Structural Engineering of Cross-Laminated Timber in South Africa

How to complete this course

The course is open access and the content is available to anyone with internet access. This course should be completed in the sequence of the numbered sections. We make use of four components in our course (a) sections with text and illustrations, (b) video clips to explain some of the concepts, (c) a short online quiz after each section to test your acquired knowledge and (d) a test, available to learners who wish to obtain a certificate of completion. Please contact us at info@thewoodapp.com with the short course name to complete this test.

To make the course accessible to most people, we’ve tried to limit the time needed to complete it (an estimated 2.5 hours). This course will give you an overview of selected building materials impact and the environment, and design strategies to incorporate timber in a sustainable way. For in-depth knowledge on some topics you might require input from other sources such as national standards, material suppliers, selected textbooks, articles, etc.
If you have any questions or suggestions, please contact us at
info@thewoodapp.com.

Structural Engineering of Cross-Laminated Timber in South Africa Video 1

4. Treatment

5. Structural fire design of CLT

5.1 Charring rate of SA Pine
5.2 CLT in fire
5.3 Connections in fire
5.4 Quiz 4 & 5

6. References

Appendix A – CLT Design Tables

Disclaimer
In the compilation of this module, free use was made of published information such as text, figures, drawings, tables, graphs, etc. As the use of such material is subject to copyright considerations, and the suitability and relevance of this content is in the process of being assessed, the content of this module is only reserved for personal use and the purpose intended. To adhere to copyright regulations, any publication of the module or parts thereof considered, is subject to obtaining the necessary copyright agreement from the publishers by the author. Photographs taken and figures, drawings, tables and graphs generated by the author, are subject to copyright as well.

Whilst all and the utmost care has been taken to ensure the accuracy of the information contained in this module, no warranty can be given regarding the use, suitability, validity, accuracy, completeness or reliability of the information, including any opinion or advice.

1. Overview of timber as a structural material

This section aims to introduce important topics on the structural behaviour of wood and CLT so that the engineer can better understand how to design with it. The explanations are kept brief, as this is not the main purpose of this course. It is recommended that the engineer does further research to educate themselves on these topics for a fuller understanding of timber as a structural material.

1.1      Structure of a tree

Wood is a plentiful and sustainable natural material that has proven to be extremely useful to human civilisation since ancient times. Humans have learnt to harness wood as a material to serve a plethora of purposes, examples include buildings, furniture, fire, paper, modes of transport (ships, planes) and weapons. However, it is worth keeping in mind that wood did not evolve with these functions in mind, but rather to serve its purpose within a tree. Therefore, it is useful to look at the structure of a tree to gain a fundamental understanding of the properties of timber.

One of the primary functions of a tree trunk (where the majority of wood is located) is to support the branches, which in turn carry the leaves which seek sun exposure for photosynthesis. Thus, the wood must withstand high compression stresses due to the mass of the tree, as well as bending (combination of axial tension and compression stress) caused by wind. It is for this reason that timber performs well in these strength properties. These forces and stresses are shown in Figure 2. Furthermore, it is in a tree’s favour to remain lightweight, and thus wood has evolved to maximise strength in the most materially efficient way to keep the weight low. The result is that timber has a higher strength-to-weight ratio than most structural materials including steel and concrete.

Another primary purpose of the trunk is to transport water from the roots to the leaves, whilst the bark transports sap from the leaves downwards. A rough analogy that is often used to explain the anatomy of wood is to imagine a bundle of straws (tracheids) that are held together with elastic bands (lignin), as shown in Figure 3. The ‘straws’ form the pathway for the water and sap to flow. With this structure in mind, it is easy to visualise the ‘straws’ being crushed or pulled apart in the direction perpendicular to grain without much force, which is a simple way to understand why the compression and tension strength are so much lower in this direction.

Figure 3: Wood structure simplified as a ‘bundle of straws’ (Wessels, 2020)

1.2      Timber species, grades, and design properties

The fact that wood is both orthotropic and heterogenous means that engineers need to use different stress values depending on what they are designing for. Table 1 gives the various characteristic stress values for different grades of SA Pine (collective term for a number of different Pinus species grown in SA). Although provision is made for four different grades, the vast majority of structural timber available in SA is S5. Whilst it is possible to source S7, it is much less common and more costly. S10 and S14 are very rare in SA pine. It is recommended that engineers specify S5 and only resort to higher grades if absolutely necessary. The exception is glue-laminated beams which often have higher grades (S7-S10). The characteristic stresses of timber are based on 5^th percentile values (i.e. statistically only 1 out of 20 planks will be weaker than this value), whilst the design modulus of elasticity (MOE) is the mean value. The rationale here is that a strength failure carries a far higher risk to human life or property than an excessive deflection, and thus more conservative values should be used. NB: for buckling calculations the 5^th percentile MOE is recommended (+-MOE_mean/1.6), as this forms part of a strength check.

Table 1: Characteristic properties of SA Pine – SANS 10163-1:2003

In addition to SA Pine, designers also have commercially-grown eucalyptus timber available for structural purposes. Whilst the tradename Saligna is commonly used, the species is actually Eucalyptus grandis or Eucalyptus grandis X urophylla. Unfortunately, there is little data on the structural properties of Saligna and no clear guidelines as to what values to use. The SA Pine grades are sometimes applied to Saligna, but it is doubtful how applicable these values are. Arguably the best source of design values for Saligna comes from a study at Stellenbosch University, shown in Table 2 (Crafford and Wessels, 2016). However, it should be noted that these values are based on young, green timber, and that except for bending and tensile parallel to grain tests, the number of samples tested was relatively small by timber research standards. Saligna used for some structural applications, especially glue-laminated beams, typically comes from older trees, and the expected strength properties can be much higher.

Table 2: Structural properties of young, dry Eucalyptus grandis (Crafford and Wessels, 2016)

NOTE: these values are merely recommendations and not validated or published in any structural codes, and the author accepts no responsibility for the use of these values

When it comes to other species of timber that are sometimes used for structural purposes (e.g. Eucalytpus cladocalyx, Oregon pine / Douglas Fir, Poplar, Balau and others), the engineer is forced to make an informed estimation, unless the timber is an imported structural timber that has been graded (e.g. spruce from Europe). More detailed information on grading and characteristic values to use can be found in the WoodApp course Stress Graded Timber.

1.3 Structural Challenges of timber

Although timber does function very effectively as a structural material, it is important to understand its limitations and material-specific challenges. These need to be taken into account if engineers want to design safe, practical and durable buildings.

The primary challenges are durability, high variability, combustibility and serviceability. These are discussed below:

1.3.1 Durability

Wood comes from the Earth, and the Earth wants it back. As an organic product there are natural phenomena that have developed alongside wood that can degrade the timber. The mechanisms by which this occurs are: a) insect attack, b) fungal decay and c) combustion.

Insect attack (e.g. wood borers, termites) may or may not be a threat, depending on geographic location and prevalence of pest. In South Africa, since the 1970s pressure-treatment of pine has been mandatory in the coastal areas, but is not required inland. However, it is recommended to always use treated timber to be safe. Termites pose a significant risk inland and great care should be taken if they are known to be a problem in the area.

Fungal decay can only occur if timber reaches and remains at a certain moisture content (rule-of-thumb 20%). As long as timber is allowed to quickly dry out after being wet (or never gets wet at all), decay should never occur as the conditions are not conducive to fungal life. The exception is a very humid environment where the moisture content in the air can be high enough for timber to reach a moisture content of +-20% (i.e. bathrooms or kitchens without adequate ventilation). Keeping the timber dry or allowing it to ‘breathe’ once wet can be ensured through the correct detailing by designers, and quality workmanship by the contractor.

1.3.2 High variability

Being a natural product there is an extremely high degree of variability in the properties of timber. Unlike concrete or steel which are designed and produced by humans with a high degree of certainty and reasonably low variation, timber can have values differing by an order of magnitude. A few (but not all) of the factors that influence the properties are:  species, age, speed of growth, location within a tree, irregularities (knots), moisture content and duration of load.

1.3.3 Combustibility

It is common knowledge that timber is a combustible material. This makes it fundamentally different to other common building materials such as concrete, masonry and steel. That said, timber buildings can be designed to be safe in the event of a fire. Timber predictably forms a char layer during a fire which protects an inner residual cross-section with unaltered structural properties that can safely carry the loads during and even after a fire event.

1.3.4 Serviceability

The fact that timber is very lightweight and relatively soft is both a blessing and a curse. These properties have a number of advantages including: smaller foundations and being relatively easy to machine and process. However, the lightweight nature and high flexibility also makes timber far more susceptible to deflections, vibrations and acoustics. It is for this reason that CLT panels (especially floors) are often governed by these serviceability criteria rather than their strength. Failure to take these into account can render the building uncomfortable or unacceptable to the end-user or inhabitant.

2. CLT Fundamentals

2.1 Terminology

When it comes to CLT there is some specific terminology that professionals may not be familiar with:

• Lamella – a layer in the CLT panel. The plural is lamellas or lamellae
• Rolling shear – shear stress leading to shear strains in a plane perpendicular to the grain direction
• Layup – the size and order of lamellas in a CLT panel
• Major strength direction – the direction parallel to the grain of the outer lamellas
• Minor strength direction – the direction perpendicular to the grain of the outer lamellas
• MOE – in timber design, the Young’s Modulus is often referred to as the MOE (Modulus of Elasticity)
• MOR – in timber design, the bending strength is often referred to as the MOR (Modulus of Rupture)

2.2 Structural use cases of CLT

CLT is typically used most often in floors and roofs where it is primarily loaded out-of-plane. CLT does however also have exceptional in-plane and axial strength and stiffness, and can be effectively used to carry vertical and shear loads in a wall configuration. For shear walls that provide lateral stability to a building, it is typically the strength of the connection, and not the panel itself, that governs the design.

Provided the connections are adequately designed, CLT floors made up of several panels can act as a membrane to transfer lateral loads throughout the structure. Again, it is the strength of the connection and not the panel that governs here.

Although a CLT panel is capable of spanning in two directions provided it is supported along all four sides, the significantly reduced minor strength direction properties make CLT far more suited to a one-way spanning system as opposed to a flat-slab type of structure. Furthermore, most floors in buildings consist of multiple panels side-by-side with the joints not being able to transfer bending moments. For these reasons, CLT floor and roof panels are usually one-way spanning and supported on beams, which then transfer the loads to the columns/walls.

CLT can also be used effectively as a beam or lintel. This will be most efficient if the outer lamellas are orientated horizontally i.e. in the direction of the span. GLT (glue-laminated timber) does however remain the preferable option for beams and columns.

2.3 Layup

CLT is almost always made up of an uneven number of lamellas (typically 3, 5 or 7). The layers are orthogonal to each other, with the outer layers always orientated in the same direction.  CLT is very sensitive to differences in layup that can have a significant effect on strength and stiffness. Engineers must be sure to specify a specific layup for a given panel thickness; merely specifying a thickness is not enough to ensure the required structural properties of the panel.

An example of this is illustrated in Table 3, in which the structural design properties of two 143mm thick CLT panels are given. The layups differ however, with a resultant difference in bending strength and stiffness. The shear properties are less affected, with zero difference in shear strength. The reason for this is that the shear strength is governed by rolling shear in the central lamella, which does not change between the layups in the example. The table is merely an example, the difference in properties between panels will vary on a case-by-case basis.

Any thickness of lamella is theoretically possible, but not necessarily practical. Lamella that are too thin are flimsy to handle in the factory and can increase the production cost due to using more adhesive and having higher handling costs. Thicker lamella can require a greater press force to overcome internal stresses if the plank is warped. Currently manufacturers in SA (XLAM and MTT) use the following standard lamella thicknesses:

• 15mm – obtained from a 38mm plank that is split with a bandsaw
• 22m – obtained from a 25mm plank
• 33-35mm – obtained from 38mm plank

The lamella thickness is intrinsically linked to the nominal dimensions of available planks (25mm, 38mm, 50mm, 76mm). The final lamella thickness will always be less than the nominal dimension due to the planing of top and bottom to achieve a uniform thickness. Whilst the standard lamella sizes above are current at the time of writing, they are subject to change and are ultimately dependent on the manufacturer. It is recommended that engineers check with manufacturers before they specify a layup. Deviating from a manufacturing norm is of course possible, but it will most likely increase timber wastage and production effort, and subsequently cost.

To enhance the structural properties of CLT in an efficient way, one option available to the engineer is to specify higher grade timber in selected lamellas only. To enhance the bending properties, the outer lamella should have a higher grade. To enhance the shear strength, the central layer should have a higher grade. This does however require a higher degree of quality control and assurance by the manufacturer, as there is no way of visually determining the grade of timber after machining (unless marked by manufacturer), and planks could easily be mixed up on the factory floor. Thus, a higher degree of trust in the manufacturer’s process is required to safely specify such layups. With all lamellas having the same grade, there is less risk.

The structural properties of a CLT panel can be mildly enhanced through the edge-gluing of planks in a lamella. However, this increases the manufacturing effort and may increase the cost as per the discretion of the manufacturer without much benefit in terms of structural performance.

2.4 Strength Directions

A fundamental concept regarding CLT is the difference between major and minor strength directions. Timber has dramatically reduced strength and stiffness perpendicular to grain, and thus there is a substantial difference in CLT structural properties depending on the directionality of the panel.

As an example of this, Table 4 compares minor strength direction properties as a percentage of major strength direction properties for three CLT layups. Three observations can be made:

1. Effective bending stiffness reduces significantly in minor direction, whilst shear stiffness is unaffected
2. Bending strength is influenced more than shear strength
3. Layups with a greater number of lamellas are proportionally less affected

For the reasons explained above, it is vital that the orientation of lamella is specified and ultimately indicated on shop drawings to ensure the panel is manufactured correctly. Two basic rules typically apply:

1. For floors/roofs, outer lamella should be orientated in the direction of the span.
2. For walls, the outer lamella should be vertically orientated to maximise the axial capacity.

2.5 Rolling shear

Rolling shear strength is a property of timber that is important for CLT design as it governs the shear strength of the panel. Engineers who are only familiar with the design of solid and glue-laminated timber may not be familiar with rolling shear, as it is only ever important for laminated timber with lamella perpendicular to the span. The greater thickness of these layers in CLT (as opposed to plywood with very thin layers) makes it more sensitive to rolling shear failure and deformation when subjected to out-of-plane loading. With the aid of Figure 6, one can visualise how the ‘straws’ of the grain ‘roll’ over one another, hence the term rolling shear.

SANS 10163-1:2003 states in Clause 13.12 that the rolling shear characteristic stress of SA Pine should be taken as 25% of that parallel to the grain. These values are given in Table 5.

Mechanical testing of CLT panels made with SA Pine was conducted as part of an MSc thesis by Jacobs (2023) at Stellenbosch University. The results from these tests provide evidence that the rolling shear values specified by SANS 10163:1 are possibly over-conservative, with the minimum rolling shear strength recorded from all samples being between 0.8-1.69 MPa (the difference in minimums depends on what point along the load-deflection curve one considers a failure). However, in-lieu of more detailed testing and results that have been verified by the professional and academic bodies, it is recommended that engineers continue to use the values as specified in the code. It is worth noting that unless the span of a slab is very short, or there are high point loads, it is unlikely that shear strength will govern the design of a CLT panel.  In most cases bending will govern.

Figure 8 shows typical rolling shear failures in two 3-ply CLT panels. Note how in both cases the failure occurs at the end of the panel (where shear stresses are highest) and how the rolling shear failure then led to a failure along the glue line.

Shear stresses are always highest in the centre of a cross-section, unlike bending where maximum stresses are at the top and bottom surface layers. It logically follows that shear failure will thus occur in the central lamella. For 3-ply and 7-ply CLT, the central layer is orientated in the minor strength direction, and is thus susceptible to rolling shear. Conversely, a 5-ply CLT panel has the central lamella in the major strength direction, which means it is not susceptible to rolling shear. One approach would thus be to use the shear strength parallel to grain in resistance checks for a 5-ply panel. However, in reality such a failure is unlikely; it is more probable that rolling shear failure will occur in one of the lamellas adjacent to the central lamella, even though the shear stress is less. Thus, it is recommended that the rolling shear strength is used regardless of the number of lamellas.

2.6 Quiz 1 & 2

3. Structural calculations for CLT

In this section the reader is introduced to some of the structural theory of CLT. The discussion is brief, however, and does not go into the detail required by the reader to fully equip themselves to perform design calculations. The Canadian CLT Handbook is recommended as a comprehensive, in-depth resource with clear explanations as well as design examples. The Swedish CLT Handbook is another useful aid. Both handbooks were recently published in 2019, and are freely available on the internet.

3.1 Analytical methods for determining CLT properties

There is currently no unified method for determining the strength and stiffness of CLT panels. Rather, various methods have been adopted across the world over the past two decades since CLT was first manufactured in the early 2000s. Empirical methods have been developed, but these are based on physical testing of full-size panels which is impractical and costly considering the multitude of ways in which one CLT panel can differ from another. Every time the layup, material, or any manufacturing parameter changes, new tests would have to be conducted. Therefore, analytical methods that are verified by physical testing present a more feasible approach. Once verified, the method can be applied to any CLT panel and layup, provided it falls within the limitations of that specific method.

The most commonly used analytical methods that are used in the industry and academia are:

1. Shear Analogy Method
2. Mechanical Jointed Beams Theory (also named the “Gamma Method”)
3. Timoshenko Beam Theory
4. Composite Theory (also named the “k-method”)

It is beyond the scope of this course to provide a detailed explanation of each of these methods. Currently the two most popular methods seem to be the shear analogy method and the gamma method. Ultimately, the engineer must decide for themselves which method to use to perform their calculations. That said, it is recommended that the shear analogy method is used, which is discussed in the next section.

3.1.1 Note on ‘Solid Block’ approximation

CLT is sometimes approximated as a solid piece of timber with the grain of all boards running in the same direction. This is undoubtedly the simplest method to calculate the structural properties, but an engineer should be cautious of using this approach. Whilst it can indeed yield relatively accurate results for a 3-ply CLT panel (only in the major strength direction), the same cannot be said for any other type of panel. With an increasing number of lamellas the behaviour of CLT increasingly deviates from that of a solid block. Furthermore, the approximation does not take shear deformation into account. This method is not recommended, especially considering the relative ease with which the shear analogy can be applied, provided a spreadsheet or other software is available to the engineer. Whilst this takes a certain amount of effort up-front if created by the engineer themself, once completed it can be used to quickly determine the structural properties of any CLT panel to a high degree of certainty.

3.2 Shear Analogy Method

The method is relatively simple and intuitive to apply, robust, and can easily be done with the aid of a spreadsheet (Figure 9). More importantly, the research by Jacobs (2022) has provided evidence backed by physical testing that suggests that the shear analogy method provides the most accurate predictions of structural properties for CLT made with SA Pine.

Another reason why the shear analogy method is recommended is that it takes into account shear deformation. This kind of deformation is present in any kind of beam/slab of any material, but is typically considered negligible by engineers because of the high shear stiffness of most materials. However, for CLT the low rolling shear stiffness (+- G_major / 10) of internal lamellas results in a proportionally much higher amount of shear deformation. Figure 10 gives the equation an engineer would use to determine the deflection of a simply-supported beam/slab under a uniformly distributed load, with separate terms for bending and shear deflection. The effective bending stiffness (EI)_eff and effective shear stiffness (GA)_eff are obtained through the calculations of the shear analogy method.

If shear deformation is neglected an engineer will under-estimate the deflection of a CLT panel under load. The under-estimation can be substantial, and in the worst case could potentially lead to a failure. The proportional influence of shear deformation decreases as the span/depth ratio increases. In other words, thin slabs that span far are not as affected by shear deformation as thick slabs with short spans. This is illustrated with an example in Table 6, which shows how a 99mm CLT panel experiences proportionally less shear deformation as the span increases.

3.3 Long-term deflection

Over longer periods of time the deflection of CLT can increase due to load duration and creep, similarly to standard timber design. Structural timber codes, including SANS 10163-1, provide methods for this. However, engineers must recognise that CLT is more prone to creep than solid and glue-laminated timber because of the orthogonal layers. Thus, the Canadian CLT Handbook recommends that a creep factor of 2 be used for long-term deflection calculations.

High moisture content can also increase creep. However, it is worth noting that CLT is not an outdoor product and should always be used in a dry service condition.

3.4 Vibrations

CLT slabs are susceptible to human-induced vibrations (primarily from foot-fall traffic), which can cause discomfort to inhabitants. The vibration response is a function of the mass, stiffness and span of the slab. The Canadian and Swedish CLT handbooks have differing approaches. The former uses a simplified method that gives a ‘vibration-controlled span’ i.e. the maximum length a panel can span before it fails. The latter is based on the Eurocode and has a more detailed approach with three checks: minimum natural frequency, maximum deflection under 1 kN point load, and impulse velocity response.

It is worth noting that similarly to deflections, there is no hard-and-fast limit or rule here, as the discomfort caused by vibrations is inherently subjective and depends on the user.

3.5 Design tables

Appendix A includes design tables giving the unfactored strength and stiffness values for standard layup CLT made with S5 SA Pine calculated using the shear analogy method. Deflection- and vibration-controlled span tables are also given for three types of beams and different live loadings. These tables can assist engineers and other designers with preliminary design and sizing of panels.

3.6 Quiz 3

4. Treatment 

The treatment of CLT in South Africa is a somewhat controversial topic. Whilst there are regulatory requirements, they were created in a time before CLT, and it is difficult/impossible to apply them. There is currently no consensus on how best to treat CLT in South Africa and it can best be described as a ‘grey area’. The status quo is discussed below.

The relevant standards (SANS 10400-A and SANS 10005) explicitly state the following:

• Structural timber of softwood species (e.g. pine) in the coastal municipalities of SA must be treated (refer to SANS 10005 for a full list of municipalities)
• Structural timber of hardwood species (e.g. eucalyptus) in all areas of SA must be treated.
• The treatment preservative and process must be approved by the standards. NB: surface-applied methods and preservatives are not included and do not comply with the standards.
• The treatment must be to the relevant hazard class. For CLT this would typically be H2 (inside above ground), as CLT is not for external use.

The regulations are easy and simple to apply for regular timber structures; the engineer specifies which hazard class is required and the appropriate timber is ordered. However, unfortunately things are not as simple for CLT – there are a number of factors that make the pressure-treatment of CLT impractical:

• CLT panels are large and do not fit inside the standard autoclaves/pressure chambers that are used for pressure treatment of timber, making the treatment of a complete CLT panel impossible.
• The next solution would be to use pressure-treated timber planks for the production of CLT. However, during the machining process a significant amount (if not all) of the treated timber will be removed, depending on the depth of penetration of the preservative. This renders the method both wasteful and ineffective. Furthermore, the wood waste that is produced (sawdust, shavings) is chemically treated, and thus cannot be use for many secondary purposes such as mulch.
• Considering the above, one might think that a solution would be to first machine the planks and then treat them. However, this is not possible because the very tight dimensional tolerances of CLT require the slabs to be pressed within a very short space of time after machining (12-24 hours) because of dimensional instability. This leaves no time for treatment. Furthermore, the treatment itself would almost certainly affect the dimensions of the timber, especially if it is water-based (e.g. CCA).
• Thus, one option that might be considered is a three-part process: 1) pre-machining to remove most of the timber, 2) pressure-treatment and drying, and 3) final ‘light’ machining before pressing. Whilst this is technically feasible, practically it is a more costly process. Research by Alade (2022) at Stellenbosch University showed that Eucalyptis grandis laminates can be treated and used for CLT manufacturing with acceptable bonding quality. The practical implementation and commercial viability of this process, however, must still be established.

It is worth noting that fungal decay of timber is caused by high moisture content, which should be prevented through correct design and detailing. A well-designed and built CLT building that prevents water ingress and keeps the timber aerated should never experience any kind of rot. Insect attack remains a threat however, which means treatment is required. The only practical solution remaining is to surface-treat the outside of CLT panels. Whilst this does not technically meet the standards, it seems to currently be the best solution to ensure the longevity of CLT buildings in South Africa. There are a number of water- and solvent-based surface treatments for wood on the market. It is imperative that the treatments are applied correctly, and that all surfaces are covered, as any gaps will remain vulnerable. Any alterations on-site during or after the construction process (cuts, holes, excessive sanding) also need to be treated to ensure a complete envelope coverage.

5. Structural fire design of CLT

The fire behaviour of timber structures is a highly complex and multi-faceted topic. A full discussion on this is outside the scope of this course. It is however very important that engineers educate themselves on this topic to ensure they design safe buildings that meet regulations.  The Fire Safe Use of Wood in Buildings by Buchanan and Östman (2022) is an excellent handbook that provides detailed guidance on the design of timber buildings in fire with reference to a variety of international standards. The Timber in Fire course on The Wood App can also be completed.

A select few topics relevant to South Africa and CLT are touched on below. It is assumed that the reader is already familiar with the basic concepts of structural fire design.

5.1 Charring rate of SA Pine

SANS 10163-1 gives guidance for the fire design of timber members in Appendix M of the code. The appendix is very brief and simplistic, and does not take into account many of the phenomena that the Eurocode (EN 1995-1-2) does (zero strength layer, two-dimensional charring, different charring phases). When it comes to timber design, and especially fire, the Eurocode is far more advanced and aligns better with modern research.

However, it cannot be assumed that the Eurocode charring rates apply to SA Pine (spruce is the standard structural timber in Europe). Furthermore, whilst phenomena are indeed not taken into account in SANS (un-conservative), the material factor is lower which accounts for the greater uncertainty (conservative). In addition, less favourable material properties are used in SANS (conservative). The most significant differences between the Eurocode and SANS are shown in Table 7. Red indicates the more conservative option relative to the other, with green indicating un-conservative. This shows that neither code is wholly more conservative than the other.

Table 7: Timber fire design differences between SANS 10163-1 and EN 1995-1-2

Research at Stellenbosch University has suggested charring rates as high as 0.95mm/min for SA pine CLT (van der Westhuysen, Walls and de Koker, 2020) and GLT (results un-published at time of writing, obtained through correspondence with researcher). It is not entirely clear why the charring rates are so much higher, but is possibly due to the lower density of some SA pine compared to European timber. In the case of CLT, it could also be due to char fall-off (delamination). The number of test samples is very low however, and is deemed not sufficient to fully justify the use of these values. Nevertheless, the tests do question the validity of the 0.8mm/min charring rate which was published in SANS 10631 in 2003.

Ultimately the engineer must decide on a charring rate that they feel is safe and applicable to SA Pine. The choice of code is also up to the engineer, keeping in mind that the ‘mixing and matching’ of codes is not recommended.

5.2 CLT in fire

According to the Eurocode and Buchanan and Östman (2022), CLT will char at a consistent rate in the same manner as a solid piece of timber provided the following conditions are met:

• The adhesive is rated to maintain glue-line integrity during a fire. If this is not the case, char fall-off will occur, which exposes pre-heated exposed timber with consequential greatly accelerated charring rates (+- double the basic charring rate) until a 25mm thick char layer forms.
• Gaps between adjacent planks in a lamella may not exceed 2mm. Larger gaps can lead to two-dimensional charring, with an associated higher charring rate.

If the conditions above are not met, the Engineer can use the methods presented in the Eurocode to take into account the accelerated charring rates.

A loadbearing 3-ply CLT panel is not well-suited to fire exposure, as the loss of only a single outer lamella reduces the panel to 2-ply, in which the original transverse layer adds no significant structural value. The panel then relies solely on the single lamella remaining in the longitudinal direction, and the resulting strength and stiffness is orders of magnitude less than the original 3-ply. Effectively this means that 3-ply CLT is generally unsuitable for loadbearing exposed panels that require a fire rating.

It is for this reason that 5- and 7-ply panels perform better in fire, because even after the complete loss of 2 lamellas, there remain 3- and 5-ply panels respectively. The outer lamella are thus effectively ‘sacrificial’ buffer layers which protect the rest of the panel for a quantifiable amount of time. The engineer must then check whether the residual panel can resist the loads during a fire event. This would fall under the ACC (accidental) limit state.

In the case of non-loadbearing partition walls, it should be noted that 3-ply CLT can be suitable. This is because only the integrity and insulation criteria need to be met, which are not as sensitive to the orientation of lamella.

5.3 Connections in fire

In timber structures the connections are often the weakest part of the building. This is especially true for fire design, because in modern timber structures the majority of connections are made of steel (screws, plates, etc), which can be problematic in a fire. Steel conducts heat well, and as it heats up it rapidly loses strength and stiffness. Furthermore, where steel protrudes into the timber (e.g. fin plates, dowels, bolts, screws), the hot steel leads to localised charring of timber. This further weakens the connection. The temperature at which charring occurs is approximately 300 ⁰C (Buchanan and Östman, 2022). For these reasons engineers need to pay special attention to load-bearing timber connections.

There are three general ways to achieve a fire-resistant timber connection (Buchanan and Östman, 2022):

1. For a 30-minute fire rating or less, partial concealment of steel with timber is possible.
2. For ratings of 30-minutes or more, complete concealment of steel with timber is required.
3. Full encapsulation of connection and surrounding area within a non-combustible board system (e.g. gypsum).

For CLT junctions self-tapping steel screws are often used to connect panels (Figure 10). Provided these are embedded deep enough in the timber to always remain behind the charring line, they can provide a full-strength fire-resistant connection. Where the head is exposed to the fire, research shows that the charring only extends 20-30mm down the shaft of the screw (Buchanan and Östman, 2022). The heightened temperature of the screw does however reduce the capacity. The Eurocode provides methods to determine the structural capacity of fire-exposed screws.

Figure 11: Typical CLT wall-to-floor connection using fully-threaded screws

Buchanan and Östman (2022) provide more guidance on the design of timber connections in fire. Figure 12 and Figure 13 are examples of such connections.

Figure 12: Steel bearing plate connector that is fully concealed with timber: (a) beam being lowered onto the connecter with timber base block pre-installed; (b) completed on-site. (Buchanan and Östman, 2022)

Figure 13: Proprietary flexible intumescent gasket (Fire Stripe Graphite) that is used to seal concealed steel plate connections (image courtesy of Rothoblaas)

5.4 Quiz 4 & 5

6. References

Alade AA, Naghizadeh Z, Wessels CB, Stolze H, Militz H. 2022. Improved adhesive-bond performance in copper azole and disodium octaborate tetrahydrate-treated Eucalyptus grandis laminates. International Wood Products Journal 13(3):139-147.

Börgstrom, E and Fröbel, J. (2019). The CLT Handbook. Swedish Wood, Stockholm.

Buchanan, A. and Östman, B. (2022). Fire Safe Use of Wood in Buildings. CRC Press, Boca Raton.

Crafford, PL. and Wessels, CB. (2016). ‘The potential of young, green finger-jointed Eucalyptus Grandis lumber for roof truss manufacturing’. Southern Forests. 78(1). pp. 61-71.

EN 1995-1-2 (2004) Eurocode 5 Design of Timber Structures. Part 1-2 General: Structural Fire Design. European Standard. CEN European for Standardization, Brussels.

Jacobs, MJ. (2023). Out-of-plane strength and stiffness prediction of SA pine cross-laminated timber. Masters Thesis. Department of Forest and Wood Science, University of Stellenbosch.

Karacababeyli, E. and Gagnon, S. (2019). Canadian CLT Handbook. FPInnovations, Point-Claire.

SANS 10163-1 (2003) The Structural use of timber Part 1: Limit-states design. Standards South Africa, Pretoria.

Van der Westhuysen, S., Walls, R. and de Koker, N. (2020). ‘Fire tests of South African cross-laminated timber wall panels: fire ratings, charring rates, and delamination’. Journal of the South African Institution of Civil Engineering. 62(1). pp. 33-41.

Wessels, CB. (2020). Wood Products Science 234 – Timber as a structural material course notes. University of Stellenbosch

Appendix A – CLT Design Tables

STIFFNESS AND UNFACTORED RESISTANCE VALUES FOR SOUTH AFRICAN PINE CROSS-LAMINATED TIMBER

CLT SLAB SPAN TABLES