Building Construction
The collapse of the Interstate 35W bridge over the Mississippi River had done major damages in Minneapolis, Minnesota. Many assumptions and speculations about the causes of the collapse of the bridge system had appeared in the public. The public was seemingly confused about the real cause of the incidents and it is their right to be informed about the state of the investigation. The closest and very logical of the causes indicated in some of the investigations are stress or fatigue failure and lack of redundancy.
Environment, Design, and Description of the I-35W bridge The I-35W bridge supports a total of eight lanes (four lanes on each direction). The average daily traffic (ADT) is given as 15,000 in each direction, with ten percent trucks. Constructed in 1967, the 581 meter long bridge has 14 ps. The main p is consist of a steel deck truss. The south approach ps are steel multi-beam. The north approach ps include both steel multieam and concrete slab p. There are two steel deck trusses. Builtup plates mostly composed the truss members.
Rolled I-beams comprised the diagonal and vertical members. The truss members undergo poor welding details with the connections as mainly riveted and bolted. According to recent evaluation and inspection before the collapse of the bridge, corrosion at the floorbeam exists and rust are forming between connection plates. The two main trussses have an 11. 6-meter cantilever at the north and south ends. Twenty-seven floor trusses spaced at 11. 6 meters are also present. These floor trusses were framed into the vertical members of the main truss.
The floor trusses consist of WF-shape members and have a 4. 97- meter cantilever at each end. The design specifications used in the bridge was the 1961 American Association of State Highway Officials (AASHTO) Specifications. During that time, most of the design uses unconservative fatigue design provisions. According to the fatigue evaluation report provided by the University of Minnesota’s Center for Transportation Studies in 2001, the approach ps had exhibited several fatigue problems promarily due to the distortion of the girders.
The bridge truss and the floor truss system also exhibited poor fatigue details. Lack of redundancy in the main truss system was also present in the design. It is stated in the evaluation report of the University of Minnesota that cracking due to fatigue cause by a future increase in loading will first appear on the floor truss. According to them these future cracks is detectable since the floor truss are easy to inspect. In the incidence that cracks are not detected, the bridge could still hold the bridge system without the entire collapse of the system.
In the report, the failure of the two main trusses of the bridge will definitely take much effect on the bridge system. Fatigue Resistance The Standard Specification and the Load and Resistance Factor Design provided by the American Association of State Highway Officials (AASHTO) contain similar provisions for the fatigue design of welded details on steel ridges. These details are designed ased on the nominal stress which can be calculated using standard design equations and does not include the effects of welds and attachments.
Since fatigue is usually present during sevice load application, the design parameters are only applied during service load conditions. Cracks due to fatigue have insignificant effect on the structures in compression but have tremendous effect on structures that experience tension. With this idea, the assessment on the cracks that propagate on such a bridge as the I-35W should only be consider to elements in tension. Structural Redundancy In all the design criteria of any structural system, loads existed in variety of paths should be significantly considered.
The strength and reliability of the system can be ensured by the existence of redundant paths or elements. Without the existence of this redundant system of elements, the failure of the entire system is much possible. Past survey of the Committee on Redundancy of Flexural System on steel highway ad railroad bridges. The report summarized that a total of 96 structures were suffering some distress. It was also taken into account that most of the failures were related to connections that were mainly welded.
The report had also collected data that indicates that few steel bridges collapse if redundancy is present. Bridge systems with no redundancy were reported to have large number. In another research conducted by Ressler and Daniels, they found that the number of fatigue sensitive details present in the structure significantly affected the bridges with no redundant elements. Theoretical and Actual Bridge Response Many studies have shown that the simplified calculations used to predict the stresses provide a much higher value compared to the actual service stresses.
Though the design calculations and load models provide appropriate results, it has great uncertainty in the maximum life of a bridge system. However, it is still beneficial to have an accurate estimate of the typical everyday stress ranges. In a large bridge, 20 Mpa is the typical value of the service live-load stress ranges. The stress ranges are typically governed by dead loads and strength design specifications. This is the reason why the stress ranges are small. Since the strength design must account for a single case loading scenario over the life of the bridge, conservative load models are used.
In addition to load conservative models, the assumptions provided in the analysis of the design can also be the cause of the large difference of the predicted stress and actual stress. A great example of the effect of the assumptions is the case of US Highway 69 in Oklahoma. Fatigue damage was said to be present upon the welding that had been used in the widening of the bridge. The design computations of the bridge illustrated that the allowable stress ranges could be exceeded at over 100 locations on the bridge.
However, when the bridge was inspected, it appeared that the measure stress ranges was only 27 percent of the allowable stress ranges. This only shows the great effect of the assumptions used in the design of a certain structural system. Moreover, another study that indicates fatigue failure to be caused by the considerable amount of corrosion takes into account. This is the case of the Bridge 4654 in Minnesota where measured stress ranges ranged from 65 to 85 percent of the calculated analysis.
These differences are to point out to the fact that analytical methods provide assumptions that neglect ways in which the structure resists loads. For example, in the study conducted y Brudette et al., more than 50 years of bridge test data were collected and examined to determine specific load-resisiting mechanisms that are ignored in the design of the system. The study concluded that lower stress ranges in a structure can be due to unintended composite action, contribution from non-structural elements, unintended partial fixity at abutments and direct transfer of load through the slab to the supports.
In another study of the Ministry of Transportation of Ontario, they conducted a program of bridge testing that included more than 225 bridges over a period of 15 years. The study noted that much of the bridges can sustain much larger loads than their estimated capacities. Observations were also made regarding the behavior of the steel truss bridge. The observations are as follows:
- the stringer of the floor system share a large tensile force thus reducing the strains experienced by the chord in contact with the floor system and
- Composite action in non-composite system was shown to exist
