achieved at the material scale [20,22,23]. These advancements are progressively extending into microscopic-level investigations. The author’s team [24], focusing on density and particle size distribution, quantitatively assessed the water vapor permeability of raw earth, revealing a significant negative correlation between the permeability coefficient and density, as well as a close relationship with the characteristics of particle size distribution. Losini et al. (2022) [25] elucidated the microstructural features of earthen materials through scanning electron microscopy and pore structure analysis, indicating that pore morphology and distribution exert a pronounced influence on both thermal conductivity and moisture transport processes. This provides empirical evidence for understanding the microscopic mechanisms of hygrothermal coupling. Losini [26] determined the thermal conductivity, density, and porosity of various clay samples, which demonstrated a significant influence of clay composition (proportions of sand, silt, and clay) and density on thermal performance. The porosity of the clay samples ranged from 23% to 40%, corresponding to a substantial variation in thermal conductivity. Synthesizing existing microscale research, it is evident that the factors influencing the hygrothermal performance of raw earth are diverse, encompassing density, particle gradation, pore structure, and mineral composition. These factors not only govern the internal moisture adsorption and desorption processes but also significantly impact the efficiency of heat and moisture transfer. A collaboration among four academic laboratories [27] conducted a series of hygrothermal tests on rammed earth from a modern building in Lyon, France, providing a comprehensive dataset of hygrothermal parameters. This work establishes a foundation for simulating coupled heat and moisture transfer in rammed earth structures and for assessing energy efficiency. Medjelekh, D et al. [28,29] integrated experimental and numerical approaches to develop a coupled heat and moisture transfer model for non-fired earth masonry. Their research found that earth masonry exhibits distinct moisture adsorption and desorption hysteresis under dynamic boundary conditions, highlighting the significant impact of pore-scale distribution and water phase change on hygrothermal behavior. Hall and Allinson [30] systematically analyzed the temperature evolution of stabilized rammed earth walls under non-constant conditions by combining transient numerical simulations with physical experiments. Collectively, these studies provide experimental insights into the hygrothermal coupling characteristics of earthen materials. However, they predominantly focus on performance testing at the specimen scale, with insufficient attention paid to the response of complete wall structures under dynamic climatic conditions. There remains a lack of quantitative research on the laws and interactions of heat and moisture transfer in real-world environments.

Field testing represents one of the most direct and reliable methodologies for investigating the hygrothermal performance of earthen buildings [31,32,33]. Existing empirical studies can be broadly categorized into two types: one focuses on the long-term monitoring of existing buildings, while the other entails more precise and systematic quantitative analysis, ranging from the material level to the construction of test cells under controlled experimental conditions. In the context of existing buildings, Wakil et al. [34] found, through field measurements of two traditional earthen structures in Morocco, that the walls effectively attenuated diurnal temperature fluctuations due to their high thermal inertia, thereby maintaining indoor thermal comfort in a hot, arid climate. Mellado M et al. [35] conducted a year-long monitoring campaign on a north-facing rammed earth wall of a traditional dwelling in Spain, revealing the dynamic interrelationship between wall moisture content, thermal conductivity, and thermal transmittance. This study validated the hygrothermal stability and nonlinear response characteristics of earthen walls under seasonal climatic variations. Collectively, these investigations demonstrate that earthen materials can provide natural regulation and assurance of comfort for the indoor environment through their inherent hygrothermal response mechanisms.

Under controlled experimental conditions, researchers often utilize test rooms or energy models to analyze the hygrothermal response mechanisms of earthen materials systematically. Numerous studies have validated the effective passive hygrothermal regulation capacity of earthen construction across different climate zones and at multiple scales. For instance, in the “Hot Summer and Cold Winter” climate zone of southwestern China, a research team [36,37], combining measurements with numerical simulation, discovered that rammed earth walls exhibit “peak-shaving and valley-filling” characteristics and hygrothermal hysteresis. These properties enable the passive regulation of the indoor environment through the processes of moisture adsorption and desorption. Similarly, Ibrahim Neya et al. [38], focusing on a tropical dry-hot region, employed energy modeling to analyze the impact of stabilized rammed earth combined with insulation on a building’s thermal inertia and adaptability, concluding that appropriate construction techniques can significantly enhance the wall’s hygrothermal performance. Serrano et al. [39] constructed experimental earthen buildings in Spain, comparing the thermal response of various stabilized and unsterilized materials. They identified that porosity and moisture content are critical factors influencing thermal inertia and heat storage capacity, thereby validating the hygrothermal coupling mechanism between the material and building scales. Idrissi Kaitouni et al. [40] systematically evaluated the hygrothermal performance of stabilized rammed earth, from material to building scale, through laboratory tests and field assessments. Based on laboratory-determined material parameters and combined with summer field data from a demonstration building and Energy Plus simulations, they evaluated the impact of passive strategies—such as wall thickness, night ventilation, and shading—on building energy efficiency and comfort. However, this study, in its simulations, still simplified the material’s hygrothermal parameters as constants, failing to adequately account for their dynamic nature as functions of temperature and air humidity in real-world conditions. This limitation in modeling is not an isolated case [41,42]. Particularly for a material like raw earth, which is highly sensitive to variations in temperature and air humidity, the non-constant nature of its hygrothermal parameters significantly influences the wall’s thermal and moisture response. Neglecting this characteristic can lead to inaccuracies in calculating heat and moisture transfer processes, which subsequently cause errors in heating and cooling load estimations, ultimately compromising the accuracy of building energy efficiency and indoor environmental assessments.

Notwithstanding the significant progress achieved by the aforementioned studies at both the material and building levels, most remain confined to single climate conditions or assume constant parameters. A systematic comparison and model integration across climate zones under long-term dynamic boundary conditions is lacking. Therefore, this study targeted rammed earth materials from actual buildings and systematically measured their hygrothermal parameters under both constant and non-constant conditions using prepared standard specimens. Based on the measured data, the obtained parameters were used as critical inputs in the WUFI-Plus simulation platform to perform year-long comparative simulations across China’s typical climate zones, quantitatively assessing the impact of different parameter settings on indoor hygrothermal environments and building energy predictions. Through a systematic quantitative investigation, this study not only uncovered the bias mechanisms of conventional constant-parameter models in performance prediction and clarified the dynamic hygrothermal regulation process of rammed earth materials, but also achieved systematic integration and validation of dynamic parameter-coupled simulations under multi-climate, year-long conditions, thereby providing a robust scientific foundation for developing a thermal design methodology for rammed earth buildings across diverse climatic regions.

2. Materials and Methods

The methodology of this study comprises two integral components: experimental testing and software simulation, as outlined in the research method and framework in Figure 1. The experimental component focuses on characterizing the hygrothermal properties at the material level. This involves systematically measuring the thermal conductivity at varying moisture contents and the water vapor resistance factor at different relative humidity levels. The objective is to acquire a comprehensive set of key hygrothermal parameters for the raw earth material under both constant and non-constant conditions.

Figure 1. Research Method and Framework.

Subsequently, the simulation study involves constructing a building model based on the WUFI Plus platform. The parameters obtained experimentally are assigned as the thermal properties of the wall assembly. This model is employed to systematically analyze the impact of employing non-constant and constant parameters on indoor hygrothermal conditions and building energy consumption across diverse climatic zones.

2.1. Sample Preparation

The rammed earth material employed in this study was sourced from the exterior wall of a traditional rammed earth residence in Yunnan. The material, formed by compacting native local soil, had a measured density of 1620 kg/m³. The collected raw soil was sieved and graded according to the NF EN ISO 17892-4 standard [43] to determine its particle size distribution, as shown in Figure 2. The compounds and elemental composition of the sand and gravel in the raw soil material were characterized using an OLYMPUS VANTA VEL-SDD handheld XRF analyzer (Olympus Corporation, Tokyo, Japan) [44,45,46,47], with the results listed in Table 1. The sample was chosen due to its regional representativeness and typical construction practice: Yunnan is among the regions in China with the most extensive and diverse rammed earth building types, and the collected soil corresponds to the most used native rammed earth in local residences. The sample was free of any chemical stabilization, preserving its natural porous structure and hygrothermal responsiveness, which is crucial for investigating its intrinsic dynamic hygrothermal coupling behavior.

Figure 2. Particle Size Distribution of the Raw Earth Material.

Table 1. Chemical composition of earth materials.

Compound Name	SiO2	AI2O3	Fe2O3	CaO	K2O	TiO2	SO3	MgO
Mass fraction (%)	50.29	21.95	12.345	7.44	3.629	2.34	0.31	0.02

To fabricate the raw earth specimens for hygrothermal parameter testing, the following procedure, illustrated in Figure 3, was implemented. First, the raw earth material sourced from the traditional dwellings in Yunnan was mixed with water and crushed, with visible impurities removed to ensure homogeneity. The pre-processed soil was then subjected to continuous drying for 24 h in a constant-temperature blast drying oven (Model: DHG-9625A, Hengyi, Shanghai, China; Temperature Control Accuracy: ±1 °C). Its mass change was periodically monitored using a precision electronic balance (Model SFA524, Lichen, Hunan, China; capacity 520 g, accuracy ±0.1 mg). The material was considered completely dry and had reached a constant mass when the difference between two consecutive weighing results fell below 0.1%. Subsequently, by controlling the mass of soil used and maintaining a constant mold volume, two types of specimens were fabricated via manual compaction to a uniform density of 1620 kg/m³: cubic specimens (L55 mm × W55 mm × H30 mm) and cylindrical specimens (R55 mm, H30 mm). All molded specimens were subsequently cured for 28 days in a controlled environment maintained at a temperature of (25 ± 2 °C) and a relative humidity of (53 ± 3)%. During the curing period, the mass of the specimens was monitored every 24 h. A state of equilibrium was deemed achieved when the mass variation remained consistently below 5%. Finally, the specimens that met these requirements were used for subsequent testing of hygrothermal performance parameters.

Figure 3. Fabrication Process of Raw Earth Specimens.

2.2. Determination of Hygrothermal Parameters

2.2.1. Thermal Conductivity Measurement

The cured cubic specimens were placed in an oven and thoroughly dried, after which they were weighed to determine their absolute dry mass. The moisture content of the specimens was controlled via the gravimetric method. The moisture content, E (%), of the specimen was calculated using Equation (1):

$$E={\displaystyle \frac{m_{\mathrm{wet}}-m_{dry}}{m_{dry}}}$$

(1)

where

$m_{wet}$ is the mass (kg) of the specimen after water spraying;

$m_{dry}$ is the mass (kg) of the thoroughly dried specimen.

This involved uniformly spraying a calculated mass of water onto the surface of the dried specimens to achieve target moisture contents of 2%, 4%, 6%, 8%, 10%, 12%, and 14%. A moisture content of 16% was excluded from the test matrix as it approximates the plastic limit of the raw earth material, beyond which its physical state changes significantly. Following water application, each specimen was tightly sealed with plastic wrap and cured for 12 h to allow for uniform moisture distribution. The thermal conductivity test was conducted under stable indoor environmental conditions, with an ambient temperature of approximately 25 ± 1 °C and a relative humidity of 55 ± 5%, while minimizing airflow and radiation interference during the testing process. Subsequently, the thermal conductivity of the specimens at each specified moisture content level was measured using a Hot Disk thermal constants analyzer (Model: TC-3000E, XIATECH, Xi’an, China, Range: 0.001–10 W/m·K, Accuracy: ±3%), in accordance with the Standard GB/T 32064-2015 [48]. The test process is shown in Figure 4. To mitigate moisture loss during the experimental procedure, key parametric tests for each specimen were performed in triplicate. The arithmetic mean of the three measurements was calculated and reported as the final valid value to ensure data reliability and repeatability.

Figure 4. Measurement of Thermal Conductivity at Varied Moisture Contents.

2.2.2. Measurement of Water Vapor Diffusion Resistance Factor

The experimental methodology employed in this study was based on the Chinese National Standard GB/T 17146-2015 [49], utilizing the desiccant method (dry-cup method) to evaluate water vapor transmission performance. A specialized test cup was filled with 50 g of anhydrous calcium chloride (CaCl₂) as the desiccant. A thoroughly dried rammed earth specimen was then mounted over the cup opening and sealed using molten wax, after which the initial mass of the assembly was recorded. The prepared test cups were subsequently placed in sealed humidity control containers. The test process, as shown in Figure 5, is based on the equilibrium between the salt solution and the air at a given temperature to maintain a constant relative humidity within the chamber. Different saturated salt solutions produce stable water vapor pressures through their solubility equilibrium, thus establishing distinct relative humidity conditions. The corresponding relative humidity values for each salt solution are listed in Table 2.

Figure 5. Determination of Water Vapor Resistance Factor at Differing Relative Humidities.

Table 2. Relative Humidity of Air Over Saturated Salt Solutions at 25 °C [50].

Chemical Formula	MgCl2	K2CO3	Mg(NO3)2	KI	NaCl	KCl	KNO3	MgCL2	K2CO2
Relative humidity (%)	32.78 ± 0.16	43.16 ± 0.39	51.40 ± 0.24	68.86 ± 0.24	75.29 ± 0.12	84.34 ± 0.26	93.58 ± 0.55	32.78 ± 0.16	43.16 ± 0.39

Driven by the water vapor pressure difference between the external and internal environments, water vapor diffused through the specimen into the cup. The water vapor diffusion resistance factor of the specimen was determined by periodically measuring the change in mass of the test cup over time. Two test cups were positioned in each humidity container as parallel replicates, and all containers were stored in a climate chamber maintaining a constant temperature of 25 ± 0.5 °C and a relative humidity of 60 ± 2%.

To ensure the reliability of the data, two parallel specimens (n = 2) were prepared for each humidity condition in this study. During the experiment, the masses of the samples and cups were recorded every 24 h; as the mass variation curves of the two specimens were highly consistent, demonstrating good internal consistency, the average value was taken as the valid data for that time point. The calculated standard deviation between the two parallel specimens was less than 0.05 g, indicating good repeatability of the experiment. When the average mass change rate over five consecutive measurements fell below 5%, it was regarded as having reached a stable state, and recording was terminated.

The water vapor diffusion resistance factor μ can be calculated using Equation (2):

$$\mu ={\displaystyle \frac{{\delta }_{air}}{\delta }}$$

(2)

${\delta }_{air}$—water vapor permeability of air.

$\delta$—water vapor permeability of the material.

The water vapor permeability of air, ${\delta }_{air}$ is dependent on the barometric pressure and temperature during the test. It can be calculated using Equation (3):

$${\delta }_{air}={\displaystyle \frac{0.083\cdot p_0}{R_D\cdot T\cdot p}}{\left({\displaystyle \frac{1}{273}}\right)}^{1.81}$$

(3)

where

$T$ is the thermodynamic temperature (K),

$p$ is the barometric pressure (hPa),

$p_0$ is the standard barometric pressure (1013.25 hPa),

$R_D$ is the gas constant of water vapor (462.10 × 10⁻⁶ N·m/(K·mg)).

The water vapor permeability of the material, δ, is defined as the mass of vapor transferred through the sample per second per unit area and is given by Equation (4):

$$\delta =w\times d$$

(4)

where

$d$ is the mean specimen thickness (m),

$w$ is the water vapor permeance [kg/(m²·s·Pa)].

The water vapor permeance, $w$ is defined with respect to the partial vapor pressure difference and is given by Equation (5):

$$w={\displaystyle \frac{G}{A\cdot {\Delta }_{P_V}}}$$

(5)

where

$G$ is the water vapor flow rate through the specimen (kg/s),

$A$ is the exposed area of the specimen (m²),

${\Delta }_{pv}$ is the water vapor pressure difference across the specimen (Pa).

The exposed area of the specimen, A, is calculated as the arithmetic mean of the free upper and free lower surface areas. The water vapor flow rate, G, is the primary measured value. It is determined as the mean of five successive determinations of the change in mass per unit time, denoted as m₁₂. The final value of G is accepted when the last five calculated m₁₂ values are within 5% of each other, indicating constant conditions. The change in mass per unit time, m₁₂, for each interval is given by Equation (6):

$$\Delta \mathrm{m}={\displaystyle \frac{m_2-m_1}{t_2-t_1}}$$

(6)

where $m_1$ and $m_2$ are the masses (kg) of the test assembly at times $t_1$ and $t_2$, respectively, $t_1$ and $t_2$ are the successive times of weighing (s).

2.3. Climatic Analysis

In accordance with the building thermal climate zoning criteria stipulated in the Code for Thermal Design of Civil Buildings (GB 50176-2016) [51], this study selected five representative cities from distinct climate zones across China. The chosen cities are Harbin (Severe Cold Zone), Beijing (Cold Zone), Nanjing (Hot Summer and Cold Winter Zone), Guangzhou (Hot Summer and Warm Winter Zone), and Dali (Temperate Zone). The geographical locations of these cities, along with their corresponding climate zones, are illustrated in Figure 6.

Figure 6. Spatial Distribution of Building Thermal Climate Zones and Representative Cities in China (Adapted from the “Code for Thermal Design of Civil Buildings” GB 50176-1993) [51].

The fundamental climatic data for each city were derived from Typical Meteorological Year (TMYx) weather files sourced from the Meteonorm v8.2.0 software [52]. This dataset is generated through multi-source fusion and long-term statistical analysis, synthesizing standard annual meteorological data by selecting 12 representative typical months. The provided continuous, hourly parameter series ensures physical consistency and is widely applicable for building performance simulation.

China’s vast territory, spanning a wide range of longitudes and latitudes, exhibits significant disparities in environmental conditions across its different climate zones. Figure 7 illustrates the specific characteristics of each city in terms of global horizontal solar irradiance, mean air temperature, mean wind speed, and relative humidity. Harbin (Severe Cold Zone) experiences extreme seasonal temperature variations, characterized by very cold winters and cool summers. The average January temperature drops to −17.3 °C, while the July average reaches 23.7 °C. Summers are cool, with spring wind speeds reaching 3–3.5 m/s. Direct solar radiation from March to September exceeds 60,100 kWh/m². Beijing (Cold Zone) has cold winters and hot summers. Winter temperatures can drop to a minimum of −3.4 °C, accompanied by relatively dry conditions (relative humidity, RH, 30–40%). Global horizontal irradiance is notably high from May to July (peak around 70 kWh/m²), and the average July temperature can reach 28 °C. Nanjing (Hot Summer and Cold Winter Zone) shows pronounced seasonal temperature fluctuations. Winter temperatures generally remain above 0 °C, with consistently high humidity levels (approximately 75% in summer and over 60% in winter). Wind speeds remain stable year-round (2.5–3 m/s), while the overall level of global horizontal irradiance is relatively low. Guangzhou (Hot Summer and Warm Winter Zone) maintains high temperatures throughout the year. Even in winter, the average temperature stays above 10 °C, with no severe cold period. Summer temperatures approach 30 °C, with prolonged high-temperature conditions, and annual relative humidity remains above 60%. Dali (Temperate Zone) exhibits minimal temperature fluctuations, with annual temperatures ranging from 9.6 °C to 20.9 °C. It receives the most abundant global horizontal irradiance among the five cities, with a peak of 109 kWh/m² in March. Wind speeds are higher in winter (approximately 3.5 m/s in January–February), while relative humidity reaches around 70% in summer and drops to 40–50% in winter.

Figure 7. Comparative Analysis of Annual Climatic Conditions Across Five Representative Cities. (a) Mean Direct Normal Irradiance, (b) Mean Air Temperature, (c) Mean Relative Humidity, (d) Mean Wind Speed.

The five major building thermal climate zones demonstrate substantial differences in key environmental parameters. These variations directly influence the hygrothermal behavior of raw earth materials, consequently determining the indoor thermal comfort, durability, and energy efficiency of earthen buildings. Therefore, a systematic investigation into the hygrothermal properties of raw earth across different climate zones is crucial for advancing the design and optimization of regionally adapted earthen architecture.

2.4. Build Simulation Model

WUFI Plus 3.5.0.1was selected as the simulation software for this study. Developed by the Fraunhofer Institute for Building Physics in Germany, this tool simulates coupled heat and moisture transfer through building enclosures and assesses the holistic hygrothermal performance of entire buildings [53,54]. Its physical core is based on the coupled heat and moisture transfer model proposed by Künzel [55], establishing it as an internationally recognized software for combined heat and moisture simulation. The accuracy of WUFI Plus in simulating heat and moisture transport within building envelopes has been validated through multiple experimental and field studies, including the IEA Annex 41 project and other benchmark experiments [56,57,58]. Notably, David Allinson and Hall [59] conducted a hygrothermal characteristics study on an 8 m² stabilized rammed earth test unit, employing this software for numerical analysis. Their findings indicated that, provided material properties are accurately characterized, the discrepancy between simulation results and experimental data can be controlled within 5%.

To ensure the generalizability and rationality of the simulation study, a single-story residential building was defined as the baseline model. This benchmark building has plan dimensions of (L8 m × W8 m × H3 m), with a total floor area of 64 m². A 2 m × 1 m window is positioned on the south facade, and a 0.9 m × 2 m door is located on the west facade. This single-story configuration was chosen to effectively reflect the influence of different climatic conditions on the indoor hygrothermal environment while reasonably managing computational complexity. Utilizing such a small-scale benchmark model helps focus the analysis on the impact mechanisms of material properties and climate variables on enclosure performance, avoiding interference from complex spatial factors such as multi-story shading or inter-floor heat transfer.

The construction details of the other enclosure assemblies used in the baseline building for this study are presented in Table 3. Except for the rammed earth walls, the thermophysical properties of all different materials were sourced from the built-in material database of WUFI Plus. The software allows users to define custom material properties, enabling the specification of either constant or non-constant hygrothermal characteristics. In this study, two rammed earth material models were constructed using software, with constant parameters assigned to one and non-constant parameters to the other. To ensure that the walls functioned as the sole variable, all other construction parameters were kept unchanged, and identical initial and boundary conditions were applied to both models, thereby ensuring the reliability and scientific rigor of the simulation results.

Table 3. Building Envelope Assembly Compositions and Key Thermal Properties.

Building Envelope	Material Layers (from Exterior to Interior)	Thermal Conductivity (W/(m·K))	Thickness (m)
Ceiling	1. PVC Roof Membrane	0.16	0.001
	2. Mineral Wool	0.04	0.08
	3. Concrete	1.6	0.15
	4. Interior Plaster	0.2	0.01
Floor	1. Concrete	1.6	0.05
	2. Mineral wool	0.04	0.08
	3. PE-Membrane	2.3	0.001
Door	Wooden door	0.343	0.04
Windows	Single-glass wooden board K = 4.7 (W/mK)

3. Results

3.1. Determination of Hygrothermal Material Parameters

3.1.1. Thermal Conductivity

The measured thermal conductivity values of the rammed earth specimens at varying moisture content levels are summarized in Table 4. Under constant ambient temperature and humidity conditions, the baseline thermal conductivity was determined to be 0.83 W/(m·K). As the moisture content increased from 0% to 14%, the thermal conductivity demonstrated a progressive rise from 0.77 W/(m·K) to 1.38 W/(m·K), representing a significant increase of 79.2%. This escalating trend was notably more pronounced within the medium-to-high moisture range (with a moisture content exceeding 4%), as illustrated in Figure 8. Similarly, under non-constant conditions, the thermal conductivity exhibited a gradual yet marked escalation with increasing moisture content, with the most substantial changes likewise observed in the medium-to-high-humidity regime.

Table 4. Thermal conductivity of rammed earth specimens at different moisture contents.

Moisture Content (%)	Thermal Conductivity (W/m·K)
Constant	0.81
0	0.77
2	0.83
4	0.92
6	1.02
8	1.15
10	1.23
12	1.38
14	1.44

Figure 8. Fitted Curve of Thermal Conductivity of Raw Earth Material at Varied Moisture Contents.

A linear regression analysis of the experimental data established the relationship between thermal conductivity λ (W/(m·K)) and moisture content w (%) as expressed in Equation (7):

$$\lambda =0.0506w+0.7383$$

(7)

The relationship demonstrates a coefficient of determination (R²) of 0.9919, indicating a statistically significant linear correlation between thermal conductivity and moisture content across the investigated range.

3.1.2. Water Vapor Resistance Factor

The water vapor resistance factor (μ) of the rammed earth specimens as a function of relative Humidity (RH) at 25 °C is presented in Table 5. As the relative humidity increased from 32.78% to 93.58%, the water vapor resistance factor decreased from 10.69 to 7.46. Analysis of this trend reveals that the μ value declined gradually when RH < 60%, whereas it exhibited a pronounced exponential decrease once RH exceeded 60%.

Table 5. Water Vapor Resistance Factor at Different Relative Humidity Levels.

Relative Humidity (%)	Diffusion Resistance Factor
constant	9.8322
32.78	10.695
43.16	10.416
51.4	10.101
68.86	9.3023
75.29	8.2305
84.34	7.8740
93.58	7.4627

3.2. Indoor Environmental Conditions

3.2.1. Summer Indoor Hygrothermal Conditions

Under summer conditions, simulation results demonstrate discernible regional variations across different climate zones, as illustrated in Figure 9. In cities with higher summer cooling loads, such as Beijing, Nanjing, and Guangzhou, the indoor average temperature simulated under the constant hygrothermal parameter assumption was consistently higher than that under the non-constant parameter model. The mean temperature difference ranged from 0.03 to 0.10 °C, with the most pronounced difference observed in Nanjing (0.103 °C). Conversely, in regions with relatively cooler summers, namely Harbin and Dali, the non-constant parameter model yielded slightly higher indoor temperatures than the constant model, with mean differences of 0.045 °C and 0.106 °C, respectively. Regarding extreme values, the maximum indoor temperature difference reached 0.488 °C in Nanjing, followed by 0.410 °C in Beijing, indicating that the parameter setting exerts a more sensitive influence on indoor temperature prediction in cities with high thermal loads.

Figure 9. Comparative Analysis of Summer Indoor Temperature and Humidity Across Representative Cities. (a) Harbin (b) Beijing (c) Nanjing (d) Guangzhou (e) Dali.

Regarding humidity simulation, the overall trend revealed that the average indoor relative humidity under constant parameter conditions was generally higher than that predicted by the non-constant parameter model. The mean difference ranged from 0.05% to 0.77%, with the most significant discrepancies observed in Beijing and Harbin. It is noteworthy, however, that Guangzhou exhibited an inverse pattern, where the non-constant parameter model predicted a higher average indoor humidity, with a mean difference of −0.293%. This deviation highlights the unique hygric behavior of buildings in high-humidity climatic conditions and underscores the importance of selecting parameters for accurate humidity prediction in such regions.

3.2.2. Winter Indoor Hygrothermal Conditions

Regarding winter temperatures, the indoor average temperature simulated using non-constant parameters was higher than that predicted by the constant e model in most cities, as shown in Figure 10, with differences ranging from 0.18 to 0.78 °C. The discrepancies were most significant in Harbin and Beijing, at 0.776 °C and 0.436 °C, respectively, indicating a more pronounced influence of diurnal temperature fluctuations and short-term solar radiation variations in severe cold climates. In contrast, the differences were relatively minor in Nanjing, Guangzhou, and Dali, all of which remained below 0.30 °C, suggesting that the constant e model provides a reasonable approximation of average indoor temperatures in temperate and warm climates.

Figure 10. Comparative Analysis of Winter Indoor Temperature and Humidity Across Representative Cities. (a) Harbin (b) Beijing (c) Nanjing (d) Guangzhou (e) Dali.

In terms of humidity, the overall trend indicated that the average indoor relative humidity under the constant parameter condition was higher than that under the non-constant parameter model, with differences ranging from 0.50% to 1.47%. Beijing and Harbin exhibited the largest discrepancies, reaching 1.35% and 1.47%, respectively; Nanjing and Dali showed moderate differences (0.50% and 0.99%); while Guangzhou demonstrated the smallest variation at just 0.59%. These results suggest that northern severe cold regions are more prone to decreased relative humidity under non-constant conditions. In contrast, the difference between constant and non-constat pronounced in the humid southern climate.

3.3. Thermal Performance of the Earth Wall

To further investigate the impact of non-constant hygrothermal parameters on building thermal performance, this study conducted a comparative analysis of the annual dynamic variations in interior surface temperature and heat flux density of the wall under the two parameter assumptions.

3.3.1. Surface Temperature

Figure 11 illustrates the distribution of the annual differences in wall interior surface temperature (constant minus non-constant) for the five representative cities under the two parameter conditions. The overall trend indicates that the median temperature difference of each town is generally close to zero with a relatively narrow fluctuation range, suggesting minor overall differences in surface temperature between the constant and non-constant parameter models. Notably, the median values for most cities are slightly negative, implying that the surface temperature under the non-constant parameter condition is marginally higher for the majority of the time. The fluctuation amplitudes, however, vary distinctly across climate zones, reflecting the influence of climatic characteristics on the non-stationarity of heat and moisture transfer.

Figure 11. Temperature Difference Distribution Across Interior Wall Surfaces.

In Harbin, the median temperature difference is −0.28 °C, with an interquartile range of −0.49 °C to 0.02 °C and extreme values ranging from −1.25 °C to 0.78 °C. The presence of numerous negative outliers indicates a more pronounced effect of non-constant hygrothermal processes on surface temperature in severe cold climates. Beijing shows a median of −0.14 °C and an interquartile range between −0.32 °C and 0.08 °C, indicating a slight overall negative bias with a relatively modest magnitude of difference. Nanjing exhibits a median value near zero (−0.01 °C) and an interquartile range from −0.18 °C to 0.14 °C, demonstrating the least pronounced annual surface temperature difference between the two modeling approaches. Guangzhou exhibits a slight overall positive bias, indicating that the surface temperature under the constant parameter condition is slightly higher than under the non-constant model in this warm and humid climate. Dali presents a median of −0.10 °C with a concentrated distribution and minor fluctuations, indicating relatively stable temperature differences in the temperate climate.

3.3.2. Heat Flux Density

Analysis of the annual simulation results comparing constant and non-constant hygrothermal parameter models, as summarized in Table 6, reveals that the standard deviation (Std) of heat flux density is generally lower under the non-constant model. This reduction in fluctuation amplitude is particularly pronounced in colder regions with greater temperature variations, where Harbin (ΔStd = 0.40) and Beijing (ΔStd = 0.30) demonstrate more significant dampening effects compared to Guangzhou and Dali.

Table 6. Comparison of annual heat flux density variations between constant and non-constant hygrothermal parameter models across different climate zones.

City	Model	Max (W/m2)	Min (W/m2)	Std (W/m2)	Mean (W/m2)
Harbin	Constant	4.63	−12.30	2.07	−1.03
	Non-constant	3.08	−9.33	1.67	−0.88
Beijing	Constant	4.08	−7.05	1.57	−0.24
	Non-constant	3.17	−5.28	1.27	−0.17
Nanjing	Constant	4.41	−7.74	1.46	0.27
	Non-constant	3.58	−5.48	1.16	0.29
Guangzhou	Constant	4.01	−4.24	1.14	0.98
	Non-constant	3.50	−4.12	0.90	0.94
Dali	Constant	2.55	−3.83	1.04	−0.07
	Function	2.06	−3.73	0.83	−0.01

Under extreme operational conditions, the non-constant parameter model moderates heat flux extremes. Specifically, minimum heat flux densities in cold regions (such as Harbin and Beijing) are higher than those predicted by the constant model. In comparison, maximum heat flux densities in the hot areas (such as Guangzhou) are correspondingly lower. Although the annual mean differences between the two parameter approaches remain below 0.2 W/m² across all regions, distinct spatial patterns emerge: cold regions (Harbin, Beijing) exhibit negative values, indicating that non-constant parameters bring the heat flux closer to zero; warm regions (Nanjing, Guangzhou) show positive values; while Dali’s near-zero difference suggests a fundamental balance between indoor and outdoor heat exchange.

4. Discussion

This study, through experimental investigation and numerical simulation, reveals the dynamic nature of the hygrothermal properties of rammed earth in response to environmental changes and confirms the significant importance of incorporating non-constant hygrothermal parameters in building energy simulation. The core finding is that, compared to the traditional constant parameter model, the non-constant model more accurately represents the true capacity of raw earth materials to regulate the indoor hygrothermal environment, consequently leading to a reassessment of building energy consumption.

4.1. Material Hygrothermal Characteristics and Their Dynamic Influence Mechanisms

The research demonstrates that the hygrothermal performance of raw earth materials is highly sensitive to variations in moisture content, primarily manifested in the dynamic changes in thermal conductivity and the water vapor diffusion resistance factor. In this study, the thermal conductivity of rammed earth exhibited a significant increasing trend, rising from 0.77 W/(m·K) to 1.38 W/(m·K) with increasing moisture content. This range of variation is largely consistent with values reported in the literature. Lovec et al. [60], for instance, reported typical thermal conductivity values of 0.7–1.25 W/(m·K) for traditional rammed earth walls, noting that specific values are influenced by soil composition, compaction density, and porosity [61]. McGregor [62] attributes this linear growth trend to the high thermal conductivity of liquid water. As moisture progressively displaces the air within pores, the heat conduction pathways between solid particles become more continuous, thereby significantly enhancing the material’s overall thermal conductivity. From a microstructural perspective, this phenomenon can be attributed to the displacement and interfacial effects of water within the pore structure. As the moisture content increases, liquid water replaces air in the pores, creating more connected paths for heat transfer. Simultaneously, the condensed water in capillary pores and the adsorbed water films on particle interfaces alter the interfacial thermal resistance, further promoting heat conduction [36,63,64].

Concurrently, the water vapor diffusion resistance factor (μ) of the rammed earth material demonstrates a decreasing trend with increasing relative humidity. This study measured a μ-value of 10.69 at 32.78% RH, which decreased to approximately 7.46 when the relative humidity increased to 93.58%. Existing studies indicate that the μ-value of raw earth materials typically falls within the range of 3–11 and can increase significantly with the addition of stabilizers, such as cement; thus, the measured results are within a reasonable range [59,65,66]. This observed pattern aligns with the findings of Hall and Allinson [67]. As ambient Humidity rises, the adsorbed water layer on pore surfaces thickens, forming continuous liquid film pathways. This shifts the dominant mode of water vapor transport from purely gaseous diffusion to include surface diffusion and liquid film transport, thereby reducing the overall resistance to diffusion.

Overall, the superior hygrothermal regulation capability of rammed earth materials can be ascribed to their intrinsic dynamic hygrothermal coupling mechanism, often termed the material’s “breathing effect.” The physical nature of this mechanism does not lie in simple variations in material parameters, but stems from the distinctive microscopic porous structure of rammed earth. Rammed earth consists of clay, sand, and gravel, where imperfect particle contacts create an interconnected porous network, enabling environmental water vapor to penetrate and move within the pores. Based on porous media heat and mass transfer theory and the Kelvin equation, water vapor in the pores can experience capillary condensation below the saturation vapor pressure [67]. Under cyclic changes in external temperature and humidity, pore water vapor undergoes periodic evaporation (heat absorption) and condensation (heat release), involving the uptake and release of latent heat [68]. This dynamic phase-change process couples heat and moisture transport, enabling rammed earth to buffer external climatic fluctuations and maintain indoor hygrothermal stability.

4.2. Impact on Indoor Environment

Annual simulation results indicate that under non-constant hygrothermal parameters, rammed earth walls exhibit more stable thermal response characteristics across different climate zones. This suggests that the hygrothermal coupling effect plays a crucial role in mitigating the impact of external climate fluctuations on the wall’s heat transfer process. This observed difference is closely related to the previously discussed heat and moisture coupling mechanism. Under non-constant conditions, dynamic variations in wall moisture content lead to a coordinated evolution of thermal conductivity and water vapor diffusion resistance. When the wall absorbs moisture, water fills partial pores, enhancing thermal conductivity while simultaneously reducing vapor diffusion resistance. This enables simultaneous moisture migration and latent heat exchange to collectively influence the heat transfer process, forming a dynamically adjustable “hygrothermal feedback mechanism” [57]. Particularly in cold climate zones like Harbin and Beijing, the dynamic adjustment of material thermophysical properties induced by humidity variations effectively buffers energy fluctuations caused by diurnal temperature differences and short-term solar radiation.

Figure 12 further reveals the dynamic thermal response characteristics of rammed earth walls under the effects of hygrothermal coupling. Taking three typical days in July in Beijing as an example, the peak temperatures of both the wall’s interior surface and indoor air lag behind outdoor air temperature by approximately 6–8 h, demonstrating a significant time-shifting effect (φ ≈ 6–7 h). Simultaneously, the amplitude of the interior surface temperature is markedly reduced compared to the exterior surface, with a significantly decreased attenuation factor (f), particularly under the non-constant hygrothermal parameter model, that indicates that the latent heat exchange process induced by dynamic moisture content variations alters the wall’s effective thermal conductivity and heat transfer rate, thereby enhancing its heat storage and peak-shaving capacity. Consequently, the wall demonstrates more pronounced phase lag and temperature attenuation effects during severe external climate fluctuations, exhibiting excellent buffering and stabilization performance.

Figure 12. Comparison of Indoor and Wall Surface Temperatures During Summer Conditions in Beijing.

4.3. Impact on Energy Consumption

Further analysis reveals that the setting of hygrothermal parameters significantly influences building energy consumption predictions. To ensure comparability of results across different climate zones, the simulation uniformly adopted a continuous operation mode for the air-conditioning system: heating was activated when the indoor temperature fell below 18 °C, and cooling was activated when it exceeded 26 °C, with annual energy consumption calculated accordingly. The results indicate that the constant parameter model generally overestimates building energy consumption compared to the non-constant hygrothermal parameter model. This overestimation primarily stems from the constant model’s inability to represent the dynamic regulatory effects arising from the heat and moisture coupling of raw earth materials in real-world conditions, leading to elevated estimates of heat transfer rates and energy demands.

As shown in Figure 13, under non-constant hygrothermal parameters, both cooling and heating energy consumption across all climate zones are lower than those predicted by the constant model. The reduction in heating energy consumption ranges between 13.6% and 16.9%, in the following order: Beijing, Harbin, Nanjing, Dali, and Guangzhou.

Figure 13. Energy Consumption Profiles of Representative Cities Across Climate Zones.

In contrast to the general reduction in heating energy, the changes in cooling energy consumption exhibit more complex regional characteristics. The reduction across climate zones ranges from 3.6% to 15.8%, with Guangzhou (−15.8%) and Nanjing (−15.2%) showing the most significant decreases, highlighting the sensitivity of the coupled heat and moisture model to energy consumption prediction in hot and humid climates. The reductions in Beijing (−12.3%) and Dali (−12.6%) are also considerable, whereas the change in Harbin (−3.6%) is minimal, primarily due to the inherently very low absolute cooling demand of its external climate. Overall, the non-constant hygrothermal parameter model consistently predicts lower energy consumption levels, demonstrating that accounting for the dynamic variation in material hygrothermal properties has a measurable impact on energy simulation outcomes.

4.4. Limitations and Future Directions

The experimental material used in this study was sourced from the wall of a traditional rammed earth dwelling located in Yunnan Province. Although the sampling process aimed to obtain the most representative specimens possible, certain geographical and compositional limitations remain. Regional soils differ in mineral composition, particle size distribution, and moisture characteristics, which consequently lead to distinct regional variations in their hygrothermal response behaviors. Therefore, the objective of this study is not to derive universally applicable parameters but to elucidate, through experimental analysis, the transfer and regulation mechanisms of the material under dynamic hygrothermal coupling, thereby providing empirical evidence and theoretical insight into the passive regulatory characteristics of rammed earth walls.

Secondly, during the thermal conductivity tests, the moisture content of the samples was adjusted using a spray-water method. This method provides satisfactory controllability and repeatability; however, it still introduces a degree of measurement uncertainty. To minimize this uncertainty, the spraying process was carefully controlled by applying small, repeated sprays at close range, thereby preventing local moisture loss or uneven infiltration. Meanwhile, after spraying, the samples were placed in a sealed environment and allowed to stand for a sufficient period to ensure internal moisture redistribution and equilibrium, thereby reducing the influence of moisture content gradients on the measurement results. Parallel specimens were employed for cross-verification, further improving the stability and reliability of the measurement results. Despite these efforts, moisture regulation remains a critical factor influencing the accuracy of thermal conductivity measurements. Future research should therefore aim to develop more refined and quantifiable methods for controlling moisture content.

Finally, to rigorously control variables and emphasize the impact of material parameter variations, this study established a simplified single-story baseline model. This model remains different from real buildings in aspects such as geometric complexity, building envelope, and actual operational conditions. Therefore, the quantitative findings in this study should be interpreted as relative prediction discrepancies induced by traditional constant hygrothermal parameter models, rather than absolute energy predictions for particular buildings. Future research can integrate long-term monitoring data from actual buildings with dynamic calibration simulations to refine model parameters, enhance prediction accuracy, and further verify and expand the mechanistic insights proposed in this study.

5. Conclusions

This study integrates experimental characterization with dynamic simulations across multiple climate zones and annual timescales to systematically quantify the non-constant hygrothermal properties of rammed earth and elucidate their influence on wall heat and moisture transfer under varying climatic conditions. The results underscore the crucial importance of accurately assessing the hygrothermal performance of earthen walls to ensure reliable predictions of building energy performance across diverse climatic conditions. The main conclusions are as follows:

• Experimental results revealed a highly significant linear positive correlation (R² = 0.9919) between the thermal conductivity of rammed earth and its mass moisture content. As the moisture content increased from 0% to 14%, the thermal conductivity rose from 0.77 W/(m·K) to 1.38 W/(m·K). Concurrently, the water vapor resistance factor (μ) of the rammed earth specimens decreased from 10.69 to 7.46 as relative humidity (RH) increased from 32.78% to 93.58%, showing a gradual decline below 60% RH and an exponential decrease above 60%.
• Simulations across different climate zones demonstrated that the non-constant hygrothermal parameter model more effectively captures the passive regulation capacity of raw earth materials. During summer conditions, indoor average temperatures under the non-constant parameter model were generally lower than those under the constant parameter model in cities with high cooling loads. Conversely, in winter, the non-constant model predicted higher indoor average temperatures, with particularly notable differences in cold regions. Regarding humidity simulation, the constant parameter model consistently overestimated the indoor average relative humidity.
• Under non-constant hygrothermal parameters, the annual fluctuation of the wall’s interior surface temperature was more moderate, and the standard deviation of heat flux density was generally reduced, leading to a moderation of extreme heat flux values.
• Compared to the constant parameter model, the non-constant model predicted lower cooling and heating energy consumption. The reduction in heating energy consumption ranged from 13.6% to 16.9%, with the most pronounced decreases observed in Harbin and Beijing (16.9% and 15.5%, respectively). The overall reduction in cooling energy consumption ranged from 3.6% to 15.8%, with the most significant decreases observed in Guangzhou (15.8%) and Nanjing (15.2%).

Therefore, for rammed earth and other highly hygroscopic porous materials, dynamic coupling models incorporating non-constant hygrothermal parameters are recommended in place of traditional constant-parameter models to more faithfully represent the materials’ actual hygrothermal response characteristics. In practical applications, the combination of rammed earth walls with impermeable materials should be avoided, as such combinations may obstruct hygrothermal transfer pathways and diminish the material’s inherent passive regulation and energy-saving potential. Through a systematic quantitative analysis, this study confirms the integrated advantages of rammed earth as a natural building material in terms of hygrothermal regulation and energy conservation, providing a solid scientific basis for life-cycle energy and carbon assessments and offering theoretical support for performance optimization and design strategies in sustainable building systems.