Summary of Released MnDOT I-35W Bridge Documents: An Ugly Read

MnDOT has released five documents on its website relating specifically to the I-35W bridge (bridge number 9340) which we have packaged into a zip file (35 MB). The files include an outside consultant review, a University of Minnesota Civil Engineering field report, two brief status summary documents, and, most troubling, a MnDOT "Fracture Critical" engineering summary which reveals in candid descriptions and shocking photographs the deterioration of many critical bridge elements. Here's my summary of the relevant parts of the documents, which we posted in our special Bridge edition of the Weekly Report last Friday:

A 299-page draft report prepared for MnDOT by the URS Corporation of Minneapolis entitled, "Fatigue Evaluation and Redundancy Analysis" for Bridge No. 9340, released July 2006.

• It recommended "steel plating of all 52 fracture critical truss members" via "additional plates bolted to the existing webs" as the best retrofit option.
  
• The bridge was designed in accordance with the 1961 AASHO Standard Specifications for Highway Bridges - a completely different fatigue design method that was revamped in the 1974 interim edition. "The poor fatigue details on the truss spans, particularly those inside the main truss tension chords that are difficult to inspect, have raised concerns on the the consequence of a possible main truss member failure triggered by a fatigue crack." The 1961 AASHO specs don't account for "bending moments" on truss members; only "axial force" is accounted for.
  
• URS developed a 3D model to simulate stress on the bridge. Pages after 169 contain bridge failure simulations.
• Page 171-174 have eerie wireframe simulations of bridge collapse: computer-generated ghosts of future events. Page 174 shows middle span failure resembling the collapse video released yesterday.
• Page 176 indicated a single upper chord main truss member collapse would cause more than a dozen floor trusses to fail under eight different simulated load conditions
• Page 231 has a drawing of major internal collapse of center simulated from a single truss chord failure
• Page 236 explores retrofit schemes: "Due to the nearly double symmetry of the structure, these eight members actually represent thirty-two truss members on the bridge." Carbon-fiber reinforced polymer plating (CFRP), which "would be expected to take over the member forces" but would be difficult to apply and "determined inappropriate" as a potential patch. Pre-tensioned bars are deemed inappropriate due to non-satisfactory performance
• Page 265 considers high-performance steel plating as the most suitable. Design loads were determined from the 3D computer model, and new plates would have to be installed extremely carefully to avoid increasing stresses.
• Page 267 says replacing the deck would provide better strength by more widely distributing the load.
• Page 299: in the conclusion someone underlined the passage: "it is more desirable to keep this symmetrical loading condition during deck replacement as much as possible," meaning that load symmetry across both sides should be maintained while work is underway, since removing either the northbound or southbound decks alone would introduce even more stress to the structure.

The brief MnDOT Structure Inventory Report (3 pages) for the last inspection on May 15, 2006, (which was generated and posted by MnDOT on August 2, 2007) indicates the deck was 6% unsound. Condition codes were: Deck: 5; Superstructure: 4; Substructure: 6; Channel: 7. Appraisal ratings were: structure evaluation: 4; Deck Geometry: 4; Underclearances: 7; Waterway Adequacy: 9; Approach Alignment: 8. The in-depth inspection for "fracture critical" status was on a 48-month cycle apparently last performed May 2003. Underwater inspections were on a 60-month cycle last performed December 2004. Bonus: 15% of the paint was unsound, it was painted in 1968, the primer type was "LEAD," and the finish was "OTHER (UNKNOWN)," though the more detailed Fracture Critical report has more information about repainting.

The Bridge Inspection Report (4 pages) has all the noted inspection issues going back to the 1970s. Current highlights include Element 412: "Relief joints need re-sealing"; Element 131: "Main truss members have numerous poor weld details (some cracked tack welds)"; Element 984: Drainage: "Pier 6: horizontal drain trough has inadequate slope (usually clogged)."

The March 2001 University of Minnesota Fatigue Evaluation of the Deck Truss of Bridge 9340 by Heather M. O'Connell, Robert J. Dexter, P.E., and Paul M. Bergson, P.E., conducted by the Department of Civil Engineering, used strain gages to measure the effects of load stress upon the bridge members.

• Researchers installed "on main trusses and floor truss to measure the live load stress ranges." While trucks of known weight crossed the bridge, researchers measured the stress gages then developed computer models to calculate stress ranges throughout the deck truss.
• The Abstract notes "the bridge's deck has not experienced fatigue cracking but it has many poor fatigue details on the main truss and floor truss system," and they concluded that "fatigue cracking of the deck truss is not likely." The project also "verified that the use of strain gages at key locations combined with detailed analysis help predict the bridge's behavior."
• Pages 54-57 have interesting charts of stress measurements
• After page 68, simulation data renders results impressively similar to their field stress tests. That's some quality computer modeling.

Finally, the most grim 50-page report came from MnDOT engineers in June 2006. Entitled "Fracture Critical Bridge Inspection: In Depth Report" for "Bridge #9340 (Squirt Bridge)". MnDOT engineers reported a profusion of "pack rust," which is essentially the formation of rusting cavities between steel members. The executive summary is grim:

If bridge replacement is significantly delayed, the bridge should be re-decked. The design of the main river spans do not allow for deck widening. Any re-decking contract should also include a complete re-painting of the superstructure, elimination of the hinge joint in span #2, and reconfiguration of the deck drainage system.
· Fatigue cracks at girder #1C (NBL), crack at the diaphragm bottom cutout, NE side measures 2" ("front face") and NW side measures 2-½" ("back face"). Fatigue cracks a girder #3 (NBL), crack at the diaphragm bottom cutout, measures 1-½" (both sides). The cracks are located in negative moment regions where the diaphragm web stiffener was not welded to the top flange and were previous fatigue cracks occurred and were repaired in 1998 and 1999. These areas should be inspected next year for any lengthening of the cracks and drilling of possible stress relief holes.
· Span 3, stringer #7 NB, has a 1-½" crack in the web with one 2" hole drilled. It is recommended to drill a 2" hole at the other end.
[....]During the 1998 inspection, numerous fatigue cracks were found in spans #3 - 5 and #9 - 10, the approach spans. The cracks were located in negative moment regions where the diaphragm web stiffener was not welded to the top flange. At one location the web had cracked through entirely. Most existing cracks were drilled out, and the fractured beam was reinforced with bolted plates. To reduce the stress levels, the diaphragms were lowered. Due to the widespread cracking, these areas should be inspected in-depth on an annual basis.

Long Term Repair Recommendations included:

· The long term plans for this river crossing need to be defined with replacement, redecking, etc. Due to the "Fracture Critical" configuration of the main river spans and the problematic "crossbeam" details, and fatigue cracking in the approach spans, eventual replacement of the entire structure would be preferable.

Under "Immediate Maintenance Recommendations":

Four-stringer connection bolts, all in the NBL, need replacement. At panel point #8, stringer #2 has 2 loose bolts, and the bearing block has rotated. This will likely require jacking the superstructure. Stringer bolts also need replacement at panel point #8, stringer #4, south side, and at panel point #11, stringer #3.
· Several strip seal joints are leaking. The glands have ripped or pulled out. Attempts were made to replace these joints during the 1998 repair contract, but the steel extrusions, which anchor the gland, had severe corrosion, and new glands could not be installed. Instead, a new product was used at the, SBL, south abutment. This utilized a hot pour seal with wire mesh reinforcing. The final product looks similar to a strip seal gland. We should monitor this joint to see how well this new gland repair performs, and consider using it at other locations.

"Areas of concern - future inspections" included:

· The truss end rocker bearings & main truss bearings should be measured for movement during each annual inspection. The truss end floor beams & approach end "crossbeams" should be closely inspected. They have section loss, had flaking rust & fatigue cracks (open finger joint).
· The hinge joint in span #2 is locked in full expansion several beam-ends are contacting, and the hinge bearings are "frozen" and no longer functioning. Consequently, pier #1 has tipped slightly to the north, and the south abutment bearings are in full contraction. This area should be thoroughly inspected.

In the section "BRIDGE DECK: NBI CONDITION CODE 5," MnDOT engineers noted:

Structural Slab: Underside of the deck has a moderate amount of transverse leaching cracks, with some areas of leaching map cracks & spalling, particularly in the south approach spans. In 1998, the median coping overhangs were replaced with steel stay-in-place forms, and the exterior copings were repaired with shotcrete. During the median slab removal, the bays adjacent to the median were damaged - some of the "stool" concrete along the stringers & beams has spalled off with exposed rebar; and in some locations, the spalling extends into the underside of the deck.

[.....]Open Finger Expansion Joints: Deck has three open finger joints, one above the hinge joint in span #2, & one at each end of the truss spans. In 1999, rubber "skirts" were installed below the truss end finger joints & the drain troughs were removed. Strip Seal Expansion Joints: Strip seal, type "H" joints at the abutments, pier #11, and at five stringer joints in the main truss spans. These were installed in 1978. Strip seal glands have pulled out, with joints leaking, in several locations. Steel extrusions, which anchor the glands, have severe section loss, making gland replacement impossible. In 1998, the south abutment, SBL, gland was patched using an experimental system. Hot poured seal with wire mesh reinforcement.
Poured Deck Joints: The deck has several transverse poured joints, from staged deck construction. All of these joints are leaching below; & at some joints the deck is spalling below.

In the section "BRIDGE SUPERSTRUCTURE: NBI CONDITION CODE 4":

Paint System: ...Currently, the overall paint system is approximately 15% unsound. The truss members have surface rust corrosion and pack rust at the floorbeam & sway frame connections, and there is paint failure & surface rust corrosion in scattered locations. The floorbeam trusses & stringer ends have surface rust corrosion at the stringer expansion joints. Some of the areas re-painted in 1999 have severe section loss. This includes the sections of the floorbeam trusses & sway bracing located below the median, and the truss end floor beams & "crossbeams", located below the open finger joints.

Main Truss Members: The two steel deck trusses are comprised of "built-up" welded members; connections include both rivets and bolts. While most truss members are welded box beams, some tension vertical & diagonal members are welded "H" beams. The truss members have numerous poor weld details. The vertical "H" beam truss members have transverse welds at the floor beam connections. The box beam truss members have welded interior stiffeners. Some of these have tack-welded tabs. Many of these tack welds have cracked. Some box beams have tack welds, or tack welded backer bars along the interior corners. The truss members have surface rust corrosion at the floor beam and sway frame connections. Pack rust is forming between the connection plates. There is paint failure, surface rust, and section loss, flaking rust in scattered locations. The interiors of the box members have severe pigeon debris. In 1999, screens were placed over openings in the truss members to prevent pigeon access. This unfortunately prevents inspection of the interiors. During the 2004 inspection, & every two years after, the plastic pigeon screens are removed on all tension and reversal members to visually inspect the member's internal diaphragms. Any questionable welding flaws discovered during this inspection were tested with magnetic particle equipment.

Floor Beam Trusses: There are 27 floorbeam trusses connecting the main deck trusses. These trusses are comprised of rolled H-beams with welded connections. The floorbeam trusses cantilever beyond the main truss on both sides. They are connected to the main truss, vertical members with bolts & rivets. The floorbeam truss members have numerous poor welding details, including plug welded web reinforcement plates, and tack welds & welded connection plates located in tension zones. Some of the top chord splices are offset vertically, up to ½" - from original construction. The splice plates are bent. The floorbeam trusses below stringer joints have section loss, severe flaking rust. There is pack rust and surface pitting at the main truss connections. In 1999, the floor beam sections below the median were re-painted. Some areas have section loss with holes.

Steel Multi-Beam Approach Spans (spans #1 - 5 & #9 - 11): [....] In span #2, multi-beam approach span, there is a cantilever expansion hinge with sliding plate bearings. The joint is closed beyond tolerable limits, possibly due to substructure movement & pavement thrust and is no longer functioning. Some beam-ends are contacting, and some bearing plates have tipped, preventing the joint from reopening. The hinge area, with open finger joint above, was re-painted in 1999. The beam-ends have section loss, moderate surface pitting.

Parts of "BRIDGE SUBSTRUCTURE: NBI CONDITION CODE 6":

Abutments: The abutments have vertical cracking, with some staining from leaking deck joints. [......]

Approach Span Piers: Piers #1 - 5 & #9 - 11, piers supporting the steel spans, consist of concrete columns with a cap. Those adjacent to railroad tracks have lower struts. The pier columns supporting the voided slab spans (piers #12 & 13) are cast directly into the slab with no cap. Pier #1 has tipped slightly to the north. This is related to the hinge failure in span #2. The east column on pier #9 has minor scrapes & spalls from a train derailment in 1969. Pier #11 has extensive shotcrete repairs from leaking deck joint above.

The next section, "Bridge Snooper Field Investigation," reads like an engineer's worst nightmare, complete with photos of rusted beams, missing pins, "tipped" sliding plates, water-saturated decking, a 90° rotated bearing block, missing bolts, two plugged drains, cracked tack welds, and other ominous findings. Along the U5/U6 truss elements (whose collapse was simulated by the July 2006 URS report), MnDOT discovered that "Truss bottom chord, bottom lateral connection plates have spread 3/16" from pack rust." On Panel Point #8' (West Truss Pier #7 Stringer Joint), "Sway bracing center horizontal has 3" x 8" severe pitting & ½" diameter hole; bottom sway bracing has a 2" x 3" hole between stringer #11 & stringer #10," as photos show. While the de-icing system installed in 1999 has thus far not been linked to the collapse, the report notes "Girder #12 has paint failure from leaking de-icing system."

A roster of "previous snooper inspections" indicated the teams have gotten continuously larger since Chester Martin checked it out by himself in 1988. The report's primary authors know the bridge well: Kurt Fuhrman inspected the bridge 11 times since 1994, and Vance Desens inspected it four times since 2001.

In Appendix A, a handy table view of diaphragm crack locations indicates that many cracks were detected in 1998 but never repaired, and in some cases deteriorated further.