SUMMARY
Recent initiatives on British Columbia Ministry of Transportation and Highways bridge projects have demonstrated that better value can be obtained through design innovation. As a result of the Ministry including Value Engineering clauses into the bridge construction contracts, the contractor was able to offer cost-saving alternative designs which were tailored to the specific constraints and requirements of each project. Five Value Engineered bridge designs were offered in 1996 and 1997. These resulted in a net savings of over $1 million. The owner obtained better value through the integration of design with construction and the incentive of the Value Engineering clauses. All five designs involved the creative use of precast concrete bridge girders.

INTRODUCTION
A major component of the B.C. Ministry of Transportation and Highways (MoTH) new construction program in recent years has been the Vancouver Island Highway. The 300 km long $1.3 billion road has been designed to four-lane arterial and freeway standards. Some 150 bridges were needed for grade-separation and watercourse crossings.

In recent years, the bridge construction contracts have included Value Engineering clauses. These clauses encourage the contractor to propose alternative designs. Where these offer the Ministry a cost or schedule advantage, with no reduction in quality, they will be considered. The incentive offered to the contractor is 50% of the net savings in addition to any material or construction technique preference.

Experience in applying the clauses suggests that savings can accrue through material competition, higher design efficiency, elimination of unnecessary features, and improved constructibility. The bridge projects described illustrate how existing structural systems can be developed and original designs revised to save costs. This was done without sacrificing durability, while maintaining the quality and elegance prevalent in successful bridges.

VALUE ENGINEERING
Recent MoTH bridge construction contracts have included Value Engineering clauses, along with an encouragement to submit any number of Value Engineering Proposals (VEPs) during the contract term. The Ministry will not accept VEPs that could negatively impact existing approvals or bridge aesthetics. The contractor cannot assume tender acceptance, or prepare a claim based on a VEP.

To be considered, a VEP must offer a tangible benefit to the owner. The substitution must be functionally equal or superior, while reducing risk, initial or life-cycle costs, or a schedule advantage. In preparing VEPs, the original design criteria must be used. A change in contract conditions does not qualify as a VEP.

VEPs may involve changes in materials, products work method, construction sequence and especially design. They are evaluated usually on the basis of net cost saving. This is calculated by adding to the negotiated cost of the VEP:

• The contractor’s cost of developing the VEP.
• The owner’s cost of evaluating the VEP.
• Any resulting administrative or management costs to the owner.

The driving force for submissions is the Value Engineering Incentive. Generally, this results in 50% of the net saving being earned by the contractor. There are, however, disincentives, which can be summarized as follows:

• The contractor’s cost in developing a VEP for which he has no certainty of recovery.
• The time taken to prepare a complete design for a VEP submission by the contractor, his suppliers and his consultant.
• The time taken by the owner to evaluate, comment on, and negotiate a VEP, over which the contractor has no control.
• The period of delay during which the contractor can only execute work unaffected by the VEP, and cannot procure materials for either the original or alternative designs.
• The uncertainty of obtaining acceptance of work carried out on the VEP prior to the supplemental contract agreement being implemented.

The following bridges involved either partial or complete redesigns. All resulted from VEPs, which were accepted following minor adjustments requested by the Ministry. All involved the substitution of superstructure systems along with either adjustments to the original substructure or a new design.

BLOEDEL CREEK BRIDGE
This bridge is a single 45.1 m span, south of Courtenay, B.C. Twin 11.4 m wide roadways on separate superstructures cross 12 m above the creek at 15 degrees of skew. The control line curvature is on 1800 m radius, which results in a deck superelevation of 3.4% (Figure 1).

The Bloedel Creek bridge was originally designed using three steel plate girders per superstructure, 1.9 m deep, at 4.2 m centres. The deck was a minimum of 275 mm thick. Most of the zone between abutments was not accessible for construction purposes because of environmental restrictions.

The contractor wanted to reduce costs by using precast concrete girders. This could only be economical if the original three-girder superstructure system was used. This would, however, result in higher loads than concrete girders typically support over this span, and would stretch the limits of precast girder design.

The selected girder is a standard 2.3 m deep I-section. At 46 m long, the precast girders weigh about 76,000 kg, well above road transportation limits. The precast concrete manufacturer’s plan was to transport the girders in one piece by barge from their plant to a loading dock near the site, and then by off-highway forestry road to the site. Erection would be carried out using two 40 m long launching trusses side by side, supported by falsework in front of the abutments.

In order to use the concrete three-girder system effectively, the superstructure needed reconfiguring. The 1 m wide girder top flanges permitted a 250 m constant thickness concrete deck to be used, in conjunction with a girder spacing of 4.5 m. To provide sufficient flexural capacity, a hybrid prestressing system is used. Each girder incorporates the following:

• 44 straight 13 mm diameter pretensioned strands.
• 12 deflected 13 mm diameter pretensioned strands.
• 36 - 15 mm diameter post-tensioned strands.

The post-tensioning consists of three draped tendons, each with twelve strands. In order to maintain adequate cover in the 150 mm thick girder web, oval-section ducts are used. The post-tensioning is applied in the field in two stages. Two tendons are tensioned prior to construction of the concrete deck, and one after the deck concrete had attained at least 20 MPa compressive strength. All tendon anchorages are located in the deck end-diaphragms which are widened for this purpose. The cast-in-place diaphragms also serve as integral ballast walls. This avoids the need for precast girder end-blocks, which simplifies fabrication and, more importantly, reduces the weight of the already very heavy girders. This is important in order to maximize the span potential of the launch trusses. Having the end-diaphragms structurally effective prior to deck construction helps to stabilize the girders. Two intermediate diaphragms are provided. To speed construction and save weight, structural steel sections are used, bolted to the concrete girders.

The multi-stage prestressing system provides sufficient girder capacity for transportation and handling. The strength is supplemented by the first stage of field post-tensioning in order to provide sufficient resistance for deck concrete placement, the largest load component. Additional post-tensioning is then added to provide sufficient capacity to resist imposed loading.

By using the hybrid prestressing system, the following is achieved which would not be possible using a conventional pretensioned girder:

• More total prestressing applied to each girder.
• Greater strand eccentricity near mid-span.
• Greater strand deflection, vertical prestress component, and shear capacity near the girder ends.
• Better control of stresses throughout the loading stages.

This approach, coupled with a high specified strength of 70 MPa at 56 days, permitted the three precast concrete girder concept to be used.

The additional 400 mm of structural depth was not an issue as the bridge opening is controlled by the highway profile. The abutments were adjusted to accommodate the revised superstructure details. The original footings were located on bridge end fill or bedrock and were adequate to support the additional superstructure weight.

The Bloedel Creek Bridge is scheduled for completion in May 1998.

WILFRED CREEK BRIDGE
The Wilfred Creek Bridge is one of the few multi-span, fully integral abutment bridges in B.C. The twin 65 m long superstructures are continuous over spans of 17 m, 29 m, and 19 m, and cross 15 m above the creek at a skew of 15 degrees. The deck widths are 11.9 m northbound and 12.9 m southbound. The control line is on a 1000 m radius which results in a 4.8% deck superelevation (Figure 2, Figure 6). The bridge is supported on driven steel pipe piles and concrete piers. The site is immediately south of Courtenay, B.C.

Originally designed with four 1.2 m deep continuous steel plate girders per superstructure and a 225 mm thick deck, the bridge was redesigned using pretensioned concrete I-girders. These are 1.7 m deep and are spaced at 3.5 m or 3.8 m centres, similar to the steel girders. To reduce the quantity of deck reinforcement, the selected deck thickness is 250 mm. As the bridge opening is governed by the highway profile, the additional superstructure depth was acceptable.

The contractor was installing the piles while investigating the Value Engineering Proposal. As a result, only superstructure changes could be contemplated, and no opportunity was available to investigate a different number of girders. With a maximum weight of 32,000 kg, the girders are readily transportable. As the girders are inherently stiff and only lightly stressed, no intermediate diaphragms are necessary. The centre span units were positioned using a steel launching truss resting on the intermediate support bents.

In order to provide the maximum structural efficiency and constructibility, the simply-supported girders were placed on the abutments and piers, while projecting reinforcement was embedded into the concrete support-diaphragms. No bearings or joints are used. After placing the support-diaphragm concrete, the girders are effectively made continuous for subsequent imposed loads. This novel approach provides a more efficient structure to resist the deck self-weight, while also stabilizing the girders. The tensile stresses near the girder ends are controlled by deflecting a high proportion of the pretensioning. Twelve of the twenty eight, 13 mm diameter strands in the centre spans are deflected. Projecting overlapping rebars resist the support moments. The concept of stiff built-in concrete girders allows the deck concrete to be placed in one continuous operation.

The girders are shorter versions of the Cowie Creek girders, and therefore, experience comparatively low stresses. A 28 day concrete strength of 40 MPa was specified. To accommodate the deeper superstructure, the abutment wall depth was increased and the intermediate pier columns were shortened. In supporting the heavier superstructure, the pile loads increased only slightly, and the pier columns were found to have adequate seismic load resistance without modification. The bridge is scheduled for completion in June 1998.

COWIE CREEK BRIDGE
This bridge is a three-span 92 m long bridge south of Courtenay, B.C. (Figure 3). Twin 11.4 m wide roadways on separate superstructures cross 20 m above the creek at a skew of 6 degrees. The main span is 37.5 m, with 27.5 m end spans. The control line is on tangent. The deck crossfall varies, however, as superelevation run-offs encroach onto both ends of the deck.

The Cowie Creek Bridge was originally designed using four 1.2 m deep continuous steel plate girders per superstructure, at 3.25 m centres. A 225 mm thick cast-in-place concrete deck is used. The zone beneath the main span is not accessible for construction purposes.

The contractor wished to replace the steel girders with prestressed concrete girders, by adapting the system used for Wilfred Creek Bridge. It was determined that a similar system could be employed for the two bridges. The 1.7 m deep prestressed girders, weighing up to 42,000 kg, are just within transportation weight and length restrictions. The centre span units could be positioned using a launching truss resting on the pier caps.

It proved to be just possible to design the precast girders for the end-moments generated by the weight of the concrete deck. Ten overlapping 30M hooked bars project from the top flanges into the concrete pier diaphragms. The precast concrete girders experience significantly higher stresses than at Wilfred Creek. Twenty six straight strands and twenty deflected strands of 13 mm diameter are used in the centre span girders, which require a 28 day strength of 56 MPa.

A single intermediate diaphragm of steel angle bracing is used at the midpoint of the centre span to assist in stabilizing the girders during construction. As the bridge opening is controlled by the highway profile, the 0.5 m additional depth of the concrete superstructure was readily accommodated by lowering the pier caps and abutment seats. The inherent stiffness of the precast girders permitted the deck concrete to be placed in one continuous operation.

Integral ballast-retaining walls are at each end of the 92 m long superstructure. No bearings are used at the monolithic pier connections, but laminated rubber bearings support the girders at the abutments.

Reanalysis of the substructure was carried out for seismic loading with the heavier concrete superstructure. No modification of the intermediate-support columns or footings was required. An increase in lateral sliding resistance of the bank-seat abutments is achieved by placing two 0.9 m deep shear keys beneath the footings. The bridge construction is scheduled for completion in June 1998.

TRENT RIVER BRIDGE
The three-span, 106 m long Trent River Bridge is located 750 m north of the Bloedel Creek Bridge (Figure 4). Twin 11.4 m wide roadways on separate superstructures cross 24 m above the creek at a skew of about 22 degrees. The 1800 m radius in plan requires a deck superelevation of 3.4%.

This bridge was originally designed using three 1.6 m deep continuous, steel-plate girders per superstructure, at varying centres of approximately 4.2 m. The deck had a minimum thickness of 275 mm. No construction access was available in the area beneath the main spans.

The contractor wished to replace the steel girders with precast concrete. A rapid assessment indicated that to achieve the maximum cost saving, the original three-girder system would have to be retained. Additionally, by increasing the girder spacing to a constant 4.4 m, a deck thickness of 250 mm could be used, reducing the superstructure weight.

The 40 m main span girders could be placed by using two launching trusses supported on the pier caps. Standard 2 m deep precast I-girders could be employed, but would need to be adapted to the high levels of loading that they would experience. Their maximum length of 39 m and weight of 53,500 kg precluded highway transportation. However, by using barges and industrial roads, the girders could be routed to site.

The challenge was to design the girders. The system used for Cowie Creek Bridge was investigated. Unfortunately, insufficient flexural capacity was available both at mid-span and at the intermediate supports. The solution was to supplement the pretensioned segments by applying field post-tensioning from end to end. In this fashion, a higher total level of prestressing could be applied, thereby increasing girder load capacity without increasing the number of girders.

Each girder has sixteen straight and twelve deflected 13 mm diameter pretensioning strands. These strands provide strength for transportation and handling while reducing the required quantity of post-tensioning. Four ducts per girders are provided, each containing seven 15 mm diameter strands. The ducts are oval-section, 55 mm by 75 mm, in order to maintain adequate cover in the 140 mm thick girder web. The ducts are field-spliced prior to casting the concrete diaphragms and subsequently post-tensioned. The tendon anchorages are located in the cast-in-place ballast retaining walls. All tendons are tensioned prior to deck construction, which is carried out in one continuous concreting operation.

The specified girder concrete strength is 70 MPa at 56 days. As well as using the maximum flexural capacity of the girders at mid-span, near the intermediate supports maximum available shear resistance is used. This is a good indication that the design has achieved maximum efficiency and effectively stretched the limits of precast concrete bridge girder technology.

To enhance stability during construction a single intermediate diaphragm of galvanized steel WT bracing is provided at the centre of the main span. The additional depth of the concrete girders was acceptable and did not require any revision to the highway profile. No bearings are used at the monolithic pier connections but laminated rubber bearings support the girders at the bank-seat abutments. Although the intermediate piers did not require strengthening to resist seismic loads from the heavier superstructure, shear keys were added beneath the footings of the bank-seat abutments.

The substructure of the Trent River Bridge was also included in the Value Engineering Proposal. An evaluation demonstrated that the spacing of the pier columns could be reduced to better balance the loading on the caps and footings. More savings were realized by reducing the size of the pier footings. The largest cost saving, however, was achieved by eliminating the forty eight, vertical 32 mm diameter threadbar ground anchors originally detailed. This was done by lowering the footings by up to 1 m, and keying them into the site bedrock. The Trent River Bridge is scheduled for completion in July 1998.

THE DUKE POINT UNDERPASS
This bridge is part of the grade-separated interchange between the Trans-Canada Highway and the spur road to the Duke Point Ferry Terminal near Nanaimo, B.C. The bridge carries three traffic lanes on a 16.7 m wide deck over the four-lane divided highway. The bridge is continuous over its 21 m, 38 m and 21 m spans. The plan alignment is part radius, part spiral, which results in a minimum of 5.6% deck superelevation.

The bridge was originally designed as two 1.5 m deep, twin-cell, cast-in-place, post-tensioned, concrete, trapezoidal-box girders supported by flared concrete piers. The contractor wanted to build a precast concrete girder bridge. Because of aesthetic considerations and depth restriction, I-section girders were not acceptable. As a result, a 1.4 m deep precast concrete trapezoidal open-top box girder was used for the Value Engineering Proposal.

By using three of the box girders at 6 m centres, a useful cost saving could be achieved. Additional costs could be saved by supporting the girders on circular discrete-column piers, while eliminating the steel pipe piles originally used to support the east pier and abutment. Instead, footing subexcavation and structural fill replacement was employed.

The superstructure design uses five segments per girder connected by cast-in-place concrete splices and field post-tensioning. The 21 m long centre span segments are lightly pretensioned, while the remainder are provided with reinforcement to control handling stresses. The16 m long pier segments include local bottom flange thickening near the piers and an integral internal diaphragm. The girders are supported by pot bearings on top of the columns and laminated rubber bearing at the abutments. No intermediate or pier diaphragms are used. The end-diaphragms retain the approach fills and accommodate the post-tensioning anchorages.

Each girder is provided with six tendons of eleven 15 mm diameter strands. One additional strand per tendon could have been added if needed in the field, although this proved to be unnecessary. Only ten strands per tendon were used in the design. The eleventh strand is included to avoid the provision for future post-tensioning specified by the Ministry. Inside the box girders, 50 mm thick precast stay-in-place forms were used. The deck was cast in one operation and then post-tensioned longitudinally.

The bridge was completed in January 1997 and opened to traffic a few months later. Despite the time taken to redesign the bridge and obtain approval for the VEP, no delay in the construction schedule occurred.

COST SAVINGS
The following are net Value Engineering savings. Under the contract Value Engineering clauses, the VEP savings are shared equally between the Ministry and the contractor. The gross design savings, i.e., prior to deduction of the cost of assessment, cost of VEP preparation, and engineering redesign, is about 30% higher.

DISCUSSION
The VEP savings on these bridges range from 6% to 20% of the tender prices. In all cases, precast concrete girders were used to replace steel girders, except for the Duke Point Underpass where cast-in-place concrete was originally used. It should be recalled, however, that for the three bridges with the lowest cost savings, only the girders changed. When this happens, an engineering evaluation of the complete structure and some re-detailing is necessary as the superstructure depth and weight is frequently different.

The largest savings occurred with the redesigns for the Duke Point Underpass and the Trent River Bridge. For these structures, the complete bridge was reengineered. Savings in the substructure were significant and were similar to superstructure savings.

Table 1 - VEP Cost Savings

Bridge	Tender Cost ($)	VEP Savings ($)
Bloedel Creek	1,400,000	97,000
Wilfred Creek	2,000,000	100,000
Cowie Creek	2,300,000	230,000
Trent River	2,800,000	400,000
Duke Point	1,500,000	300,000
Total	10,000,000	1,127,000

If the VEP savings are viewed as the gross difference in cost of the materials changed, the savings are highly significant. That this can be achieved is perhaps surprising. The bridges, it should be realized, have already been subject to careful evaluation and design optimization by the original designer. On top of this, the Ministry subjects all consultant designs to rigorous cost-control scrutiny. Savings of over 10%, however, are appreciated as this enables more bridges to be built for a given budget.

Why, therefore, are any savings possible? The answer, we believe, is because there is no infallible universal system for designing the most cost-effective bridge. The incentive of 50% of the VEP savings will drive an aggressive contractor to find a better way. In this case, it involved stretching the recognized limits of precast concrete girder technology. The contractor, however, takes a significant risk in deferring construction while he prepares a VEP, during which time affected work is either delayed or performed speculatively.

By allowing the contractor to take ownership of the design via Value Engineering, the cost and schedule advantages of design/build are available. The resulting stimulus to be creative and develop a solution geared to the contractor’s available resources often permits significant savings to be found. By opening up the bridge design as well as construction to competition, better value can be delivered by industry to the owner. Always desirable, it is especially important for publicly-funded projects.

ACKNOWLEDGMENTS
The author wishes to thank the B.C. Ministry of Transportation ad Highways, Bridge Branch, Director Peter Brett, P.Eng., for permission to publish this paper. The following are gratefully acknowledged for their contributions to the project:

Owner’s Representative: Vancouver Island Highway Project Management Team.

Contractor: Wilfred Creek Bridge, Cowie Creek Bridge: Carlson Construction Ltd.

Bloedel Creek Bridge, Trent River Bridge, Duke Point Underpass: Emil Anderson Construction Ltd.

Precast Concrete Manufacturer and Erection: Con-Force Structures Ltd.