Sabo Bridge Report

MARTIN OLAV SABO PEDESTRIAN BRIDGE Cable Diaphragm Plate Fracture Investigation
Minneapolis Bridge #6925, State Bridge #27B35 Minneapolis, Minnesota
Final Report June 28, 2012 WJE No. 2012.0901
Prepared for: Ms. Debra Brisk Assistant County Administrator Public Works Hennepin County A-2303 Government Center 300 South Sixth Street Minneapolis, Minnesota 55487
Ms. Heidi Hamilton Deputy Director, Public Works City of Minneapolis 350 S. 5th Street, Rm 127 Minneapolis, Minnesota 55415
Prepared by: Wiss, Janney, Elstner Associates, Inc. 330 Pfingsten Road Northbrook, Illinois 60062 847.272.7400 tel | 847.291.4813 fax
EXECUTIVE SUMMARY
Cable Diaphragm Plate 9 of the Martin Olav Sabo Pedestrian Bridge over Hiawatha Avenue fractured on the evening of Sunday February 19, 2012. The fracture was oriented so as to allow the portion of the plate connected to the two stay cables to fall to the ground, with the cables still attached. Upon review, fractures and cracks were also discovered in Cable Diaphragm Plate 8. The Diaphragm Plate 8 damage left each of the associated stay cables connected to the pylon. Winds were reported to be out of the southsoutheast at 6 mph to 13 mph for several hours prior to the incident. The temperature was approximately 23 degrees Fahrenheit. In response to this incident, Wiss, Janney, Elstner Associates, Inc. (WJE) was retained by Hennepin County and the City of Minneapolis to perform the following tasks: assist on site during stabilization of the structure; inspect the remaining in-service diaphragm plates using nondestructive test methods; determine the cause of the cable diaphragm plate failures; perform an independent review of the bridge design for Spans 1 through 3; and develop retrofit options to address any noted deficiencies. During the investigation, nondestructive testing of the sixteen remaining in-service diaphragm plates revealed welding defects or cracks in two plates. A welding defect was removed from one plate, and a temporary redundancy fixture was installed over the other to provide load path redundancy. The other fourteen cable diaphragm plates were found to not contain cracks. Samples from Cable Diaphragm Plates 8 and 9 were removed from the structure and analyzed to determine the cause of cracking and fracture. In addition, materials tests for physical, chemical, and toughness properties were performed. The plate samples were found contain evidence of extensive fatigue cracking prior to fracture. The steel plate material was found to be consistent with the requirements of the project specifications. It should also be noted that the cable diaphragm plate assembly geometry contains welded details that provide poor fatigue resistance. An instrumentation program was developed and implemented to determine stresses and other structural responses to wind-induced cable vibrations. Wind speed and direction was recorded to correlate wind effects to this data. Low to moderate wind speeds with a direction approximately transverse to the structure alignment result in cable vibrations that induce damaging stress cycles at the fatigue sensitive details of the cable diaphragm plates. These vibrations occur at high frequency, resulting in a large accumulation of stress cycles over short periods of wind-induced vibration. An independent review of the original bridge design was performed. The results from an independent finite element model revealed that the primary bridge elements were adequately designed for the reported loads. The effects of wind-induced cable vibrations were not included in the original design calculations package. However, a monitoring program was implemented following bridge construction to determine if stay cable vibrations were problematic. The results of the monitoring work indicated that stay cable vibrations were not a structural concern. Failure of Cable Diaphragm Plates 8 and 9 occurred due to fatigue cracking that initiated at weld toes with low fatigue resistance. These welds connected the 3/4 inch thick diaphragm plates to the pylon flanges, and to the reinforcing ring plates used at the cable connection points. The fatigue cracks extended to critical size in both plates due to wind-induced cable vibrations that resulted in unstable fractures. Retrofit options to upgrade the cable diaphragm plate connections are presented.
ES-i
TABLE OF CONTENTS
Executive Summary ................................................................................................................................. ES-i 1.0 Introduction ............................................................................................................................................. 1 2.0 Background ............................................................................................................................................. 1 2.1 Structure Description ........................................................................................................................ 1 2.2 Parties Involved in the Design and Construction ............................................................................. 2 2.3 Summary of Diaphragm Plate Failures and Structure Stabilization ................................................. 2 3.0 Scope of Work ........................................................................................................................................ 3 4.0 Nondestructive Testing ........................................................................................................................... 3 4.1 Nondestructive Test Methods ........................................................................................................... 3 4.2 Nondestructive Test Results ............................................................................................................. 4 5.0 Physical Measurements ........................................................................................................................... 5 5.1 Cable Diaphragm Plate Alignment ................................................................................................... 5 5.2 Pylon Flange Plastic Rotation at Cable Diaphragm Plate 8 ............................................................. 5 6.0 Fracture Investigation ............................................................................................................................. 6 6.1 Summary of Dr. Fishers Assessment .............................................................................................. 6 6.2 Summary of Dr. Fishers Findings ................................................................................................... 6 7.0 Instrumentation and Monitoring Program............................................................................................... 7 7.1 Instruments and Installation Locations ............................................................................................. 7 7.2 Data Acquisition System .................................................................................................................. 8 7.3 Pluck Tests ....................................................................................................................................... 8 7.4 Monitoring Period ............................................................................................................................ 9 7.5 Effective Stress Range Calculation ................................................................................................ 10 8.0 Peer Review of Original Bridge Design ............................................................................................... 11 9.0 Review of Wind-Induced Stay Cable Vibration Phenomena ................................................................ 13 10.0 Structural Analysis of Diaphragm Plates ............................................................................................ 14 10.1 Stresses Due to Cable Vibrations ................................................................................................. 14 10.2 Stresses Due to Plate Misalignment ............................................................................................. 14 11.0 Discussion ........................................................................................................................................... 15 12.0 Conclusions and Recommendations ................................................................................................... 17 12.1 Conclusions and Items for Further Study ..................................................................................... 17 12.2 Cable Diaphragm Plate Retrofit Options ...................................................................................... 17 Appendix A - John W. Fisher Report Appendix B - Instrumentation Plan Drawings Appendix C - Effective Stress Range Histograms Appendix D - Preliminary Retrofit Drawings
TOC-i
MARTIN OLAV SABO PEDESTRIAN BRIDGE Cable Diaphragm Plate Fracture Investigation
Minneapolis, Minnesota
1.0 INTRODUCTION
The main span of the Martin Olav Sabo Pedestrian Bridge is a cable-stayed structure consisting of a reinforced concrete deck supported by eighteen stay cables arranged in nine pairs and extending from the main pylon to deck anchorages placed along the 215 foot span. During the evening of Sunday, February 19, 2012, the diaphragm plate connecting Cables 9L and 9R fractured, disconnecting the cables from the pylon. Both cable clevises remained attached to the portion of the plate that separated from the pylon, and therefore, the cable ends descended together, landing in a grassy area east of Hiawatha Avenue. Assessment of the pylon cable connection was made by City forces immediately following discovery that Cables 9L and 9R had detached from the pylon. The City inspectors discovered that the diaphragm plate connecting Cables 8L and 8R was also cracked and fractured into two segments. However, both segments remained attached to the pylon. Following these events, Hennepin County and the City of Minneapolis jointly engaged Wiss, Janney, Elstner Associates, Inc. (WJE) to investigate the cause of the failure and provide engineering services needed to return the structure to service. This report describes the investigation into the cause of the failure and summarizes the findings of an independent peer review of the structural design. Also, the results of WJEs inspection of the remaining diaphragm plates are presented along with preliminary options for retrofitting the cable-to-pylon connections.
2.0 BACKGROUND 2.1 Structure Description

The Martin Olav Sabo Pedestrian Bridge consists of a total of ten spans and extends over a total length of 686 feet 8 inches. An overall view of the bridge, looking northeast, is shown in Figure 1. Located southeast of downtown Minneapolis, the structure carries the Midtown Greenway and Hiawatha Bike Trail over Minnesota Trunk Highway 55 (Hiawatha Avenue) and two sets of light rail train tracks. The main span of the bridge (Span 3), oriented in an approximate east-west direction, is 215 foot long and is a cable-stayed structure. The remaining nine spans consist of conventional reinforced concrete deck and substructures that vary in length from 51 to 55 feet. The bridge deck for the cable-stayed unit (Spans 1 through 3) was constructed using conventionally reinforced concrete construction, varying in thickness from 8 inches to 2 feet. A 2-foot thick curb section is also present along each edge of the concrete deck. A typical deck section is 18 feet wide, between curbs, and flares out to approximately 30 feet wide over Pier 2. An overall plan drawing of the bridge deck for Spans 1 to 3 is shown in Figure 2. The reinforced concrete deck in Span 3 is supported by nine pairs stay cables (fore stays) that are attached to deck anchorages placed along the 215 foot span. Each stay cable is manufactured using multiple wires to achieve a diameter of 1 1/2 or 2 inches. At their upper end, each stay cable is attached to cable diaphragm plates at the east side of the steel pylon set on Pier 2. An additional nine pairs of stay cables (back stays) are attached to the west side of the steel pylon, using similar cable diaphragm plates. These back stays are attached to ground anchorages adjacent to Span 1, and are used to balance the fore stay
Martin Olav Sabo Pedestrian Bridge Cable Diaphragm Plate Fracture Investigation June 28, 2012 Page 2
loads applied to the opposite side of the pylon. Typical cable attachments at the concrete deck, pylon, and earth anchorage are shown in Figures 3, 4, and 5, respectively. The cable diaphragm plate anchorages are identified as Cable Diaphragm Plates 1 through 9 (bottom to top) on the east side of the pylon, and Cable Diaphragm Plates 10 through 18 (bottom to top) on the west side of the pylon, as shown in Figure 6. Each stay cable is identified by the cable diaphragm plate to which it attaches, and its position within the structure; the left side (L) of the structure is taken as north and the right side (R) of the structure is taken as south. The pylon was constructed at a 20 degree angle from vertical and was fabricated as a built-up section using Grade HPS 70W steel for the flange and web plates. Tab plates were fillet welded to the edge of each pylon flange along the majority of its height to provide a connection for pylon screening panels. Each cable diaphragm plate attached to the pylon was cut from 3/4 inch thick steel plate conforming to ASTM A572 Grade 50, and was welded to the inside surfaces of the 1 1/2 inch thick pylon flanges using 3/8 inch fillet welds around the entire perimeter of each interface. Holes for the clevis pins are reinforced by ring plates on each side of the diaphragm plate. Each ring plate was fillet welded (1/4 inch fillet) to the diaphragm plate all around its outside edge. A typical cable diaphragm plate is shown in Figure 7. The concrete deck in Span 3 acts as a compression strut to resist the horizontal compression forces between the fore stays and the pylon. Compression struts were also constructed between Pier 2 and the earth anchorage blocks to resist the horizontal compression forces generated between Pier 2 and the backstay cables.
2.2 Parties Involved in the Design and Construction

Design and construction of the Martin Olav Sabo Pedestrian Bridge was commissioned by Hennepin County and the City of Minneapolis. URS Corporation (URS) designed the structure, provided review of shop drawings, and was retained by Hennepin County to provide contract and construction administration duties. Ames Construction, Inc. (Ames) was awarded the construction project and acted as General Contractor. Fabrication of the pylon and other steel components was subcontracted to PDM Bridge, LLC. Shop inspection of the steel portions of the structure was performed by the Minnesota Department of Transportation (MnDOT). Since construction, the City of Minneapolis has carried out maintenance and inspection operations.
2.3 Summary of Diaphragm Plate Failures and Structure Stabilization

The diaphragm plate connecting Cables 9L and 9R to the pylon fractured on the evening of Sunday, February 19, 2012. During their immediate response to the incident, City forces discovered that the diaphragm plate connecting Cables 8L and 8R to the pylon was also cracked, and had fractured into two segments. In addition, the left flange of the pylon had rotated out of alignment in the vicinity of the connection for Cables 8L and 8R. This rotation caused the force in Cables 8L and 8R to be reduced somewhat. Winds at the nearest major airport (MSP) were reported to be out of the south-southeast at 6 mph to 13 mph for several hours prior to the incident. The temperature was approximately 23 degrees F. No personal injuries were reported as a result of this incident. The immediate response to stabilize the structure included erection of shoring towers underneath the reinforced concrete deck of Span 3 to support the deck in the area where Cables 8 and 9 were connected. Following installation of adequate shoring, the remaining tension in Cables 8L and 8R was released using
a hydraulic jacking system and winch system. The cables were then disconnected from the fractured diaphragm plate at the pylon. Following these operations, the pylon screening panels in the vicinity of the cable diaphragm plates were removed to allow access for detailed nondestructive testing of the remaining in-service diaphragm plates. Defects or cracks were located in Cable Diaphragm Plates 7 and 5. The defect in Cable Diaphragm Plate 7 was removed. However, two cracks observed in Cable Diaphragm Plate 5 prompted the design and installation of a temporary redundancy fixture to provide load path redundancy for the cracked diaphragm plate.
3.0 SCOPE OF WORK

WJE was retained by Hennepin County and the City of Minneapolis to perform the following tasks: 1. Assist on site during stabilization of the structure. 2. Inspect the remaining in-service diaphragm plates using nondestructive test methods. 3. Determine the cause of the cable diaphragm plate failures. 4. Perform an independent review of the bridge design for Spans 1 through 3. 5. Develop retrofit options to address any noted deficiencies. This report describes the investigation into the cause of the failure and summarizes the findings of the peer review of the structural design. Also, the results of WJEs inspection of the remaining diaphragm plates are presented along with preliminary options for retrofitting the cable-to-pylon connections. The assignment of fault for any party responsible for the design, fabrication, erection, or maintenance of the structure is beyond this scope of work.
4.0 NONDESTRUCTIVE TESTING

Various nondestructive test methods were used to assess the sixteen remaining in-service cable diaphragm plates to determine if these plates contained unknown welding defects or cracks. These examinations took place between February 23 and 25, 2012.
4.1 Nondestructive Test Methods

Visual, magnetic particle, and ultrasonic test methods were used for condition evaluation of all cable diaphragm plates, except for Cable Diaphragm Plates 8 and 9 where fractures had already occurred. Visual examinations and magnetic particle testing (MT) were aided by flashlights for illumination of dark areas and a mirror for observation in areas with restricted viewing access. MT was limited due to constraints of access. The upper sides of all plates were sufficiently accessible for MT examination, but portions of lower sides of many plates were not accessible. This condition was most prevalent at the west side of the pylon, where lower plate surfaces were inaccessible except for visual examinations. MT was performed in accordance with the provisions of ASTM E-709. An alternating current yoke was used for the induction of magnetic fields, in conjunction with application of an inspection medium of red dry particles for examinations of welded joints and diaphragm plate regions. Magnetic fields were oriented to the fullest possible extent for detection of cracks or defects, although ideal positioning of the yoke was not always possible due to the constraints of the assembly. Figure 8 shows a typical application
for evaluation by MT of a fillet welded joint. All examinations were performed with paint coatings in place, except for local removal where indications were detected. Ultrasonic testing (UT) was performed using conventional pulse echo equipment. Calibrations of the test apparatus followed the provisions of AWS D1.1 for linearity and sensitivity. A scanning plan was implemented for detection of potential crack initiations using multiple probe angles and orientations. Ultrasonic scans were directed toward the toes of all fillet welded joints, as well as toward respective regions of the diaphragm plates along the fracture surfaces exhibited by Cable Diaphragm Plates 8 and 9. In general, all scans were introduced from the top surface of the subject diaphragm plate. Figure 9 shows UT in progress, scanning the diaphragm plate toward the fillet welded joint connecting a representative ring plate. A summary of primary equipment used in nondestructive examinations is provided in Table 1.
4.2 Nondestructive Test Results

Examinations of the cable diaphragm plates emphasized detection of crack initiations. Use of multiple nondestructive methods facilitated complete coverage of potential crack initiation sites by at least one method, and in many instances confirmed the assessment based on the complimentary methods. The magnetic particle method is sensitive to surface cracks and may enhance detection of cracks that are too fine for visual observation. In addition, the penetrating nature of ultrasonic waves permitted volumetric examination through the thickness of the subject plates. The fabricated edges of the diaphragm plates resulted in a variation of gaps at the connection, with a wider gap at the lower side of the plates. In general, the fillet welded joints appeared to be of uniform size and profile, although no actual sizing of the fillets was carried out using gages. UT indicated additional, but inconsistent fusion below the theoretical root in portions of some joints. In many instances, a fillet welded joint for attachment of the pylon screen tab plate intersects with the termination of the diaphragm plate fillet weld completion around the pylon flange tip. Discontinuities were detected in two of the in-service cable diaphragm plates. In Cable Diaphragm Plate 7, a crack was detected at the right flange tip by MT, as shown in Figure 10. This defect was removed by grinding less than 1/4 inch of material with a die grinder. This discontinuity appeared to be the result of non-fusion within the welded joint, but was located in a prime location for crack initiation. In addition, cracks originating in Cable Diaphragm Plate 5 near the toe of the diaphragm plate-to-pylon flange welds were detected. These cracks extended along the bottom plate surface and exhibited lengths of 3/4 inches and 1 1/2 inches at the right and left pylon flange tips, respectively. These defects are shown in Figures 11 and 12, respectively. Removal of the painted coating revealed a slight length increase in comparison with the initial observations. UT corroborated these findings and confirmed that no further extension was present. Embedded discontinuities and cracks were generally discernible with UT, except for very small indications at the pylon flange tips. These could not be distinguished from the component geometry using UT methods. In these flange tip locations, MT provided an adequate evaluation. With the volumetric penetration of UT, and MT aided by visual inspection of accessible areas, virtually all regions of potential crack initiation were examined.
5.0 PHYSICAL MEASUREMENTS

Measurements were taken to define the approximate alignment for each cable diaphragm plate and to define the extent of plastic pylon flange deformations of the left flange plate at its connection to Cable Diaphragm Plate 8.
5.1 Cable Diaphragm Plate Alignment

WJE carried out three independent methods of measurement to define the cable diaphragm plate alignments with respect to the pylon web plate. The first method included use of a tape measure (direct measurements) to determine the distance from the face of the pylon web to the top and bottom corners of the subject cable diaphragm plate, or use of a protractor type measurement device. The direct measurements, if used, were used to create a scaled drawing of the plate alignment in AutoCAD from which the alignment angle was reported. As an alternate measurement method, a Bosch Model DAF220 digital miter finder was used to directly report the angle between the top of each cable diaphragm plate and the pylon web plate. This instrument has a digital output that reports the angle to the nearest 0.1 degrees. Lastly, a Stabila Type 96 digital level was used to measure the absolute attitude of each cable diaphragm plate and the pylon web adjacent to each plate. The plate alignment angle was then calculated using these values. The Stabila instrument has a digital output that reports the angle of the instrument to the nearest 0.1 degrees. The results of the three forms of cable diaphragm plate alignment measurement are included in Table 2. For reference, the cable diaphragm plate alignments included in the design and shop drawings for each plate and deviations from these values are included in the table. It should be noted that the accuracy of the reported measurements is anticipated to be within 0.5 degrees due to geometric constraints of the structure, paint thickness or debris in the vicinity of the measurement locations, and limitations of the measurement devices. Design alignment values were given as the angle between the center of the pylon web and the center of the cable diaphragm plate to a precise value including degrees, minutes, and seconds. For example, the intended alignment of Cable Diaphragm Plate 9 was given as 492134 on both the design and shop drawing sets. All but three of the eighteen cable diaphragm plates were found to be aligned within about 0.5 degrees of the design alignment values. Cable Diaphragm Plates 6, 9, and 10 were determined to out of alignment by about 2.87 degrees, 4.46 degrees, and -4.56 degrees, respectively. (Positive values of misalignment represent rotation up.) Review of the allowable fabrication tolerances that make up the cable diaphragm plate alignment indicates that an acceptable plate alignment would be within about 1 degree of that specified in the design drawings.
5.2 Pylon Flange Plastic Rotation at Cable Diaphragm Plate 8

The distance between the pylon flange tips was measured at several locations in the vicinity of Cable Diaphragm Plate 8 following fracture. It was determined that the outstanding edge of the 1 1/2 inch thick left pylon flange had been deformed out if its original alignment by a maximum of about 7/8 inches. This deformation took place over an approximate distance of 19 inches along the flange in the vicinity of the cable diaphragm plate connection.
6.0 FRACTURE INVESTIGATION

An in-depth assessment was carried out for the fracture surfaces of Cable Diaphragm Plates 8 and 9. The plate segment from Cable Diaphragm Plate 9 that separated from the pylon was sent to WJEs subconsultant, John W. Fisher and Associates, Inc. Dr. Fisher indicated upon initial review of the plate segment that the fracture surfaces had been damaged during the fall to the ground. In addition, the fracture path for each plate was different. A segment of Cable Diaphragm Plate 9 including the ring plates and cable clevises separated from the pylon, as shown in Figures 13 and 14. The Cable Diaphragm Plate 8 fracture path included the right side plate connection weld to the pylon flange tip and the back of the right ring plates. This fracture then extended up through the center section of the plate, as shown in Figure 15. This fracture path allowed Cables 8 to remain connected to the pylon. However, the left pylon flange was plastically deformed out of plane as shown in Figure 16. WJE developed and implemented a fracture sample removal plan to recover the remaining portion of Cable Diaphragm Plate 9 and the two attached segments of Cable Diaphragm Plate 8 from the pylon. This removal plan included coring the pylon flange tips to completely remove and preserve the end of the weld terminations connecting the cable diaphragm plates to the pylon flanges. The removed segments of Cable Diaphragm Plates 8 and 9 are shown shortly after removal in Figures 17 and 18, respectively. All fracture samples were sent to Dr. Fisher for fractographic assessment. A summary of Dr. Fishers evaluation and findings is included below. The complete report document, titled Report on the cracking that developed in the Cable Stay diaphragm plates of the Martin Olav Sabo Pedestrian Bridge, authored by John W. Fisher and Eric J. Kaufmann, and dated May 24, 2012 (John W. Fisher Report), is included as Appendix A.
6.1 Summary of Dr. Fishers Assessment

Dr. Fisher performed a visual examination of each fracture surface; a fractographic evaluation of selected areas of the fracture surfaces from each of Cable Diaphragm Plates 8 and 9; materials tests for physical, chemical, and toughness properties of the cable diaphragm plate material; a fractographic assessment of selected fracture surfaces using a scanning electron microscope (SEM); a review of the fatigue resistance of the welded details; and a metallographic assessment of a representative set of ring plate-to-diaphragm plate welds.
6.2 Summary of Dr. Fishers Findings

A summary of the findings from the John W. Fisher Report are provided below. The welded details used on the cable diaphragm plates provide low fatigue resistance, consistent with Category E or E. Prior to its fracture, Cable Diaphragm Plate 8 sustained fatigue cracking with approximately 7 1/2 inches of stable growth occurring at the right side of the plate between the lower end of the pylon flange-to-diaphragm plate weld and the right ring plate-to-diaphragm plate weld, initiating in both locations. These two fatigue cracks coalesced and then extended in an unstable manner to intersect the top edge of the plate, dividing the plate into two pieces. A similar but shorter fatigue crack existed at the left side of the plate, initiating at the lower end of the pylon flange-to-diaphragm plate weld, but there were no signs of significant unstable crack extension. Cable Diaphragm Plate 9 sustained extensive fatigue cracking before it fractured. A stable crack extension of approximately 11 1/4 inches occurred at the right side of the plate between the pylon flange-to-diaphragm plate weld and the right ring plate-to-diaphragm plate weld, initiating in both
locations. A separate stable crack extension of about 9 inches was found at the left ring plate weld. These fatigue cracks in Cable Diaphragm Plate 9 extended in an unstable manner, separating the complete front section of the diaphragm plate, including the ring plates, from the remainder of the plate. The noted fatigue cracks initiated from the top surface of both cable diaphragm plates, indicating that the applied cable forces and plate geometry created some degree of initial bending in the plates. The noted fatigue cracks initiated and extended due to cyclic stresses caused by wind-induced cable vibrations. Welding defects including embedded slag and porosity existed in the portion of the fillet welded connection to the pylon flanges that wrapped around the lower end of the subject cable diaphragm plates. However, these welding defects did not influence the weld toe cracks that formed in the diaphragm plates. The steel plate material conformed to the design specifications and provided good toughness.
7.0 INSTRUMENTATION AND MONITORING PROGRAM

An instrumentation and monitoring program was developed and implemented to measure stresses, cable vibrations, and other structural responses at selected locations. The instrumentation measured the structural response to pluck testing of instrumented cables as well as ambient conditions over a monitoring period of about 18 days.
7.1 Instruments and Installation Locations

Accelerometers, strain gages, displacement transducers, thermocouples, and wind anemometers were used to determine structural responses to environmental and induced loadings. All sensors were installed during the week of April 16, 2012. A description of each type of sensor and its installation location is included below. Instrumentation plan drawing Sheets Si-1 through Si-5 are included as Appendix B. Strain Gages. A total of 24 bondable strain gages, Type CEA-06-125UN-350 manufactured by Vishay Micro-Measurements, Inc., were used for this installation. Strain gages were installed on Cable Diaphragm Plates 7 and 6 back-to-back (top side and bottom side) at the locations shown on Sheets Si-2 and Si-3 in Appendix B, for a total of 12 strain gages on each plate. A typical strain gage is shown in Figure 19 prior to application of environmental protection for water resistance. Accelerometers. Eight uniaxial accelerometers, Model 353B33 manufactured by PCB Piezotronics, Inc., were mounted on Cables 6L, 6R, 7L, and 7R. Two accelerometers were mounted on each cable using a mounting cube fastened to an aluminum ring used to clamp the stay cable, as shown in Figure 20. These accelerometers were positioned to measure lateral and vertical cable vibrations. The mounting locations along the cables were determined on-site during a period of cable vibrations. The accelerometers mounted on Cables 7 and Cables 6 were positioned approximately 9 feet and 7 1/2 feet from the pylon cable anchor point (measured from the center of the clevis pin), respectively. Displacement Transducers. A total of two linear variable displacement transducers (LVDTs), Model 242-000, 0.25 inch, manufactured by Trans-Tek, were used to monitor the displacement of the outstanding edge of Cable Diaphragm Plates 7 and 6. Special fixtures were fabricated for each installation location to allow the displacement transducer armature to be mounted to the bottom surface of the subject diaphragm plate near the tip of the plate as close as possible to the plate centerline, as shown in Figure 21.
Out-of-plane (up and down) displacements were measured at each location. Installation details are given on Sheets Si-4 and Si-5 in Appendix B. Wind Speed and Direction. Two wind anemometers, Model 5103 manufactured by R. M. Young Co., were used to measure wind speed and direction at selected locations on the structure. One wind monitor was mounted 5 feet above the top of the pylon using a fixture attached to the pylon cap plate. The other wind monitor was mounted 10 feet above the top surface of the left curb between Cables 3L and 4L, as shown in Figure 22. For reference, project north (0 degrees) was taken as perpendicular left of the longitudinal centerline of Span 3. Increasing degrees were taken clockwise looking down like a compass. Temperature. A total of three thermocouples were used to monitor the temperature of the concrete at the pylon base connection, the steel pylon web between Cable Diaphragm Plates 7 and 6, and the air temperature at the data acquisition enclosure. Type T thermocouple wire manufactured Omega was used for this installation.
7.2 Data Acquisition System

A CR9000X data logger manufactured by Campbell Scientific, Inc., was used to monitor responses from all instruments. This high speed data acquisition system provides precision measurements and is rugged for field use applications. The system was outfitted with eight I/O modules to accommodate the number of sensors used for this installation: one CR9050 Analog Input w/ RTD, one CR9060 Excitation Module, and six CR9052DC Filtered Analog Input Modules with DC excitation. This data logger was supplemented using a Model 483B07 ICP Signal Conditioner manufactured by PCB Piezotronics, Inc. to excite and pass signals from the accelerometers. The data acquisition equipment is shown in Figure 23. The data acquisition system was installed in a steel tool box enclosure mounted on the bridge deck near the base of the steel pylon. All instrumentation wire was routed from the instruments to the data acquisition system and was attached to the structure using a combination of cable ties and beam clamps.
7.3 Pluck Tests

A total of four pluck tests were performed to determine structural responses to known loading. One pluck test was performed on each of Cables 7L, 7R, 6L, and 6R. For each pluck test, a 250 pound force was applied to the subject cable approximately perpendicular to the cable alignment in the vertical plane. A hydraulic tension ram was used in line with a load cell to apply the force into the cable. Following load application, a quick-release connector was used to instantaneously release the applied force, and the structural response was recorded. Cable vibrations due to ambient winds were minimal at the time the pluck tests were performed. The loading equipment for the four pluck tests was attached to the subject cables as follows:
Cable 6L 6R 7L 7R Total Cable Length (feet) 180.588 180.664 204.712 204.782 Approximate Pluck Test Location (feet from pylon anchor point) 58 38 62 42
Recorded stress ranges for the cable diaphragm plates in response to the pluck tests are included in Table 3. Note that the stress range values reported in Table 3 were taken from the first major stress cycle following load release. In general, the location of the highest stress range for each respective cable diaphragm plate and pluck test was similar. The highest stress ranges were recorded at the strain gages positioned perpendicular to the ring plate welds at the midpoint between the cable alignment and the weld wrap around where the plate is attached to the pylon flange on the same side of the plate where the loaded cable connects. These stress ranges were 3.1 ksi to 3.4 ksi for Cable Diaphragm Plate 7 and 3.3 ksi to 3.8 ksi for Cable Diaphragm Plate 6. Note that the digital strain measurements collected from each strain gage were converted to stress using the equation , where is a strain range determined from the response of the gage and E is the modulus of elasticity for steel, which was taken as 30,000,000 psi. All stress readings were stored on the data acquisition equipment in units of pounds per square inch (psi). Representative stress data plots for these tests are included in Figures 24 through 31. Accelerations recorded during the pluck test indicate that multiple modes of vibration participated in the cable response. Figure 32 includes a normalized Fast Fourier Transform (FFT) analysis plot of the vertical acceleration data from the pluck test performed on Cable 7L. This data was double-integrated to determine the magnitude of approximate displacements at the accelerometer locations. Figure 33 shows a plot of approximate displacements at the vertical accelerometer location for Cable 7L (Accelerometer Accel01) during the Cable 7L Pluck Test. The peak-to-peak displacement of the cable during this test was approximately 0.3 inches at the location of the accelerometer. No approximation of maximum cable amplitude or location for this maximum can be made using data from this one location along the cable due to the participation of multiple vibration modes.
7.4 Monitoring Period

The monitoring period consisted of a continuous 18 day period over which data was recorded to determine structural responses to ambient winds. During this time, data from all strain gages, displacement transducers, accelerometers, and wind anemometers was recorded during triggered events, when threshold values set in the data acquisition program were exceeded. In addition, the data acquisition system was used to perform real-time analysis of the stress range data to allow calculation of the effective stress range for each instrumented location. About 225 triggered events were stored throughout the monitoring period. Each triggered event represents a period of about one minute during which stress data from selected strain gages exceeded set limits. The threshold was adjusted to capture events with large stress ranges without filling the data acquisition memory allotment. Data from ten representative events was reduced and evaluated. A summary of the event number, time, average wind speed, average wind direction, maximum stress range, and corresponding plate tip displacement is included in Table 4. The wind direction values for these triggered events indicate that cable vibration is most active for the instrumented cables when the wind blows within 30 degrees of north (0 degrees) or south (180 degrees). The range of average wind speeds is 6 mph to 25 mph for these ten events. Note that no significant rain was reported during the occurrence any of the selected events. Maximum stress range values for the strain gage indicated in the table ranged from 5.9 ksi to 7.6 ksi. Note that the highest wind speed for this group occurred at the top of the pylon during Event 224 with a wind speed of 38 mph.
Selected stress data plots from Events 007 and 225 are included as Figures 34 and 35, respectively. As seen in the stress data presented, the vibration characteristics vary with wind speed and direction. The stress data presented for Event 007 is affected primarily by the 11th, 12th, and 14th modes of vibration of Cables 7, as shown in the normalized FFT of this stress data shown in Figure 36. Note that this FFT mode signature represents a superposition of mode shapes from both Cables 7L and 7R. Figures 37, 38, and 39 include FFT data for Cable 7L (Accelerometer Accel01), 7R (Accelerometer Accel03), and the superposition of these two, respectively. Note that the dominant mode signatures from Figures 36 and 39 are similar. The stress data presented for Event 225 is affected primarily by the 7th mode of vibration of Cables 7. FFT data for Strain Gage SG11 (stress data presented in Figure 35 above) and the superposition of FFT data from Accelerometers Accel01 and Accel03 are shown in Figures 40 and 41, respectively. These two FFT data plots are similar. Due to the dominance of cable mode 7 in the data presented for Event 225, the available accelerometer data for this event was double-integrated to produce a plot of approximate displacements at the accelerometer locations. Figures 42 and 43 are plots of approximate vertical displacements at Accelerometers Accel01 (Cable 7L) and Accel03 (Cable 7R). Maximum displacements at the accelerometer locations for Accel01 and Accel03 were 0.16 inches and 0.43 inches peak-to-peak, respectively. This data indicates that the high stress ranges shown in Figure 35 can be generated due to small displacements at higher modes of vibration. The higher vibration modes essentially shorten the length of the cable between nodes and therefore increase the rotation demand at the pylon anchorage for similar peak-to-peak displacements.
7.5 Effective Stress Range Calculation

The recorded stress data from each strain gage was analyzed to produce a stress histogram for each instrumented location. Rainflow analysis in accordance with ASTM E1049 was performed by the data acquisition system on-board and was used to produce stress histograms with stress bin sizes of 150 psi. A representative stress histogram for Strain Gage SG02 is shown in Figure 44. Stress histograms for all strain gages are included in Appendix C. Miners Rule, given below, was then used to calculate the effective stress range, , at each gage location. The purpose of the effective stress range calculation is to provide a weighted average of the amount of damage caused by a random loading pattern at a particular instrumented detail. The effective stress range provides a means of relating random variable stress cycles to constant amplitude stress cycle data. Miners Rule for a variable stress spectrum is as follows:
where: is the effective stress range is the ratio of the number of occurrences (counts) of Sr to the total number of stress cycles is the actual measured stress range at the detail under consideration Calculation of the effective stress range at each location was performed using the stress histograms produced by the rainflow analysis with the bin cut-off set at 50 percent of the constant amplitude fatigue limit for the Category E details. This equates to a bin cut-off of 2.25 ksi. The fifty percent cutoff has been
recommended by Dr. Fisher and serves to eliminate the contribution of stress cycles that are smaller in magnitude. A summary of the calculated effective stress range for all strain gaged locations over the 18 day monitoring period is included in Table 5. The Sre values ranged from 2.5 ksi to 3.9 ksi across all locations. It should be noted that cycle count data from the majority of the instrumented locations includes more than 0.01 percent of the total cycle counts greater than the constant amplitude fatigue limit (CAFL) for Category E details (4.5 ksi). In addition, about 2,000,000 stress cycles accumulated at strain gages SG02, SG05, and SG08 during the 18 day period.
8.0 PEER REVIEW OF ORIGINAL BRIDGE DESIGN

The tasks performed by WJE included a peer review of the original design for Spans 1 to 3 of the bridge structure. This task included a review of the structural calculations, prepared by URS, for the primary structural elements, including the cables, pylon, deck, and piers. The following documents were used for the basis of the peer review task: Design Drawings for Bridge No 27B35, Sheets B1 through B98, dated April 18, 2006 29th Street Greenway, Pedestrian Bridge Spans 1 to 3 Calcs, Volume 1 of 3, Dated 4-21-2006 29th Street Greenway, Pedestrian Bridge Spans 1 to 3 Calcs, Volume 2 of 3, Dated 4-21-2006 29th Street Greenway, Pedestrian Bridge Spans 1 to 3 Calcs, Volume 3 of 3, Dated 4-21-2006 29th Street Midtown Greenway Cable-Stayed Bridge, Inspection and Maintenance Manual, Version 1, February 2008 Response to WJE Project Summary of Evaluation of Sabo Bridge Cable Failure, authored by Dr. Steven L. Stroh of URS Corporation, transmitted to WJE on June 25, 2012
The original structural design drawings and applicable portions of the calculations were reviewed for general compliance with the structural requirements with the design standards. Sheet B1 of the original design drawings indicates that the following design standards were used: American Association of State Highway and Transportation Officials (AASHTO), Guide Specifications for Design of Pedestrian Bridges, August 1997 AASHTO, Standard Specifications for Highway Bridges, 2002 and Interims, Load Factor Design (LFD) Method CEB/FIP Model Code for Concrete Structures 1978 Post-Tensioning Institute (PTI), Recommendations for Stay Cable Design
The peer review effort focused on selected details for the primary structural elements. In addition, an independent finite-element analysis model was created and independent calculations were performed to verify the results of the original analysis and design. This review did not include a review of every detail or component of the design; therefore, in no way does this review relieve the designer of record from responsibility for the design. The peer review approach and findings are summarized in the following paragraphs. WJE developed an independent finite-element model of Spans 1 through 3 to evaluate forces in the primary elements of the bridge structure. This finite-element model, shown in Figure 45, was created
using SAP2000 computer analysis software. Linear frame elements were used to represent the cables, pylon, and piers, while quadrilateral shell elements were used for the concrete deck. Each element in the model was assigned properties based on the properties of the corresponding member in the bridge structure. The independent finite-element model considered dead loads (self-weight of the bridge elements) and staged construction modeling techniques were utilized to generate dead load cable forces as specified on the original design drawings. After the structural model was completed, an 85 pounds per square foot (psf) pedestrian live load was applied to the deck surface. Under certain loading scenarios, the pedestrian live load was reduced to 65 psf in accordance with the applicable design standards. In addition, H-10 vehicle loading was considered. Static forces resulting from wind and thermal loads were also evaluated. The peer review focused on evaluating member forces and capacities for the as-designed bridge structure; however, the Inspection and Maintenance Manual for the structure states that The stay cable anchorage brackets are not considered fracture critical elements since the failure of any one bracket will not jeopardize the stability of the bridge system. This Manual goes on to state that the steel cable diaphragm plates are also not considered fracture critical elements since the loss of any one plate will not jeopardize the stability of the bridge. Therefore, the WJE analysis model assumed failure of a single cable or a pair of cables at various locations in order to determine the increase in force at the remaining cables. The cable failures in the analysis were considered when the bridge is subjected to the full design live load. The results of the independent analysis model were summarized to determine forces in various primary structural members for the controlling load cases. The independently-calculated forces were used to verify forces reported in the original design calculations package. Once the forces were verified, a review of the original design calculations package was performed. This review concentrated on strength limit states of the bridge elements; however, selected serviceability limit states were also checked, such as live load deflections of the bridge deck and pedestrian-induces footfall vibrations on the bridge deck. The peer review and independent analysis performed by WJE revealed the following: Forces in the primary structural elements determined from WJEs independent finite element analysis model are in general agreement with those reported in the original design calculations package. The design for static loads was in general accordance with the specified standards. The primary structural bridge elements are adequate for the reported design loads. Although adequately designed for strength, the member design was often efficient as there were not significant levels of reserve capacity. Under full live load, the deflections of the concrete deck were in the range of L/500. A modal analysis performed using the WJE independent finite-element model revealed the fundamental frequency of the bridge structure to be less than 3 Hz, which is not in accordance with the AASHTO Design Standards. URS considered several approaches to evaluate pedestrian-induced footfall vibrations of the bridge deck and found them to be acceptable.
Wind-induced stay cable vibrations were not included in the original design calculations package, but they were considered in the original design1. Failure of a single cable or pair of cables redistributes the dead and live loads to the remaining cables. The WJE independent analysis revealed that the adjacent cables have a significant increase in cable force. However, it should be noted that the adequacy of these adjacent cables was determined using a 33% increase in allowable stress as suggested by the project design standard authored by PTI. Checks of other members after a failure of a single cable or pair of cables, such as the pylon, pylon anchorage plates, deck anchorages, and the reinforced concrete deck did not appear to be included in the original design calculations package.
9.0 REVIEW OF WIND-INDUCED STAY CABLE VIBRATION PHENOMENA

Wind-induced stay cable vibration can be attributed to several phenomena. The four most prevalent phenomena to be considered during evaluation of stay cables are vortex shedding, excitation due to interactions of rain and wind, wake galloping of cable groups, and galloping of individual inclined cables2. A brief discussion of each is included below: Vortex Shedding. During occurrence of this phenomenon, the shedding of alternating wind vortices occurs as air flows past the subject stay cable. This effect can cause a cable to resonate in one of its fundamental modes if the frequency of alternation of the wind vortices matches the fundamental frequency of the cable, which typically occurs when the wind flow is perpendicular to the cable alignment. The amplitude of cable vibrations in response to vortex shedding is typically small and occurs at low wind speeds. This phenomenon can typically be controlled with even a small amount of damping. Rain-Wind Interaction. Rain-wind interaction occurs when moderate rainfall causes rivulets of water to form on the surfaces of cables. These water rivulets cause a change in the aerodynamic properties of the cable by fundamentally changing the overall shape of the cable as it is exposed to wind action. Moderate wind speeds have been documented to cause large amplitude cable vibrations due to this phenomenon. This phenomenon can typically be controlled by altering the surface roughness of the smooth cable to prevent water rivulet formation. Wake Galloping of Cable Groups. This phenomenon occurs at high wind speeds and results in large amplitude cable oscillations when cables are in the wake of other elements, causing variations in drag and wind forces. This phenomenon is not typically a consideration for separated cable alignments. Galloping of Individual Inclined Cables. The inclination of long stay cables can theoretically cause large amplitude vibrations due to high wind speed with appropriate direction. Recent wind tunnel testing performed to study this phenomenon indicates that a small amount of damping can prevent resulting cable vibrations.
URSs Response to WJE Project Summary of Evaluation of Sabo Bridge Cable Failure, dated June 25, 2012, indicates that URS considered wind-induced vibrations during design but deferred a final decision on providing a stay cable damping system until after construction. After periodically monitoring cable vibrations over a period of 30 days, URS concluded that the observed vibrations were acceptable and stay cable dampers were not necessary. 2 Kumarasena, S., Jones, N., Irwin, P., Taylor, P., Wind-Induced Vibration of Stay Cables, Publication No. FHWA-HRT-05083, McLean, VA. : U.S. Dept. of Transportation, Federal Highway Administration, Research, Development, and Technology, Turner-Fairbank Highway Research Center, August 2007 (Citation applies for all of Section 9)
Other Phenomena. Stay cable vibrations can be attributed to the effects of other phenomena. These include live load effects, deck and stay cable interaction, and natural effects such as ice formation on the stays, which would alter the aerodynamic properties of the cable (similar to the effects of rain-wind interaction). The effects of all wind related phenomena on the magnitude of stay cable vibration amplitudes are typically controlled by increasing the mass of the cable, increasing the damping of the cable system, providing cable cross-ties, and/or altering the cable surface roughness. As discussed in Section 11, the observed stay cable vibrations are the result of vortex shedding.
10.0 STRUCTURAL ANALYSIS OF DIAPHRAGM PLATES

A finite element analysis (FEA) model of a cable diaphragm plate was developed and analyzed using SAP 2000, a proprietary software program developed for static and dynamic analysis of structures using finite elements. The FEA model was used to evaluate stresses due to cable vibrations and plate misalignment. A rendering of the FEA model is shown in Figure 46. The model was based on the dimensions of Cable Diaphragm Plate 7, which are similar to the dimensions of Cable Diaphragm Plates 8 and 9.
10.1 Stresses Due to Cable Vibrations

Wind-induced vibrations in the vertical plane cause cyclical bending moments in the diaphragm plates. The analysis focused on Event 225 (refer to section 7.4), which caused steady-state vibrations of Cables 7L and 7R in the seventh mode, with a zero-to-peak amplitude of 0.217 inches at a frequency of about 11 Hz. Note that the reported zero-to-peak amplitude value was for Cable 7R and this displacement was measured at the location of Accelerometer Accel03. The measured zero-to-peak stress is approximately 3.5 ksi at the location of Strain Gage SG11, as can be seen in Figure 35. The measured zero-to-peak deflection of the diaphragm plate was 0.006 inches. To provide an independent correlation between the measured vibrations and diaphragm plate deflections, a bending moment was applied in the model at each clevis pin hole until the plate deflections matched the measurement of 0.006 inches. The resulting maximum bending stresses at the surface of the plate are plotted in Figure 47. The strain gage locations are also shown. As can be seen, the stress predicted by the model at Strain Gauge SG11 is approximately 2.1 ksi, which is comparable to but somewhat less than the measured stress. Note that the amplitude of vibration for Cables 7L and 7R during Event 225 was not equal although the vibrations were in phase. Therefore, direct comparison of the measured stresses to the results of an evenly loaded finite element model is somewhat difficult.
10.2 Stresses Due to Plate Misalignment

If the angle of the cable is not aligned with the diaphragm plate, cable tension will induce bending stresses in the diaphragm plate. The bending force is approximately equal to the tension force in the cable times the sin of the misalignment angle times the distance from the hole to the end of the clevis assembly. Diaphragm Plate 9 was misaligned by approximately 4.5 degrees, as shown in Table 2. The project drawings indicate a long-term dead load force of 79.8 kips in Cables 9L and 9R. The calculated bending moment due to the misalignment is 210 inch-kips. The resulting maximum bending stresses at the surface of the plate are plotted in Figure 48. As can be seen in this figure, there is a small region of relatively high stress near the end of the pylon flange, and the stresses approach or exceed the yield stress of 60 ksi
between the end of the pylon flange and back of the ring plate. The significance of the misalignment stresses is discussed in Section 11. The project drawings also indicate a temperature force of 8.2 kips in Cables 9L and 9R, which is about 10 percent of the dead load force. (The live load force in Cables 9L and 9R is only 0.5 kips, according to the drawings.) Thus, the combination of misalignment and extreme temperature change results in stresses equal to about 10 percent of those plotted in Figure 48. Actual stresses over the last five years are somewhat less because temperatures in Minneapolis have been well within the design extreme temperatures of -30 degrees F and 120 degrees F. Regardless, thermal stresses are comparable to those measured and calculated due to wind-induced vibrations, but the number of cycles are less by several orders of magnitude.
11.0 DISCUSSION
The stay cable system used for the Martin Olav Sabo Pedestrian Bridge includes a total of eighteen cable diaphragm plates to connect the cables to the steel pylon. Two of the eighteen plates fractured, resulting in Cables 9L and 9R detaching from the pylon, and the need to relieve and remove Cables 8L and 8R. An additional two of the remaining sixteen in-service cable diaphragm plates were found to contain defects or cracks that had to be addressed following the discovery of the two fractured plates. At Cable Diaphragm Plate 7, the noted weld defect was removed by grinding to remove about 1/4 inch of material. A redundancy fixture was installed at Cable Diaphragm Plate 5 to provide load path redundancy around the cracks located at the edges of the plate. The discontinuities discovered in Cable Diaphragm Plates 7 and 5 were located using appropriate nondestructive test methods. No other discontinuities of significance were detected in the other in-service diaphragm plates. Comprehensive evaluations of the cable diaphragm plates included visual, magnetic particle, and ultrasonic methods. The combined application of visual, magnetic particle, and ultrasonic examination methods includes overlapping and complimentary coverage, and extends confidence to these assessments. The cable diaphragm plate assemblies include thick reinforcement rings that are fillet welded to the comparatively thin diaphragm plates. In turn, the diaphragm plates are fillet welded to the thick pylon flanges. These welded connection details provide poor fatigue resistance in conformance with Category E or E. Physical measurements of the alignment for all cable diaphragm plates were performed. The results of this work indicate that the majority of the cable diaphragm plates were installed within about 0.5 degrees of the alignment prescribed on the design drawings. The alignment of Cable Diaphragm Plates 6, 9, and 10 was found to be 2.87 degrees, 4.46 degrees, and -4.56 degrees away from the design alignment, respectively. These values are outside of the accepted tolerance of about 1 degree. A representative cable diaphragm plate was analyzed for the combined effects of misalignment and temperature change. This analysis considered the static effects on the stress field within the diaphragm plate due to the misalignment. Also considered was the calculated fluctuation in cable tension due to temperature fluctuations. The effects of these increased static stresses on the cable diaphragm plates over those which were anticipated during design may have attributed to the path along which fatigue cracks driven by wind-induced cable vibrations extended. Thermal stresses are comparable to those measured and calculated due to wind-induced vibrations, but the number of cycles are less by several orders of
magnitude. As such, thermal stresses are not considered an important contributing factor to the fatigue damage. Examinations of the fracture surfaces for Cable Diaphragm Plates 8 and 9 revealed the presence of significant stable fatigue crack growth prior to fracture. Cable Diaphragm Plate 8 sustained fatigue cracking with approximately 7 1/2 inches of stable growth occurring at the right side of the plate between the lower end of the pylon flange-to-diaphragm plate weld and the right ring plate-to-diaphragm plate weld, initiating in both locations. These two fatigue cracks coalesced and then extended in an unstable manner to intersect the top edge of the plate, dividing the plate into two pieces. A similar but shorter fatigue crack existed at the left side of the plate, initiating at the lower end of the pylon flange-todiaphragm plate weld, but there were no signs of significant unstable crack extension. Cable Diaphragm Plate 9 also sustained extensive fatigue cracking before it fractured. A stable crack extension of approximately 11 1/4 inches occurred at the right side of the plate between the pylon flangeto-diaphragm plate weld and the right ring plate-to-diaphragm plate weld, initiating in both locations. A separate stable crack extension of about 9 inches was noted at the left ring plate weld. These fatigue cracks extended in an unstable manner, separating the front section of the diaphragm plate, including the ring plates, from the remainder of the plate. Materials tests for physical, chemical, and toughness properties performed on the diaphragm plate material indicate that it conformed to the design specifications. This material provided good toughness. Embedded welding defects including slag and porosity existed in the portion of the fillet welded connection to the pylon flanges that wrapped around the lower end of Cable Diaphragm Plates 8 and 9. However, these defects did not influence the weld toe cracks that formed in the diaphragm plates. An instrumentation program was implemented to determine the magnitude of stresses in representative cable diaphragm plates in response to wind-induced cable vibrations over an 18 day monitoring period. Cable accelerations, plate tip displacements, and wind speed and direction were also recorded. During the monitoring period, rainflow analysis of the stress data was performed to produce stress range histograms for the data from each strain gage. This information was recorded continuously. The effective stress range (Sre) calculated for instrumented regions of the cable diaphragm plates adjacent to the terminations of the pylon flange-to-diaphragm plate welds, and ring plate-to-diaphragm plate welds is in the range of 2.5 ksi to 3.9 ksi for the ambient events monitored. The number of stress range cycles accumulated over the 18 day monitoring period was as high as 2,000,000 at some locations. In addition, selected structural responses to wind events resulting in high stress ranges were recorded. The stress data reported for the individual events indicates that stress ranges on the order of 5.9 ksi to 7.6 ksi are generated in response to wind events with average wind speeds ranging from 6 mph to 25 mph. The wind direction for these recorded events was primarily transverse to the bridge alignment. The measured structural responses are the result of wind-induced cable vibrations consistent with the phenomenon of vortex shedding. Stress data correlated to wind speed and direction, plate tip displacement, and accelerometer displacement data calculated by double-integrating accelerations was presented for Event 225. For this event, the average wind speed and direction was 7 mph to 8 mph and 25 degrees to 30 degrees, respectively. Maximum accelerometer displacements were on the order of 0.16 inches to 0.43 inches peak-to-peak. In addition, steady state cable vibration in the 7th mode at about 11 Hz was apparent. It is evident that low
wind speeds with appropriate direction can induce high stress range cycles at high frequency in the cable diaphragm plates due to low amplitude cable displacements. An independent peer review of the original bridge design was performed for Spans 1 through 3 of the structure. The model was used to verify forces reported in the original design calculations package. Structural checks for strength and serviceability limit states were performed for selected primary structural elements, including the steel pylon, the cables, the bridge deck, and the piers. A review of the original design calculations package and the design drawings was also performed. Forces in the primary structural elements determined from WJEs independent finite element analysis model are in general agreement with those reported in the original design calculations package. In addition, the primary structural bridge elements are adequate for the reported design loads.
12.0 CONCLUSIONS AND RECOMMENDATIONS 12.1 Conclusions and Items for Further Study
The following conclusions and items for further study are drawn from this fracture investigation: The fracture of Cable Diaphragm Plates 8 and 9 resulted from extensive fatigue crack propagation that initiated at multiple sites. These fatigue cracks coalesced and led to sudden and unstable fracture. The cable diaphragm plates incorporate welded details that offer poor fatigue resistance and are sensitive to cyclic stresses due to wind-induced cable vibrations under ambient conditions. The stress ranges measured during the 18 day monitoring period are of sufficient magnitude to cause fatigue cracking. The diaphragm plate misalignment exceeded reasonable tolerances at three cable diaphragm plates and caused high static stresses within the plates. The misalignment influenced the location of the fatigue cracks in Cable Diaphragm Plate 9, but is not an important contributor to plate fracture. The fillet welded joint between Cable Diaphragm Plates 8 and 9 and the pylon flanges exhibited inclusions and other welding defects, but the defects did not initiate cracks or significantly diminish fatigue resistance. Other than the vulnerability of the cable diaphragm plates to fatigue damage from wind-induced vibration, the design is structurally adequate. The existing cable diaphragm plates should be retrofitted to enhance fatigue resistance, including the backstay plates. Cable damping systems should be considered to avoid user concerns and protect the cables themselves. The lower cable anchorages and sockets should be reviewed for fatigue susceptibility and damage.
12.2 Cable Diaphragm Plate Retrofit Options

Three retrofit options have been developed and should be considered to retrofit the pylon cable diaphragm connections. Retrofit detail Sheets S-0 through S-4 are included as Appendix D. Option 1 - Mechanically Fastened Plates to the Existing Connection Plate and Pylon Flanges. Sheets S-1 and S-2 (or S-2A) detail this retrofit option complete with a Procedure for installation and General Notes. To allow the fit-up the additional plates, removal of the connection ring plates is required.
Placement of two 1 3/4 inch thick plates will provide the correct clevis connection dimension. The two new diaphragm plates will be fastened through the existing 3/4 inch thick plate using twelve 1 inch diameter ASTM A325 high strength bolts. Four angles will be used to connect the cable diaphragm plates to the pylon flange plates. Each connection angle is fastened with nine 1 inch diameter ASTM A325 high strength bolts. It should be noted that all faying surfaces for this option are prepared to Category B slip resistance. This retrofit provides a mechanically fastened connection to the pylon that eliminates welding, results in increased section properties, and provides built-in redundancy. Option 2 - Mechanically Fastened Doubler Plates to Pylon Flanges to Complete Cable Connections. For installation of this retrofit the cable diaphragm plate will be removed and existing welds will be ground off the pylon flange plates. At each pylon flange, this new connection is made up of three bent plates consisting of a 1 1/2 inch thick fill plate and two 1 3/8 inch thick side plates. Sheet S-3 shows the retrofit connection installed on the pylon complete with General Notes. A two piece HP14x73 cut off section will be installed as a strut between the pylon flanges. In total, fifteen 1 inch diameter ASTM A325 high strength bolts connect the three bent plate group to each pylon flange. Fabricating this three bent plate group is considered to be difficult. In addition, this option includes modification of the original cable alignment and therefore may require modification to the distance between cable anchor points. Option 3 - Mechanically Fastened Weldment to Pylon. For installation of this retrofit the cable diaphragm plate will be removed and existing welds will be removed by grinding. This two piece weldment is made up of multiple plates that are connected to the pylon flanges as shown on Sheet S-4. The main connection of the cables to the weldment uses two 4 3/8 inch thick plates. These plates will be installed vertically on the proper alignment of the cables. However, the cable clevis at the pylon has been rotated 90 degrees. Note that the weldment must be fabricated in two pieces to allow placement at any connection elevation without total removal of all cable diaphragm plates during installation. In total, sixteen 1 inch diameter ASTM A325 high strength bolts connect the weldment to the pylon. It is recommended that any of the three retrofit options should be carried out by qualified and experienced machine type metal workers that have documented experience working closely with a fabricator that is qualified in accordance with MnDOT and AASHTO requirements to produce fracture critical bridge members. Although the cable diaphragm plates are not considered fracture critical, a fabricator with the stated qualifications is preferred. Construction period services and field inspection should be carried out by an engineering firm with documented experience in retrofit design and physical modification of fracture critical bridge details. Proper fit-up of each piece is of utmost importance and high quality workmanship is required. Based on the retrofit options, the detail shown on Sheets S-1 and S-2 (or S-2A) is the preferred choice for the installation at this time. This detail can be installed at one location at a time with only one pair of cables removed from the pylon for the work period. In addition, Option 1 is the most adaptable and straight-forward connection that can be used to upgrade all 18 cable diaphragm plate connection locations.
TABLES
Table 1. Primary Equipment for Nondestructive Examination of Cable Diaphragm Plates

Test Equipment Ultrasonic test instrument All ultrasonic scans Transducer 1 Straight Beam Scans Transducer 2 Straight Beam Scans Transducer 3 Shear Wave Scans Transducer 4 Shear Wave Scans Transducer 5 Shear Wave Scans Calibration Standard Couplant Magnetic Particle Yoke Magnetic Medium Examination General Description G.E. Krautkramer USN 58 L Ultrasonic Flaw Detector G.E. - 1.00 inch diameter, 2.25 MHz - 0 (Longitudinal) G.E. - 0.50 inch diameter, 5.00 MHz - 0 (Longitudinal) G.E. - 0.50 inch diameter, 3.50 MHz - 45 (Shear) G.E. - 0.50 inch diameter, 2.25 MHz - 60 (Shear) G.E. - 0.50 inch diameter, 2.25 MHz -70 (Shear) IIW Type 2 Ultrasonic Reference Standard Sonotech - Ultragel II (modified with windshield washing fluid to resist freezing) Parker Research Model B310 Mini-Magnetic particle yoke Dry red particles Primary Function Production & detection ultrasonic pulses Initial straight beam scans of
Straight beam scans
Diaphragm Plates - Shear wave scans Diaphragm Fillet Weld - Shear wave scans Ring Plate Fillet Weld - Shear wave scans Horizontal linearity transducer sensitivity Coupling for all scans Induction of magnetic fields Enhanced exhibition discontinuities of and
Table 2. Cable Diaphragm Plate Alignment Measurements

Cable Diaphragm Plate 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Design Angle Value Protractor and/or Direct Measurement (deg) Value Design 17.5 -0.08 28.3 35.0 39.0 42.0 47.5 46.5 49.0 54.0 54.0 55.5 53.0 50.5 47.5 45.5 43.0 40.5 39.0 0.00 0.01 -0.45 -0.60 2.57 -0.22 0.70 4.64 -4.41 -0.11 0.09 0.19 -0.32 0.08 -0.12 -0.42 0.32 Bosch Instrument (deg) Value 17.0 28.3 34.9 39.6 42.5 47.9 46.9 48.5 53.7 53.6 55.7 53.0 50.6 47.7 45.4 42.7 40.7 38.7 Design -0.58 0.00 -0.09 0.15 -0.10 2.97 0.18 0.20 4.34 -4.81 0.09 0.09 0.29 -0.12 -0.02 -0.42 -0.22 0.02 Stabila Instrument (deg) Value 18.0 28.6 35.2 39.5 42.7 48.0 46.9 48.8 53.8 54.0 56.1 53.4 50.9 48.2 45.8 43.2 41.1 39.2 Design 0.37 0.25 0.16 0.05 0.05 3.07 0.13 0.50 4.39 -4.46 0.44 0.44 0.59 0.33 0.38 0.03 0.18 0.47 Avg Design (deg) -0.10 0.08 0.03 -0.08 -0.22 2.87 0.03 0.47 4.46 -4.56 0.14 0.21 0.35 -0.03 0.15 -0.17 -0.15 0.27
DMS 17 35 02 28 18 13 34 59 26 39 27 00 42 36 01 44 55 55 46 43 16 48 17 55 49 21 34 58 24 18 55 36 29 52 54 36 50 18 45 47 49 00 45 25 09 43 07 17 40 55 15 38 40 42
Deg 17.58 28.30 34.99 39.45 42.60 44.93 46.72 48.30 49.36 58.41 55.61 52.91 50.31 47.82 45.42 43.12 40.92 38.68
Table 3. Summary of Pluck Test Data

Pluck Test Cable 7L Stress Gage Range (psi) 1,380 SG07 3,070 SG08 1,960 SG09 860 SG10 1,770 SG11 600 SG12 Wind 6-10 Direction (Deg) Pluck Test Cable 6L Stress Gage Range (psi) 980 SG19 3,670 SG20 2,500 SG21 1,210 SG22 2,260 SG23 1,190 SG24 Wind 6-10 Direction (Deg) Stress Range (psi) 1,460 3,110 2,200 650 1,920 1,530 98 Pluck Test Cable 7R Stress Gage Range (psi) 830 SG07 2,140 SG08 930 SG09 2,360 SG10 3,160 SG11 1,270 SG12 Wind 7-9 Direction (Deg) Pluck Test Cable 6R Stress Gage Range (psi) 620 SG19 2,610 SG20 1,270 SG21 2,770 SG22 3,550 SG23 2,200 SG24 Wind 6-10 Direction (Deg) Stress Range (psi) 880 2,200 1,100 1,860 3,400 3,230 114
Gage SG01 SG02 SG03 SG04 SG05 SG06 Wind Speed (mph)
Notes: 1. 2. 3. 4. 5. 6. 7. 8.
Stress Range (psi) 2,490 3,300 2,650 1,240 2,370 1,930 103
Stress Range (psi) 1,540 2,340 1,350 2,490 3,770 3,470 108
All pluck tests performed using a 250 lb load perpendicular (down) to the cable alignment. Load applied at approximately 58 ft, 38 ft, 62 ft, and 42 ft from the top Anchor Point for Cable 6L, 6R, 7L, and 7R, respectively. Load applied using tension ram and load cell to verify force. Quick release equipment used to instantaneously release load. Reported stress range is from first cycle following load release. Reported wind speed and direction are average values throughout the test period. Wind-induced cable vibration minimal during test period. Wind direction reference is 0 Degrees perpendicular left (north) of the structure with increasing increments clockwise looking down.
Table 4. Representative Triggered Event Data

Event 007 014 067 072 076 123 150 223 224 225 Date/Time 4/21/12 4/21/12 4/22/12 4/27/12 5/1/12 5/1/12 5/1/12 5/1/12 5/1/12 5/6/12 9:34 17:27 16:18 17:11 09:35 11:40 15:54 19:46 20:04 14:58 Average Wind Speed (mph) 10-14 10-13 6-9 6-8 13-14 12-13 19-23 16-19 20-25* 7-8 Average Wind Direction (deg) 195-200 175-190 5-15 20-35 180-185 190-205 180-185 190-205 190-200 25-30 Max Stress Range Gage 2 11 11 11 2 2 11 2 2 11 Max Stress Range (ksi) 6.2 6.8 7.6 7.1 6.3 5.9 6.0 7.5 7.2 6.9 Max Tip Displacement (inches) 0.009 0.013 0.015 0.014 0.009 0.012 0.012 0.014 0.013 0.014
* max of 38 mph at pylon top
Table 5. Summary of Effective Stress Range Data

Gage SG01 SG02 SG03 SG04 SG05 SG06 SG07 SG08 SG09 SG10 SG11 SG12 SG13 SG14 SG15 SG16 SG17 SG18 SG19 SG20 SG21 SG22 SG23 SG24 Total Counts 123,237 2,058,369 512,472 939,634 1,855,078 79,735 107,528 1,988,854 348,232 1,244,034 1 321,260 62,359 98,028 326,342 24,152 85,642 254,300 30,881 304,871 Sre (psi) 2,720 3,660 2,900 3,060 3,810 2,560 2,710 3,320 2,870 3,900 2,330 3,290 2,710 2,650 3,050 2,530 2,790 3,160 2,560 2,980 Comments Percent of Counts for Sr Above 4.5 ksi 0.000% 0.965% 0.036% 0.179% 1.802% 0.000% 0.001% 1.734% 0.032% 3.100% 0.000% 0.209% 0.001% 0.001% 0.060% 0.000% 0.003% 0.108% 0.000% 0.037%
No Data No Data No Significant Data
No Data No Data
FIGURES
Figure 1. Side view of the Martin Olav Sabo Pedestrian Bridge looking northeast.
PIER 3
PIER 1
Figure 2. Overall plan drawing of Spans 1 through 3 taken from the original design drawings.
WEST ABUTMENT
Figure 3. Typical cable anchorage assembly at the concrete deck.
Figure 4. Typical cable anchorage at the pylon showing the cable diaphragm plate and clevis pin connection.
Figure 5. Typical cable earth anchorages.
Figure 6. Elevation drawing of Spans 1 through 3 looking north taken from the original design drawings.
WEST ABUTMENT
PIER 1
PIER 3
Figure 7. Typical cable diaphragm plate. Cable Diaphragm Plate 6 shown.
Figure 8. Magnetic particle testing (MT) in progress showing test equipment.
Figure 9. Ultrasonic testing (UT) in progress.
Figure 10. Area at the right side diaphragm plate-to-pylon flange weld wrap around for Cable Diaphragm Plate 7 showing the area from which the noted indication was removed.
Figure 11. Crack located in the bottom surface of Cable Diaphragm Plate 5 at the right side diaphragm plate-to-pylon flange weld. The crack is approximately 3/4 inches long. It is shown highlighted with red MT powder before chipping the paint off.
Figure 12. Crack located in the bottom surface of Cable Diaphragm Plate 5 at the left side diaphragm plate-to-pylon flange weld. The crack is approximately 1 1/2 inches long. It is shown highlighted by red MT powder after chipping the paint off.
Figure 13. View of Cable Diaphragm Plate 9 after fracture.
Figure 14. Segment of Cable Diaphragm Plate 9 that separated from the pylon following plate fracture.
Figure 15. View of Cable Diaphragm Plate 8 following fracture and removal of Cables 8L and 8R.
Figure 16.View of the left pylon flange alignment following local plastic rotation of the flange at the attachment elevation for Cable Diaphragm Plate 8.
Figure 17. Segments of Cable Diaphragm Plate 8 shown shortly after removal from the pylon.
Figure 18. Segment of Cable Diaphragm Plate 9 shown shortly after removal from the pylon.
Figure 19. Typical strain gage installation near a diaphragm plateto-ring plate weld toe.
Figure 20. Typical accelerometer installation on a stay cable. The two accelerometers are attached to block surfaces normal to one another in order to measure vertical and lateral accelerations.
Figure 21. Typical LVDT installation at the tip of Cable Diaphragm Plate 7.
Figure 22. Wind monitor mounted 10 feet above the left curb between Cables 3L and 4L.
Figure 23. Data acquisition system shown after attachment of instrument wiring.
6000 SG02 SG08 4000
2000
Stress (psi)
-2000
-4000
-6000 0 50 100 150 200 250 300 350 400 450 500 550 600 650
Seconds
Figure 24. Plot of stress data from selected strain gages in response to the Cable 7L pluck test.
6000 SG02 SG08 4000
2000
Stress (psi)
-2000
-4000
-6000 167
168
169
170
171
Seconds
Figure 25. Zoom of stress data at load release for the data shown in Figure 24.
6000 SG05 SG11 4000
2000
Stress (psi)
-2000
-4000
-6000 0 50 100 150 200 250 300 350 400 450 500 550 600 650
Seconds
Figure 26. Plot of stress data from selected strain gages in response to the Cable 7R pluck test.
6000 SG05 SG11 4000
2000
Stress (psi)
-2000
-4000
-6000 200
201
202
203
204
Seconds
6000 SG14 SG20 4000
2000
Stress (psi)
-2000
-4000
-6000 0 50 100 150 200 250 300 350 400 450 500 550 600 650
Seconds
Figure 28. Plot of stress data from selected strain gages in response to the Cable 6L pluck test.
6000 SG14 SG20 4000
2000
Stress (psi)
-2000
-4000
-6000 70
71
72
73
74
Seconds
6000 SG17 SG23 4000
2000
Stress (psi)
-2000
-4000
-6000 0 50 100 150 200 250 300 350 400 450 500 550 600 650
Seconds
Figure 30. Plot of stress data from selected strain gages in response to the Cable 6R pluck test.
6000 SG17 SG23 4000
2000
Stress (psi)
-2000
-4000
-6000 33
34
35
36
37
Seconds
1
FFT Accel01
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 32. Normalized FFT plot showing multiple modes of vibration during response to the Cable 7L pluck test.
0.3 Accel01 0.2
Displacement (inches)
0.1
-0.1
-0.2
-0.3 167
168
169
170
171
Seconds
Figure 33. Approximate displacement data generated by doubleintegration of acceleration data from Accelerometer Accel01 at load release during the pluck test for Cable 7L.
6000 SG02 4000
2000
Stress (psi)
-2000
-4000
-6000 0 1 2 3 4
Seconds
Figure 34. Plot of stress responses from Strain Gage SG02 during Event 007.
6000 SG11 4000
2000
Stress (psi)
-2000
-4000
-6000 0 1 2 3 4
Seconds
Figure 35. Plot of stress responses for Strain Gage SG11 during Event 225.
1
FFT SG02
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 36. Normalized FFT plot of the stress data from Strain Gage SG02 during Event 007 showing the dominance of cable modes 11, 12 and 14. Note that the approximate first mode of cable vibration is 1.53 Hz.
1
FFT Accel01
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 37. Normalized FFT plot of the acceleration data from Accelerometer Accel01 mounted on Cable 7L during Event 007.
1
FFT Accel03
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 38. Normalized FFT plot of the acceleration data from Accelerometer Accel03 mounted on Cable 7R during Event 007.
1
FFT Super Accel01 + Accel03
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 39. Superposition of normalized FFT plot for the acceleration data from Accelerometers Accel01 and Accel03, mounted on Cables 7L and 7R during Event 007. Similar to plot shown in Figure 36.
1
FFT SG11
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 40. Normalized FFT plot of the stress data from Strain Gage SG11 during Event 225 showing the dominance of cable mode 7. Note that the approximate first mode of cable vibration is 1.53 Hz.
1
FFT Super Accel01 + Accel03
0.75
0.5
0.25
0 0 5 10 15 20 25 30 Frequency, Hertz 35 40 45 50
Figure 41. Superposition of normalized FFT plot for the acceleration data from Accelerometers Accel01 and Accel03, mounted on Cables 7L and 7R during Event 225. Similar to plot shown in Figure 40.
0.3 Accel01 0.2
0.1
-0.1
-0.2
-0.3 1 2 3 4 5
Seconds
Figure 42. Approximate displacement data generated by doubleintegration of acceleration data from Accelerometer Accel01 during Event 225.
0.3 Accel03 0.2
0.1
-0.1
-0.2
-0.3 1 2 3 4 5
Seconds
Figure 43. Approximate displacement data generated by doubleintegration of acceleration data from Accelerometer Accel03 during Event 225.
Figure 44. Stress range histogram for Strain Gage SG02. The calculated effective stress range for this gage over the 18 day monitoring period is 3.7 ksi.
Figure 45. View of finite element model of Spans 1 through 3 created using SAP2000.
Figure 46. Finite element model of cable diaphragm plate.
Figure 47. Maximum bending stresses at diaphragm plate surface due to simulated wind-induced vibration.
Figure 48. Bending stresses at diaphragm plate surface due to 4.5 misalignment.
Appendix A
John W. Fisher Report
Report on the cracking that developed in the Cable Stay diaphragm plates of the Martin Olav Sabo Pedestrian Bridge Minneapolis, Minnesota By John W. Fisher and Eric J. Kaufmann 5-24-12
1. Introduction a) Cable Diaphragm Plate 9 The discovery of a fractured diaphragm plate for Cables 9L and 9R occurred when the severed diaphragm and attached cables dislodged from their pylon connection. Figure 1 shows the crack surface of the failed plate end with the ring plates welded to the diaphragm plate surfaces. The entire cross section is cracked. Fatigue cracks had formed at four distinct locations: Two at each end of the diaphragm-flange weld terminations and the other two forming at the weld toes of the ring plates on the top surface of the diaphragm plate. These fatigue cracks coalesced into two distinct conditions, as evident from the corrosion covered surfaces, until the final fracture which developed at the two clean looking segments occurred.
Although the transverse distance across the diaphragm plate at the ends of the flange welds is 23in., the total crack path was longer at about 25in, as it followed the principal stress toward the ring plates from the end welds and around the ring plate weld toes as can be seen in Figure 2 of the diaphragm plate segment that remained in the pylon. The edge crack on the north (left) side (right side of Figure 2) is clearly visible and has propagated following the principal stress behind the primary crack and shielded that end of the primary crack from the north ring plate weld toes.
The weld toe terminations seen at each end of the diaphragm in Figure 2 of the diaphragm-flange welds are both Category E fatigue details with a fatigue limit of 2.6ksi(1). Likewise, the weld toes of the ring plates to the diaphragm plates provides a
Category E detail for the axial and bending components of cyclic stress introduced into the diaphragm plate. This detail is at the upper bound margin of diaphragm plate thickness so Category E may also apply. The stress field across the diaphragm width is not uniform and is influenced by the applied cable forces and shear lag.
The segment of Cable Diaphragm Plate 9 that remained in the pylon was removed along with the weld ends and small segments of the pylon flanges as shown in Figure 3. The figure also shows the section from which mechanical properties (tensile and CVN tests) were obtained. The red lines indicate where crack surface segments would be cut to permit cleaning and fracture surface examination.
b) Cable Diaphragm Plate 8
Cable Diaphragm Plate 8 was also discovered to be cracked and can be seen in Figure 4 after Cables 8L and 8R were removed. It is apparent from the failure that the diaphragm was not as symmetrically stressed across the diaphragm to the degree that Cable Diaphragm Plate 9 was. This resulted in the fatigue crack propagation from the south (right) side weld toe of the diaphragm to flange weld. This edge crack intersected and coalesced with a crack that developed at the adjacent ring plate weld toe, and the combined crack reached an edge crack length of about 7.5in at which time it fractured, following the principal stress as a result of the eccentric load acting on the diaphragm plate. Although a similar shorter crack formed at the north (left) flange weld end as seen in Figure 4, it was not through the diaphragm plate thickness. The eccentricity caused by the fracture plastically deformed the north pylon flange out of its plane, as is apparent in Figure 4.
The two diaphragm segments were removed from the pylon along with small pieces of the pylon flanges and can be seen in Figure 5. Both segments of the crack surface were heavily corroded. Segments of the diaphragm plate marked in red were removed for mechanical property tests. The diaphragm-flange segment on the left side of the plate was also removed so that the crack surface could be exposed.
2. Examination of the Crack Surface Segments
a) Cable Diaphragm Plate 9
Cable Diaphragm Plate 9 with attached flange pieces was cut into segments as shown in Figure 6. The segments are identified as Segments 9A through 9F from left to right (north to south). These segments were all cleaned in an ultrasonic bath, first using acetone to remove the petroleum jelly that had been applied to the crack surface. After that it was cleaned with Alconox to remove corrosion product without damaging the crack surface so that micro features could be examined in the Scanning Electron Microscope (SEM).
Figure 7 shows the cleaned crack surface of Segment 9D (the middle segment). Near the middle is where the cleaned semi-elliptical crack extending from the left of the surface in the figure intersects with the inner fracture segment. The beach marks on both the left and right ends of the surface in the figure show semi-elliptical fatigue cracks that originate from the top surface of the diaphragm plate. The adjacent smaller segment which shows the crack surface at the south (right) ring plate welds can be seen in Figure 8. The surface shows extensive corrosion product that could not be removed. Similar corrosion product is shown in Figure 7. Beach marks are visible on the right side of Figure 8 and multiple initiation surface cracks at the weld toe are apparent in the crack offsets seen in Figure 8.
The adjacent crack surface of Segment 9F at the flange weld can be seen in Figure 9. This edge crack from the south (right) flange weld is more heavily corroded than the other crack segments. The linkage of these crack segments provides a total edge crack length of the coalesced fatigue crack that is 11.25in long up to the middle of the diaphragm plate where it intersects with the clean fracture surface. This can be easily seen in the right side mating crack surface shown in Figure 10.
At the north (left) side of the diaphragm plate, cleaned Segment 9C at the left ring plates is seen in Figure 11. This shows beach marks and evidence of multiple crack initiation
sites at the surface of the diaphragm ring plate weld toe that have coalesced into a common crack.
Figure 12 shows a view of the crack in Segment 9B which is adjacent to Segment 9C shown in Figure 11. This marks the boundary of the fatigue crack that originated at the north (left) ring plates. An overview of the left side mating surface is provided in Figure 13. This shows the through thickness crack bounded on each side by fractured segments.
Finally the north (left) edge crack from the diaphragm-flange weld end, as seen in Figures 2 and 6, was exposed to examine the Segment 9A crack surface. Figures 14 and 15 show the mating crack surfaces of this crack that extended behind the primary failure surface. Crack initiation is from the end of the weld as expected. This crack extended across the diaphragm plate more quickly from the top surface which is more apparent at the end of the surface where the segment was fractured to separate the surfaces.
b) Cable Diaphragm Plate 8
Figure 16 shows the segments cut from Cable Diaphragm Plate 8 (see Figure 5 for the components removed from the pylon). One segment was removed from the north (left) side of the plate and is indicated as Segment 8L. Four segments were removed from the south (right) side of the plate and are indicated as Segments 8R1, 8R2, 8R3, and 8R4 from back to front. Note that Segment 8R4 is the mating piece to Segment 8R3. The edge crack initiated from the south (right) diaphragm-flange connection in Segment 8R3 as shown in Figure 17. The mating crack surfaces provide the same characteristics and show that the fatigue crack initiated at the weld end and progressed toward the south (right) ring plates with the top surface advancing faster than the bottom surface of the crack. This edge crack coalesced with the weld toe cracking at the ring plate as seen in Figure 18. A brittle fracture was triggered as a result of an eccentric load from the two cables that completely separated the diaphragm plate as evident in Figure 19.
At the North (left) diaphragm-flange weld end a crack had formed at the top diaphragm plate surface in Segment 8L as seen in Figures 4 and 5. This segment was removed and the fracture surfaces exposed by breaking it open at a reduced temperature and the separated pieces are shown in Figure 20. The bottom surface of the diaphragm plate can be seen in Figure 21. This shows that an independent fatigue crack had initiated in the bottom surface of the diaphragm plate about 1/2in from the larger primary crack whose exposed surfaces are shown in Figures 22 and 23. The segment shown in Figure 23 was cleaned in an ultrasonic bath and the cleaned surface is shown in Figure 24. Fatigue crack beach marks show that the crack was advancing as a semi-elliptical surface crack toward the North (left) ring plate weld. The crack in the bottom surface was also extending parallel to the top surface crack. The bottom of the top surface crack appears to be rubbed and abraised from contact indicating some compression likely from bending.
No evidence of weld toe cracking was observed at the North (left) ring plate weld toe. It seems likely that much of the crack growth seen at Segment 8L occurred after the fracture of the diaphragm plate. The behavior at Cable Diaphragm Plate 8 suggests that the two cable loads were not the same when introducing cyclic forces into the diaphragm. Crack development was significantly different than observed at Cable Diaphragm Plate 9 where nearly equal behavior was seen in the crack development at both sides of the plate.
A segment of the North (left) ring plate and diaphragm plate was saw cut and removed from the connection so that the fillet weld toes on the diaphragm plate could be polished and etched and further examined for the possibility of weld toe cracks. The region removed can be seen in Figure 25.
3. Material Properties
Material properties of Cable Diaphragm Plates 8 and 9 were evaluated to determine their chemical composition, tensile properties and CVN properties. Tensile and CVN specimens were oriented in both the longitudinal direction of the plates (loading direction) and the transverse direction. Test specimens were located in areas of the plates where no evidence of plastic deformation had occurred during fracture.
The test specimen locations are shown in Figures 3 and 5. Since the fracture path in Cable Diaphragm Plate 8 bisected the plate, test specimens were extracted from two separate areas of the plate. Sketches of the test specimen layout are shown in Figures 26 and 27. One 0.505 in. dia. tensile specimen was fabricated in each plate orientation. Twelve standard size CVN specimens were fabricated in each plate orientation positioned at the plate quarter thickness. The CVN specimens were tested in triplicate over a range of test temperatures to define the transition behavior of the steel. Results of the chemical composition, tensile, and CVN testing are provided in the laboratory test report in Appendix A.
Results of the chemical composition analyses of Cable Diaphragm Plates 8 and 9 showing the primary constituents are reproduced in Table 1 with ASTM A572 Gr. 50 requirements for comparison. Table 1
Diaphragm Plate 8 9 ASTM A572 Gr. 50 C 0.16 0.15 0.23 Max. Mn 1.11 1.12 1.35 Max. P 0.02 0.02 0.04 Max. S 0.01 0.01 0.05 Max. Si 0.31 0.32 0.40 Max. Wt% Ni Cr 0.19 0.48 0.19 0.48 Mo 0.016 0.015 V 0.029 0.029 0.010.15 Max. Nb 0.002 0.002 Cu 0.23 0.23 0.20 Min.* Al 0.021 0.022
*When specified
Both plates conformed to ASTM A572 Gr. 50 steel, which is the steel specified for the diaphragm plates. The close similarity in composition of the two plates strongly suggests they were fabricated from the same heat of steel. It is also noted that the plates also
conform to ASTM A588 steel. The residual alloy elements included in the full analyses (Sn, W,Zn, Zr ) were found to be at normal residual levels that would have no measurable effect on the material properties.
Results of tensile tests of the diaphragm plates are shown in Table 2. Both plates satisfied the tensile requirements of ASTM A572 Gr. 50. The tensile properties of the plates were identical, consistent with the chemical composition analyses indicating they were fabricated from the same heat of steel. The properties were also nearly identical in both the longitudinal and transverse directions.
Table 2
Diaphragm Plate 8 9 ASTM A572 Gr. 50 Orientation Yld. Strength (0.2% Offset) (ksi) 61.0 60.0 60.5 59.5 50 min. Tensile Strength (ksi) 88.0 88.5 88.5 89.0 65 min. Elong. (2) (%) 27 27 26 26 21 min. R.A. (%) 71 72 65 68
Long. Trans. Long. Trans.
The CVN tests were conducted in both the longitudinal and transverse orientation of the diaphragm plates. Tests were performed in triplicate at temperatures selected to encompass the transition temperature range of the steel. Tests were performed at room temperature (70 F), +10 F, corresponding to the AASHTO Zone 3 test temperature for Gr. 50 steel, 40 F, and -30 F. Results of the tests are provided in Appendix A and shown plotted in Figure 28.
The results of the CVN tests show that both plates have similar CVN properties and both satisfy the AASHTO Zone 3 CVN requirement for Gr. 50 steel for both non-fracture critical and fracture-critical members (15 ft-lbs @ 10F and 25 ft-lbs @ 10 F, respectively)
in both orientations tested. The difference in CVN energy in both orientations is small, consistent with the small difference in tensile properties measured in the two orientations. The measured CVN toughness is very high in both plates and explains the large fatigue crack sizes that were observed prior to brittle fracture of the plates.
4. Fractographic Assessment and Crack Growth
a) Segment 8L The crack surface seen in Figure 29 shows beach marks of fatigue crack growth over much of its width as the fatigue crack initiated from the end of the weld toe at the diaphragm-flange connection. The crack extended across the top of diaphragm plate toward the adjacent ring plate as a semi-elliptical surface crack. The striations seen in Figure 30 were located about 3mm (1/8in) below the top surface of the diaphragm plate.
Figure 30 shows the fatigue crack growth striations that were found at location x at a magnification of 4000X. These represent individual crack growth increments and are seen to be at a very high growth rate of 2m (7.874 x 10-5in/cycle) to 5m (19.6 x 10-5 in/cycle). These high levels of crack growth are in Region III and cannot be used to estimate K levels from the usual Region II relationship da/dN = 3.6 x 10-10 K3.
It seems logical to assume that these high levels of crack growth occurred after the fatigue crack that developed at the right side of the plate coalesced with the fatigue crack from the adjacent bushing weld toe and fractured the diaphragm plate. Hence, only the remaining section was resisting the forces from the two cables. The stress cycles were stress reversals and may have exceeded the yield point of the diaphragm plate as the paint in the vicinity of the 8L crack showed evidence of yielding with cracked and missing paint.
Figures 31 and 32 show SEM views of the weld toe fatigue crack growth origin in Segment 8L at 11X and 300X. These photos show the existence of corrosion product at the crack origin and no evidence of weld toe defects.
b) Segment 9C
Figure 33 shows the location on the fracture surface of Segment 9C where possible striation or beach mark features were detected at a magnification of 1000X as seen in
Figure 34. The location shows visually obvious beach marks as the crack extended out of the ring plate weld toe region. The spacing of 5m to 10m is well into Region III of fatigue crack growth.
The K value is well above the 60ksiin level of crack growth for ferrite-pearlite steels shown in Figure 9.12 of Barsom & Rolfe for Region II crack growth (2). The through thickness crack at the left side of Cable Diaphragm Plate 9 is about 9in. long (see Figure 13) between the two fracture segments on each side of the crack that developed at the bushing weld toe which is a low fatigue strength Category E detail. The crack development involves both tensile axial stress and bending in the weak axis of the 3/4in. diaphragm plate. Near the ends of this through thickness crack, a crude estimate of K is
K = Sr a F(a/b) Where a 4.5in , F(a/b) 1.5 and a probable stress range of 10-15ksi or higher results in K 85ksiin which is compatible with the rapid growth of the large crack implied by the features in Figure 34. Unfortunately, the crack surfaces at the weld toes of the north (left) ring plates is severely corroded as illustrated in Figures 35, 36, and 37. The fatigue crack growth features are not apparent in beach marks nor micro level characteristics in Figure 35. However Figures 36 and 37 show end of life features of beach marks.
10
5. Fatigue Resistance
The two basic welded details used at the diaphragm cable connection are both very low fatigue resistant details (1). The fillet welded connections between the diaphragm plate and the pylon flanges are both long attachments that are classified as Category E fatigue details. Shear lag also influences crack growth from these details as the cracks follow the principal tensile as the crack is initiated and propagates from the weld toe termination. The lower bound fatigue limit for Category E details, is 2.6ksi and occurs at about 20 million cycles.
The second fatigue critical detail is the weld toe of the ring plate to diaphragm plate connections. These details are on the margin between Category E and E as the diaphragm plate thickness of in is at the limit for Category E. Its fatigue limit is slightly higher at 4.5ksi and occurs at about 10 million cycles.
The applied loads from the cables do not provide a uniform stress field across the diaphragm width. There is a shear lag influence that elevates the principal stress at the end welds and at the ring plate weld toes near the end welds. This is reflected in the crack development at the diaphragm-flange end weld and at the adjacent ring plate weld toe.
Limited measurements by WJE on April 20, 2012 on Cable Diaphragm Plate 7 show that the end of the diaphragm plate is deformed up and down by equal amounts from vibration of the cable at low wind speeds of 5-8mph. The resulting stress cycles are repeatable narrow band cycles for this wind speed at a frequency of 11Hz. The stress range at the end welds (the gage was slightly offset and 0.3in from the weld toe) was a narrow band with a maximum stress range of 2.8ksi on the top and bottom diaphragm plate surfaces as illustrated by the response at Gage SG01 seen in Figure 38. This cyclic stress is always in tension as a result of the axial force in the cables. At the weld toe this cyclic stress is likely greater. Since the variable stress range spectrum exceeds the fatigue limit by more than 0.01%, stress cycles above 0.5 x Fatigue Limit will contribute to fatigue damage at Cable Diaphragm Plate 7.
11
Strain gages were also installed at 0.3in from the ring plate weld toe at half the distance between the line of action of the cable and the diaphragm plate to pylon flange weld, where fatigue crack growth was dominate in both Cable Diaphragm Plates 8 and 9. The narrow band maximum stress range was 6.3ksi which also exceeds the fatigue limit of Category E by more than 0.01% as is visually apparent in Figure 39 as the stress cycles above 0.5 x Fatigue Limit will all contribute to fatigue damage at the top and bottom ring plate weld toes at Cable Diaphragm Plate 7.
There is every reason to expect that the response of Cable Diaphragm Plates 8 and 9 will have experienced cyclic stresses that are greater than those measured at Cable No 7. Fatigue crack growth and eventual failure would be expected at the stress ranges that developed at the end welds and adjacent ring plates. At 11Hz, 1million stress range cycles would accumulate in about 28hrs of comparable wind conditions. Higher wind speeds are likely to result in higher stress cycles at higher modes and frequency.
From the extensive corrosion on the crack surfaces, portions of these fatigue cracks likely reached full diaphragm plate thickness several years ago. Cable vibration is the cause of high frequency stress cycles and it is clear that these cycles exceed the fatigue limit at both welded details. At Cable Diaphragm Plate 9, the crack development is nearly symmetrical leading to crack growth at all 4 locations. At the south (right) side of Cable Diaphragm Plate 9 the two cracks joined providing an 11.25in edge crack at the final failure. At the north (left) side of Cable Diaphragm Plate 9 the two cracks were offset so that the ring plate weld toe crack extended to a through thickness crack about 9in long. The left side edge crack extended over 6in behind the through thickness crack which shielded the larger ring plate crack at that end.
The gross section capacity at Cable Diaphragm Plate 9, based on yield strength of the diaphragm plate, is about 203kips, as the remaining section was about 3.4in2 prior to fracture. The reported total dead service load cable force at Cable Diaphragm Plate 9 is about 162kips. Hence, the observed brittle fracture is consistent with the high toughness
12
of the diaphragm material that allowed such large fatigue cracks to form at Cable Diaphragm Plate 9.
At Cable Diaphragm Plate 8, the fatigue crack growth was not symmetrical as no detectable cracking was found at the left ring plate and much of the cracking that developed at the left end weld may have occurred after the failure originating at the right side of the plate. This edge crack fracture suggests that the cable loading was not symmetric at Cable Diaphragm Plate 8. The critical edge crack length at fracture was about 7.5in. The high frequency vibration observed at Cable Diaphragm Plate 7 provides a loading rate much higher than that associated with the 1 second loading rate used for bridge structures subjected to truck traffic. This is an order of magnitude greater which contributes to the brittle fracture in the lower fracture toughness direction provided by TL CVN tests.
13
6. Metallography
Weld macro-cross-sections were prepared from the segment removed from the north (left) ring plate of Cable Diaphragm Plate 8 (see Figure 25) to determine if crack development had occurred at the fillet weld toe. Figure 40 shows the etched surfaces of the segment. No weld toe cracking was detected in either of the cross-sections.
A weld cross-section was also prepared from the diaphragm-flange weld at the right side of Cable Diaphragm Plate 9 to characterize the weld geometry. The cross-section was located near the weld termination. Figure 41 shows the etched cross-section. The weld joint consists of two multi-pass fillet welds. No weld defects were observed in the crosssection.
14
7. Summary and Conclusions
1. Fatigue crack growth developed at Cable Diaphragm Plates 8 and 9 as a result of low fatigue strength weld details at four locations.
2. The cyclic stress range driving the crack development and growth was due to the vibrations of the stay cables from wind. The high frequency response resulted in large numbers of stress cycles in only a few years of service that exceeded 0.5 of the Fatigue Limit. This was verified by strain gage measurements on Cable Diaphragm Plate 7.
3. Since the cable tension forces were always present, all stress cycles were in tension at high R ratios at the top and bottom surfaces of the diaphragm plates.
4. Both Cable Diaphragm Plates 8 and 9 showed evidence of crack development mainly from the top surface of the diaphragm plates. This indicates that the applied cable forces created some initial bending in the diaphragm plates due to the plate geometry.
5. At the end welds that were wrapped around the diaphragm plate some embedded slag and porosity existed. These did not influence the weld toe cracks that formed in the diaphragm plate. As observed at left side of Cable Diaphragm Plate 8 these toe cracks could form in more than one plane.
6. Although corrosion damage prevented detection of crack growth striations near the weld toes on the micro level, fatigue beach marks and the crack surface characteristics verified that stable fatigue crack propagation existed over most of the crack surface at Cable Diaphragm Plate 9 and at the north (left) side of Cable Diaphragm Plate 8 up to the left ring plate.
7. The plate material used for the diaphragms had tensile properties that were in compliance with the Specifications. The CVN test results showed that the material exceeded the toughness required for Zone 3.
15
8. The brittle fracture that occurred at Cable Diaphragm Plate 8 suggests that an eccentric load contributed to fatigue crack growth and initiated a brittle fracture in the lower toughness direction.
9. The high frequency of the fatigue loading is a significant order of magnitude greater than normally considered for bridge structures. This has reduced the fracture resistance of the diaphragm material.
16
8. References 1) AASHTO LRFD Bridge Design Specifications, Section 6.6, Table 6.6.1.2.3-1 section 7, 2009 (see also the AISC Specifications, Appendix 3 Fatigue Design, 2005) 2) Barsom, J.M. and Rolfe, S.T., Fracture and Fatigue Control in Structures, 3rd Edition, ASTM, West Conshohocken, PA 1999
17
Figure 1: View of fracture surface of Cable Diaphragm Plate 9 showing crack growth from weld toes of ring plates and end welds
Figure 2: View of cracked Cable Diaphragm Plate 9 before removal from pylon
18
9A 9B 9C 9D 9E
9F
Figure 3: View of Cable Diaphragm Plate 9 segments with ring plates. Note markings for mechanical test samples (black marks) and fractographic samples (red marks). Fractographic sample segments are indicated 9A through 9F from left to right.
Figure 4: View of Cable Diaphragm Plate 8 after cables were removed showing fracture, crack, and flange plastic distortion
19
8R4 8L 8R3
8R2
8R1
Figure 5: View of Cable Diaphragm Plate 8 after removal from pylon. Note markings for mechanical and fractographic test samples. Fractographic test sample segments are indicated as 8L, 8R1, 8R2, 8R3, and 8R4.
Edge Crack Primary Crack Surfaces
9A 9E
9F
9B
9C
9D
Figure 6: View of segments of Cable Diaphragm Plate 9 looking at the top surface
20
9D
Figure 7: Cleaned crack surface of Segment 9D (middle segment) with the top surface corresponding to the top surface in Figure 6.
9E
Figure 8: Cleaned crack surface Segment 9E (the smaller segment at the ring plate weld toe adjacent to the surface seen in Figure 7).
21
9F
Figure 9: View of crack surface of Segment 9F propagating from the end weld of diaphragm-flange weld
Mating to 9D
Mating to 9E
Mating to 9F
Figure 10: Mating crack surface of right side of Cable Diaphragm Plate 9 showing coalesced edge crack. The approximate mating surfaces of Segments 9D, 9E, and 9F are indicated.
22
9C
Figure 11: View of cleaned surface of Segment 9C at the left ring plate weld toe showing multiple initiation sites and coalescence of the cracks
9B
Figure 12: View of primary crack surface in Segment 9B on the left side of Cable Diaphragm Plate 9 in front of edge crack
23
Mating to 9B Mating to 9C
Mating to 9D
Figure 13: Overview of mating crack surface at left ring plates for Cable Diaphragm Plate 9 showing fracture segments on each side of the primary crack. The approximate limits of the mating surfaces for Segments 9B, 9C, and 9D are indicated.
24
9A
Figure 14: Exposed and cleaned crack surface of edge crack in Segment 9A behind the primary crack showing corrosion product on surface and crack growth from the weld toe
Back of 9B
Figure 15: Mating crack surface to Segment 9A of the edge crack in Segment 9B behind the primary crack. 25
8L
8R3
8R2
8R1
Figure 16: View of segments cut from Cable Diaphragm Plate 8
8R3
Mating 8R3 Surface (Segment 8R4)
Figure 17: View of cleaned crack surface of diaphragm to flange edge crack in Segment 8R3 and the mating crack face (Segment 8R4) showing the end weld at the flange
26
Fatigue Crack Growth 8R2
Fracture
Figure 18: Cleaned fatigue crack surface of Segment 8R2 at the south (left) ring plate and adjacent fracture
8R2
Figure 19: Overview of the unstable fracture that broke the diaphragm plate in Segment 8R2
27
Figure 20: Primary crack surfaces of Segment 8L separated at reduced temperature
Figure 21: Secondary crack on bottom surface of Segment 8L parallel to the primary crack surface. Note this is a view of the bottom side of the lower piece in Figure 20.
28
Figure 22: View of primary exposed fatigue crack surface of Segment 8L originating at the weld toe and propagating as a elliptical surface crack from the top surface
Figure 23: View of matching crack surface of Segment 8L; secondary crack seen in Figure 21 lies behind the primary crack surface and propagates from the bottom surface
29
8L
Figure 24: Cleaned crack surface of Segment 8L showing fatigue beach marks and rubbing near bottom of crack
Figure 25: showing where segments were cut from the left side of Cable Diaphragm Plate 8 at the ring plates and crack face from the right side of the plate
30
#9
R
Long. CVN (12)
Trans. CVN (12)
9L1-12
~8
9T1-12 Trans. 0.505 Tensile (1) 9T Long. 0.505 Tensile (1) 9L
~23 Figure 26: Test Specimen Layout for Diaphragm Plate 9.
31
L
~8
#8
Long. CVN (12)
#8
Trans. CVN (12) 8T1-12
8L1-12
~5
Trans. 0.505 Tensile (1) 8T Long. 0.505 Tensile (1) 8L
~10
~9
Figure 27: Test Specimen Layout for Diaphragm Plate 8.
32
Figure 28: Diaphragm Plate CVN Test Results.
33
100 mm 22 mm 3 mm x
8L
Figure 29: Cleaned crack surface of Segment 8L with location of region where fatigue striations were detected is identified with an x
34
2 5
Figure 30: SEM view of the fatigue crack surface of the Segment 8L crack at the location shown in Figure 29. Magnification 4000X
Figure 31: SEM view of Segment 8L at weld toe crack origin @11X
35
Figure 32: Enlarged view of Figure 31 showing corrosion product on crack surface 8L @300X
5m 2m x
9C
Figure 33: View of crack surface of Segment 9C at ring plate weld showing beach marks and multiple initiation sites from weld toe on top surface
36
Figure 34: SEM (1000X) of crack surface of Segment 9C at point x in Figure 33
Figure35: SEM view of weld toe region @10X of Segment 9C showing multiple initiation sites of fatigue crack growth
37
Figure 36: SEM view of crack initiation from the weld toe and beach marks at Segment 9C @ 11X
Figure 37: SEM view of crack surface in Figure 36 showing fatigue fracture surface @300X
38
4000 SG01
2000
Stress (psi)
-2000
-4000 9:06:02 AM
9:06:03 AM
9:06:04 AM
9:06:05 AM
9:06:06 AM
Time
Figure 38: Stress response from strain gage SG01 located adjacent to the left end diaphragm plate to pylon flange weld toe at Cable Diaphragm Plate 7 for a wind speed of 5 mph to 8 mph at a direction of about 20 degrees. Note 0 degrees is taken as perpendicular left of the structure with increasing increments clockwise looking down.
4000 SG02 SG08
2000
Stress (psi)
-2000
-4000 9:06:02 AM
9:06:03 AM
9:06:04 AM
9:06:05 AM
9:06:06 AM
Time
Figure 39: Stress response from strain gages SG02 and SG08 for the wind event shown above. SG02 is positioned on top of Cable Diaphragm Plate 7 normal to the ring plate weld toe at the midpoint between the line of action of the cable and the diaphragm plate to pylon flange weld termination. SG08 is positioned on the bottom of the plate opposite SG02.
39
Figure 40: Polished and etched sections of the diaphragm plate and ring plate weld toes
Figure 41: Cross section of the diaphragm-flange welds at the right side of Cable Diaphragm Plate 9 (Segment 9F) showing weld profile
40
APPENDIX A
41
42
43
44
45
46
47
Appendix B
Instrumentation Plan Drawings
Appendix C
Effective Stress Range Histograms
Martin Olav Sabo Pedestrian Bridge Stress Range Histogram Data

Bin (psi) SG01 SG02 SG03 SG04 SG05 SG06 SG07 SG08 SG09 SG10 125 10879221 10561097 11017861 2487208 5011781 10434023 1496533 1862868 846270 225 16957203 5593941 3921012 2707615 3605996 3258355 2135496 2575925 1524606 375 1538857 1457515 1293846 1097399 1250238 1193037 1119642 1171466 740484 525 1035787 849274 868486 683149 766490 836739 937415 744065 531266 675 878703 601253 758811 499993 569604 674244 775013 550897 467547 825 690680 478765 661781 431348 452993 528986 600540 463148 419582 975 539842 438561 543004 395594 402514 410101 480999 429150 358394 1125 453882 405973 448163 357874 376229 331629 429557 397599 297242 1275 388080 384203 376640 328406 347930 292177 367077 382162 250696 1425 312993 361844 331966 281334 324742 250921 288511 354685 216486 1575 231351 333938 289711 243838 295115 196507 206853 326536 199949 1725 171958 299684 252858 217039 272170 137302 156611 290487 170159 1875 125472 267264 210135 189764 239988 101257 111777 260560 147463 2025 83435 243729 171197 176739 211392 68783 71610 240473 120817 2175 56609 223571 143815 168477 185686 45992 50072 219583 94009 2325 38805 203514 120975 150375 167431 30781 33818 199198 75558 2475 26416 192051 94492 140088 154386 19990 22995 191875 63596 2625 18137 187248 70619 128109 138651 12732 15905 188762 51757 2775 12032 176701 55207 101846 127783 7674 10554 174953 36973 2925 8131 162181 43416 81607 125129 4583 6919 158936 28759 3075 5697 150092 33091 70622 122174 2162 5082 148332 24245 3225 4098 137042 25158 58581 113735 947 3721 133251 18365 3375 3062 122108 18071 46460 103478 457 2769 117640 13936 3525 2215 104594 12917 37568 99073 209 1975 99605 10554 3675 1659 87900 9195 28875 94663 111 1392 82016 8031 3825 1088 75792 6669 23972 84913 58 1007 72219 5790 3975 783 67621 5196 18064 71592 21 599 65239 3651 4125 483 59302 4025 14193 59858 7 372 55658 2222 4275 304 50849 3038 10849 52022 3 230 47455 1588 4425 174 44276 2485 8330 48884 0 97 41210 1037 4575 82 37423 1975 6361 44538 0 50 33803 737 4725 46 30400 1521 4161 38213 0 20 26541 477 4875 10 24284 1232 3392 31858 0 16 21626 368 5025 6 20278 873 2449 27621 0 5 18638 238 5175 8 17958 710 1570 23046 0 2 16485 156 5325 1 15367 519 879 19283 0 0 14178 81 5475 0 13213 393 492 16944 0 0 11852 60 5625 0 11351 266 265 15317 0 0 10551 28 5775 0 9892 169 195 12751 0 0 9023 13 5925 0 8497 117 133 10509 0 0 7565 7 6075 0 7177 67 80 9128 0 0 6306 3 6225 0 6116 39 54 7640 0 0 5434 1 6375 0 5225 18 31 6600 0 0 4668 1 6525 0 4422 6 18 5487 0 0 3934 0 6675 0 3790 5 8 4539 0 0 3221 0 6825 0 3042 6 4 3642 0 0 2674 0 6975 0 2677 2 2 3055 0 0 2361 0 7125 0 2255 0 1 2645 0 0 2102 0 7275 0 2026 0 0 2153 0 0 1781 0 7425 0 1831 0 0 1676 0 0 1513 0 7575 0 1446 0 0 1199 0 0 1308 0 7725 0 1309 0 0 903 0 0 1178 0 7875 0 1144 0 0 642 0 0 995 0 8025 0 1016 0 0 502 0 0 852 0 8175 0 864 0 0 364 0 0 787 0 8325 0 724 0 0 263 0 0 585 0 8475 0 669 0 0 188 0 0 533 0 8625 0 514 0 0 163 0 0 409 0 8775 0 439 0 0 115 0 0 398 0 8925 0 409 0 0 69 0 0 292 0 9075 0 292 0 0 64 0 0 193 0 9225 0 236 0 0 53 0 0 182 0 9375 0 193 0 0 40 0 0 140 0 9525 0 164 0 0 28 0 0 112 0 9675 0 110 0 0 24 0 0 81 0 9825 0 95 0 0 10 0 0 63 0 9975 0 80 0 0 7 0 0 36 0 10125 0 42 0 0 14 0 0 27 0 10275 0 39 0 0 3 0 0 16 0 10425 0 22 0 0 2 0 0 20 0 10575 0 16 0 0 2 0 0 11 0 10725 0 19 0 0 3 0 0 9 0 10875 0 9 0 0 1 0 0 5 0 11025 0 6 0 0 1 0 0 6 0 11175 0 3 0 0 1 0 0 6 0 11325 0 7 0 0 0 0 0 3 0 11475 0 4 0 0 0 0 0 0 0 11625 0 2 0 0 0 0 0 0 0 11775 0 0 0 0 0 0 0 2 0 11925 0 1 0 0 0 0 0 0 0 12075 0 0 0 0 0 0 0 0 0 SG11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 864833 1628493 808117 520307 372985 295635 255786 222577 205311 189736 172335 158565 151343 133751 114282 103046 97997 90658 78217 73027 77495 75521 70199 62251 59938 57618 53039 45617 38224 33732 31973 29141 25637 21299 18603 15767 12559 11154 10799 9476 7504 6069 5571 4359 3636 3159 2447 1868 1552 1453 1179 819 540 375 205 125 84 40 33 17 3 4 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 SG12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Page 1 of 2
Martin Olav Sabo Pedestrian Bridge Stress Range Histogram Data

Bin (psi) SG13 SG14 SG15 SG16 SG17 SG18 SG19 SG20 SG21 125 796002 2395682 1751353 1454193 3700324 7472074 1536725 2242135 225 1370457 3012589 2041722 1756740 3458512 3171206 1951819 2732852 375 678138 1014500 838810 690557 970689 802260 829588 944521 525 270028 599071 571922 437234 550615 466187 561488 583538 675 137823 436209 430151 318479 389705 354004 422871 428591 825 77730 349140 315099 273880 290328 289903 338343 349605 975 42717 294804 209917 231220 237097 253153 245598 285089 1125 22261 244659 137362 214052 209396 190639 168705 243866 1275 10434 215209 97825 185233 179852 132400 112639 205613 1425 4267 181300 74285 143593 166319 93906 83856 163273 1575 1214 144589 58211 106719 154359 69026 65065 126934 1725 307 115735 44988 81907 134330 50020 52326 94186 1875 109 87479 35795 64856 110976 36434 41681 72097 2025 32 68098 28452 50288 90288 26000 32784 58665 2175 8 56247 22034 38451 71124 17591 26608 49385 2325 1 47442 17174 29425 58307 10763 20913 41653 2475 0 40658 13113 22622 48688 6184 16195 35249 2625 0 34993 9624 16763 41165 3343 12911 30347 2775 0 30706 7144 12090 33793 1792 9890 25543 2925 0 26353 5131 7174 28259 979 7546 21761 3075 0 22588 3693 4074 23480 528 5569 18113 3225 0 19313 2465 2421 19889 236 4157 15127 3375 0 16588 1708 1376 16471 160 2906 12578 3525 0 13969 1019 853 13282 62 2115 10340 3675 0 11810 555 531 10856 47 1354 8770 3825 0 9984 293 278 8301 37 868 7334 3975 0 8569 175 171 6225 10 476 5890 4125 0 7190 105 100 4719 9 297 4948 4275 0 6099 57 49 3683 1 154 3945 4425 0 5085 50 38 2657 1 108 3127 4575 0 4187 23 38 1902 0 61 2458 4725 0 3463 18 11 1413 0 45 1970 4875 0 2803 4 8 960 0 34 1476 5025 0 2221 4 4 666 0 18 1114 5175 0 1802 3 2 514 0 12 796 5325 0 1472 1 0 335 0 6 536 5475 0 1104 0 0 233 0 2 376 5625 0 817 0 0 165 0 4 265 5775 0 609 0 0 95 0 1 157 5925 0 377 0 0 88 0 0 116 6075 0 311 0 0 46 0 0 82 6225 0 206 0 0 40 0 0 63 6375 0 136 0 0 29 0 0 47 6525 0 101 0 0 19 0 0 31 6675 0 76 0 0 22 0 0 25 6825 0 59 0 0 19 0 0 21 6975 0 42 0 0 8 0 0 14 7125 0 28 0 0 3 0 0 10 7275 0 27 0 0 4 0 0 9 7425 0 24 0 0 5 0 0 1 7575 0 12 0 0 0 0 0 3 7725 0 14 0 0 1 0 0 3 7875 0 8 0 0 0 0 0 1 8025 0 5 0 0 0 0 0 1 8175 0 1 0 0 0 0 0 0 8325 0 3 0 0 0 0 0 0 8475 0 3 0 0 0 0 0 0 8625 0 1 0 0 0 0 0 0 8775 0 1 0 0 0 0 0 0 8925 0 0 0 0 0 0 0 0 9075 0 0 0 0 0 0 0 0 9225 0 0 0 0 0 0 0 0 9375 0 0 0 0 0 0 0 0 9525 0 0 0 0 0 0 0 0 9675 0 0 0 0 0 0 0 0 9825 0 0 0 0 0 0 0 0 9975 0 0 0 0 0 0 0 0 10125 0 0 0 0 0 0 0 0 10275 0 0 0 0 0 0 0 0 10425 0 0 0 0 0 0 0 0 10575 0 0 0 0 0 0 0 0 10725 0 0 0 0 0 0 0 0 10875 0 0 0 0 0 0 0 0 11025 0 0 0 0 0 0 0 0 11175 0 0 0 0 0 0 0 0 11325 0 0 0 0 0 0 0 0 11475 0 0 0 0 0 0 0 0 11625 0 0 0 0 0 0 0 0 11775 0 0 0 0 0 0 0 0 11925 0 0 0 0 0 0 0 0 12075 0 0 0 0 0 0 0 0 SG22 SG23 7220388 3106022 477319 222792 148117 125138 109876 108208 94772 72985 50963 35654 26509 19268 13953 10958 8180 5773 3116 1421 718 334 192 102 52 16 13 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 SG24 3334555 3166097 918085 520488 367799 275389 229836 205196 177830 167044 154981 132816 107734 87797 68834 56855 48253 39712 32864 26795 22462 18630 14938 11770 9463 6902 4808 3502 2439 1744 1122 817 596 377 258 169 124 82 52 33 33 15 29 6 10 5 4 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Page 2 of 2
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
Note: 18 Day Monitoring Period
SG01
Stress Bin (psi)
Sre = 2,720 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG02
Stress Bin (psi)
Sre = 3,660 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG03
Stress Bin (psi)
Sre = 2,900 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG04
Stress Bin (psi)
Sre = 3,060 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG05
Stress Bin (psi)
Sre = 3,810 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG06
Stress Bin (psi)
Sre = 2,560 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG07
Stress Bin (psi)
Sre = 2,710 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG08
Stress Bin (psi)
Sre = 3,620 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG09
Stress Bin (psi)
Sre = 2,870 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG11
Stress Bin (psi)
Sre = 3,900 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG13
Stress Bin (psi)
Sre = 2,330 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG14
Stress Bin (psi)
Sre = 3,290 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG15
Stress Bin (psi)
Sre = 2,710 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG16
Stress Bin (psi)
Sre = 2,650 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG17
Stress Bin (psi)
Sre = 3,050 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG18
Stress Bin (psi)
Sre = 2,530 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG19
Stress Bin (psi)
Sre = 2,790 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG20
Stress Bin (psi)
Sre = 3,160 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG22
Stress Bin (psi)
Sre = 2,560 psi
Counts 100000 120000 140000 160000 180000 200000 20000 80000 40000
60000
0 125 375 675 975 1275 1575 1875 2175 2475 2775 3075 3375 3675 3975 4275 4575 4875 5175 5475 5775 6075 6375 6675 6975 7275 7575 7875 8175 8475 8775 9075 9375 9675 9975 10275 10575 10875 11175 11475 11775
SG24
Stress Bin (psi)
Sre = 2,980 psi
Appendix D
Preliminary Retrofit Drawings

Sabo Bridge Report

Uploaded by

Copyright:

Available Formats

Sabo Bridge Report

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sabo Bridge Report

Uploaded by

Copyright:

Available Formats

MARTIN OLAV SABO PEDESTRIAN BRIDGE Cable Diaphragm Plate Fracture Investigation

Minneapolis Bridge #6925, State Bridge #27B35 Minneapolis, Minnesota

Final Report June 28, 2012 WJE No. 2012.0901

2.0 BACKGROUND 2.1 Structure Description

2.2 Parties Involved in the Design and Construction

2.3 Summary of Diaphragm Plate Failures and Structure Stabilization

3.0 SCOPE OF WORK

4.0 NONDESTRUCTIVE TESTING

4.1 Nondestructive Test Methods

4.2 Nondestructive Test Results

5.0 PHYSICAL MEASUREMENTS

5.1 Cable Diaphragm Plate Alignment

5.2 Pylon Flange Plastic Rotation at Cable Diaphragm Plate 8

6.0 FRACTURE INVESTIGATION

6.1 Summary of Dr. Fishers Assessment

6.2 Summary of Dr. Fishers Findings

7.0 INSTRUMENTATION AND MONITORING PROGRAM

7.1 Instruments and Installation Locations

7.2 Data Acquisition System

7.3 Pluck Tests

7.4 Monitoring Period

7.5 Effective Stress Range Calculation

8.0 PEER REVIEW OF ORIGINAL BRIDGE DESIGN

9.0 REVIEW OF WIND-INDUCED STAY CABLE VIBRATION PHENOMENA

10.0 STRUCTURAL ANALYSIS OF DIAPHRAGM PLATES

10.1 Stresses Due to Cable Vibrations

10.2 Stresses Due to Plate Misalignment

12.2 Cable Diaphragm Plate Retrofit Options

Table 1. Primary Equipment for Nondestructive Examination of Cable Diaphragm Plates

Straight beam scans

Table 2. Cable Diaphragm Plate Alignment Measurements

Table 3. Summary of Pluck Test Data

Table 4. Representative Triggered Event Data

* max of 38 mph at pylon top

Table 5. Summary of Effective Stress Range Data

No Data No Data No Significant Data

Figure 3. Typical cable anchorage assembly at the concrete deck.

Figure 5. Typical cable earth anchorages.

Figure 7. Typical cable diaphragm plate. Cable Diaphragm Plate 6 shown.

Figure 8. Magnetic particle testing (MT) in progress showing test equipment.

Figure 9. Ultrasonic testing (UT) in progress.

Figure 13. View of Cable Diaphragm Plate 9 after fracture.

6000 SG02 SG08 4000

6000 SG02 SG08 4000

6000 SG05 SG11 4000

6000 SG05 SG11 4000

6000 SG14 SG20 4000

6000 SG14 SG20 4000

6000 SG17 SG23 4000

6000 SG17 SG23 4000

0.3 Accel01 0.2

6000 SG02 4000

6000 SG11 4000

0.3 Accel01 0.2

0.3 Accel03 0.2

Figure 46. Finite element model of cable diaphragm plate.

John W. Fisher Report

b) Cable Diaphragm Plate 8

2. Examination of the Crack Surface Segments