Modeling Moisture Accumulation and Decay Potential in Cross-Laminated Timber Wall Assemblies Exposed During the Construction Phase

Abstract

This study examines decay risks in cross-laminated timber (CLT) wall assemblies with built-in moisture, aiming to develop a simulation-based methodology to assess moisture dynamics during the construction phase. Differing from previous research, this study focuses on the central regions of CLT wall panels. Moisture distribution within the panel, especially in the exposed layer, is critical for understanding potential degradation. A series of simulations were conducted to determine the necessary level of detail for moisture profiling, comparing approaches that use a single average value, layer-specific averages, and a refined profile that distinguishes the outer 5 mm from the remaining material. The influence of factors such as wood type, glue type, delivery moisture content, orientation, and rain exposure was systematically evaluated to define realistic moisture profiles at the end of the construction phase. Subsequent degradation assessments incorporated these profiles along with variations in insulation materials to evaluate the time of wetness, dose accumulation, and heat flux increases. Results indicate that a detailed moisture profile is essential for accurately predicting decay risk and that trade-offs exist between moisture management and thermal performance depending on the insulation used. These findings provide a framework for predicting decay risks in CLT assemblies and offer insights for designing more durable and energy-efficient structures.

1. Introduction

Cross-laminated timber (CLT) is an engineered wood product composed of multiple timber layers bonded at right angles, offering benefits such as enhanced aesthetics, a reduced carbon footprint, faster construction, lighter weight, and lower thermal conductivity compared to traditional materials like brick and concrete. However, moisture poses a significant technical challenge. Since CLT is typically made from untreated softwood, it is particularly susceptible to mold and wood rot—a risk that is heightened during the construction phase when CLT panels are exposed to rain [1,2,3,4,5] and may be enclosed in the structure before fully drying.

While previous research on wood durability [6,7,8] and the impact or presence of built-in moisture in CLT structures has mainly focused on areas near exposed edges or moisture traps [1,2,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23] or horizontal surfaces, [3,4,24,25,26] the central regions of CLT wall panels are often overlooked despite being exposed to outdoor conditions and potentially accumulating moisture. On top of that, the moisture distribution within wooden panels is only rarely documented and, if taken into account, often based on single-sided soaking [5,27,28] instead of based on how rain would actually fall onto a wall. In Belgium, where CLT is a rather new construction material without local technical guidelines, several litigation cases came to the front, for which it was shown that exposure to rain during the construction phase ultimately led to structural failure. On top of that, previous research has shown that in Europe, Belgium is one of the locations with the highest risk of wood decay [29].

The use of hygrothermal simulations is promising for estimating moisture contents and, by extension, calculating degradation [30]. The study uses the HAM-modeling software Delphin 6.1.6. The software’s model configuration, solver parameters, and calculation of results were validated by Sontag et al. [31]. The transport model itself was validated by performing tests on different test houses and is further detailed in the dissertation of Scheffler et al. [32].

This study investigates the decay risks associated with built-in moisture in CLT wall assemblies during construction (step 3). Differing from previous research, this study focuses on the central regions of CLT wall panels. However, to accurately assess the risk of degradation, it is first necessary to understand how moisture is distributed within the panel (step 2) and whether that distribution significantly impacts decay potential (step 1).

2. Materials and Methods

This study focuses on a 5-layer 100 mm thick (5 × 20 mm) CLT wall panel. The wall panel consists of 5 layers of spruce wood with 0.25 mm of MUF glue between each layer. See Figure 1.

The panel was exposed to climate conditions according to a moisture reference year (MRY) [33] based on Belgian climate data since previous research has shown that in Europe, Belgium is one of the locations with the highest risk of wood decay [29].

The CLT panel was assigned an initial moisture content of 16% when arriving on site (delivery moisture content) and a rain exposure coefficient of 0.5. The CLT panel is only exposed to rain and solar radiation from one side (West-orientation), imitating an outdoor CLT wall during the building phase. The other side of the CLT panel (East-Orientation) is protected from rain and solar radiation during the construction phase. However, it is still exposed to the temperature and relative humidity of the outside environment since this would also be the case during the assembly period.

Simulations were performed using Delphin 6.1.6, a simulation program for combined heat, air, moisture, and pollutants simulations in porous building materials. As no pollutants are studied here, the abbreviation HAM simulation can be used. In this paper, only one-dimensional (1D) simulations are done as the simulated configurations mainly consist of planar elements.

The methodological framework can be divided into three steps (see also Figure 2):

1

Determine the necessary grid size for moisture profiles:

1.a

Define the extreme values and gradients of moisture contents in the CLT panel,

1.b

Define the most coarse grid that is admissible;

2.

Define moisture profile types:

2.a

Determine which parameters have the highest influence on moisture contents,

2.b

Define reference moisture profile types at the end of the construction phase;

3.

Assess degradation risks for wall assemblies with the reference moisture profiles.

These different steps are discussed below.

2.1. Determine the Necessary Grid Size for Moisture Profiles

To balance simulation detail with effort and relevance, the most coarse grid admissible is defined. This is done by defining the extreme values and gradients based on the first set of simulations (1.a, see Figure 2). This consists of 12 individual simulations, one for each month of the year, where the CLT panel is exposed to the corresponding climate conditions recorded within the MRY for each month, beginning at 7:00 a.m. on the 1st and finishing at 7:00 pm on the final day of the month. Moisture values for each mm through the depth of the CLT panel are then extracted from the simulations for every hour over the period. The CLT panel is simulated with the hygrothermal properties as described in

Table 1. Hygrothermal properties of materials in simulation set 1.a.

Material	ID 1	d mm	ρ kg m−3	cp J kg−1 K−1	λ W m−1 K−1	µ -	ϴ80 kg m−3	ϴsat kg m−3	Aw kg m−2 s−1/2
CLT panel
Spruce tangential	713	5 × 20	393.7	1843	0.106	487.7	59.8	728.1	0.005
MUF adhesive 2	713 2	4 × 0.25	425.0	1245	0.079	36.5	0.00	72.6	590.2

¹ ID is the identification code used in the Delphin 6.1.6 software material database. ² Material properties derived from experiments at Ghent University.

.

To assess whether these moisture profiles lead to degradation, another set of simulations must be executed to define the required modeling approach (1.b, see Figure 2). This set of simulations takes into account the CLT panel, a polyurethane insulation board with an aluminum facer and a brick facade cladding. The moisture content values within the CLT panel, which were derived from simulation set 1.a, are then used as moisture content values at the end of construction.

However, when assessing the possible degradation of a CLT wall panel on a building site, it is not feasible to measure the moisture content in every mm of the CLT panel. To explore how detailed the moisture profile needs to be to allow a meaningful decay risk assessment, simulations are conducted with varying degrees of detail. This varies from using the average moisture content for the entire panel to modeling moisture content for each millimeter of the CLT panel. This step helps determine whether a simplified approach is sufficient or if a more granular moisture profile is necessary.

To obtain information about how coarse the assessed grid can be before losing the required accuracy, the time instance with the widest range of moisture contents within the CLT panel is found:

• The moment with the greatest difference in moisture content between the maximum and minimum values within layer A (exterior layer of the CLT panel);
• The moment with the greatest difference in moisture content between the center of the CLT panel (i.e., the center of layer C) and the outer edge of the CLT panel;
• The moment with the greatest difference between the average moisture content of layer A (exterior layer of the CLT panel) and the average moisture content of layer C (middle layer of the CLT panel).

The hours outside of the 7 am to 7 pm range are left out of scope as these are not part of the working hours, and therefore, the CLT panel will not be enclosed during this time.

The moisture content values found for these hours are chosen as the starting conditions for simulation set 1.b. These starting conditions are implemented according to the following degrees of resolution (see Figure 2):

• Every millimeter in the CLT panel was assigned the moisture content output values from the first simulation;
• Every 5 mm in the CLT panel is grouped (creating 4 groups per layer of wood), and each group is assigned the average moisture content of these 5 mm groups;
• Every 10 mm in the CLT panel is grouped (creating 2 groups per layer of wood), and each group is assigned the average moisture content of these 10 mm groups;
• Every layer of wood in the CLT panel is assigned the average moisture content value of the layer;
• All the wood in the CLT panel is assigned the average moisture content values of all the wood in the CLT panel;
• Every layer of wood in the CLT panel is assigned the average moisture content values of this layer, with the exception of layer A, where the distinction is made between the outermost 5 mm and the other 15 mm.

The specifications of the wall setup in simulation set 1.b can be found in

Table 2. Hygrothermal properties of materials in simulation set 1.b.

Material	ID 1	d mm	ρ kg m−3	cp J kg−1 K−1	λ W m−1 K−1	µ -	ϴ80 kg m−3	ϴsat kg m−3	Aw kg m−2 s−1/2
CLT panel
Spruce tangential	713	5 × 20	393.7	1843	0.106	487.7	59.8	728.1	0.005
MUF adhesive 2	713 2	4 × 0.25	425.0	1245	0.079	36.5	0.00	72.6	590.2
Insulation
PUR	194	140	37	1500	0.031	65.0	0.0001	0.001	945.0
Aluminum facer 3	1763	-	-	-	-	-	-	-	-
Façade cladding
Air gap (vertical)	17	40	1.3	1050	0.222	0.25	0.00	1000	0.00
Brick	512	100	1786.3	889	0.548	18.01	13.4	319.4	0.20

¹ ID is the identification code used in the Delphin 6.1.6 software material database. ² Material properties derived from experiments at Ghent University. ³ Contact condition with S_d-value of 3521.23 m (ID1763 ‘BARRIER ALU NET SD1500’ in the Delphine database).

. The climate data used is again one of the moisture reference years by Vandemeulebroucke et al. [33], where the indoor temperature is limited to values between 20 and 25 °C and the indoor relative humidity varies between 35 and 65%. The rain exposure coefficient is set to 0.5. A water source of 7% of the rain flux is assigned to the outer surface of the insulation board to take rain leakage through the facade cladding into account [34]. The air change rate in the air cavity is set at 5 air changes per hour [35].

In each of these simulations, the time of wetness TOW25/10 (the number of hours where a moisture content higher than 25% occurs during a period with a temperature higher than 10 °C) is calculated to get an idea of the possible degradation in different locations within the CLT panel. The TOW25/10 is calculated in the first millimeter exposed to the outside environment (e); in the first millimeter exposed to the protected environment (i); and for the average temperature and moisture contents in layers A, B, C, D, and E, as shown in Figure 1 (during service life).

2.2. Define Moisture Profile Types

Once the required level of detail is established, the next step is to estimate realistic moisture profiles at the end of the construction phase. Various parameters (discussed in detail below), such as wood type, glue type, delivery moisture content, orientation, and rain exposure coefficient, can have an impact on the moisture levels obtained during the construction phase.

To assess the impact of each of these parameters, a set of simulations is carried out (2.a, see Figure 2). For this set of simulations, a One Factor at a Time (OFAT) analysis is executed using the same baseline setup as described in Table 1. Each parameter is varied to get an overview of which parameters have the biggest influence on the moisture content. The differences in moisture content are evaluated according to the degree of detail found in step 1: ‘determining the necessary level of detail for moisture profiles.’ The date and time evaluated are based on the highest moisture contents found in these locations.

2.2.1. Wood Type

All spruce wood in the Delphin 6.1.6 database with a radial or tangential orientation is considered in this analysis. Their specifications are shown in

Table 3. Hygrothermal properties of varied wood types (part of simulation set 2.a).

Wood Type	ID 1	ρ kg m−3	cp J kg−1 K−1	λ W m−1 K−1	µ -	ϴ80 kg m−3	ϴsat kg m−3	Aw kg m−2 s−1/2
Spruce tangential (from Saxony)	713	393.7	1843	0.106	487.7	59.8	728.1	0.005
Spruce radial (from Saxony)	712	393.7	1843	0.112	487.7	59.8	728.1	0.012
Spruce radial	696	437.6	1391	0.122	474.7	62.6	445.1	0.013
Spruce tangential	695	393.7	1252	0.106	187.7	70.7	728.0	0.005
Spruce radial/tangential	626	425.0	1245	0.079	73.0	72.6	590.2	0.002
Spuce SW Radial	460	519.9	1189	0.130	236.2	67.6	692.0	0.058

¹ ID is the identification code used in the Delphin 6.1.6 software material database.

. Based on previous research at Ghent University [3], CLT with wood with ID713 was closest to measurements and was therefore chosen for the baseline scenario.

2.2.2. Glue Type

The µ and Sd values used in the base scenario were derived from tests with melamineurea-formaldehyde (MUF) glue at Ghent University [3].

First of all, it is assessed whether it makes a difference to simulate the glue layer as a material with a thickness of 0.25 mm and µ values of 3368 (dry cup) and 426 (wet cup) or to simulate the glue layer as a resistance with Sd values of 0.842 (dry cup) and 0.104 (wet cup) or to simulate the CLT panel as a whole wooden element with altered µ values taking the µ value of the glue into account. An overview of these scenarios is shown in the three first rows of Table 4. Note that these three scenarios have the same vapor diffusion resistance when assessing the CLT panel in its entirety.

According to Brandner et al. [36], the main types of adhesives currently used are melamineurea-formaldehyde (MUF), polyurethane (PUR) and emulsion polymer isocyanate (EPI). The other simulations (rows 4 to 6 of Table 4) were based on the µ of the pure adhesion films provided by Niemz et al. [37]. All adhesives are assumed to have a thickness of 0.25 mm [38].

Table 4. Hygrothermal properties of varied glue types (part of simulation set 2.a).

Glue Type	µdry [-]	µwet [-]	Sd,dry [m]	Sd,wet [m]
MUF simulated by altering the µ-value of the CLT panel 1,3	513.20	485.69
MUF simulated by adding its resistance 2,3			0.104	0.842
MUF simulated as an additional material with a thickness of 0.25 mm 3	3368	416
MUF material 4	1993	357
PUR material 4	1411	1020
EPI material 4	7143	311

¹ Material properties calculated through weighted average between wood and glue properties. Properties of the glue based on experiments at Ghent University. ² Conditions implemented as an interface resistance instead of a material layer, properties based on experiments at Ghent University. ³ Material properties derived from experiments at Ghent University. ⁴ Material properties derived from Niemz et al. [37].

Table 4. Hygrothermal properties of varied glue types (part of simulation set 2.a).

Glue Type	µdry [-]	µwet [-]	Sd,dry [m]	Sd,wet [m]
MUF simulated by altering the µ-value of the CLT panel 1,3	513.20	485.69
MUF simulated by adding its resistance 2,3			0.104	0.842
MUF simulated as an additional material with a thickness of 0.25 mm 3	3368	416
MUF material 4	1993	357
PUR material 4	1411	1020
EPI material 4	7143	311

¹ Material properties calculated through weighted average between wood and glue properties. Properties of the glue based on experiments at Ghent University. ² Conditions implemented as an interface resistance instead of a material layer, properties based on experiments at Ghent University. ³ Material properties derived from experiments at Ghent University. ⁴ Material properties derived from Niemz et al. [37].

2.2.3. Delivery Moisture Content

The delivery moisture content (dMC) is the term used to describe the moisture content of the CLT panel at the beginning of the assembly period. Delivery moisture contents may vary depending on the production environment, transportation time, protection during transportation and storage conditions. RDH Building Science (2021) [39] and RISE Research Institutes of Sweden (2019) [40] mention a moisture content of a maximum of 16% upon delivery. That is why this percentage is chosen for the baseline scenario.

Other delivery moisture contents assessed in this research are 12%, 14%, 18%, 20%, 25%, 30% and 40%. This is based on literature stating that CLT is usually made with a target moisture content of 12% and that the moisture content of the CLT will come into equilibrium with the relative humidity of the surrounding air creating variations between 4–5% over the year (Gustafsson, 2019) [40]. The Swedish CLT Handbook (Gustafsson, 2019) [40] also suggests a maximum moisture content of 18% to avoid fungal growth and rot. 20% and 25% are also often used as the cut-off values when executing different degradation calculations, such as TOW20/5 or TOW25/10 [6]. The values of 30% and 40% were considered reasonable based on measurements by Ghent University on a building site where a CLT construction was exposed for a prolonged time.

2.2.4. Orientation

Research has shown that the orientation plays a significant role in the degradation since the orientation not only impacts the rain exposure but also the exposure to wind and solar radiation, impacting the wetting and drying of the structure. The following orientations are taken into account: North (N), North-East (NE), East (E), South-West (SE), South (S), South-West (SW), West (W), North-West (NW). The West orientation is considered in the baseline scenario since this is a common wind direction in Belgium.

2.2.5. Rain Exposure Level

The rain exposure coefficient (RE) in Delphin is calculated through the following equation:

$$RE={\displaystyle \frac{C_R\ast C_T\ast O\ast W}{C_{R,fixed}\ast C_{T,fixed}\ast O_{fixed}\ast W_{fixed}}}$$

(1)

with:

• C_R = factor for terrain variations (0.665421, 0.771443 (fixed value), 1.00668, 1.174318);
• C_T = factor for topography;
• O = obstacle factor (0.2, 0.3, 0.4, 0.5; 0.6, 0.7, 0.8 (fixed value), 0.9, 1);
• W = wall factor (mean values for a whole wall are 0.3 and 0.4 (fixed value), distribution over the entire wall also include factors such as 0.2 and 0.5.

For each factor, the value is varied based on EN ISO 15927-3 [41]. The values can be found above, in between brackets. Note that no variations regarding nearby slopes (C_T) are taken into account since too many unknown factors play a role in the determination of this value. The value used for this study is 1, which is also the fixed value C_T,fixed.

Table 5. Rain exposure coefficients (part of simulation set 2.a).

	Rain Exposure Coefficient
Minimum	0.1
Q1	0.3
Median	0.5 (baseline)
Q3	0.8
Maximum 1	1.5
Maximum 2	1.8

¹ The maximum found when only considering the values for the whole wall in general (W = 0.3 or 0.4). ² The maximum found when also considering the wall factor W = 0.5. Note that values above 1.5 can only be obtained for the highest wall factor W = 0.5.

gives an overview of the selected rain exposure coefficients based on the found minimum, Q1, median, Q3, and the maximum of the above describe variations.

2.2.6. Define Reference Moisture Profile Types at the End of the Construction Phase

Based upon the One Factor at a Time analysis, as described above, the values of these parameters that lead to the highest moisture contents are determined. These parameters are combined in different ways and used as input for another set of simulations (2.b, see Figure 2), in order to define moisture profile types for further use in step 3 of the study.

2.3. Assess Degradation Risks for Wall Assemblies with the Reference Moisture Profiles

In the third and final part of this study, the moisture profile types found in step 2 are used as starting conditions in simulation set 3. These simulations start where the previous one stopped, i.e., at the moment where previously the highest moisture contents were found. The climate data used is again one of the moisture reference years by Vandemeulebroucke et al. [33], where the indoor temperature is limited to values between 20 and 25 °C and the indoor relative humidity is limited to values between 35 and 65%.

This set of simulations considers not only the nine defined moisture profiles as starting conditions (step 2) but also different wall assemblies with different insulation materials (both with different vapor permeability and hygroscopic behavior) (see

Table 6. Hygrothermal properties of materials in simulation set 3.

Material	ID 1	d mm	ρ kg m−3	cp J kg−1 K−1	λ W m−1 K−1	µ -	ϴ80 kg m−3	ϴsat kg m−3	Aw kg m−2 s−1/2
CLT panel
Spruce tangential	713	5 × 20	393.7	1843	0.106	487.7	59.8	728.1	0.005
MUF adhesive 2	713 2	4 × 0.25	425.0	1245	0.079	36.5	0.00	72.6	590.2
Insulation—PUR
PUR	194	140	37.0	1500	0.031	65.0	0.0001	0.001	945.0
Aluminum facer 3	-	-	-	-	-	-	-	-	-
Insulation—MW
Mineral wool	731	160	67.0	840	0.035	1.0	0.1	900.0	0.00
Insulation—CEL
Blown-in Cellulose	580	186	55.2	2544	0.048	2.05	6.3	780.0	0.56
Wood fiber board	693	22	160.8	1662	0.039	3.45	17.9	550.0	0.00
Façade cladding
Air gap (vertical)	17	40	1.3	1050	0.222	0.25	0.00	1000	0.00
Brick	512	100	1786.3	889	0.548	18.01	13.4	319.4	0.20

¹ ID is the identification code used in the Delphin 6.1.6 software material database. ² Material properties derived from experiments at Ghent University. ³ Contact condition with S_d-value of 3521.23 m (ID1763 ‘BARRIER ALU NET SD1500’ in the Delphine database).

). The wall assemblies selected for this study all have a similar U-value (i.e., between 0.15 and 0.16 W/m²K).

A full factorial assessment sheds light on the degree to which the wall assemblies are prone to potential decay due to built-in moisture. This is evaluated based on data calculated by a 4 year simulation period for the following elements:

• The Time of Wetness with a moisture content higher than 20% during a period with a temperature higher than 5 °C (TOW20/5). This should be a maximum of 720 h (1 month) to avoid mold growth [6];
• The Time of Wetness with a moisture content higher than 25% during a period with a temperature higher than 10 °C (TOW25/10). This should be a maximum of 168 h (1 week) to avoid wood rot [6];
• Number of daily doses, calculated through the product of a moisture-induced component and a temperature-induced component. This calculation is based on the dose-response model and should remain below the limit of 451 doses to avoid white rot and below the limit of 229 doses to avoid brown rot [7,8].

These degradation calculations are executed at different locations in the CLT panel: in the first mm exposed to the outside environment (e), in the first mm exposed to the protected environment on the inside (i), and for the average temperature and moisture contents in layers A, B, C, D, and E (see Figure 1, during service life).

Besides the above decay assessment of the CLT panel, the extent to which the insulation capacity of the insulation layer is impacted by the built-in moisture is also evaluated. This evaluation is done based on the following elements:

• Heat flux increase of the wall assembly with built-in moisture compared to the same wall assembly with the least amount of built-in moisture during the first heating season (from the start of the simulation, which occurs from the beginning of winter, until the end of April).
• Heat flux increases in the wall assembly with built-in moisture compared to the same wall assembly with the least built-in moisture during the second heating season (about one year after the end of construction from the beginning of November until the end of April).

3. Results

3.1. Determine the Necessary Grid Size for Moisture Profiles

Figure 3 provides a first insight into possible moisture profiles found during an exposure time of one month during the different months of the year.

The highest difference in moisture content between the lowest and highest moisture content in layer A and the highest difference in moisture content between the highest and lowest moisture content between the middle and the edge of the CLT panel were both found on 9 May at 8:00. The highest difference in average moisture content between layers A and C is found on 22 November at 7:00.

Table 7. Moisture content at the moments selected for set 1.b of simulations.

Date and Time	Average MC [%]			Difference in MC [%points]
	Layer A	Layer C	Total	Layer A	Middle—Edge of CLT Panel	Avg Layer A—Avg Layer C
10 May 8:00	29.32	14.20	16.55	160.44	160.66	15.11
22 November 7:00	37.07	13.98	18.29	124.71	129.27	23.10

gives an overview of the average moisture content and differences in moisture content at these moments.

The values found at these times were used as initial conditions for the second set of simulations. In this second set of simulations, the further evolution of moisture content was calculated under the condition that insulation and façade cladding were added. These simulations ran until each cell of the CLT panel dried out to a moisture content value below 25%, which happened within the year. The TOW25/10 was calculated and was found to be equal to zero for layer B, layer C, layer D, layer E, and the inside surface of the CLT panel. Figure 4 shows the calculated TOW25/10 for the outside surface of the CLT panel and in layer A for the different degrees of detail, both for the simulations starting on 10 May at 8:00 and 22 November at 7:00.

Figure 4 shows that the calculated TOW25/10 for the degree of detail ‘total’ (where all the wood in the CLT panel is assigned the average moisture content values of the wood in the CLT panel) does not match the calculated TOW25/10 for detailed version (where every millimeter in the CLT panel was assigned the moisture content output values from the first simulation). The degree of detail where each layer of wood in the CLT panel is assigned the average moisture content value of that layer also does not give an accurate approximation. However, the grid size is called ‘layer A,’ where every layer of wood in the CLT panel is assigned the average moisture content values of this layer with the exception of layer A, where the distinction is made between the outermost 5 mm and the other 15 mm, is an acceptable simplification. It is concluded that the average in each of the layers (especially of layer A) and the average found in the outermost 5 mm are the locations to look into in further assessment.

3.2. Define Moisture Profile Types

The impact on moisture content resulting from differences in wood type, glue type, delivery moisture content, orientation, and rain exposure coefficient are discussed below. The evaluated period is the month of November since this is the month where the highest total average moisture content of the CLT panel occurs (23 November at 7:00, see Figure 3).

3.2.1. Wood Type

As can be seen in Figure 5, the wood type has an impact on the moisture content in layer A. The biggest differences can be found on 20 November at 7:00 with a difference of 16.72 %points, where the highest value of 36.91% is found for wood with ID 460 and the lowest value of 20.19% for wood with ID 695. On 23 November at 7:00, the highest and lowest average moisture contents of layer A are found for wood with ID 712 (48.86%) and wood with ID 695 (32.78%), respectively. This impact is even more noticeable in the outermost 5 mm of this layer, where the highest and lowest average moisture contents are found for wood with ID 713 (106.20%) and wood with ID 695 (61.79%), respectively.

The difference in moisture content in the other layers is less noticeable. On 23 November at 7:00, the difference for layer B, layer C, layer D, and layer E are limited to 2.22, 2.70, 2.75, and 4.42 percentage points, respectively.

3.2.2. Glue Type

The different glue types result in limited differences in moisture content for the different layers, with the biggest difference found in layer B, which has a difference of 2.7 %points. It should however be noted that the moisture content values in the CLT panel, where the panel was simulated as a whole wooden element (with altered µ values taking the µ values of the glue into account), always deviated the most from the moisture content values in the CLT panels where the glue was implemented as a material layer or resistance. For layers A and E, the moisture content values found are lower in case the CLT panel was simulated as a whole wooden element compared to the panels where the glue was implemented as a separate element. For layers B, C, and D, the moisture content values found are higher in case the CLT panel was simulated as a whole wooden element.

3.2.3. Delivery Moisture Content

As can be seen in Figure 6, the effect of delivery moisture content diminishes over time. On 11 November, the difference in moisture content values in layer A was already reduced to less than 8 %points. On 23 November at 7:00, there is only a difference of 5.6 %points found between the highest (49.30% for dMC of 40%) and lowest (43.71% for dMC of 12%) average moisture contents of layer A. A similar impact is found for the outermost 5 mm of this layer.

The delivery moisture content has little impact on the outside layers, especially when the delivery moisture content values only vary between 12 and 20%. In layer A, this initial difference (of 8 %points) is already reduced to 2.30 %points by 11 November and is further reduced to 1.81 %points by 23 November at 7:00. In layer E, this initial difference of 8 %points is reduced to 1.86 %points by 23 November at 7:00 (see Figure 7).

However, in the middle layers of the CLT panels, the difference remains visible for longer periods of time (see Figure 7). After one month, the spread in average moisture content in layers B, C, and D drops down to 15.73 %points, 26.23 %points, and 18.32% points, respectively. When only considering the delivery moisture values varying between 12 and 20%, the impact of this initial moisture content is limited as well. In this case, the spread in average moisture content after one month in layers B, C, and D drops already down to 2.98 %points, 3.32 %points, and 2.93 %points, respectively.

3.2.4. Orientation

As can be seen in Figure 8, the orientation has an impact on the moisture content in layer A. On 23 November at 7:00, the difference in layer A is about 24.83 %points between the different orientations. This is even more pronounced in the outer 5 mm, where this difference is almost 84.26 %points. In layer B, the difference between the different orientations showed to be limited to 1.75 %points, and in layers C to E, it was even less.

3.2.5. Rain Exposure Level

As can be seen in Figure 9, the rain exposure level has an impact on the moisture content in layer A. The biggest differences can be found on 23 November at 7:00 with a difference of 40.26 %points where the highest value of 61.80% is found for a rain exposure of 1.8 and the lowest value of 21.54% for a rain exposure of 0.1. This impact is even more noticeable in the outermost 5 mm of this layer, where the highest (153.21%) and lowest average moisture content (25.16%) create a difference of 128.05 %points.

As the exposure time increases, slight differences in moisture content in the other layers become noticeable: after about 2 weeks of exposure, slight differences occur in layer B, and after about 3 weeks of exposure, slight differences occur in layer C. Whereas the differences in layer A clearly fluctuate with the rainfall, the differences in the deeper layers do not decrease as fast after a period without rain. These differences remain limited and are the highest at the end of the simulation period (after one month) with a difference of about 4.4 %points in layer B, about 1.6 %points in layer C and less than 1%point in layers D and E.

3.2.6. Define Reference Moisture Profile Types at the End of the Construction Phase

The wood type, orientation, and rain exposure level have the highest impact on moisture content in the outer 5 mm and, to a lesser extent, also in the following 15 mm of layer A. The highest moisture contents (and also the highest differences in moisture content) in these zones are, in general, found on 23 November at 7:00. Figure 10 shows the moisture contents found in the different layers at this moment. The moisture content in layers B, C, and D is influenced by differences in delivery moisture content. Note that the differences were higher at the beginning of the building phase and decreased significantly by 23 November. The moisture content in layer E is barely impacted by any of the variations of parameters.

• Based on the results in Figure 10, nine profiles were defined using the parameters shown in

  Table 8. Input for the parameters wood type (Wood), delivery moisture content (dMC) in %, orientation (Orient), and rain exposure level (RE) to determine the moisture profile types.

  Moisture Profile Type		Wood	dMC	Orient	RE
  a	Baseline	ID713	16	W	0.5
  b	Baseline with increased rain exposure	ID713	16	W	1.8
  c	Baseline with increased delivery MC	ID713	40	W	0.5
  d	Worst case for outer 5 mm 1	ID713	40	SW	1.8
  e	Worst case for layer A 2	ID712	40	SW	1.8
  f	Baseline with decreased rain exposure	ID713	16	W	0.1
  g	Worst case for outer 5 mm 1 with lowest rain exposure	ID713	40	SW	0.1
  h	Worst case for layer A 2 with lowest rain exposure	ID712	40	SW	0.1
  i	Best case	ID695	12	E	0.1

  ¹ Worst case based on the highest average moisture contents found for the outermost 5 mm. ² Worst case based on the highest average moisture contents found for layer A. Variables in bold indicated that these inputs differ from the baseline scenario (a).

  .
• Moisture content profile type a is based on the baseline scenario;
• Moisture content profile type b is the baseline scenario with an increased rain exposure level (RE = 1.8) since Figure 10 shows that this parameter has the highest influence on the moisture content in the outermost 5 mm and in layer A in general.
• Moisture content profile type c is the baseline scenario with an increased delivery moisture content (dMC = 40%) since Figure 10 shows that this parameter has the strongest influence on the moisture content in layers B, C, and D.
• Moisture content profile type d is based on the highest average moisture contents found for the outermost 5 mm.
• Moisture content profile type e is based on the highest average moisture content found for layer A.
• Moisture content profile type f is the baseline scenario with a decreased rain exposure level (RE = 0.1) since Figure 10 shows that this parameter has the highest influence on the moisture content in the outermost 5 mm and in layer A in general.
• Moisture content profile type g is based on the highest average moisture contents found for the outermost 5 mm, except the rain exposure level. A decreased rain exposure level (RE = 0.1) since Figure 10 showed that this parameter has the strongest influence on the moisture content in the outermost 5 mm and in layer A in general.
• Moisture content profile type h is based on the highest average moisture contents found for layer A, except the rain exposure level. A decreased rain exposure level (RE = 0.1) since Figure 10 showed that this parameter has the strongest influence on the moisture content in the outermost 5 mm and in layer A in general.
• Moisture content profile i is based on the lowest average moisture contents found for the outermost 5 mm and layer A in general.

Note that the glue type is not varied. It is for all profile types the same as in the baseline profile type.

Table 9. Moisture content [%] in the different layers of the CLT panel was used as input values for the fourth set of simulations for the different moisture profile types.

	Layer A Outer 5 mm	Layer A Inner 15 mm	Layer B	Layer C	Layer D	Layer E
a	106.20	23.99	18.70	17.78	17.80	19.62
b	153.21	31.33	20.43	18.34	17.97	19.72
c	112.43	28.26	35.16	40.87	38.35	25.38
d	143.43	38.10	37.92	40.87	38.17	25.35
e	157.08	76.66	39.80	40.31	33.84	25.51
f	25.16	20.33	17.42	17.36	17.64	19.45
g	28.11	24.98	32.94	40.87	37.90	25.02
h	28.76	26.61	30.02	39.98	33.50	25.23
i	21.64	19.84	18.49	18.49	18.48	19.92

gives an overview of the moisture content obtained for the different layers in the CLT panel after the CLT panel was exposed from 1 November 7:00 until 23 November 7:00. The values were extracted on 23 November at 7:00 since this is the moment where the highest total average moisture content of the CLT panel occurs (Figure 3).

3.3. Assess Degradation Risks for Wall Assemblies with the Reference Moisture Profiles

The moisture profiles (see Table 9) are used as starting conditions in a fifth batch of simulations. These simulations start on 23 November at 7:00 a.m., the moment when the insulation and façade cladding were added to the CLT panel.

This set of simulations considers not only the 6 defined moisture profiles as starting conditions but also different insulation materials (with different vapor permeability and hygroscopic behavior).

Figure 11 gives an overview of the damage evaluation for the different types of insulation materials at the different locations in the CLT panel.

Wall assemblies with moisture profile i remain at zero for every degradation calculation and are therefore not further discussed.

3.3.1. TOW20/5

All wall assemblies uphold a TOW20/5 of zero at the interior surface. At the exterior surface of the CLT panel, wall assemblies with PUR insulation only stay below the threshold of 720 h that should be respected to avoid mold growth [6] in case of moisture profile type f (41 h) and moisture profile type i (1 h)—i.e., the moisture profile types based on exposure conditions with a reduced rain exposure level and without an increased delivery moisture content. Wall assemblies with cellulose insulation do not exceed the 720-h threshold under the scenarios for the lowest rain exposure of 0.1 (moisture profile types f, g, h, and i) and in case of an average rain exposure of 0.5 in combination with a delivery moisture content of 16% (moisture profile type a). Wall assemblies with mineral wool only exceed this limit at the exterior surface for the moisture profile types based on the worst cases for the outermost 5 mm (type d) and for layer A (type e).

Note that, in the case of mineral wool and cellulose insulation, the moisture content in the exterior layer drops to an acceptable moisture content more quickly than deeper in the CLT panel. In layer B, the values exceed the limit of 720 h in a few more cases: the wall assembly with mineral wool and moisture profile type g (740 h), the wall assembly with cellulose insulation and moisture profile type g (966 h) and the wall assembly with cellulose insulation and moisture profile type h (926 h).

3.3.2. TOW25/10

As can be seen in Figure 11, the TOW25/10 only stays below 168 h for wall assemblies with mineral wool with moisture profile type f and i. In all other assemblies, this limit is exceeded in at least one layer of the CLT panel.

The time of wetness only remains below the threshold of 168 in layer B for all insulation types in the case of moisture profile types without an increased built-in moisture content (moisture profile types a, b, f, and i). The longest time of wetness (4068 h, i.e., approximately 5.5 months) can be found for the wall assembly with PUR insulation and an initial moisture profile type e. For the wall assemblies with mineral wool or cellulose insulation, it takes the structure about 2 months (1478 h) or 3 months (2096 h), respectively, to dry to moisture contents below 25%. This is a notably shorter period of time, but still way above the 1-week (168 h) threshold that should be respected to avoid wood decay [6].

A similar trend is visible in layer C, for which the TOW25/10 remains zero for wall assemblies with moisture profile types a, b, f, and i but exceeds 431 h for all other moisture profile types. The TOW25/10 even reaches values above 1386 h for moisture profile type e.

A similar trend with lower values is visible in layer D, for which the TOW25/10 reaches values around 265 h in the case of moisture profile type e. The time of wetness remains zero or one in layer E in all cases.

3.3.3. Daily Doses

When counting the number of daily doses [7,8], the highest number can again be found for the wall assembly with PUR insulation and moisture profile type e. The amount of doses in layer A is 115.5 doses, which is still well below the limit of 451 doses to avoid white rot and below the limit of 229 doses to avoid brown rot. Although this value is still below the threshold and is therefore considered ‘safe,’ it is clear that this type of wall assembly has a higher degradation risk than similar assemblies with mineral wool (42.3 doses in layer B and 34.2 doses in layer A) and cellulose insulation (between 56 and 57 doses in layer A and B).

Note that the time of wetness and number of doses are, in general, the highest in the exterior mm and in layer A. However, for the higher delivery moisture contents in the wall assemblies with mineral wool insulation, the highest values can be found in layers B or C.

3.3.4. Heat Flux Increase

In general, based on TOW and the number of doses, wall assemblies with mineral wool look like the best option. This is probably due to the low vapor diffusion resistance of this insulation material. However, based on the heat flux increase during the first heating season, the wall assemblies with mineral wool would be considered the worst option (with an increase up to 70% for moisture profile type e), followed by wall assemblies with cellulose insulation (with an increase up to 37% for moisture profile type e). Wall assemblies with PUR insulation suffer the least from an increased heat flow. For wall assemblies with PUR insulation, the increase only climbs above 5% for moisture profile types h, c, d, and e. For wall assemblies with mineral wool or cellulose insulation, the heat flux increase only stays below 5% for moisture profile types f and i.

However, this heat flux increase is only significant during the first heating season.

4. Discussion

The results of this study provide a comprehensive look into how the simulation detail and varying moisture profiles affect the evaluation of decay risks in CLT wall assemblies. Several key observations emerged from the analysis:

4.1. Determine the Necessary Grid Size for Moisture Profiles

The simulations revealed that using a single average moisture content for the entire panel or even for each layer of wood fails to capture critical variations—especially in the outermost 5 mm of layer A. The “layer A” approach, which distinguishes between the outer 5 mm and the remaining 15 mm in layer A, more accurately approximates the moisture distribution needed for degradation assessments. This finding underscores the importance of selecting an appropriate grid size when modeling moisture behavior, as oversimplified grids may lead to underestimation of decay risk.

4.2. Determination of Moisture Profiles Types

The study examined how factors such as wood type, glue type, delivery moisture content, orientation, and rain exposure level influence the moisture content, particularly in layer A.

• Wood type: Significant variations in moisture content were observed in layer A (with a difference of 16 %points), especially in the outermost 5 mm (with a difference of 44 %points), indicating that material properties strongly affect how moisture is absorbed and retained.
• Glue type: Although differences were generally limited (with the highest difference of 2.7 %points found in layer B), the simulation showed that the way of simulating the glue layer (whether integrated or modeled separately) could influence the moisture dynamics in the CLT panel.
• Delivery moisture content: The initial moisture content had a diminishing impact over time, though its influence persisted longer in the middle layers compared to the surface. The delivery moisture content has little impact in the outside layers, especially when the delivery moisture content values only vary between 12 and 20%: the spread in average moisture content in layer C drops to 3.32 %points, the spread in average moisture content in layers B and D this drops to less than 3 %points and the spread in moisture content in layers A and E even drops to 1.58 %points after one month. When including the delivery moisture contents of 25, 30 and 40%, the spread in average moisture content in layers A and E drops to less than 7 %points; in layers B and D, the difference in average moisture content remains more than 15 %points, and in layer C this difference remains above 26 %points.
• Orientation: The orientation of the CLT panel had pronounced effects on the moisture content in layer A and the outermost 5 mm of the CLT panel. The highest average moisture content in the different layers is found in the West, South-West, and South orientations. A difference in average moisture contents in layer A was found to be about 24.83 %points between the different orientations. In the outer 5 mm, the difference is almost 84.26 %points. Both emphasize that exposure conditions during the construction phase critically affect the moisture profile. The difference between the different orientations in the other layers is limited to 1.75 %points.
• Rain Exposure: The rain exposure level had the strongest influence on the moisture content in the outermost 5 mm (with a difference of 128.05 %points) and in layer A in general (with a difference of 40.26 %points), underscoring the need for protection against the outdoor climate during the construction period. Only slight differences in moisture content in the other layers become noticeable after about 2 weeks (layer B), 3 weeks (layer C), or even later (layers D and E).

The moisture content in layer A is mainly influenced by the rain exposure level (creating a difference of 40.26 %points), followed by orientation (creating a difference of 24.83 %points) and wood type (creating a difference of 16 %points). The moisture content in layers B, C, and D is influenced by differences in delivery moisture content (with a difference of 15.73 %points, 26.23 %points, and 18.32 %points, respectively). The moisture content in layer E is barely impacted by any of the variations of parameters.

The implications of the above findings are that the necessary precautions have to be taken to limit the delivery of moisture content and to limit the exposure to rain during the construction phase by, for example, adding a protective cover.

4.3. Degradation Assessment of Different Wall Assemblies with Reference Moisture Profiles

The subsequent degradation simulations, which incorporated the reference moisture profile types and various insulation materials, revealed clear trade-offs:

• TOW20/5:
  At the interior surface, all wall assemblies achieved a TOW20/5 value of zero, indicating effective moisture control in this zone. However, at the exterior surface, performance varied markedly with insulation type and moisture profile. For instance, all but two wall assemblies with PUR insulation exceeded the 720-h threshold in all cases (except for moisture profile types f and i), while those with mineral wool insulation breached the limit solely under the worst-case scenarios (moisture profile types d and type e). Wall assemblies with cellulose insulation did not exceed the 720-h threshold under the scenarios for the lowest rain exposure of 0.1 (moisture profile types f, g, h, and i) and in case of an average rain exposure of 0.5 in combination with a delivery moisture content of 16% (moisture profile type a).
• TOW25/10:
  The TOW25/10 metric remained below the critical 168-h threshold exclusively for wall assemblies under moisture profile types f and i. In all other configurations, at least one layer within the CLT panel experienced prolonged wetness. This shows wood decay (calculated through the TOW25/10 method) would only occur for moisture profiles for which the delivery moisture content is limited (dMC = 12 or 16% and not 40%) in combination with a decreased rain exposure level (RE = 0.1 and not 0.5 or 1.8).
  PUR-insulated assemblies with moisture profile type e exhibited a TOW25/10 of 4068 h—approximately 5.5 months—far exceeding the acceptable limit. Assemblies with mineral wool or cellulose insulation dried more rapidly (approximately 2 and 3 months, respectively). Although these durations still surpassed the one-week (168-h) guideline for preventing decay [6], all wall assemblies have dried out by mid-May at the latest.
• Daily doses:
  While the absolute number of daily doses remained below the critical thresholds for the development of white or brown rot, the highest doses were observed in wall assemblies with PUR insulation, suggesting a comparatively higher degradation risk while, however, still being considered “safe.”
• Heat flux increase:
  A noticeable trade-off was observed in thermal performance. Assemblies with mineral wool, although effective in reducing moisture retention, exhibited increased heat flux during the first heating season, compromising energy performance. On the other hand, PUR insulation provided the smallest increase in heat flow. However, this heat flux increase no longer exists as from the second heating season.

Together, these results emphasize that a nuanced approach to moisture simulation is vital. The selection of simulation detail and careful consideration of material and exposure parameters are critical to accurately assess decay risks.

Results regarding the calculation of the daily doses show that there is no degradation due to white or brown rot to be expected. Since this evaluation was based on several ‘worst case’ initial moisture content profiles, one could conclude that moisture contents in wall assemblies don’t exceed critical values after one month of exposure during the construction phase. Since this study only takes the middle of wall assemblies into account, the findings could be different near wall edges and possible moisture traps.

Moreover, the interaction between moisture retention, drying rates, and thermal performance suggests that design strategies for CLT wall assemblies must balance durability and energy efficiency.

5. Conclusions

This study evaluates the decay risks in CLT wall assemblies by simulating moisture profiles with varying degrees of detail and by assessing the impact of several influencing parameters. The main conclusions are as follows:

• A sufficiently detailed moisture profile grid—particularly one that distinguishes the outermost 5 mm and the wooden layer exposed to rain (layer A) from the rest of the panel—is necessary to accurately predict the evolution of moisture content and subsequent decay risk. Simplified profiles that use an overall average or even layer-specific averages are insufficient.
• Moisture content values in CLT assemblies are significantly affected by wood type, initial moisture levels, orientation, and rain exposure. Of these, the outer region of the panel (layer A) is most sensitive to these factors, highlighting the need for targeted mitigation strategies during the construction phase.
• The moisture content in layer A is mainly influenced by the rain exposure level (creating a range of 40.26 %points), followed by orientation (creating a range of 24.83 %points) and wood type (creating a range of 16 %points). The moisture content in layers B, C, and D is influenced by differences in delivery moisture content (with a range of 15.73 %points, 26.23 %points, and 18.32 %points, respectively). The moisture content in layer E is barely impacted by any of the variations of parameters.
• The degradation calculations reveal that all assemblies maintained a TOW20/5 of zero hours at the interior surface, confirming effective moisture control. However, at the exterior surface, only configurations with low rain exposure (RE = 0.1) and low delivery moisture content (12–16%) could keep the TOW25/10 below the critical 168-h threshold. The longest time of wetness was found for wall assemblies with PUR insulation, which dried in about 5.5 months. Wall assemblies with mineral wool dried in about 2 months and those with cellulose in about 3 months.
• Although the TOW calculations indicate a clear risk of decay, the accumulated degradation doses (a more detailed calculation approach) remained below the safety thresholds (451 doses for white rot and 229 for brown rot). Even though the number of doses remained below the thresholds PUR-insulated assemblies showed a comparatively higher number of doses.
• While mineral wool achieved superior moisture management, it also led to a heat flux increase during the first heating season (increase up to 70%). Built-in moisture in wall assemblies with cellulose insulation also led to a notable heat flux increase during the first heating season (increase up to 37%). Wall assemblies with PUR insulation suffer the least from an in-creased heat flow (increase up to 8%). This heat flux increase is nonexistent in the following heating seasons since the structure dried out sufficiently within the first year.

The findings provide a methodological framework for balancing simulation detail and practical considerations in decay risk assessments.

It should be noted that this research only considers wall assemblies at a sufficient distance from wall edges and possible moisture traps. Although these other locations often have a higher risk of decay, this research shows that the middle of walls should not be ignored either. However, a similar methodology could be put forward for assessing moisture accumulation at edges and moisture traps, horizontal building elements, and other additional design modifications or alternative materials.

As this study only took into account Belgian climatological conditions, it would be useful to assess other climate regions.

Even though the simulations in this study were based on the validated software model Delphin 6.1.6, future work could focus on validating these specific simulation results with experimental data.