Timber Building in Australia

Concrete Enhanced Timber

Peter J. Yttrup
University of Tasmania
Australia.

The potential of timber-concrete composite construction is well known and proven. The methods of achieving composite behaviour have been significantly influenced by steel-concrete detailing with mechanical connectors such as nails, spikes, bolts, dowels and similar being used. Glued blocks and cut notches have also been used. The investigation reported in this paper explores the achievement of composite action by embedment of the beam prepared with small scale "dimples" into the timber, completely eliminating steel connectors. This achieved enhanced structural performance, and hence the term "Concrete Enhanced Timber" CET.

The research into the behaviour of timber bridges in Tasmania, Australia, reported by Yttrup and Nolan (1996), has demonstrated the superior strength and durability of timber bridges protected by a concrete deck or overlay. The exclusion of water from the timber and the connections leads to significant improvement in performance and a major reduction in maintenance cost, for only a minor increase in construction costs. Constructing a "roof" over the bridge in the form of a concrete deck improves performance, so why not also exploit composite behaviour to enhance structural strength and stiffness.

The comprehensive testing programme reported by Baldcock and McCullough (1941) demonstrated that composite timber and concrete bridge construction was technically and economically feasible. This investigation used steel dowels, steel rings and notch type shear connectors. Many of the bridges built using this technology are still in service, Classic Wood Structures (1989).

The use of timber-concrete composite floor systems in buildings was reported by Murthy (1989), where timber beams were embedded into the concrete floor and horizontal steel dowels run through the beams to provide shear connection. Ahmadi and Saka (1993) reported similar experimentation with timber-concrete composite floor systems using vertical nails and screws for shear connectors.

The research at the University of Tasmania is focussed on achieving full composite action between timber and concrete with simple details that do not require the use of steel connectors. This is considered to be appropriate for economic and buildability reasons, as well as structural efficiency.

The degree of composite action that can be achieved depends on the effectiveness of the shear transfer between the timber beam and concrete deck. Small amounts of slip between the beam and deck cause a disproportionate loss in composite action.

Mechanical connectors can be used to promote composite action, as can embedded beam details.

Composite action induced by mechanical connectors probably mimics steel-concrete composite construction. The obvious physical connector is probably comforting to some structural engineers but is not particularly efficient. The slip needed to mobilise their resistance can compromise composite action.

To demonstrate this, consider the concrete decked beam shown in Figure 1, where dowels provide the shear connection. With full composite behaviour the beam only deflects about 0.2 times that of the timber beam by itself. The effect of connector shear stiffness is shown in Figure 2, as can be seen, fully effective shear connectors will need to be very stiff. The "relative deflection" in Figure 2 is the deflection of the beam with composite action divided by the deflection of the timber beam acting alone.

The stiffness of 20mm diameter steel dowels, snug fitting in seasoned jarrah and spaced at 200mm centres, provides a stiffness of 200 to 300 kN/mm/m. These connectors provide only about 50% of full composite action in the above example.

Incomplete composite action leads to larger deflections, non-recoverable deflections, and greater internal action effects from applied loads. Mechanical shear connectors, such as steel dowels, may also be prone to fatigue, as experienced with shear studs in steel-concrete composite bridges.

Figure 1 - Example, timber - concrete composite beam

Figure 2 - Effect of connector stiffness on composite behaviour.

To avoid mechanical fasteners the University of Tasmania research is investigating "embedded beam" concepts. The timber beam is set into the concrete deck, or slab, to generate frictional, adhesive and mechanical interlock shear connection. To enhance the shear transfer directly at the timber concrete interface, roughness in the form of shallow drill holes is used. These "dimples" are typically 20mm in diameter and a maximum of 10mm deep. The short dimples ensure a very rigid load transfer, and being fully confined are very strong. Dimple spacing of 100mm was used in these trials. Indications are that much closer spacing could be used.

The shear stress at the timber-concrete interface is usually not as high as the shear stress in the timber beam immediately below the deck. This is because the neutral axis is generally close to the underside of the concrete deck. The length of the embedded timber surface resisting shear flow is greater than the width of the beam. The small scale and close spacing of the dimples is to approximate a surface transfer rather than by discrete connectors. This is in keeping with good timber engineering practice where many small fasteners are preferable to a few large ones.

Three timber-concrete composite specimens manufactured with pairs of beams, as shown in Figure 3, have been load tested. Two have radiata pine LVL beams and the third jarrah glue laminated beams. All the timber beams are embedded 30mm into the concrete deck, one LVL has dimples the other does not. The jarrah beam is provided with dimples.

The timber beams were load tested prior to the concrete placement to measure the timber beam stiffness. The completed beams with concrete deck were then tested with deflection and strain measurement to failure. The LVL specimens had a span of 5.4 metres and the jarrah 5.7 metres, all were loaded with a central point load.

Figure 3 - Embedded timber-concrete beam specimens.

Figure 4 - Load deflection curves.

Figure 5 - Strain profile measured on LVL beam.

Figure 6 - Strain profile measured on jarrah beam

The load deflection curves for the three specimens are shown in Figure 4. The enhanced stiffness is about three to four times.

The strain profiles measured on the LVL beam are presented in Figure 5, and in Figure 6 for the jarrah beam.

The LVL beam without dimples failed by slip, as did the LVL beam with dimples but at a 2.5 times higher load. The jarrah beam failed abruptly from flexural tension failure in the outer laminates.

The strain profile for the LVL and jarrah beams in Figures 5 and 6 both show that the neutral axis is located just above the bottom of the deck, and not moving significantly during the test. The theoretical location of the neutral action, assuming fully composite action, is about the top of the embedded timber beam. This agrees very closely to the strain profiles measured during the tests.

The LVL beam without any active shear connection still achieved full composite behaviour to a load of about 100kN due only to "friction". When slip did occur the post slip load deflection curve was almost parallel to the timber beam stiffness alone, suggesting a friction type connection did exist.

The influence of the shear dimples on the LVL beam was to prevent slip to a load of about 250kN. The dimpled LVL behaviour post failure was ductile due to the dimples causing fibre buckling type crushing failures in the LVL. Such ductility may be advantageous in some structures.

The jarrah beam failed at a tensile stress of about 35MPa, which is very similar to the strength of individual beams tested to failure. The jarrah is much harder than the pine LVL, providing very effective composite action with no apparent failures of the dimples.

The embedded beam details tested suggest that full composite timber concrete beam action is achievable without the use of any steel connectors. The small scale dimples used as shear connectors are able to provide effective shear transfer.

When full composite action is present simple beam theory is valid. The strain profile, deflections and neutral axis depth calculated by simple beam theory agreed very closely to that observed.

CET is suitable for low cost bridges, floor systems and similar structural applications.

Ahmadi B.H. and Saka, M.P. (1993). "Behaviour of Composite Concrete Timber Floors", Journal of Structural Engineers, Vol. 119, No. 10, November 1993, pp 3111-3130.

A.S.S.E (1989). "Classic Wood Structures". Task committee on classic wood structures, Committee of Wood Structural Division, American Society of Structural Engineers, New York.

Baldcock R.H. and McCullough C.B., (1941). "Loading tests on a new composite type short span highway bridge combining concrete and timber in flexure", Oregon State Highway Department, Technical Bulletin No. 1, Salem, 1941.

Murthy C.K. (1984). "Timber concrete composites for low cost housing", Housing Science, Vol. 8, No. 2, 1984, pp 209-215.

Yttrup P.J. and Nolan G. (1996). "Performance of timber beam bridges in Tasmania, Australia". International Wood Engineering Conference, New Orleans, Louisiana, USA, October 1996