The U.S. Environmental Protection Agency’s (EPA) Water Infrastructure Resiliency and Finance Center, in collaboration with the ...
When opened to traffic on Dec. 2, 1961, the Duluth-Superior High Bridge was recognized as an overdue transportation modernization. The Twin Ports of Duluth, Minn., and Superior, Wis., are situated on opposite shores of the St. Louis River at the western end of Lake Superior, and travel between them had been a considerable inconvenience for over a hundred years. Now, motorists can make the crossing without driving on timber bridge planks or experiencing delays caused by passage of the thousands of vessels using the harbor.
Early in the bridge's life it became apparent that the curb height median was insufficient to prevent vehicles from crossing over onto the opposing side. About the same time the median was replaced with a traffic barrier, the bridge was dedicated to its chief promoter, and officially named the John A. Blatnik Memorial Bridge. The new barrier effectively separated opposing traffic lanes but failed to block the annoying nighttime glare from oncoming headlights. Elliptical blinds bolted to the top of the barrier were difficult to maintain.
Then there was the traffic-safety dilemma caused by the absence of shoulders the bridge. For emergency service providers responding to accidents, the lack of shoulders seriously plagued rescue efforts. Similarly, maintenance and repair operations caused major traffic delays because of the lack of space to establish convenient work-zone traffic bypasses.
Added to these problems the bridge was showing the effects of one of the nation's most brutal winter environments. Temperatures often dip to 30 deg below zero, and application of deicing chemicals is necessary to keep the bridge safe for winter travel. So, in the mid-1980s transportation officials from Minnesota and Wisconsin began developing plans for the reconstruction of the aging and deficient Blatnik Bridge. The states take turns being the lead agency on their shared interstate bridges. This time the lead agency status went to Minnesota.
Many issues needed to be addressed in developing the scope and details of the rebuilding. On the engineering side, the most obvious goal was to add safety shoulders. This would require an additional outboard girder with several support options to evaluate. Unfortunately the shoulders couldn't be added through the truss, but 81% of the bridge could be safer. Also of a safety concern, the 52 aged pin and hanger-joint assemblies accommodating expansion/contraction forces in the plate-girder floor-beam portion of the bridge could be replaced. Finger joint expansion devices had allowed corrosive deicing chemicals to attack structural components beneath the joints.
The deteriorated members could be repaired or replaced and waterproof modular box-seal expansion devices installed. An extensive condition survey would be required. Shear connectors could be added to those deck supports not so originally equipped. Oddly, only the nine spans of rolled beams at the Wisconsin end had welded shear connectors and now these spans were slated for replacement with heavier rolled beams. Three spans of multiple plate girders on the Minnesota end of the bridge could be replaced with heavier members. An acceleration lane could be incorporated for Duluth motorists entering Ip;535 southbound, and a deceleration widening added on the opposite lane.
A new structural slab with epoxy-coated reinforcement, a 2 in. thick low-slump concrete wearing course, new electrical systems and navigational lighting would complete the update, but be largely invisible to the motorists. They would be aware of the barrier railings, median-mounted high-pressure sodium illumination, a new paint job, and the addition of a decorative lighting system to the arch, plus the disappearance of the increasingly frequent potholes in the southbound deck.
The Minneapolis office of Parsons Brinckerhoff Quade and Douglas was retained as the project's designer. Because of great concern about the project's economic and social impacts on the Duluth and Superior communities, the designer was asked to give top priority to developing various scenarios for accomplishing this massive reconstruction that would be taken to the citizens and businesses for review, discussion and input.
The first public information meetings to discuss the reconstruction were held during the fall of 1990. It drew over of 400 citizens, business leaders, local and state elected officials, and the news media from throughout the region. At the meeting, various construction staging alternatives were put on the table for public comment.
Initially, the preferred alternative pointed toward closing the bridge for two construction seasons and reopening it during the winter months. The fundamental precept of this option was that it would offer the highest quality workmanship in the shortest time, at the lowest cost and in the safest manner. Although most supported these four principles, the required two-year time frame for the project met substantial community resistance.
Recognizing the vital importance of listening to the communities, Minnesota DOT's (MnDOT's) Duluth office formed a special public/private task forces to refine the consultant's initial recommendations. Composed of members of the two transportation departments, local business and community representatives from Duluth and Superior and staff from the consultant, the Blatnik Bridge Task Force was charged with the following objectives: evaluate cost, sequencing and minimum time necessary for the major construction operations; recommend an orderly construction staging schedule that will minimize bridge closure; and recommend an incentive-disincentive package that should be included in the construction contracts.
The task force's report to top management of the two transportation departments wisely recommended dividing this massive undertaking into three phases. The first phase would include the substructure extensions and additions as well as fabrication and partial erection of the structural steel that would have little or no effect on traffic. By specifying that this work would take place between mid-May 1992 and early September 1993, the entire bridge could be used unimpeded for two tourist seasons.
Next, the bridge would enter the second phase, which would necessitate its complete closure for deck replacement and arch repainting. But, it wouldn't close until after the 1993 Labor Day weekend, which signals the end of the busy summer tourist season, reopening to traffic in November 1994.
Closure for only one full tourist season reduced the size of the bitter pill to where most, reluctantly, swallowed it. Enthusiasm was high for the probability of success, especially with the enhancement of this phase being bid under the A + B or Houston method. It was anticipated that additional refinement of the scheduling and aggressive prosecution of the work would result from employing this scheme. The daily cost to the public for maintaining the detour, driving the extra distance, special law enforcement, and accident investigation on the double-traffic detour now carrying an additional 25,000 vehicles per day was calculated at $21,000. The $21,000 would be multiplied by the calendar days bid in determining the low bidder in the A + B aspect of the contract. It would also be the figure in the incentive/disincentive package for restoring four-lane traffic to the bridge.
The final phase of the reconstruction, the painting of the approach spans, would follow in a third contract that would not affect traffic. Parsons' engineers began a detailed inspection and evaluation of the aging bridge, determining the minute details of the reconstruction then organizing them into the three contracts for which support was being gathered.
Contract A was let on April 3, 1992 to Johnson Bros. Corp. of Litchfield, Minn. Even while the wheels of contract approval machinery ground through the tedious details, another innovation was discussed and then implemented. The snowballing concept of partnering was decidedly a perfect fit for this high-stakes project, and was made part of Contract A by supplemental agreement.
Designated personnel from MnDOT's District 1 and Central Office; Johnson Bros. and their key suppliers; and the consultant's designers met for a three-day retreat under the leadership of FMI, Inc. The precepts of partnering, while not universally embraced and sometimes misused, were to be important in maintaining the owner-contractor relationship throughout the construction.
Work began on encasing 39 pier caps on each face with 18-in. to 20-in. thick structural concrete additions bearing the longitudinal post-tensioning ducts, and locked together by drilled inplace PT rods that had to slither between the tightly spaced original cap longitudinal reinforcement. Electronic location of the original bars proved to be less reliable than probing their location with an electric drill and masonry bit. Four of the pier caps were extended on one end to support the new fascia girder, 32 were extended on both ends. These massive extensions were held to the cap ends solely by the clamping force of the PT strands in the encasements, typically twelve 0.6-in. diam tendons on each side, locked off at 0.7 G.U.T.S.
Eight new hammerhead piers accommodating the additional girders for the acceleration and deceleration lanes were built on the Minnesota end. Two were deleted in their entirety and moved to the second contract. Interference with the bridge rail and deck overhang would be eliminated with second-phase deck removal, permitting driving of the HP 14 x 73 piles that were battered away from the bridge. The alternative was to drive five 115-ft-long piles in 5 ft lengths.
Concrete and post-tensioning were the priorities for the summer months of 1992 and 1993, with the winter between them being used for items like removing the old drainage system and crosswalks, plus jacking up the bridge and renovating 209 bearing assemblies.
Another critical part of the renovation was deferred to the second contract. The plans called for replacing 52 of the existing pin and hanger assemblies. The original assemblies were designed with the bushings in the girder webs and the pins solidly fitted to the link plates. The plates would have rusted tightly to the pins long ago, so the design did not promote inspection and replacement of worn components. These assemblies are now designated as fracture critical and would be replaced with elements made from the latest alloys, with the bushings in the hanger plates and pins tight in the girders.
Existing pin holes in the girders were to be rebored to a diameter 0.75 in. larger than original, with a tolerance of .0005 in. Performing this critical operation under traffic would be undesirable unless absolutely necessary, because of the effect on accuracy and the risk to the public. However, an opportunity existed for the installation to be quickly, easily, and safely done following phase-two deck removal. The materials were fabricated and stored on site, but the installation was moved into the second contract.
Contract A provided for the fabrication and erection of 4 million lb of structural steel, with an additional 2 million lb to be fabricated and delivered for use in Contract B. The new fascia beams and cross frames were to be erected, except for the three spans on the Minnesota end and the nine spans on the Wisconsin end where all girders were being replaced in Contract B. The steel was shop painted with an epoxy zinc-rich primer, an epoxy intermediate coat and a catalyzed urethane finish coat. As the steel was delivered and erected, the judgment slowly formed that the prepainting was a mistake on a fast track project of this size. The men simply got tired of being careful, but the majority of finish coat damage occurred during deck replacement activities.
Contract B was being prepared with a wary eye to the pitfalls anticipated with doing business differently. Hours were spent analyzing every sentence that defined the qualifying criteria upon which payment of the incentive would be based. It was foreseen that some minor operations could require temporary traffic control and to delay a reopening simply to accomplish them all was not consistent with the efforts in making this an aggressive fast-track project. So, precise definitions of the extent of completion necessary were developed. Another major concern was to assure the quality construction, which had been promised to the community.
There could be an overwhelming tendency for the contractor to accelerate all work to the maximum degree possible, putting in a minimum of workmanship in order to achieve maximum production and gain another day of incentive pay. To this end, quality specifications previously never applied to a bridge project were developed for the smoothness of the deck wearing course and the allowable drop below the concrete surface to the top of the expansion devices. Incentive/disincentive schedules were interfaced with the results obtained, very similar to the concrete paving specs that have been judged responsible for improving and maintaining ride quality.
Another area of intense effort in the contract-development process was that of worker safety, particularly fall protection. The contractor was required to submit a specific plan with details showing how static line installation would be integrated with deck removal so that the safety equipment installation was safe. An on-site safety engineer was also required, to review the start of each major operation and make necessary adjustments. Designated safety representatives would fill in when the safety engineer wasn't present.
Finally , on June 25, 1993, Contract B was let to Johnson Bros. Corp. at $19,767,470. Knowing that the contract was on the fast-track approval process, Johnson Bros. immediately started intensive planning and mobilization. The contractor selected three of his most seasoned superintendents to oversee crews working in four zones on the bridge. Anticipating natural tendencies for each to work to his own advantage, the Johnson Bros.' president committed himself to be the first-phase on-site coordinator. The contract specified this project to be partnered, so the Wisconsin partners were brought into these sessions. At the partnering sessions, an invitation was extended to MnDOT to attend the superintendents' meetings.
These planning sessions, often running late into the evening, dealt with details of all sizes. The goal in having an outsider's ears listen to the process, was that maybe an overlooked item would be thought of or a different method found to be better. A side effect was that a greater understanding of the depth of preparation was gleaned from these intensely focused meetings.
Contract B was nonrestrictive in the way the deck replacement was to be done, except for while working in the truss area. The truss-arch area and anchor arms required unloading and loading sequences to prevent torsional destruction of rotation of the cantilevers. But elsewhere, the deck removal and replacement was at the contractor's discretion, either full width or one roadway at a time. A permissible longitudinal construction joint over either of the interior girders would allow removal and replacement of either lane without disturbing the adjacent roadway and still permit adequate projection of transverse- deck reinforcement.
Johnson Bros.' scheme for the fall of 1993 was to first remove, then totally replace, the northbound deck from the Minnesota abutment to pier 14, and the same from pier 19 to pier 43. The 480-ft suspended deck in the arch would be removed and replaced, full width for loading considerations. The cantilevers, anchor arms and one span of counterbalancing weight for the ends of the anchor arms would be left to enable painting equipment to have a staging area. The southbound lane would be available for access to the work areas until completion of the season's operations. From pier 43 to the Wisconsin abutment the northbound deck and structural steel would be removed in preparation for the total reconstruction of those nine spans.
For the first time in any MnDOT contract, an advertising agency was retained by the state for purposes of informing the public of the bridge closure, construction progress, and detour and traffic management information. An upbeat campaign was developed and readied as the Labor Day weekend approached. Local and regional media were enthusiastic in their coverage as this monumental disruption of a 31-year-old Duluth-Superior connection loomed ever closer.
Meanwhile, the contractor was mobilizing his materials and workforce. Over 300 employees would be engaged in the three-shift removal operations and daylight shift replacement work. Orientation and information meetings for perspective employees were held, at which an explanation of the project was given and a detailed, project-specific safety video was shown. The contractor had developed and tested a static-line support system with a pass-through feature that permitted moving a safety lanyard past a post without disconnecting. The video demonstrated the system and a safety harness, not a belt, that every employee would wear-in addition to a hard hat-at all times. The meetings made it very apparent that safety would be a major issue and cause for dismissal if not followed. Meanwhile huge plywood boxes labeled with the superintendent's name and the contents began arriving from the contractor's shop. Every adjustable deck-joist hanger the contractor could locate was rented and the total wasn't enough for the first phase, so the contractor designed his own variation and had several thousand fabricated. All materials were planned and targeted to its specific area of use.
As the contractor prepared for the shutdown on Sept. 7, the public information campaign kicked off its detour promotion with the unveiling of the "Travelin' Bug." This "vehicle" could at best be described as a distinctive collage of vehicular parts. The campaign wanted to demonstrate that the detour would be adequate, safe, and not all that inconvenient. The public information campaign was waged on radio, TV, newspapers, and billboards.
The Contract B traffic provisions permitted a series of 750-ft long barrier-protected lane closures separated by 500-ft long pull-off zones to begin preliminary work on deck removal. This proved to be essential to completing Phase 1 work. The contractor had planned to make a longitudinal full-depth sawcut midway between the stringers and then saw the deck into 15 ft lengths transversely. As the longitudinal sawcut was made, slight offsets could be seen as traffic passed by. When the transverse cuts were started, the combination of traffic-caused deflections and vibrations bound the saw blades in the cuts, quickly ruining expensive blades. The contractor was forced to use backhoe-mounted hydraulic hammers to cut the deck concrete, followed by cutting the reinforcement with an acetylene torch. The rails and curbs were crushed inplace with hydraulic shears, also mounted on a backhoe. This work progressed in three locations in preparation for the traffic shutdown, but much slower than if the sawing had worked.
The traffic detour was placed in effect at 12:01 a.m. Sept. 7, 1993 as planned. The clock started on the 327 days until the reopening, but nobody noticed. Crews immediately moved floodlights to the four main work areas and removal operations began. The temporary barriers were unbolted and hauled off. Adjacent deck sections were transversely cut and the median barrier ground to bits with the hydraulic shears. The saws were found to still be unusable transversely, since construction operations shook the bridge as much as ever. Forty-ton hydraulic cranes centered themselves over the stringers and began lifting 8-ton slabs of deck and lowering them to the ground, neatly flanking the bridge. Three holes drilled through the slabs were used to pick the deck sections using two-part, diagonally cut cylindrical wedging devices.
Removal activities went on around the clock for approximately three weeks to gain a sufficient lead on the carpenters and ironworkers eager for a place to work. In the main span, the deck was sawed into four strips, then cut with hoe-rams into 15-ft pieces. Two 40-ton hydraulic cranes began synchronized, balanced removal of deck pieces at the Minnesota end of the span. Each crane loaded two slabs onto a 30-ton mine truck behind it, then backed up 15 ft as the pair of trucks drove simultaneously off the span at the end of the Wisconsin anchor arm and dropped their loads over the sides. As those 8-ton slabs hit Connor's Point 130 ft below, all north end Superior could tell that work was being done on a Saturday. With the deck weight off, there was 81/2 in. of additional clearance over the water. Seven inches of arch shape change and hanger cable relaxation combined with a 11/2 in. rotational rise of the cantilevers to present a unique challenge to the establishment of new deck forming grades on the suspended span.
Virtually as soon as an open space was available, deck forming began in each of the four work zones. The pin and hanger replacement operation went well, easily not interferring with deck replacement. Deck reinforcing followed installation of shear studs, with bar delivery scheduled to the jobsite like an assembly line. The first deck section was concreted just 17 calendar days after traffic was removed. All deck concrete was placed using concrete pumps except for the main arch where a belt system was necessary. Strike off was done with a self propelled vibrating screed that rode the top mat of reinforcing, except in the arch where the full width placement justified using a twin-roller finishing machine. Deck concreting took place in as many as three locations simultaneously, reaching a peak during the week ending Oct. 15. Barrier-rail slip forming followed as soon as the deck unit was cured, with 6,757 lin ft cast before winter clamped down on the region. The Phase 1 goal had been met.
The season's end also found that only half of the Minnesota anchor arm had been repainted, delayed both by difficulties in setting up the fiberglass panel enclosure and slower than anticipated progress in abrasive blasting removal of the old paint. The platforms were moved to the suspended span before the cranes were blocked by the ice forming rapidly in the river. Next spring, the suspended deck beams could be painted in heated enclosures before vessel traffic resumed, eliminating a prickly navigational clearance problem.
A reduced workforce stayed home very few times that winter, one during which 35 deg below zero temperatures were recorded on site.
The cold temperatures revealed a major error. A misunderstanding of a plan by the modular joint manufacturer lead to an incorrect gapping table on his shop drawings. One side of nine joints already installed in the northbound structural slab would require jack hammering the embedded iron loose and recasting to a tighter setting. Expansion/contraction measurements taken at the joints also revealed some erratic measurements due to stiffness of bearing assemblies and tall, flexible piers. In order to assure that movements remained in the range the joints were capable of accommodating, a restrainer system was designed and installed. Fortunately, all this corrective work had no direct effect on the progress as a whole. The columns of piers 44 through 51 were shortened and new, deeper, longer caps were cast. The abutment widening and caps accommodated new rolled beams. Decking operations quickly followed steel erection.
On March 1, 1994, the moderating temperatures permitted casting a 406-ft segment of northbound deck, which of necessity was insulated. It would be the end of the month before it was warm enough to cast the next segment. April was deck casting month for the southbound lane, completing 3,807 lin ft. This deck, because of the offset construction joint, was 41 ft 8 in. wide as compared to the 33 ft 8 in width of the northbound. All structural slabs were inplace by May 11.
In April, rail slipforming resumed, alternating priorities with the median barrier now that the centerline of the deck was replaced. In a single shift, over 800 ft of top-quality rail was cast. The painters, now working on the arch, competed with the decking operation for space on the anchor arms and cantilevers. The low-slump concrete wearing course gradually closed in on the hub of activity-the truss area. A shot-blasting subcontractor moved a shiny new machine into town to prepare the deck for overlay, cleaning a 6 ft wide swath faster than walking speed, with no dust and little engine noise. High spots in the structural slab could be lowered 1/2 in. by slowing it down a little. This was one of the most impressive innovations demonstrated on the project, and at a fraction of the cost of dust abated sandblasting.
Work became intensely focused towards a late June reopening. At 11 p.m. on June 29, 1994-only 296 days after closure-hundreds of citizens awaited their first opportunity to travel the new, safer Blatnik Bridge. Over 18,200 cu yd of concrete now solidified the deck rejoining the two northern states. Only one month of tourist season had been sacrificed and 31 days of incentive bonus was happily awarded to the contractor.