Two of the major topics of interest to those designing taller and larger wood buildings are the susceptibility to differential movement and the likelihood of mass timber components drying slowly after they are wetted during construction. The Wood Innovation and Design Centre in Prince George, British Columbia provides a unique opportunity for non-destructive testing and monitoring to measure the ‘As Built’ performance of a relatively tall mass timber building. Field measurements also provide performance data to support regulatory and market acceptance of wood-based systems in tall and large buildings.
This report first describes instrumentation to measure the vertical movement of selected glulam columns and cross-laminated timber (CLT) walls in this building. Three locations of glulam columns and one CLT wall of the core structure were selected for measuring vertical movement along with the environmental conditions (temperature and humidity) in the immediate vicinity. The report then describes instrumentation to measure the moisture changes in the wood roof structure. Six locations in the roof were selected and instrumented for measuring moisture changes in the wood as well as the local environmental conditions.
All sensors and instrumentations, with the exception of one, were installed and became operational in the middle of March 2014, after the roof sheathing was installed. The other instrumentation was installed in July 2014. This report presents performance of the building during its first year as measured from topping out of the structure. In the end, the one-year period covers six months of construction and six months of occupancy. This is the first year of a planned five-year monitoring.
The first year’s monitoring showed that the wood inside the building had reached moisture content (MC) of about 4-6% in the heating season, from an initial MC of 13% during construction. Glulam columns were extremely dimensionally stable given the changes in MC and loading conditions. With a height of over 5 m and 6 m, respectively, the two glulam columns measured in this study showed very small amounts of vertical movement, each below 2 mm. The cumulative shortening of the six glulam columns along the height of the building would be about 8 mm, not taking into account deformation at connection details or effects of reduced loads on upper floors. The CLT wall was found to be also dimensionally stable along the height of the building. The measurements showed that the entire CLT wall, from Floor 1 to Floor 6, would shorten about 14 mm. The CLT floors, however, had considerable shrinkage in the thickness direction, and therefore should be taken into consideration in the design and construction of components, such as curtain walls, which are connected to the floors. In terms of the roof performance, two locations, both with a wet concrete layer poured above the plywood sheathing, showed wetness during construction but dried slowly afterwards. The good drying performance must be attributed to the interior ventilation function designed for the roof assemblies by integrating strapping between the sheathing and the mass timber beams below. Overall this monitoring study shows the differential movement occurring among the glulam columns and the CLT wall is small and the wood roof has good drying performance.
Ajoutez cet article à votre liste de sélections pour demander le PDF - Add this item to your selection list to request the PDF
Based on the data from this study and a literature review, there are two distinct trajectories for hemlock wood moisture content, depending on if the tree was felled before or after May. Hemlock trees felled before May gain the full benefit of spring drying according to the ambient conditions of their local micro-climate. Trees felled after May suffer from a physiological spike in moisture content that the tree generates to promote its growth and survive the summer soil drought.
The problem of second-growth western hemlock (Tsuga heterophylla) sinking when watered continues to plague the coastal logging industry of British Columbia. A study conducted by FPInnovations in 2015-16 concluded that felled hemlock logs took two distinct trajectories in their drying patterns through time, depending on whether they were felled before or after May.
A study was initiated in late 2012 to assist in designing panelized wood roofs and other wood roof assemblies as part of the work supporting the British Columbia (BC) Advisory Group on Advanced Wood Design Solutions. The test aimed primarily to assess and compare the drying performance of different roof assemblies with built-in moisture under worst-case scenarios in the coastal climate of BC. A range of roof assemblies were built using plywood, OSB, cross-laminated timber (CLT), or laminated veneer lumber (LVL) as the roof deck. Each deck specimen was wetted using hourly spray of water for 15-20 days in the laboratory. Either self-adhered impermeable membrane, or standard roofing felt plus asphalt shingles, was used to cover the wetted top surface. Closed-cell spray foam insulation was applied on the bottom surface in several roof assemblies to assess its impact on the drying performance. Subsequent drying was carried out under three ambient conditions for three groups of specimens, beginning in February or March, 2013. Limited drying forces, except natural moisture diffusion and evaporation, were created, except for one roof assembly placed under controlled temperature gradients. The drying rates of each wood deck were monitored by weighing the roof assemblies periodically, followed by calculating the changes in the average moisture content (MC) of each wood deck. In total, 111 assemblies were tested to compare the drying rates between different wood decks and between different assemblies.
The major test results included:
The MCs of the plywood, OSB, CLT, and LVL specimens, after hourly spray in the laboratory for 15 to 20 days, were all highly comparable with those of the naturally exposed reference specimens after about two months of natural exposure from January to March. This validated the use of such laboratory wetting to simulate severe wetting scenarios on construction sites.
The four types of roof deck materials in this study showed very different water absorption and wetting potentials. The CLT specimens, as a mass timber product tested in this study, showed the lowest wetting potential, closely followed by LVL. The plywood and the OSB specimens tested had the highest MCs after wetting under both conditions.
In general the higher the material’s wetting potential, the faster it would dry under conditions allowing drying.
The CLT and LVL assemblies, covered with impermeable membrane on the wetted surfaces, showed little drying during the test periods. The use of two layers of plywood as a roof deck also showed much slower drying compared with the one-layer assemblies, when covered with an impermeable membrane on the wetted surfaces.
The combination of 15-pound roofing felt and asphalt shingles showed only slightly better vapour permeance, as indicated by the slightly better drying performance of roof assemblies, compared with the impermeable membrane used in this test.
All of the roof deck materials tested had very poor drying performance when enclosed in materials with low vapour permeance, such as impermeable membranes and closed-cell spray foam.
Among the three main drying conditions used in this study, the plywood and OSB assemblies in the shed showed the lowest drying, and those in the lab showed the fastest drying, with those in the conditioning chamber providing intermediate drying rates. Such differences appeared to correlate well with the vapour pressure differences between the saturated pressures and the partial pressures of the environments. The shed was intended to simulate covered on-site conditions in this climate.
This test indicated that, when roof panels get wet, drying during construction can be improved fairly easily for relatively permeable panels or assemblies, such as one layer of plywood or OSB, one layer of plywood or OSB covered with other materials only on one side. Favourable weather conditions and mechanical methods (e.g., ventilation and space heating) can accelerate drying. However, the ambient environment may not have much effect on the drying performance of relatively impermeable products or assemblies, including massive wood members and wood enclosed in impermeable materials. The test data also indicates that drying could take a long time, typically weeks and even months depending on environmental conditions, once wood panels get severely wet. This could allow mould growth and even decay to occur under severe circumstances. This study confirms the importance of keeping wood dry in the first place, and drying wood before enclosure when wetting occurs as the second strategy, in order to minimize moisture-related risks. Guidelines are needed to recommend best practices for on-site wood protection and moisture management in wet climates, for the use of innovative engineered wood products in larger building projects in particular.
The idea that a wicking treatment – leaving tops, branches, and needles attached to the stem after felling – would reduce stem moisture content and lead to reduced loss of western hemlock logs from sinking during watering was tested. This study did not show that wicking produces a different result from normal bucking in average stem moisture content if both groups are treated equally. Further, after curing for two months logs did not increase in moisture content during watering.