Design-build procurement of the Hampton Roads Bridge-Tunnel Expansion
The $3.86 billion Hampton Roads Bridge-Tunnel Expansion (HRBT) project in southeastern Virginia developed rapidly from concept selection in December 2016 to contract award in April 2019. During this time, the Virginia Department of Transportation (VDOT) evaluated whether to structure the procurement as a privately financed P3 or publicly financed designbuild; engaged regional stakeholders and the tunneling community to choose between immersed-tube and boredtunnel construction alternatives; identified and retired cost-driving risks during procurement; and continuously refined the project scope. This ultimately entailed a 2.4-km (1.5-mile) four-lane tunnel crossing, 4 km (2.5 miles) of marine bridges and 8 km (5 miles) of highway widening.
Hampton Roads is the second-largest metropolitan area in Virginia, with a regional population of around 1.7 million people across 15 localities. The region takes its name from the maritime roadstead at the confluence of the Elizabeth, Nansemond and James Rivers as they enter the Chesapeake Bay and ultimately the Atlantic Ocean.
This maritime influence is significant for Hampton Roads’ extensive tunneling history, since the region is home to numerous military installations including the Naval Station Norfolk, the largest naval base in the world. Due to U.S. Navy requirements, no bridges are permitted above the major navigable waterways leading to the Atlantic Ocean, resulting in the large number of marine tunnels in this area.
Due to the region’s soft soils, though, only immersedtube tunneling was feasible in this geology for many years. Relative to the widespread use of bored tunnels, the immersed-tube construction method is comparatively newer and less frequently employed. The ITA Catalogue of Immersed Transportation Tunnels (Rasmussen and Grantz 1997) records the first such tunnel as having been built in Detroit, MI in 1910. As of 1950, there were only 10 examples in existence worldwide, seven of which were in the United States.
Even so, Hampton Roads was an early adopter of immersed-tube tunneling due to the constraints noted above and nine subaqueous tunnels comprising 10 two-lane tubes have been constructed there since 1952 (Table 1). By 1957, when the Hampton Roads Bridge-Tunnel (HRBT) opened, it was only the 15th immersed transportation tunnel in the world.
Unlike the 1952 Downtown Tunnel, which crossed the Elizabeth River directly shore-to-shore, the 1957 Hampton Roads tunnel could not span the entire 5.6 km (3.5-mile) mouth of the Virginia harbor. Tunnel-ventilation technology at the time was not sufficient to support this length for a highway tunnel without intermediate shafts. Since the Navy’s bridge-building restrictions applied only to the navigation channel, engineers developed the idea to construct artificial islands bordering the channel, immerse the tunnel between these islands and then connect this roadway to shore via lowlevel trestle bridges over shallower, nonnavigable water.
Given the width of the navigation channel, though, the necessary length of the resulting tunnel was still extraordinarily long for its day. When it opened in 1957, the HRBT was the longest subaqueous tunnel in the world, with an immersed length of 2,091 m (6,859 ft) and a total portalto- portal length of 2,280 m (7,479 ft), including the cut-andcover segments at each end of the tunnel.
The HRBT’s 1957 opening also marked the first time worldwide that a marine tunnel had been constructed between artificial islands (Fig. 1). This concept was replicated successfully throughout coastal Virginia, with the 1964 Chesapeake Bay Tunnels, the 1976 Hampton Roads Tunnel (Fig. 2) and a 1992 Monitor-Merrimac Tunnel also incorporating manmade islands connected to land via trestle bridges. Southeastern Virginia remains one of the world’s densest concentrations of immersed-tube tunnels, particularly those featuring artificial islands.
The new Hampton Roads Bridge-Tunnel
Contract scoping. The twin Hampton Roads tunnels had become heavily congested by the early 2000s, with traffic on the four-lane crossing averaging nearly 90,000 vehicles per day and exceeding 100,000 daily vehicles during peak summer travel. When Virginia’s Commonwealth Transportation Board recommended an HRBT expansion in December 2016, VDOT moved forward with shaping the contract scope and procurement approach for this project.
The project’s concept design included not only a 2.4-km (1.5-mile) four-lane tunnel crossing, island expansions and 4 km (2.5 miles) of marine bridges, but also 8 km (5 miles) of conventional landside highway widening. To confirm market interest in this scope, VDOT conducted an industry sounding in April 2017 with a formal request for information and pre-procurement one-on-one meetings with interested contractors and developers.
With preliminary studies having estimated the cost of this scope at more than $3 billion, feedback from the industry sounding indicated that certain bidders would be more likely to pursue the work if it were segmented into smaller contracts. VDOT considered this input but ultimately determined to procure this scope as a single contract, due to the complexities of coordinating interfaces between multiple contracts while safely maintaining high volumes of live interstate traffic.
Given the size and complexity of this scope, coupled with the agency’s prior experience in alternative delivery methods, VDOT determined from the outset not to use design-bid-build procurement. As such, the agency advanced the concept design only to the level necessary for identifying preliminary right-of-way impacts, conducting the required design public hearings and developing cost estimates to help ensure the scope remained within budget limits.
Procurement method. The April 2017 industry sounding also provided valuable input on the question of whether to structure the procurement as a privately financed P3 or a publicly financed design-build project. Both were conceptually feasible, with three teams having submitted unsolicited public-private partnership (P3) proposals for an HRBT expansion in 2010 and 2011, and with the parallel Midtown Tunnel having been procured as a P3 in 2011.
Since then, though, Hampton Roads had developed its own transportation funding stream via a regional tax dedicated to major congestion-relief projects. The Hampton Roads Transportation Accountability Commission, which was founded in 2014 to administer revenues from an increased 0.7 percent sales tax and 2.1 percent fuel tax, had sufficient financial capacity to fund 95 percent of HRBT project costs. In addition, state policy determined that only the new HRBT capacity could be tolled, with the existing four-lane capacity remaining free of charge for travelers.
Based on these considerations, industry participants felt the business case for a publicly financed design-build approach was stronger than for a privately financed P3. VDOT incorporated this feedback and issued a design-build request for qualifications in December 2017, shortlisted three teams in April 2018 and issued a request for proposals (RFP) in May 2018.
Risk management during procurement
Although VDOT encouraged proposers to develop innovative solutions within the design-build context, the agency recognized widely differing interpretations of certain performance-based specifications could introduce unintended risks — either for the proposer, the state or the public — into the project. In these areas, VDOT set prescriptive specifications instead of the performance-based specifications that are traditionally used in design-build contracts.
The goal was to help ensure proposers competed on their merits, rather than on differences in risk appetite. Although a higher-risk proposal could potentially offer a lower bid price, VDOT preferred not to incentivize this strategy nor to encourage a design that reduced construction costs at the expense of greater operations and maintenance effort in the future. To address this, VDOT also structured the technical portion of the best-value selection criteria, which assigned 60 percent of the evaluation score for price and 40 percent for technical merit, to reward lower-risk approaches.
In cases where a prescriptive specification interfered with an innovative approach that a proposer felt had particular merit, the proposer could submit the idea as an alternative technical concept (ATC) and explain its risks and benefits in confidential one-on-one meetings. The ATC process allowed VDOT to evaluate the risk profile of these concepts and make a determination whether to allow, or conditionally allow, each concept.
Immersed tube or bored tunnel? One example of the tension between prescriptive and performancebased specifications involved the selection of the tunnelconstruction method for the new HRBT. Although all existing tunnels in coastal Virginia are immersed tubes (e.g., Fig. 3), the Parallel Thimble Shoal Tunnel procurement had received three TBM tunneling proposals in early 2016, indicating soft-ground TBM technology had advanced sufficiently to enable bored tunneling in this region.
Although VDOT had initially assumed the new HRBT would be an immersed-tube tunnel, the agency decided — based on this new information — to give proposers the option to select the method they believed was best suited to project conditions and proposers’ individual strengths. Accordingly, VDOT’s concept design was silent regarding tunneling approach, and the request for proposal contained performance-based specifications that accommodated both immersed-tube and bored tunneling methods. Proposers were instructed to evaluate the available information and declare their chosen method in July 2018.
To aid in this decision, VDOT provided extensive geotechnical data, including 1953 and 1969 boring logs from the existing tunnels, and the spring 2018 ground investigations at the HRBT islands. Proposers also received input from regional maritime stakeholders, who noted the 8,800 vessel movements per year across the HRBT and expressed concern regarding the potential mutual risks to shipping and to immersed-tube tunneling operations sharing an active navigation channel.
Based on this information, one of the three shortlisted teams, which had submitted qualifications only for immersed-tube tunneling, declined to continue with the procurement. Both remaining proposers declared their intention to advance with a bored-tunneling approach, and VDOT subsequently amended the RFP to prescribe this construction method.
Soft soils at South Island. The soft soils at HRBT’s South Island presented another example in which VDOT recognized that differing interpretations of performancebased specifications could introduce unintended risks. Despite the proximity of the North and South Islands, geotechnical conditions vary greatly between them, as noted by Kuesel et al. (1973): “The North Island is founded on sands and silty sands and presents no substantial settlement or stability problems. At the South Island, however, about 80 feet of normally consolidated clay overlies sandy soils.”
The challenging soil conditions at the South Island had been known since the 1820s, when the U.S. Army’s attempt to construct heavy masonry fortifications nearby was hindered by unexpected cracking and settlement. From geotechnical borings, engineers identified thick layers of soft clay, silt and organic materials (Fig. 4), which were excavated by dredging prior to construction of the 1957 tunnel.
For the 1976 tunnel, though, a similar dredging program would have destabilized the 1957 tunnel unless the new tubes were built prohibitively far away. Instead, the design prescribed sand drains and 8 m (26 ft) of surcharge above the island’s finished elevation. Fifteen months after the surcharge reached its full height, Kuesel et al. (1973) reported a maximum settlement of 4 m (13 ft) under the highest fill near the north end of the island, “where the clays are thickest and more plastic than average.” The South Island footprint ultimately measured 0.1 sq km (24 acres), significantly larger than the North Island, to accommodate this temporary surcharge mound.
Given the ample unoccupied space on the South Island, the proposers for the HRBT Expansion found it attractive to locate the TBM launch shaft there. Engineering judgment differed, though, on appropriate solutions for maintaining the tunnel’s vertical alignment along the 5 percent grade as the TBM’s heavy cutterhead — approximately 14 m (46 ft) in diameter — passed through these very soft layers. Because ground improvement along this portion of the alignment would be a significant cost item, VDOT recognized the proposers could be tempted to gamble on the risk in this location. Hence, the agency mandated a conservative program of ground improvement in the RFP technical requirements: “Wherever any part of the tunnel below the springline would lie within the soft clays and organics (layers Qf and Qo as defined and baselined in the GBR), the Design-Builder shall provide ground improvement to these in-situ soils in advance of tunneling.
“This ground improvement shall extend from the tunnel springline to a minimum depth of one half of the tunnel outside diameter below the tunnel invert, or 2 feet below the base of the soft clay and organics layers, whichever is less. In addition, the ground improvement shall be performed over the width of the TBM plus a minimum of 5 feet on either side.”
Although this approach was conservative, it helped lessen the bidding pressure for proposers to develop a risky engineering solution in an attempt to gain a pricing advantage. VDOT advised the teams that the selected proposer would be welcome to submit a value-engineering proposal after award, informed by additional geotechnical investigations along its specific alignment, to reduce the contract’s prescriptive ground-improvement requirements.
Tunnel diameter. The question of TBM diameter illustrated another example where VDOT intervened to address unintended risks arising from different interpretations of the specifications. Initially, the RFP requirements for tunnel dimensions were performance-based, allowing proposers to set their own tunnel diameter as long as it accommodated given vehicle-clearance envelopes plus requirements for ventilation, egress, utilities, finishes and other spaceproofing considerations.
From the one-on-one meetings and written questions submitted during procurement, it became clear the RFP’s original provisions were incentivizing proposers to reduce their bid prices by minimizing the tunnel diameter to the smallest possible dimensions. This, however, was not VDOT’s intent: a smaller diameter would subtly, but inevitably, complicate operations and maintenance (O&M) tasks throughout the structure’s 100-year design life. Even though the individual requirements were accurate per se, the sum of these performance-based specifications did not fully reflect the agency’s needs.
In response, VDOT amended the RFP to specify a minimum interior tunnel diameter of 12.6 m (41.5 ft), from concrete liner to concrete liner, in order to eliminate this dimension as a point of competition between the teams. This aligned with other prescriptive O&M guidance already in the RFP: although locating a jet fan directly above the roadway centerline could be efficient from a spaceproofing standpoint, for instance, this was not permitted because all maintenance had to be possible from within a single-lane closure, and servicing a fan directly above the centerline would require closing both lanes of tunnel traffic for safety reasons.
Conclusion
Following award of the HRBT Expansion contract in April 2019, the project team looks forward to continued progress in delivering the next crossing of the Hampton Roads. Throughout this effort, the inputs from past and current industry partners and stakeholders have been invaluable in beneficially shaping the project’s procurement structure, design scope, construction method, technical requirements and risk profile. All of these influences have contributed to make this a better project and are gratefully acknowledged.
References
Kuesel, T.R., Schmidt, B, and Rafaeli, D. 1973. Settlements and strengthening of soft clay accelerated by sand drains. Highway Research Record. 457:18-26. Rasmussen, N., and Grantz, W. 1997. Catalogue of immersed tunnels. Tunneling and Underground Space Technology. 12(2):163-316.
