I-35W Mississippi River Bridge Collapse and the G/P Phenomenon
About 6:05 p.m. central daylight time on Wednesday, August 1, 07, the eight-lane, 1,907-foot-long I-35W Highway Bridge over the Mississippi River in Minneapolis, Minnesota, experienced a catastrophic failure in the main span of the deck truss. As a result, 1,000 feet of the deck truss collapsed, with about 456 feet of the main span falling 108 feet into the 15-foot-deep river. A total of 111 vehicles were on the portion of the bridge that collapsed. Of these, 17 were recovered from the water. As a result of the bridge collapse, 13 people died, and 145 people were injured. On the day of the collapse, roadway work was underway on the I-35W Bridge, and four of the eight travel lanes (two outside lanes northbound and two inside lanes southbound) were closed to traffic. In the early afternoon, construction equipment and construction aggregates (sand and gravel for making concrete) were delivered and positioned in the two closed inside southbound lanes. The equipment and aggregates, which were being staged for a concrete pour of the southbound lanes that was to begin about 7:00 p.m., were positioned toward the south end of the center section of the deck truss portion of the bridge and were in place by about 2:30 p.m. About 6:05 p.m., a motion-activated surveillance video camera at the Lower St. Anthony Falls Lock and Dam, just west of the I-35W Bridge, recorded a portion of the collapse sequence. The video showed the bridge center span separating from the rest of the bridge and falling into the river.” 2
The I-35W Mississippi River Bridge was constructed in 1967 to serve Interstate W35 traffic in Minneapolis, Minnesota. The bridge was provided with eight lanes for traffic which in 2007 numbered about 140,000 vehicles a day. The bridge was designed by Sverdrup & Parcel and Associates of St. Louis, Missouri in accordance with the 1961 AASHO (American Association of State Highway Officials) specifications. It consisted of two, three-span continuous deck trusses, six continuous steel-girder approach spans, two steel-beam approach spans, and three concrete voided-slab approach spans. Abutments and piers were of reinforced-concrete construction. On August 1, 2007, after nine days of high temperature that averaged about 90º, the bridge suddenly collapsed.
After this collapse, the National Transportation Safety Board (NTSB) was given the responsibility for investigating this national tragedy. The report of the board was completed and published about a year later. The abstract of the report described the event as follows.
The content and especially the conclusion of the NTSB Report on the collapse of the I-35W Bridge is shocking to an engineer who has spent almost his entire fifty-year career evaluating the behavior of real bridges. Upon examining the contents of NTSB report, the author discovered that the investigation was elementalistically focused on the behavior of only one part of the complete or “composite structure,” its deck truss system (see EXHIBIT A). Missing from the investigation was any awareness of the behavior of the other parts of the composite structure (six steel-girder approach spans, two steel-beam approach spans, three concrete voided-slab approach spans, two approach embankments, two approach concrete highway pavements, substructure foundations, etc.), or how the behavior of these other parts of the structure, taking into account the forces associated with them and their various application orientations, would impact the behavior of the overall structure. Consequently, the NTSB investig- ation appears to have been merely an academic exercise that focused only on the response of a deck-truss system to hypothetical vertical loads.
Also, what appeared to be missing from the NTSB report was any awareness that the conclusions of the report could have life or death consequences for other travellers if bridge maintenance engineers were not given suitable recommen- dations to guide them in their own evaluation and possible modifications of other structures similar to the real I-35W Bridge.
Unfortunately, this was not the first time that such a short-sighted structural investigation was performed on this bridge. In 2001, the Minnesota bridge maintenance forces were concerned about web cracks and out-of-plane distortion of the bridge’s transverse approach-span support girders. The investigation of the bridge by others (3) unfortunately consisted in only a fatigue evaluation of the bridge’s deck trusses responding to hypothetical vertical traffic loads. Entirely missing from that investigation was any holistic awareness of the behavior of the ten approach spans, and the behavior of the two approach roadway pavements, etc., all with respect to how their behavior would affect the performance of the transverse support girders.
Neither of these investigations accomplished the real purpose of the NTSB. And little of real substantial value was learned from either of these investigations. Actually, in a sense, these investigations were somewhat counter productive in that they failed to discover the primary causes of the structural distress that led to the collapse of the bridge. Consequently, they did not contribute fundamental knowledge to the bridge engineering profession. They in effect have left bridge design and maintenance forces holding the bag containing a structural mystery, so to speak, and have kept the traveling public, who in the past have come to trust bridge engineers, at risk of injury and/or death. Because of this sad situation, this author urges that the NTSB investigation of the I-35 Bridge be reopened so that a comprehensive investigation of the composite structure (4, page 158) can be carried out. In such an investigation, hopefully, all of the primary forces that led to the collapse of this bridge will be identified and described. Based on such a conclusion, all bridge maintenance forces can then be given the kind of directions they need that will enable them to modify their own structures as necessary to protect the integrity of their structures and most importantly, to ensure the safety of the traveling public.
Based on the symptoms, signs, drawings and photographs that have appeared in the bridge plans, bridge inspection reports, NTSB Report, etc., that have been made available on the internet (1), and other records that the author was able to find, the author is convinced that the collapse of the I-35W Bridge was not “caused”by weak and corroded truss gusset plates. Instead, he believes that the collapse was initiated” primarily” by the pavement Growth/Pressure Phenomenon. The remainder of this paper will be devoted to describing how this bridge was gradually and periodically being squeezed and/or compressed by the growth and pressures being generated in the bridge’s jointed concrete approach pavements.This conclusion will not seem strange to State, County and City pavement and bridge maintenance engineers, engineers who have been dealing with this scourge periodically for most of their pavement and bridge maintenance lives. However, it may seem odd to public, consultants and university bridge design engineers who have not had the opportunities of continually dealing with real or physical bridges surviving in the transportation environment.
For instance, it has been noted in the bridge inspection records that Pier 1, a pier with fixed bearings, was tilted to the north. Also, as described below, a photograph is presented showing that the movable deck joint to the rear of Pier 2 had been jammed shut (5,p.17). Obviously, the superstructure of spans one and two has also been moved or jammed to the north. This behavior is typical of bridges located adjacent to jointed concrete pavements. Typically, the growth and pressures generated by the G/P Phenomenon against the south abutment would force the abutment forward or northward enough to close the movable deck joint at the abutment (a 2 inch (50mm) movement). After this 2 inch growth of pavement, the G/P Phenomenon probably began to generate pressures of sufficient magnitude to eventually force the south abutment, the superstructure of spans 1 and 2, and Pier 1 northward sufficient to close the movable deck joint to the rear of Pier 2 (another 2 inch (50mm) movement). This author is convinced, but he does not know for a fact, that it was this movement of the south approach structures, and probably the north approach structures as well, that motivated the state bridge maintenance forces to modify the damaged tops of the abutments so that pressure-relief joints could be installed in the bridge approaches. Presumably, this modification was accomplished in 1986 or earlier when the pavements and bridge were about nineteen years old.
Obviously, after this second movement, the generation of pavement pressures would have commenced again until the entire south and north approach spans were jammed tight against the bearings of the two deck trusses. At this stage, a knowledgeable engineer would be tempted to ask how this continual pressure could have been generated against this structure, especially since pressure-relief joints had been installed in both bridge approaches. Unfortunately, that appears to be the crux of the story behind the collapse of this structure. The type of pressure-relief joint chosen to protect this structure could function effectively for only two years or three years at the most before they became totally compressed and ceased to function (4, Figure A 1.9). The inspection records, at least those from 1997 to 2006 for the north approach pavements, indicate that these relief joints were only 1 inch (25mm) wide (essentially closed and ineffective). It appears however from the maintenance records that the relief joints at the south approach had been changed in 2004 since these joints measured slightly less than 4 inches (100mm) at that time. It is also instructive to recognize that by 2007, the pressure relief joints in both approaches would have been closed and ineffective in protecting the structure from the continual pressures being generated by the G/P Phenomenon. See below for details of the pressure-relief joints and also inspection records about their condition.
Figure 3 is a file photograph of the crossbeam between the north end of the truss spans and the south end of the north approach spans, at the west side of the bridge (5, p. 29). It appears clear that the crossbeam at this location has been modified a number of times over the years with structural additions to stiffen the cross beam against continual longitudinal pressure being exerted by the approach spans and the cross beam against the two inoperative bearings. For example, the peculiar riveted stiffener and the longitudinal beam brace in Figure 3 appear to be stiffening additions to the original crossbeam. Presumably, similar stiffening additions were also installed in the other three corners of the truss spans. In Figure 3, also notice that two holes near the right edge had been drilled into the web of the crossbeam to terminate cracks that had been induced into the horizontal stiffener weld, presumably by the force of the bearing against the crossbeam web at this location. The bridge records are full of comments about distress at these four junctions between the primary truss and secondary approach spans of this bridge. For example, consider the following statements recorded in the records:
“In 1986, the southeast rocker bearing “froze,” resulting in damage to the crossbeam with two cracked web stiffeners. The rocker bearing pin was replaced…. The crossbeam was repaired and the cracks in the web stiffeners were welded, crack ends drilled out, and stiffeners reinforced with angle plates. Installing braces between the crossbeam and beams #2&3 also reinforced the connection.” 5, p. 13.
When this damage was occurring to the crossbeams of the I-35W Bridge in 1986, it would not come as a surprise to the author to discover that damage to the crossbeams occurred on the exact same day that damage was occurring to the concrete pavements on other stretches of the I-35 pavement. The Minneapolis Star Tribune Newspaper published an article about the local consequences of the heat generated pavement pressures. An abstract of the article dated June 1, 1986 stated: “Hot popping pavement on Interstate Hwy. 35W in Mound View causes traffic delays and a few flat tires late Saturday afternoon [June 1stt], the Minnesota Patrol said.” Obviously, nothing was reported in the newspapers that day about the I-35W Bridge damage since bridge maintenance engineers would probably not be aware of such damage until sometime later. Also,crossbeam damage to the I-35 Bridge would not of particular interest to the traveling public.
See Exhibit B for copies of Newspaper articles about other hot temperature events that occurred in Minneapolis in 1998, 2000, and 2007 where it appears that the ubiquitous G/P Phenomenon generated unsustainable pavement pressures. Also refer to Exhibit C for other comments (1986, 1992, 1997, 2000 and 2006) from inspection records about pressure damage and repairs at the junction of the transverse support crossbeams and the deck-trusses’ malfunctioning corner bearings.Note especially that in the report of 2006, a year before the collapse of the bridge, the inspection records documented the fact that the “SW rocker bearing has no movement.” 5, p 30.
Based on the meager evidence that this author has been able to gather (Because of on-going litigation associated with the I 35W Bridge tragedy, he was not permitted to question state design or maintenance engineers, or the members of the State Highway Patrol), it appears that the I-35W Bridge has, for a long period of time, been subjected to significant longitudinal forces (more on this later) that have been generated by the bridge’s approach pavements. These forces appear to have been transmitted through the approach spans (girders and deck slabs), and, by way of the transverse approachspans’ support girders (crossbeams) and inoperative truss bearings, into the main structure’s end spans (truss and deck slab), and because of the five deflection joints in the deck slabs of the deck trusses, into the top chords of the three main spans as well.
04-03-2000: (1991) All four approach panels have transverse cracks (Relief joints need sealed)
09-26-2001; 05-17-2002; 05-13-2003; 06-10-2004; (same comment as above).
06-10-2005: North Approach:
26 LF SBL Ramp 2" wide
48 LF SBL 1" wide
48 LF NBL 1" wide
South Approach:
52 LF SBL 4" wide
52 LF NBL 3-1/2" wide
06-10-2006: Same as above
Figure 4 above, showing both existing and proposed pressure-relief joints for the roadway approach panels, provides some evidence that the bridge design and maintenance engineers of Mn DOT were aware of and were trying to provide protection for the I-35W Bridge against the pavement forces generated by the pavement G/P Phenomenon. Notice that the figure indicates that the bridge was somewhat protected in the past by joints that were originally 4 inches (100mm) wide. However, this choice of joint type was unfortunate because in an aged pavement, they could only be effective for two to three years at the most; but in aged interstate pavements, probably not more than two years at most. Actually, it is this author’s opinion that such joints should only be used for emergency repairs and should be replaced within a year or two with joints that are”at least” 4 ft. (1.22m) wide.
The yearly inspection reports (See the tabulation above) contain comments on the condition of these pressure-relief joints. Unfortunately, the most important thing that the inspectors could think to say about the relief joints in the yearly reports for 2000 to 2004 was that they needed sealed. However in the 2005 and 2006 reports, the inspectors provided joint measurements, suggesting that some bridge maintenance engineer in authority was concerned and paying attention. The reports indicated that at the North Approach to the structure, both the north-bound and south-bound pavements contained 4 inch (100 mm) wide joints that had been compressed to 1 inch (25mm), a clear indication that the approach pavements to this structure were generating sufficient pressures to grow the pavementand close not only the pressure–relief joints but some of the movable bridge joints as well. At a 1 in (25 mm) width, such joints should be considered closed and ineffective in preventing the transmission of pressure from the approach pavements into the structure. Presumably, these particular joints had been 1 inch (25 mm) wide for many years indicating that the structure was unprotected and was allowing the approach pavements to generate greater and greater pressures against this structure. These 1 inch (25m mm) wide joint measurements were probably the reason why they were being replaced in the rehabilitation contract. Unfortunately, they were being replaced in kind. If the structure had survived,these joints would have been effective for two years at most. In the opinion of this author, it appears that these relief joints were the reason why this structure was being subjected to great pavement forces periodically for many years. It is also of interest to state that the NTSB investigators appeared to be unaware of these pressure-relief provisions,or if they were aware of them, they appear to have ignored their significance.
The South Approach to the structure was somewhat protected with originally 4 inch (100 mm) wide pressure-relief joints. Based on the 2005 inspection report, these joints had compressed to 3½ inches (87.5 mm) in the north-bound lanes and had remained 4 inch (100 mm) wide in the south-bound lanes, suggesting to the author that these relief joints were newly installed.The author is left to wonder why the relief joints in both the south and north approaches were not replaced at the same time? Presumably, if this had been done, the pavement forces against this structure would have been considerably less and such a change might have saved the structure for a few more years until major modifications of the bridge could have been accomplished. However, with respect to these South Approach relief joints, the joint dimensions were exactly the same in the 2005 and 2006 reports. They had not changed a half inch (12mm) in a year’s time. These reported joint dimensions seem hardy credible to the author, and in addition, the apparent lack of joint movement suggests to the author that the inspectors neglected to measure these joints in 2006 and instead repeated the dimensions that had been measured the year before. If the author had been the supervisor of an inspector who turned in such duplicate dimensions, and since these dimensions were critical for the welfare of the structure, and since it would have been practically impossible for them to remain the same after a year’s time, that inspector would have been fired.
Presumably, the south approach pressure-relief joints would have compressed to 2 inches (50mm) in 2006 and to 1 inch (25mm) in 2007. Consequently, it appears that the relief joints on both approaches to this bridge were effectively closed in 2007, and the bridge would have been unprotected from the G/P Phenomenon at the time of the collapse. Nevertheless, the disparity between the relief-joint dimensions on the north and south approaches suggests that the amount of pressure generated by the G/P Phenomenon would be significantly greater at the north approach to the structure making it more probable that with everything else being equal, the greater pressures at the north approach would suggest that the collapse of the structure initiated at that end of the structure.
Along with the discovery of these bowed gusset plates, indicating that significant forces had been applied to the structure, the local Minneapolis Star Tribune Newspaper reported on July 15 of the same year that, “The extreme heat is causing pavement to buckle on some highways forcing the Minneapolis Department of Transportation to make emergency repairs – even during rush hour…”(emphasis added by author). Does this appear to be a coincidental timing of bridge distress and pavement distress, and are these events both directly related to the same hot temperature event? Or, to be bolder, does the appearance of major structure distress and pavement distress during the same high temperature events suggest that the pavement G/P Phenomenon in Minneapolis pavements in July of 1998 is blowing up pavements in some areas of the city and pressuring bridges in other areas of the city? Does the G/P Phenomenon care which structure it damages just so long as it can continue to generate growth or pressure, and periodically and continuously so?
It is also instructive to note that the NTSB Report (2, p 143)also stated: “The Safety Board was unable to fully evaluate the possibility that some type of usual loading was generated in the gusset plates as the L9(‘)/U10(‘) diagonal was connected to the U9(‘)/U10(‘) upper chord member but considers it unlikely that such loading could have caused bowing deformation at all four U10(‘) plates”(emphasis added by author). The author would hasten to suggest the possibility that the unusual loading may have been the longitudinal pressures generated by the G/P Phenomenon on both bridge approaches.
When considering the pressures being generated in the approach pavements by the G/P Phenomenon, and recognizing that by 1998 almost of all of the bridge’s movable joints in both approaches were closed or locked, it does not seem difficult to imagine that the pressure from the pavements had finally entered the deck truss, and that in conjunction with the restrained expansion of the bridge, had been creating higher than usual localized stresses in deck-truss members. And these high localized stresses would with certain exceptions have little trouble reaching all eight of the bowed gusset plates.
Bill Kallman, a consulting engineer and formerly a long-term staff member of the New York Department of Transportation with extensive experience with troubled pavements and bridges,said that he had examined the pavement approaches of the collapsed I-35W Bridge and stated that “The very warm weather preceding August 1, 2007…. The “frozen” bearings, cracked and misaligned approach span members, early deck repairs, the tilted north pier, distressed condition of the on-grade highway pavement joints leading to the bridge, all point to severe pavement shove as the cause of the collapse.” 6(emphasis added by the author)
It is this author’s opinion that the Critical Flaw in the collapse of the I-35W Mississippi River Bridge of Minneapolis, Minnesota, was not the undersized and corroded gusset plates, not the weight of construction materials and equipment on the bridge, and not the streams of vehicular traffic that were on the bridge at the time of the collapse. Instead the critical flaw in the collapse of this bridge was the type of pressure-relief joint that was chosen and installed in the pavement approaches to protect this bridge from the pavement Growth/Pressure Phenomenon (4, Appendix 1). It was the choice of this joint type and its apparent neglect that permitted the approach pavements of this bridge to generate unsustainable longitudinal pressures (forces) on the structure. These forces, plus those generated by the restrained expansion of the greatly compressed structure itself, the weight of the structure and all super imposed loads,appear to have coalesced to overstress the undersized and corroded gusset plates and collapse the structure.
Unfortunately, the type of pressure-relief joint that was chosen (a 4 inch (100 mm) wide elastomer-filled joint) a type of joint that, depending upon the age of pavement where and when it was installed, would only be effective in minimizing the amount of pressure transmitted through the relief-joint and into the bridge for two years or possibly three years at most. After such time, the joint would essentially be closed (1 inch (25mm) width) and incapable of minimizing further pressure transmission. Presumably, the engineers who chose this type of joint must have assumed, because of the joint’s “name,” that the joints would be continuously effective in minimizing pressure. How else could one account for the fact that these closed joints were allowed to remain in place year after year, after their effectiveness had ceased.For example, each time the inspectors examined theseclosed joints (1 inch (25mm) width) on a yearly basis, all that they could think to report was that the joints needed sealed. And that comment was allowed to be repeated year after year. Consequently, the I-35WBridge was essentially unprotected for many years from the effects of the G/P Phenomenon, the phenomenon that eventually succeeded in compressing the various spans of the structure until the bridge eventually collapsed. Subsequently, the G/P Phenomenon was free to generate growth instead of pressure.
If instead of choosing the 4 inch (100 mm) wide joints, the Minneapolis bridge engineers had chosen 4 ft. (1.22 m) wide or wider asphalt-filled joints for pressure relief, the I-35W Bridge would probably still be serving traffic. Since such a wide joint would allow the pavement to generate growth instead of pressure, the pavement contraction joints within the first ten to fifteen hundred feet of pavement would gradually widen, presenting the pavement maintenance engineers with the problem of keeping the contraction joints well sealed. This would be a continuous problem but one that could be justified to protect a bridge from periodic damage and from possible collapse. Otherwise, bridge and pavement engineers should consider constructing the first fifteen to twenty hundred feet of pavement adjacent to bridges, continuously reinforced. Such construction should obviate the need for installing pressure-relief joints in bridge approaches. The activity of the G/P Phenomenon would then be confined to the further reaches of the pavement, and the bridges would at last be at last free of this particular scourge.
With respect to the NTSB investigation of the “central spans” of the I-35W Bridge, a statement is made to the effect “…that the change in temperature on the day of the accident did not play a role in initiating of the collapse.” 1, p. 126. The NTSB Report could not be more wrong. As indicated throughout this paper, temperature was a primary determinant by way of the pavement G/P Phenomenon in the generation of longitudinal compressive forces on the I-35W Bridge; temperature was also the primary determinant in the generation of compressive forces by way of the restrained expansion of the structure itself. The NTSB Report identified the truss gusset plates as the structure’s weak link, but it failed to recognize these huge temperature generated forces that helped to break the chain. Consequently, with respect to the NTSB investigation and Report that was published about the collapse of the I-35W Bridge, the author urges that the investigation of the collapse be reopened so that the complete composite structure (4, page 158) can be included in the investigation, and the report suitably supplemented. Bridge design and maintenance engineers depend upon the recommendations that are given in damage or collapse reports. Consequently, it is imperative that such reports reflect the results of a comprehensive investigation. Otherwise these engineers will be misled, their bridges will suffer, and the individuals that use their structures will be exposed to possible injury or death.
Finally, this author, realizing that the pavement Growth/Pressure Phenomenon was at first ignored in the investigation of the collapse of the I-35W Mississippi River Bridge, wonders how many other bridges in the past were either severely damaged or collapsed by this same unfamiliar phenomenon. He “knows” that any bridge constructed adjacent to jointed concrete pavements, unprotected by adequate pressure-relief joints, will be subjected to longitudinal pavement forces, and depending upon the age of the pavement and a number of other factors, these forces can be huge. Consequently, in the investigation of any distressed or collapsed bridges (i. e., the collapsed Canadian River Bridge of near Tulsa, Oklahoma, the damaged I-90 Grand River Bridges of Lake County, Ohio, the partial collapse of the River des Peres Bridge of St. Louis, Missouri, etc.),bridge investigators should always consider the role that the G/P Phenomenon may have played in distressing or collapsing such bridges.
“In 1986, the southeast rocker bearing “froze,” resulting in damage to the crossbeam with two cracked web stiffeners. The rocker bearing pin was replaced. This required closing I-35W and jacking up the span. The crossbeam was repaired and the cracks in the web stiffeners were welded, crack ends drilled out, and stiffeners reinforced with angle plates. Installing braces between the crossbeam and beams #2 & 3 also reinforced the connection.” 5,p.13.
“In 1992, a crack was found in a crossbeam stiffener weld above the northeast rocker bearing, which was drilled out.” 5, p. 13.
“In 1997, at the same location [above the northeast rocker bearing] a weld between a vertical & horizontal stiffener was found cracked through entirely. Cracks were also discovered at the end of horizontal stiffeners near the northeast and southwest rocker bearings. Strain gages were installed to analyzed stresses, crack ends were drilled out, and installing bracing between the crossbeam and 2 stringers reinforced the northeast connection.” 5, p. 13.
In 2000, the following observation was given with respect to the reinforced openings in the cross beams: “It appeared that resistance to movement of the bearings was causing significant out-of-plane forces and distortions of the cross girder [crossbeams], leading to cracks forming at the termination of the stiffeners reinforcing the opening, The cross girder was retrofit by drilling holes at the tips of the cracks and adding struts from the reinforcing stiffeners back to the girders to reduce the distortions. This retrofit has been successful so far in preventing further crack propagation.” 3, pp. 28 & 29
In the 2006, the following observations were given: “Panel Point #0 (EndFloorbeam End of West Truss): …[1997] Floorbeam horizontal stiffener is bent directly above the rocker bearing. [1998/1999] Floorbeam repainted, side facing finger joint has section loss, pitting. [2004] Truss, top chord exterior connection plate has 1/8 “ deep section loss with pitting. SW rocker bearing has no movement.”5, p. 30
In all of the above, no speculations were proffered as to what could have caused the bridge joints to close, or periodically caused the distress and distortion of the transition girders (crossbeams) between the approach spans and the deck trusses. Presumably, they must have thought that this distress was due to corrosion-locked bearings and the thermal response of the steel deck trusses and adjacent girder spans to thermal changes. This author wonders if any of them conceived of the possibility that the distress that the bridge was suffering was due to the fact that the bridge was periodically being compressed primarily by the longitudinal forces being generated by the G/P Phenomenon of the bridge approach pavements?
He was born in Pittsburgh, Pennsylvania in 1926. After serving with the 36th and 125th USN Construction Battalions in the Pacific during World War II, he attended the University of Pittsburgh where he received a Bachelor’s Degree in Civil Engineering. He became a career employee of the Bureau of Bridges, Ohio Department of Transportation, retiring there as Assistant Engineer of Bridges in 1982. Until 2002, he was employed by Burgess & Niple, Engineers and Architects, as Bridge Design Consultant. He is widely known for his interest in Bridge Aesthetics and in Integral and Semi-Integral bridges, and has promoted these subjects both in the United States and abroad. His latest book Integral and Semi-Integral Bridges was published in 2009.
Sir Norman Angell
“If the world has nearly destroyed itself, it is not from lack of knowledge… but is due to the fact that the mass of men have not applied to public policy knowledge they already possess, which is indeed of almost universal possession, deducible from the facts of everyday life. If this is true – and it seems inescapable – then no education which consists mainly in the dissemination of “knowledge” can save us. If men can disregard in their policies the facts they already know, they can just as easily disregard new facts which they do not at present know. What is needed is the development in men of that particular type of skill which will enable them to make social use of knowledge already in their possession; enable them to apply simple, sometimes self-evident, truths to the guidance of their common life.”
Martin P. Burke Jr., PE
Figure 1: I-35W Mississippi RiverBridge(1967 – 2007) of Minneapolis, MN.1
The I-35W Mississippi River Bridge was constructed in 1967 to serve Interstate W35 traffic in Minneapolis, Minnesota. The bridge was provided with eight lanes for traffic which in 2007 numbered about 140,000 vehicles a day. The bridge was designed by Sverdrup & Parcel and Associates of St. Louis, Missouri in accordance with the 1961 AASHO (American Association of State Highway Officials) specifications. It consisted of two, three-span continuous deck trusses, six continuous steel-girder approach spans, two steel-beam approach spans, and three concrete voided-slab approach spans. Abutments and piers were of reinforced-concrete construction. On August 1, 2007, after nine days of high temperature that averaged about 90º, the bridge suddenly collapsed.
Figure 2: Collapsed sections of the I-35W
Mississippi River Bridge.
Photo courtesy of Tim Davis.
Mississippi River Bridge.
Photo courtesy of Tim Davis.
The content and especially the conclusion of the NTSB Report on the collapse of the I-35W Bridge is shocking to an engineer who has spent almost his entire fifty-year career evaluating the behavior of real bridges. Upon examining the contents of NTSB report, the author discovered that the investigation was elementalistically focused on the behavior of only one part of the complete or “composite structure,” its deck truss system (see EXHIBIT A). Missing from the investigation was any awareness of the behavior of the other parts of the composite structure (six steel-girder approach spans, two steel-beam approach spans, three concrete voided-slab approach spans, two approach embankments, two approach concrete highway pavements, substructure foundations, etc.), or how the behavior of these other parts of the structure, taking into account the forces associated with them and their various application orientations, would impact the behavior of the overall structure. Consequently, the NTSB investig- ation appears to have been merely an academic exercise that focused only on the response of a deck-truss system to hypothetical vertical loads.
Also, what appeared to be missing from the NTSB report was any awareness that the conclusions of the report could have life or death consequences for other travellers if bridge maintenance engineers were not given suitable recommen- dations to guide them in their own evaluation and possible modifications of other structures similar to the real I-35W Bridge.
Unfortunately, this was not the first time that such a short-sighted structural investigation was performed on this bridge. In 2001, the Minnesota bridge maintenance forces were concerned about web cracks and out-of-plane distortion of the bridge’s transverse approach-span support girders. The investigation of the bridge by others (3) unfortunately consisted in only a fatigue evaluation of the bridge’s deck trusses responding to hypothetical vertical traffic loads. Entirely missing from that investigation was any holistic awareness of the behavior of the ten approach spans, and the behavior of the two approach roadway pavements, etc., all with respect to how their behavior would affect the performance of the transverse support girders.
Neither of these investigations accomplished the real purpose of the NTSB. And little of real substantial value was learned from either of these investigations. Actually, in a sense, these investigations were somewhat counter productive in that they failed to discover the primary causes of the structural distress that led to the collapse of the bridge. Consequently, they did not contribute fundamental knowledge to the bridge engineering profession. They in effect have left bridge design and maintenance forces holding the bag containing a structural mystery, so to speak, and have kept the traveling public, who in the past have come to trust bridge engineers, at risk of injury and/or death. Because of this sad situation, this author urges that the NTSB investigation of the I-35 Bridge be reopened so that a comprehensive investigation of the composite structure (4, page 158) can be carried out. In such an investigation, hopefully, all of the primary forces that led to the collapse of this bridge will be identified and described. Based on such a conclusion, all bridge maintenance forces can then be given the kind of directions they need that will enable them to modify their own structures as necessary to protect the integrity of their structures and most importantly, to ensure the safety of the traveling public.
Based on the symptoms, signs, drawings and photographs that have appeared in the bridge plans, bridge inspection reports, NTSB Report, etc., that have been made available on the internet (1), and other records that the author was able to find, the author is convinced that the collapse of the I-35W Bridge was not “caused”by weak and corroded truss gusset plates. Instead, he believes that the collapse was initiated” primarily” by the pavement Growth/Pressure Phenomenon. The remainder of this paper will be devoted to describing how this bridge was gradually and periodically being squeezed and/or compressed by the growth and pressures being generated in the bridge’s jointed concrete approach pavements.This conclusion will not seem strange to State, County and City pavement and bridge maintenance engineers, engineers who have been dealing with this scourge periodically for most of their pavement and bridge maintenance lives. However, it may seem odd to public, consultants and university bridge design engineers who have not had the opportunities of continually dealing with real or physical bridges surviving in the transportation environment.
South Approach Spans
Obviously, after this second movement, the generation of pavement pressures would have commenced again until the entire south and north approach spans were jammed tight against the bearings of the two deck trusses. At this stage, a knowledgeable engineer would be tempted to ask how this continual pressure could have been generated against this structure, especially since pressure-relief joints had been installed in both bridge approaches. Unfortunately, that appears to be the crux of the story behind the collapse of this structure. The type of pressure-relief joint chosen to protect this structure could function effectively for only two years or three years at the most before they became totally compressed and ceased to function (4, Figure A 1.9). The inspection records, at least those from 1997 to 2006 for the north approach pavements, indicate that these relief joints were only 1 inch (25mm) wide (essentially closed and ineffective). It appears however from the maintenance records that the relief joints at the south approach had been changed in 2004 since these joints measured slightly less than 4 inches (100mm) at that time. It is also instructive to recognize that by 2007, the pressure relief joints in both approaches would have been closed and ineffective in protecting the structure from the continual pressures being generated by the G/P Phenomenon. See below for details of the pressure-relief joints and also inspection records about their condition.
Transverse Support Girders (Crossbeams)
The spans of both approaches to this bridge terminated adjacent to the deck trusses by being integrated with transverse girders (called crossbeams in the records). These two crossbeams were in turn supported at the ends of the two trusses on two rocker bearings per cross beam. The records are full of statements describing the problems that the maintenance staff had with these bearings, primarily freezing of the bearings probably due to crevice corrosion. Probably another reason why these bearings ceased to function was due in part to the abnormal rotation of the curved superstructure. In other words, these bearings appear to have been inoperative as intended for most of the life of the structure. And, in addition, the author is convinced that many of the inspections of these rockers were in error when the records indicated that the rockers were moving, when in fact they were probably frozen in position. The author has had many experiences where inspectors who had examined these types of bearings were mistaken when questioned closely about the evidence that convinced them that bearings were functioning as intended by the design. The following comments from inspection records should illuminate this negative aspect of bearing behavior.Figure 3: Truss bearing at the NW corner of the deck truss.5
“In 1986, the southeast rocker bearing “froze,” resulting in damage to the crossbeam with two cracked web stiffeners. The rocker bearing pin was replaced…. The crossbeam was repaired and the cracks in the web stiffeners were welded, crack ends drilled out, and stiffeners reinforced with angle plates. Installing braces between the crossbeam and beams #2&3 also reinforced the connection.” 5, p. 13.
When this damage was occurring to the crossbeams of the I-35W Bridge in 1986, it would not come as a surprise to the author to discover that damage to the crossbeams occurred on the exact same day that damage was occurring to the concrete pavements on other stretches of the I-35 pavement. The Minneapolis Star Tribune Newspaper published an article about the local consequences of the heat generated pavement pressures. An abstract of the article dated June 1, 1986 stated: “Hot popping pavement on Interstate Hwy. 35W in Mound View causes traffic delays and a few flat tires late Saturday afternoon [June 1stt], the Minnesota Patrol said.” Obviously, nothing was reported in the newspapers that day about the I-35W Bridge damage since bridge maintenance engineers would probably not be aware of such damage until sometime later. Also,crossbeam damage to the I-35 Bridge would not of particular interest to the traveling public.
See Exhibit B for copies of Newspaper articles about other hot temperature events that occurred in Minneapolis in 1998, 2000, and 2007 where it appears that the ubiquitous G/P Phenomenon generated unsustainable pavement pressures. Also refer to Exhibit C for other comments (1986, 1992, 1997, 2000 and 2006) from inspection records about pressure damage and repairs at the junction of the transverse support crossbeams and the deck-trusses’ malfunctioning corner bearings.Note especially that in the report of 2006, a year before the collapse of the bridge, the inspection records documented the fact that the “SW rocker bearing has no movement.” 5, p 30.
Based on the meager evidence that this author has been able to gather (Because of on-going litigation associated with the I 35W Bridge tragedy, he was not permitted to question state design or maintenance engineers, or the members of the State Highway Patrol), it appears that the I-35W Bridge has, for a long period of time, been subjected to significant longitudinal forces (more on this later) that have been generated by the bridge’s approach pavements. These forces appear to have been transmitted through the approach spans (girders and deck slabs), and, by way of the transverse approachspans’ support girders (crossbeams) and inoperative truss bearings, into the main structure’s end spans (truss and deck slab), and because of the five deflection joints in the deck slabs of the deck trusses, into the top chords of the three main spans as well.
Approach Panel and Pressure Relief Joint Records1
02-04-1997; 04-11-1998, 04-15-1999: No Comments04-03-2000: (1991) All four approach panels have transverse cracks (Relief joints need sealed)
09-26-2001; 05-17-2002; 05-13-2003; 06-10-2004; (same comment as above).
06-10-2005: North Approach:
26 LF SBL Ramp 2" wide
48 LF SBL 1" wide
48 LF NBL 1" wide
South Approach:
52 LF SBL 4" wide
52 LF NBL 3-1/2" wide
06-10-2006: Same as above
Figure 4: Pressure relief joint details abstracted from the I-35W,2007 Bridge project plans.1
Figure 4 above, showing both existing and proposed pressure-relief joints for the roadway approach panels, provides some evidence that the bridge design and maintenance engineers of Mn DOT were aware of and were trying to provide protection for the I-35W Bridge against the pavement forces generated by the pavement G/P Phenomenon. Notice that the figure indicates that the bridge was somewhat protected in the past by joints that were originally 4 inches (100mm) wide. However, this choice of joint type was unfortunate because in an aged pavement, they could only be effective for two to three years at the most; but in aged interstate pavements, probably not more than two years at most. Actually, it is this author’s opinion that such joints should only be used for emergency repairs and should be replaced within a year or two with joints that are”at least” 4 ft. (1.22m) wide.
The yearly inspection reports (See the tabulation above) contain comments on the condition of these pressure-relief joints. Unfortunately, the most important thing that the inspectors could think to say about the relief joints in the yearly reports for 2000 to 2004 was that they needed sealed. However in the 2005 and 2006 reports, the inspectors provided joint measurements, suggesting that some bridge maintenance engineer in authority was concerned and paying attention. The reports indicated that at the North Approach to the structure, both the north-bound and south-bound pavements contained 4 inch (100 mm) wide joints that had been compressed to 1 inch (25mm), a clear indication that the approach pavements to this structure were generating sufficient pressures to grow the pavementand close not only the pressure–relief joints but some of the movable bridge joints as well. At a 1 in (25 mm) width, such joints should be considered closed and ineffective in preventing the transmission of pressure from the approach pavements into the structure. Presumably, these particular joints had been 1 inch (25 mm) wide for many years indicating that the structure was unprotected and was allowing the approach pavements to generate greater and greater pressures against this structure. These 1 inch (25m mm) wide joint measurements were probably the reason why they were being replaced in the rehabilitation contract. Unfortunately, they were being replaced in kind. If the structure had survived,these joints would have been effective for two years at most. In the opinion of this author, it appears that these relief joints were the reason why this structure was being subjected to great pavement forces periodically for many years. It is also of interest to state that the NTSB investigators appeared to be unaware of these pressure-relief provisions,or if they were aware of them, they appear to have ignored their significance.
The South Approach to the structure was somewhat protected with originally 4 inch (100 mm) wide pressure-relief joints. Based on the 2005 inspection report, these joints had compressed to 3½ inches (87.5 mm) in the north-bound lanes and had remained 4 inch (100 mm) wide in the south-bound lanes, suggesting to the author that these relief joints were newly installed.The author is left to wonder why the relief joints in both the south and north approaches were not replaced at the same time? Presumably, if this had been done, the pavement forces against this structure would have been considerably less and such a change might have saved the structure for a few more years until major modifications of the bridge could have been accomplished. However, with respect to these South Approach relief joints, the joint dimensions were exactly the same in the 2005 and 2006 reports. They had not changed a half inch (12mm) in a year’s time. These reported joint dimensions seem hardy credible to the author, and in addition, the apparent lack of joint movement suggests to the author that the inspectors neglected to measure these joints in 2006 and instead repeated the dimensions that had been measured the year before. If the author had been the supervisor of an inspector who turned in such duplicate dimensions, and since these dimensions were critical for the welfare of the structure, and since it would have been practically impossible for them to remain the same after a year’s time, that inspector would have been fired.
Presumably, the south approach pressure-relief joints would have compressed to 2 inches (50mm) in 2006 and to 1 inch (25mm) in 2007. Consequently, it appears that the relief joints on both approaches to this bridge were effectively closed in 2007, and the bridge would have been unprotected from the G/P Phenomenon at the time of the collapse. Nevertheless, the disparity between the relief-joint dimensions on the north and south approaches suggests that the amount of pressure generated by the G/P Phenomenon would be significantly greater at the north approach to the structure making it more probable that with everything else being equal, the greater pressures at the north approach would suggest that the collapse of the structure initiated at that end of the structure.
Bowed Gussed Plates
In the NTSB Report, mention is made about the discovery of bowed U10 and U10’ gusset plates (See EXHIBIT A). The report stated that “Before the collapse, seven of the eight U10 and U10’ gusset plates had noticeable distortions in the form of bowing along one unsupported edge of each plate. This distortion was visible in photographs taken in 1999 and 2003…The Safety Board found no pre-1998 photographs of the gusset plates.” (2, page 143)Exhibit A1
Nodes (connection points) on main trusses of I-35W bridge, designated by number. The nodes were further referenced by "U" for upper nodes (connecting members at the upper chord) and "L" for lower nodes (connecting members at the lower chord). Location of nodes on the east or west main truss was designed by "E" or "W," respectively. (Source: Adapted from Mn/Dot graphic).
Along with the discovery of these bowed gusset plates, indicating that significant forces had been applied to the structure, the local Minneapolis Star Tribune Newspaper reported on July 15 of the same year that, “The extreme heat is causing pavement to buckle on some highways forcing the Minneapolis Department of Transportation to make emergency repairs – even during rush hour…”(emphasis added by author). Does this appear to be a coincidental timing of bridge distress and pavement distress, and are these events both directly related to the same hot temperature event? Or, to be bolder, does the appearance of major structure distress and pavement distress during the same high temperature events suggest that the pavement G/P Phenomenon in Minneapolis pavements in July of 1998 is blowing up pavements in some areas of the city and pressuring bridges in other areas of the city? Does the G/P Phenomenon care which structure it damages just so long as it can continue to generate growth or pressure, and periodically and continuously so?
It is also instructive to note that the NTSB Report (2, p 143)also stated: “The Safety Board was unable to fully evaluate the possibility that some type of usual loading was generated in the gusset plates as the L9(‘)/U10(‘) diagonal was connected to the U9(‘)/U10(‘) upper chord member but considers it unlikely that such loading could have caused bowing deformation at all four U10(‘) plates”(emphasis added by author). The author would hasten to suggest the possibility that the unusual loading may have been the longitudinal pressures generated by the G/P Phenomenon on both bridge approaches.
When considering the pressures being generated in the approach pavements by the G/P Phenomenon, and recognizing that by 1998 almost of all of the bridge’s movable joints in both approaches were closed or locked, it does not seem difficult to imagine that the pressure from the pavements had finally entered the deck truss, and that in conjunction with the restrained expansion of the bridge, had been creating higher than usual localized stresses in deck-truss members. And these high localized stresses would with certain exceptions have little trouble reaching all eight of the bowed gusset plates.
Distressed Approach Pavements
The distressed condition of the bridge approach pavements was another sign of the presence of accumulating pressures generated by the G/P Phenomenon. For example, in the bridge’s yearly inspection records for the years 2000 through 2004, the following statement was recorded regarding the approach pavement panels. “All four approach panels have transverse cracks.”(1) For pavement engineers, such transverse cracks are a sure sign of high pavement pressures and the cracks themselves could be harbingers of a pavement blowup, and a blowup at that particular crack location.Bill Kallman, a consulting engineer and formerly a long-term staff member of the New York Department of Transportation with extensive experience with troubled pavements and bridges,said that he had examined the pavement approaches of the collapsed I-35W Bridge and stated that “The very warm weather preceding August 1, 2007…. The “frozen” bearings, cracked and misaligned approach span members, early deck repairs, the tilted north pier, distressed condition of the on-grade highway pavement joints leading to the bridge, all point to severe pavement shove as the cause of the collapse.” 6(emphasis added by the author)
Conclusion
The evidence that this author has gathered above should help to prove that the G/P Phenomenon was active in the approaches to the I-35W Bridge. However, such documentation should not be necessary to prove of the existence of the G/P Phenomenon in jointed concrete approach pavements. All bridge design and maintenance engineers should realize that the phenomenon is active in “all” except freshly placed jointed concrete pavement. The events recorded in exhibit B should be proof enough. Consequently, if there is a bridge in distress and it is served by jointed concrete pavement, that bridge distress is probably due to partly or wholly to the growth and pressures generated by this ubiquitous phenomenon. And the I-35 Bridge of Minneapolis, Minnesota is no exception. Actually, that bridge has now become the prime example of the great trouble and grief that can be associated with a bridge where the presence and pressure magnitude of this phenomenon is not respected, or where the means to blunt its potentially destructive behavior are neglected.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
It is this author’s opinion that the Critical Flaw in the collapse of the I-35W Mississippi River Bridge of Minneapolis, Minnesota, was not the undersized and corroded gusset plates, not the weight of construction materials and equipment on the bridge, and not the streams of vehicular traffic that were on the bridge at the time of the collapse. Instead the critical flaw in the collapse of this bridge was the type of pressure-relief joint that was chosen and installed in the pavement approaches to protect this bridge from the pavement Growth/Pressure Phenomenon (4, Appendix 1). It was the choice of this joint type and its apparent neglect that permitted the approach pavements of this bridge to generate unsustainable longitudinal pressures (forces) on the structure. These forces, plus those generated by the restrained expansion of the greatly compressed structure itself, the weight of the structure and all super imposed loads,appear to have coalesced to overstress the undersized and corroded gusset plates and collapse the structure.
Unfortunately, the type of pressure-relief joint that was chosen (a 4 inch (100 mm) wide elastomer-filled joint) a type of joint that, depending upon the age of pavement where and when it was installed, would only be effective in minimizing the amount of pressure transmitted through the relief-joint and into the bridge for two years or possibly three years at most. After such time, the joint would essentially be closed (1 inch (25mm) width) and incapable of minimizing further pressure transmission. Presumably, the engineers who chose this type of joint must have assumed, because of the joint’s “name,” that the joints would be continuously effective in minimizing pressure. How else could one account for the fact that these closed joints were allowed to remain in place year after year, after their effectiveness had ceased.For example, each time the inspectors examined theseclosed joints (1 inch (25mm) width) on a yearly basis, all that they could think to report was that the joints needed sealed. And that comment was allowed to be repeated year after year. Consequently, the I-35WBridge was essentially unprotected for many years from the effects of the G/P Phenomenon, the phenomenon that eventually succeeded in compressing the various spans of the structure until the bridge eventually collapsed. Subsequently, the G/P Phenomenon was free to generate growth instead of pressure.
If instead of choosing the 4 inch (100 mm) wide joints, the Minneapolis bridge engineers had chosen 4 ft. (1.22 m) wide or wider asphalt-filled joints for pressure relief, the I-35W Bridge would probably still be serving traffic. Since such a wide joint would allow the pavement to generate growth instead of pressure, the pavement contraction joints within the first ten to fifteen hundred feet of pavement would gradually widen, presenting the pavement maintenance engineers with the problem of keeping the contraction joints well sealed. This would be a continuous problem but one that could be justified to protect a bridge from periodic damage and from possible collapse. Otherwise, bridge and pavement engineers should consider constructing the first fifteen to twenty hundred feet of pavement adjacent to bridges, continuously reinforced. Such construction should obviate the need for installing pressure-relief joints in bridge approaches. The activity of the G/P Phenomenon would then be confined to the further reaches of the pavement, and the bridges would at last be at last free of this particular scourge.
With respect to the NTSB investigation of the “central spans” of the I-35W Bridge, a statement is made to the effect “…that the change in temperature on the day of the accident did not play a role in initiating of the collapse.” 1, p. 126. The NTSB Report could not be more wrong. As indicated throughout this paper, temperature was a primary determinant by way of the pavement G/P Phenomenon in the generation of longitudinal compressive forces on the I-35W Bridge; temperature was also the primary determinant in the generation of compressive forces by way of the restrained expansion of the structure itself. The NTSB Report identified the truss gusset plates as the structure’s weak link, but it failed to recognize these huge temperature generated forces that helped to break the chain. Consequently, with respect to the NTSB investigation and Report that was published about the collapse of the I-35W Bridge, the author urges that the investigation of the collapse be reopened so that the complete composite structure (4, page 158) can be included in the investigation, and the report suitably supplemented. Bridge design and maintenance engineers depend upon the recommendations that are given in damage or collapse reports. Consequently, it is imperative that such reports reflect the results of a comprehensive investigation. Otherwise these engineers will be misled, their bridges will suffer, and the individuals that use their structures will be exposed to possible injury or death.
Finally, this author, realizing that the pavement Growth/Pressure Phenomenon was at first ignored in the investigation of the collapse of the I-35W Mississippi River Bridge, wonders how many other bridges in the past were either severely damaged or collapsed by this same unfamiliar phenomenon. He “knows” that any bridge constructed adjacent to jointed concrete pavements, unprotected by adequate pressure-relief joints, will be subjected to longitudinal pavement forces, and depending upon the age of the pavement and a number of other factors, these forces can be huge. Consequently, in the investigation of any distressed or collapsed bridges (i. e., the collapsed Canadian River Bridge of near Tulsa, Oklahoma, the damaged I-90 Grand River Bridges of Lake County, Ohio, the partial collapse of the River des Peres Bridge of St. Louis, Missouri, etc.),bridge investigators should always consider the role that the G/P Phenomenon may have played in distressing or collapsing such bridges.
References
- Minnesota Dept. of Transportation, 2007 Bridge Rehabilitation Project Plans, Internet at www.state.mn.us/I35W bridge/index.html, 2007.
- National Transportation Safety Board, Collapse of I-35W Highway Bridge, Minneapolis, Minnesota, August 1, 2007, NTSB, Washington, D. C., 2008, pp.162.
- O’Connell, H., Dexter, R. J., Bergson, P. M., Fatigue Evaluation of the Deck Truss of Bridge 9340, University of Minnesota, Minneapolis, Minnesota, 2001, pp. 28 and 29.
- Burke, M. P. Jr., Integral and Semi-Integral Bridges, Wiley-Blackwell, Chickester, United Kingdom, 2009, pp. 255.
- Minnesota Dept. of Transportation Metro District, Fracture Critical Bridge Inspection, In Depth Report, Bridge #9340 [I-35W Bridge], Minneapolis, Minnesota, 2006, pp. 50.
- Kallman, B., I-35W Mississippi River Bridge #9340 Collapse, Internet, 2007.
Exhibit C5 Crossbeams & Rocker Bearings
At both ends of the two deck trusses, the steel girder approach spans were integrated with steel girder “crossbeams.” These two crossbeams were supported at the ends of the two deck trusses by two cast-steel rockers, one rocker at each end of each deck truss. Three of these rocker bearings and their truss supports penetrated through reinforced web openings or cutouts in the crossbeams while the forth bearing, because of the superelevation of the bridge deck at the south end of the structure, was located below the bottom flange at the east end of the crossbeam. Various inspection records for the bridge contain comments about the problems that the bridge maintenance staff encountered with these two crossbeams and their bearings.“In 1986, the southeast rocker bearing “froze,” resulting in damage to the crossbeam with two cracked web stiffeners. The rocker bearing pin was replaced. This required closing I-35W and jacking up the span. The crossbeam was repaired and the cracks in the web stiffeners were welded, crack ends drilled out, and stiffeners reinforced with angle plates. Installing braces between the crossbeam and beams #2 & 3 also reinforced the connection.” 5,p.13.
“In 1992, a crack was found in a crossbeam stiffener weld above the northeast rocker bearing, which was drilled out.” 5, p. 13.
“In 1997, at the same location [above the northeast rocker bearing] a weld between a vertical & horizontal stiffener was found cracked through entirely. Cracks were also discovered at the end of horizontal stiffeners near the northeast and southwest rocker bearings. Strain gages were installed to analyzed stresses, crack ends were drilled out, and installing bracing between the crossbeam and 2 stringers reinforced the northeast connection.” 5, p. 13.
In 2000, the following observation was given with respect to the reinforced openings in the cross beams: “It appeared that resistance to movement of the bearings was causing significant out-of-plane forces and distortions of the cross girder [crossbeams], leading to cracks forming at the termination of the stiffeners reinforcing the opening, The cross girder was retrofit by drilling holes at the tips of the cracks and adding struts from the reinforcing stiffeners back to the girders to reduce the distortions. This retrofit has been successful so far in preventing further crack propagation.” 3, pp. 28 & 29
In the 2006, the following observations were given: “Panel Point #0 (EndFloorbeam End of West Truss): …[1997] Floorbeam horizontal stiffener is bent directly above the rocker bearing. [1998/1999] Floorbeam repainted, side facing finger joint has section loss, pitting. [2004] Truss, top chord exterior connection plate has 1/8 “ deep section loss with pitting. SW rocker bearing has no movement.”5, p. 30
In all of the above, no speculations were proffered as to what could have caused the bridge joints to close, or periodically caused the distress and distortion of the transition girders (crossbeams) between the approach spans and the deck trusses. Presumably, they must have thought that this distress was due to corrosion-locked bearings and the thermal response of the steel deck trusses and adjacent girder spans to thermal changes. This author wonders if any of them conceived of the possibility that the distress that the bridge was suffering was due to the fact that the bridge was periodically being compressed primarily by the longitudinal forces being generated by the G/P Phenomenon of the bridge approach pavements?
Acknowledgment
The development of this paper would not have been possible without the I-35W Bridge documentation that was made available by the Minnesota Department of Transportation. Hopefully, that magnanimous decision and this paper may help to avoid the collapse of similar structures by the ubiquitous presence and destructive potential of the G/P Phenomenon.Martin P. Burke Jr., PE
Sir Norman Angell
“If the world has nearly destroyed itself, it is not from lack of knowledge… but is due to the fact that the mass of men have not applied to public policy knowledge they already possess, which is indeed of almost universal possession, deducible from the facts of everyday life. If this is true – and it seems inescapable – then no education which consists mainly in the dissemination of “knowledge” can save us. If men can disregard in their policies the facts they already know, they can just as easily disregard new facts which they do not at present know. What is needed is the development in men of that particular type of skill which will enable them to make social use of knowledge already in their possession; enable them to apply simple, sometimes self-evident, truths to the guidance of their common life.”
NBMCW September 2011
| < Prev | Next > |
|---|