1
TIMBER VEHICULAR BRIDGES IN GHANA – THE PROGRESS TO DATE
K. A. Solomon-Ayeh Building and Road Research Institute (BRRI) Kumasi, Ghana [email protected]
Abstract
The infrastructure in Ghana is relatively low because of, among other factors, the lack of crossing structures like bridges. Existing bridges are predominantly of reinforced concrete and as the cost of the imported steel and cement became expensive, research, particularly at the Building and Road Research Institute (BRRI), started on the possibility of using Lesser Used Timber Species (LUS) for vehicular bridges, at least on low-volume feeder roads in Ghana. A prototype 7.6 metre-span, I-sectioned timber girder vehicular bridge was built by the BRRI over the Subin stream at Kaase, in Kumasi, Ghana, in 1991. This bridge is successfully performing its designed functions to date. The structural forms used then and the likely spans achievable, brought to the fore likely problems should spans beyond 8 metres be bridged and, together with the Timber Research and Development Association (TRADA), the possibility of the use of timber truss bridges was looked at. The form of the prototype I-girder bridge, the technical problems inherent in taking this structural form beyond 8 metres and the consequent truss bridge proposed, are presented in this paper. Problems that arose during the design of the truss bridge were those to d with the size f transverse beams, the design of couplers and the need to proof test the integrity f the proposed joints. Design solutions t some of the foregoing problems are given. However, lack of relevant equipment such as loading rams and strain/displacement transducers have meant that the research efforts at validating the design options adopted for the joints have not been possible. Future collaborative research and testing efforts are required if more bridges are to be provided as crossing structures on low-volume roads, if more efficiency in use is to be made of LUS and a likely industry in pre-fabricated timber bridge components engendered in Ghana. Keywords: Bridges, girder, timber, truss, couplers
1.0 INTRODUCTION
In 1991, following research carried out at the Building and Road Research Institute (BRRI)
on scaled timber I-girders (Allotey, 1990, 1992), a prototype timber bridge was built over
the Subin Stream at Kaase, in Kumasi, Ghana. Considerable problems arose with the
construction of the bridge and these problems and their solutions are briefly highlighted.
2
The bridge is functioning structurally well though some challenges have arisen with
regards to the rate of maintenance. However, the main aim of proving that Lesser Used
Timber Species (LUS) can be used for bridges has been achieved. Replicating this
particular bridge configuration in the design for crossings wider than 8 metres brought to
the fore short-comings, mainly to do with the rather deep sizes of I-girders.Timber truss
bridges were looked at in a bid to overcome the challenges with I-girder bridges. At the
design stages of this alternative bridge type, there are problems to do with the material
and size of transoms, couplers at truss node points, lateral bracing methods and the need
to proof test the integrity of joints. Where solutions were proposed, these are highlighted in
this paper. The successful solution to the identified problems should increase the choice of
forms of crossing structures available, make better structural use of LUS, introduce skills
in the construction of timber bridges, engender an industry in pre-fabricated timber bridge
components and improve commerce by the access provided by better transport
infrastructure.
2.0 HIGHLIGHTS OF I-GIRDER BRIDGE
The bridge was designed to AASHTO HS-20-44 and American Interstate Highway loads. It
is simply-supported over a span of 7.6m. The superstructure is made up of seven
mechanically built I-girders, 1.143m deep. The girders are braced by timber cross-braces.
The deck over the girders is made up of transverse 50x150mm timber members laid on
edge and toe-nailed to previous members to form a continuous plate. Asphaltic concrete
was applied over the deck and this forms the running surface of the bridge. The parapets
are also timber. The I-girder is made up of a web of two layers of diagonally crossed
planks (25x100mm) with the flanges bolted to the outside of the web members. The web-
flange system is fastened together by mechanical fasteners made up of screws and bolts
(from 20mm and 25mm mild steel reinforcement bars). The deck is fastened to the
girders by toe-nailing and by bolts connected intermittently through 150mm long, 75x75
steel angle plates.
3
The substructure is made up of 22 end-bearing timber piles which also act as abutment
and wing wall braces (columns). The abutments and wing walls are of timber planks and
the piles are held together at their tops by timber pile caps. The pile cap provides a seat
for the I-girder.
The I-sections were fabricated off site and given a preservative treatment by the use of
chromated copper arsenate (CCA) under pressure. Treatment of site cuts and additional
treatment of timber in direct contact with the soil was by a mixture (by volume) of 60%
creosote and 40% bitumen, heated till almost boiling.
The mechanical properties of the timber used in various sections of the bridge are given in
Table 1.0. A sketch of the arrangement of the girder at the abutment is shown in Fig. 1.0
and pictures of the existing bridge are shown in Figs. 2 to 7.
Table 1.0: Mechanical Properties of Timber Species for I-girder Bridge
Section
Timber Species Density Kg/m
3
Mechanical Properties (Nmm2)
Compressive Strength (Parallel)
Bending Stregnth
Modulus of Elasticity
Pile
Afina (Strombosia glaucescens)
900
24.5
25.0
18,500
Wing/Retaining Wall
Kusia (nauclea diderrichii)
760
17.0
17.6
14,500
Pile Cap
Dahoma (Piptadenistrum Africanum)
720
12.5
12.5
10,600
Girders
Kusia
76.0
17.0
17.6
14,500
Floor Deck
Dahoma
720
12.5
12.5
10,600
Parapet
Dahoma
720
12.5
12.5
10,600
At the time of construction, the cost of the timber bridge, at $72,000 compared favourably
with an assumed equivalent reinforced concrete crossing (4.0 x 4.0m, box culvert, 11.50m
long) which was $92,000 (Allotey, 1992).
4
5
FIG 2
FIG 3
FIG 5
FIG 4
6
FIG 6
FIG 7
7
3.0 PROBLEMS WITH I-GIRDER
Problems of the prototype I-girder were those to do with construction and the structural
configuration as regards its use with spans longer than 7.6m (Allotey, 1992).
3.1 Construction Problems
(i) Slanted Piles – The piling machine used was defective and had no feature for
maintaining the poles in line and in a plumb position. Some piles became slanted and
though the slants were not excessive they were likely to:
- introduce considerable moments at the base of the piles by the applied vertical
load at the pile head and the eccentricity of the slant;
- result in the pile cap not sitting centrally on the piles;
- make the backs of the piles to receive the retains/abutment out of line.
The foregoing problems were solved by a combination of use of braces, (using same
material as the pile) in-between piles to form a truss-like support, finding the line of best fit
between the two end piles of an abutment and arranging the centre of the width of the pile
cap to lie on this line and fastening timber packing in the gaps at the backs of piles till a
plane back is achieved to receive the abutment planks.
(ii) Size of I-girders – Because of the design loads used, AASHTO H20-44, the I-
girders were rather deep (1.143m) for the span considered (7.6m). Such deep girders
required the use of heavy hoisting equipment (cranes) and resulted in major (and costly)
earthworks for the approach roads to align with the bridge deck. Also, an increase in span
beyond 8m will call for even deeper I-girders or the use of midspan piers; a situation which
will be expensive on account of the construction of piers, the prolonged use of heavy lifting
equipment and the inconvenience of working with large-sized sections (DFID, 1999,
Solomon-Ayeh et al., 2002).
8
If timber bridges are to be used as crossings for spans larger than the foregoing then the
identified shortcomings of large girder depth, weight and size, number of superstructure
timber members and the very many fasteners must be overcome. The choice of a
structural configuration different from the I-girder should be looked at and truss bridges
and stress laminated decks come readily to mind. Stress-laminated decks offer
considerable equipment and technical difficulties if opted for in the Ghanaian setting since
it’s a novel technology and the funds outlay in the acquisition of the relevant equipment
can only be justified if more of such decks are envisaged.Timber truss bridges were
looked at in collaboration with TRADA Technology, U.K., because they have potential
advantages such as:
- fewer superstructure materials;
- flexibility in spans and permissible loads by the use of modular construction using
pre-fabricated components;
- use of fewer fasteners;
- use of a variety of medium-sized hoisting equipment;
- fabrication technology and equipment needs within Ghana.
4.0 TIMBER TRUSS BRIDGE
4.1 Highlights
The initial truss bridge proposed was a Warren through truss of 6m triangular modules,
fastened at their nodes by steel couplers. Timber (Afina) round poles, of 300mm were
proposed for the members of the truss. The transverse transoms were to be either steel
533x210x92 universal beams (UB) or 1500x250 (300kg/m) mechlam assembly. Stringers
were to be 254x146x46 UB or 525x250 (105kg/m) mechlam. The running deck was to be
of 50x100 dahoma (D50) toe-nailed lumber and 50x250 dahoma running decks (DFID,
2001).
9
The design depended on U-frame action for bracing and the same 300mm Afina poles
were utilized as members of a horizontal truss bracing at the top of the through bridge.
The configuration was arrived at from designs to nominally span 20 metres and to full HA
loading and 20 HB (abnormal) loading.
Mainly due to objection to a through truss bridge and the insistence of higher design loads
by Ghana Highway authorities (Section 4.2), the configuration of a through truss was
changed to a modified Warren half-through truss of 3.0m triangular modules to span
distances ranging from 12m to 25.5 metres (in 3-metre modules) (Solomon-Ayeh, 2002,
2004). Other features of the modified bridge were:
- transverse bracing using 300mm Afina poles diagonally placed between the top
chord of the truss and the horizontal extensions from the transverse transoms;
- flitch beam (steel/timber) transverse transoms;
- 150x150mm stringers.
Round timber (Afina) poles were preferred for the truss members (of both truss designs)
because these poles:
- are available in the required lengths;
- are inherently stronger than processed timber;
- have a relatively low taper;
- come in manageable girths (approx. 300mm).
Some design and potential fabrication problems were identified and these are presented.
10
4.2 Design Challenges and Their Solution
4.2.1 Design loads
The bridges were meant for low-volume roads. Such bridges were predominantly in rural
areas which happen to be the source of cash and food crops (cocoa, maize, roots and
tubers) and timber logs. These items are heavy. Bridges built on any road should be such
as to carry such heavy loads and all others. Also, enforcement of loading regulations on
roads in Ghana is not regular and virtually non-existent on rural roads. Thus, the advice
of the bridge authorities was for a full HA loading (normal) and a 45 units HB abnormal
loading (by British Standards). Such design loads will produce very deep and heavy timber
members, and deep trusses of multiple chord members. For the transverse transoms, the
depth of the timber members can only be minimized by the use of steel universal beams
(UB). Thus the simplicity, use of LUS, economic cost, compatibility and aesthetic appeal
inherent in the use of timber would have been lost.
Where design loads for bridges are concerned, the ‘magnitude’ of the values depend not
only on the envisaged weight of material to be carried, but more importantly, on the
frequency of their passage on the crossing structures. A heavy load which is infrequently
passed on a bridge will produce little distress to the bridge. Also, a single-lane bridge
(typical for low-volume roads) puts less load on bridges. TRADA/BRRI took this view and
hence designed the original bridge to 20 units HB loading, and to be of single-lane to
reduce member sizes. Even with this, steel UB’s were specified for the transverse
transoms.
The modified truss bridges, were in some instances, to be sited in more populated areas
and thus the abnormal load (HB load) was put at 30 units, though the single-lane
11
carriageway was maintained. Also, the transoms were spaced at 1.5m centres to reduce
their sizes (though their numbers were increased).
4.2.2 Dimensions of Chord Members and Transoms
For design to spans of between 20 and 27 metres and for 30 units of HB loading, the
required capacity of the top and bottom chords suggested 4No. 300mm Afina poles each
(Solomon-Ayeh, 2002, 2004).
The transoms will require a capacity to satisfy both I (moment of inertia) of 275.1x106mm4
and Z (modulus of elasticity of 34.6x106mm3. If round Afina poles are to be used, four (4)
poles will be required. The use of mechlam timber (300x75mm basic unit) will require a
beam of 300x750mm size. Both solutions will result in unwieldy members, will be heavy
and in the former case, be particularly difficult to couple into the truss structure. A
composite (fliched) beam was resorted to for the transom, made up of two confining 12mm
thick x 225mm wide steel plates and 4No., 225x75mm timber (D50) pieces. The overall
dimensions of the flitched beam was 225x375mm (Table 2.0).
4.2.3 Connection at Nodes
Connection at the nodes of the truss in the original TRADA design were of three types,
depending on the method of load transfer (DFID, 2001).
(i) Transfer of tension forces from pole to nodes
High tensile forces (40 tonnes) are transferred between the timber member through steel
couplers. A single plate, set in a kerf cut along the centerline of the pole, is secured with
one row of steel dowels (bolts) of 38 to 50mm
(ii) Transfer of tension forces to nodes (from pole)
Poles will terminate in axial steel connecting plates and the node will comprise of steel-to-
steel connection and bolts will be used. This configuration is simple and minimizes or
avoids the tendency of the movement in the truss due to ‘loose’ joints.
12
13
14
FIG 10
15
Transfer of compression forces from pole to nodes
The end-bearing capacity of the timber pole (proved the use of Afina as piles) is exploited
by designing a single plate which will bear on the end of the pole and is welded to a
second plate set in a kerf cut along the centerline of the pole and is secured by bolting.
The modified truss bridge design (Solomon-Ayeh, 2002, 2004) uses the attachment to the
pole through a kerf cut set in the centerline of the pole, as above, in case of spans of
below 12 metres. For spans above this, and particularly for more than 24m, where four (4)
poles are to be used in both chords, a torso-shaped coupler will be used where two poles
will be attached to each ‘arm’ of the ‘torso’ ; one on each side of the ‘arm’. Two bolts (32 to
50mm size) will secure the poles through the steel ‘arm’. The base of the coupler will be
attached (by bolting) to the confining bracket of the flitched beam transom. The
arrangements are shown in Figs. and 9.
The size and types of fastening envisaged in both truss types call for proof testing of the
nodal joints to validate the design options chosen. The tests are to be carried out at the
BRRI when the necessary equipment becomes available. It is hoped that at the end of this
activity, vital skills in proof testing of joints in timber and other materials will have been
established at the BRRI as part of any collaborative effort.
4.2.4 Vehicle Containment and Through Truss Bridges
The initial TRADA design (DFID, 2001) was for a through truss bridge. For such a bridge,
the vertical and diagonal members are vulnerable to vehicle impact and given the
generally undisciplined nature of driving and lax enforcement procedures in Ghana, this
option was found unacceptable by the bridge authorities. A half-through truss was
therefore proposed (Section 4.2.3). Whist this option is possible, there is a concern for the
torsional rigidity of this truss and the large number of connections required. The former
concern is partly negated by the proposed use of diagonal pole braces at alternate
transom positions (i.e. at 3.0m centres).
16
The risk of vehicle impact on truss members is to be minimized by ensuring adequate
vehicle containment. The containment must safely direct or re-direct if necessary, the path
of the vehicle travelling across the bridge. Timber handrail (with diagonal bracing) will be
used for vehicle containment (Fig. 10).
5.0 ERECTION AND TENTATIVE COSTS
The proposed half-through bridge, being modular and with individual components of
manageable sizes and weight can be assembled quite easily on site on trestles or erected
on prepared ground behind one abutment. Launching can be by crane or tractors
commonly used to haul timber logs in the bush. For launching, the deck, which accounts
for about a third of the total self-weight of the bridge can be temporarily omitted.
The estimated cost of a typical25.5m bridge in 2004, was $57,000, made up of:
$30,000 - Cost of materials, off and on-site assembly
6,000 - Excavation, materials and construction of abutment.
5,000 - Launching of bridge.
7,000 - Transportation, tools and equipment
9,000 - Supervision (off and on-site)
$57,000
The estimated cost of the proposed through truss bridge was $46,000 in 2001.
6.0 CONCLUSION
Structural designs using realistic design loads have shown that it is possible to use Afina
round poles as main members of truss bridges in Ghana, to span crossings above 8
metres. The designed spans are recommended to be to a maximum of 24 metres.
Distances above 24m will need to be spanned by the introduction of piers in the crossings
and the provision of simply-support truss bridges in modulus of 12m, 18m, 21m, 24m, etc.
17
Configuration of nodal couplers have been given to support the large loads and the many
members typical of half-through trusses and the high abnormal (HB) loading used. The
integrity of these couplers have to be validated by proof test loads, preferably at the BRRI,
if the requisite equipment is made available. The proof test loads will equip local personnel
with the valuable knowledge in proof testing of joints of timber and other materials
Overall, the success to be achieved when these bridges are constructed will include the
engendering of an industry and market in the fabrication of modular bridge components
and making of the transportation of people, goods and services safer and less expensive.
6.0 REFERENCES
ALLOTEY, I.A. (1990). Laboratory tests on mechanically built-up timber girder for bridge
construction. Materials and Structures, Vol.23, No. 137; pp. 372-380.
ALLOTEY, I.A. (1992). Design and construction of a new timber Bridge in Ghana:
Problems and Solutions. ASAE Technical Paper No. 924558. International Winter Meeting
of the American Society of Agricultural Engineers, Nashville.
DFID PROJECT CNTR 986147 (1999). Timber bridges a review. TRADA Technology,
High Wycombe, 40pp.
DFID PROJECT CNTR 986147 (2001). Bridge specification and design options. TRADA
Technology, High Wycombe, 12pp.
SOLOMON-AYEH, K.A., ALLOTEY, I.A., FOLLETT, P., JAYANETTI, L. (2002). Advances
made in transportation structures by the design and construction of timber vehicular
bridges in Ghana. Proceedings of IABSE Symposium, Melbourne, pp. 442.
SOLOMON-AYEH, K.A. (2002). Proposed design of timber bridges on feeder roads in 5
regions for the Department of Feeder Roads. Unpublished design report for the
Department of Feeder Roads, Ghana.
18
SOLOMON-AYEH, K.A. (2004). Final (Draft) bridge design and specifications for the
design of three timber vehicular bridges at the Mole National Park. Ministry of Lands and
Forestry (Game and Wildlife Division) (Unpublished).