The Persistence of Failure Historic Bridge Collapses

Henry Petroski Distinguished Professor Emeritus of Civil Engineering Duke University

• • • Historic Failures of Iron and Steel Bridges

• What are the Landmark Cases? • Do They Reveal a Pattern? • Can It Predict Future Collapses? Landmark Bridge Failures

(some still being revisited) North British Railway, Edinburgh-to-Dund ee, ca. late 1870

Chester and Holyhead Railway, Ca. 1845 Dee Bridge Collapse 1847 Collapsed Dee Bridge Trussed Girder

Revisiting the Dee Bridge Failure

Peter Lewis, Open University, ca. 2004 Tay Bridge Collapse 1879

Tay Bridge, 1879 Tay Bridge, 1879

Dec. 28, 1879 Note smokestacks still standing in Dundee.

In 2002 Peter Lewis digitized 1880 photos

Replacement Tay Bridge (left), 1887

Bridge Proposed to Cross the Firth of Forth (Design Abandoned after Tay Bridge Collapse) Forth Railway Bridge

Human Model, late 1880s Forth Rail Bridge, 1890

Quebec Bridge Collapse 1907 Quebec Bridge, 1907

(Second) Quebec Bridge, 1917 Tacoma Narrows Bridge Collapse 1940 Tacoma Narrows Bridge, 1940

Tacoma Narrows Bridge, 1940 Collapse 1967 Silver Bridge, 1928 Silver Bridge Collapse December 15, 1967

Milford Haven Bridge Collapse 1970 Milford Haven Bridge

June 1970 Cleddau (Milford Haven) Bridge,1975

Westgate Bridge

October 1970

367-foot span being lifted onto piers Landmark Bridge Failures Is there a pattern?

• 1847 - Dee Bridge, trussed girder • 1879 - Tay Bridge, through truss • 1907 - Quebec Bridge, cantilever • 1940 - Tacoma Narrows Bridge, suspension • 1967 - Silver Bridge, eye-bar suspension • 1970 - Milford Haven Bridge, box girder

Landmark Bridge Failures Occur at 30-year Intervals (Paul Sibley & Alastair Walker, 1977)

• 1847 - Dee Bridge 0 • 1879 - Tay Bridge 32 • 1907 - Quebec Bridge 28 • 1940 - Tacoma Narrows Bridge 33 • 1967 - Silver Bridge 27 • 1970 - Milford Haven Bridge 30 Why 30 Years?

• Time for a design type to become routine

• Tasked to inexperienced engineers

• Lax oversight by senior engineers

• Design assumptions forgotten or unknown • Time for an innovation to reach limits

• Range of a model’s validity exceeded • Time for lessons learned to be forgotten

• Time span of an engineer’s career

• Time frame of institutional memory Pedestrian Bridges Passerelle Solferino, 1999 Ironically, innovative designs can fare better than supposedly tried-and-true ones. Gateshead Millennium Bridge, 2001

The London Millennium Bridge is a . CABLE SAG (ratio of drop to span length)

Typical: 1/8 to 1/12 Millennium Bridge: >1/60

London Millennium Bridge, opened 2000 (closed after 3 days; reopened 2002)

Failure Without Collapse

Landmark Bridge Failures Continuation of a 30-year Period

• 1847 - Dee Bridge, trussed girder • 1879 - Tay Bridge, through truss • 1907 - Quebec Bridge, cantilever • 1940 - Tacoma Narrows Bridge, suspension • 1967 - Silver Bridge, eye-bar suspension • 1970 - Milford Haven Bridge, box girder • 2000 - London Millennium Bridge, pedestrian Minneapolis (I-35W) Bridge Collapse, 2007 Minneapolis I-35W Bridge

Collapsed Aug. 1, 2007 Gusset plate Replacement Bridge

St. Anthony Falls Bridge opened Sept. 18, 2008 Chronology of Landmark Bridge Failures

• 1847 - Dee Bridge, trussed girder • 1879 - Tay Bridge, through truss • 1907 - Quebec Bridge, cantilever • 1940 - Tacoma Narrows Bridge, suspension • 1967 - Silver Bridge, eye-bar suspension • 1970 - Milford Haven Bridge, box girder • 2000 - London Millennium Bridge, pedestrian • 2007 - I-35W Minneapolis, deck truss • 2025 to 2030 ??? Time Interval (in years) Between Landmark Bridge Failures

• 1847 - Dee Bridge • 1879 - Tay Bridge 32 1907 - Quebec Bridge 28 • 1940 - Tacoma Narrows Bridge 33 • 1967 - Silver Bridge 27 • 1970 - Milford Haven Bridge 30 • 2000 - London Millennium Bridge 30 • 2007 - I-35W Minneapolis 37 • ? 2030 to 2035 - ??? ? Cycle of Success and Failure

• Failed bridge type made suspect • New designs introduced with caution • Continuing success gives confidence • Caveats and fears forgotten or ignored • Technology pushed beyond limits • Failure provides wake-up call • Failure analysis and recriminations • Design changes called for and made • The pattern repeats What Will Be the Next Landmark Bridge Failure? (1990s; NASA JSC, 2009)

• Cable-Stayed Bridge?

■ Instability During Construction?

■ Cable Fatigue? • Precast Concrete Box-Girder Bridge?

■ Unbalanced Cantilever Construction?

■ Post-Tensioning Instability? Cable-Stayed Bridges

• Post-War Development in Europe • Intended for Moderate (<1200-ft) Spans • Allow for Creative Cable Arrangements • Tend to be Signature Spans • Increasingly Longer Spans Built • Problems with Cable Vibration Persist • Behavior Incompletely Understood A Gallery of Cable-Stayed Bridges Sunshine Skyway Bridge, 1987 1,200 ft Sunshine Skyway Bridge, 1987 (1,200-ft)

Dames Point Bridge, Jacksonville, Florida 1989, 1,300-ft main span Alamillo Bridge (Santiago Calatrava), 1992 Pont de Normandie

1995

856 m (2,800 ft) main span

Michel Virlogeux, engineer Zakim-Bunker Hill Bridge, Boston 2003; 745-ft main span; 183-ft wide Zakim-Bunker Hill Bridge

Millau Viaduct, 2004 1,122 ft. longest span 1,104 ft. max pylon height Cooper River Bridges, Charleston, S.C.

Ravenel Bridge 2005 1,546-ft main span Penobscot Narrows Bridge & Observatory 2006, 1,161-ft main span Penobscot Narrows Bridge & Observatory Cycle of Success and Failure

• New designs introduced with caution • Continuing success promotes confidence • Caveats and fears forgotten or ignored • Technology pushed to limits and beyond • A failure provides a wake-up call • Failure analysis, recriminations occur • Design changes called for • New designs introduced with caution From Success to Failure

Henry Petroski Distinguished Professor Emeritus of Civil Engineering Duke University “Those who cannot remember the past are condemned to repeat it.”

--George Santayana Suspension Bridges

Success Through Failure Union Chain Bridge

England-Scotland, 1820

450-ft main span Brighton Chain Pier, ca. 1824 (c. 250-ft. spans) Brighton Chain Pier, 1837 Menai Suspension Bridge, 1827 (577 ft.) John Roebling (1806-1869) “Storms are unquestionably the greatest enemies of suspension bridges.”

--John Roebling 1841 Niagara Gorge Railway Suspension Bridge 1854 (825 ft.) What made the Niagara Bridge work for railway traffic?

“Weight, Girders, Trusses, and Stays”

--John Roebling, Final Report (1855) weight, stiffness, stays “If your bridge succeeds, mine is a magnificent blunder”

--Robert Stephenson to John Roebling British

American Bridge, 1867 (1,057 ft.) additional steel cables, new truss added, 1890s (const. 1869-1883) Emily Warren Roebling (1843-1903)

Washington A. Roebling (1837-1926) Pneumatic caisson, 1870 1883

1,595 ft.

Williamsburg Bridge, 1903 (1,600 ft., no stays)

Manhattan Bridge, 1909 (lightweight, no stays) Benjamin Franklin Bridge, 1926 (1,750 ft.) Ambassador Bridge (Detroit), 1929 (1,850 ft.) George Washington Bridge

Othmar Ammann

1931

3,500 ft. lightweight, no truss, no stays (3,500 ft.-span) GWB with lower deck added (1962) GoldenGolden Gate Bridge,Gate Bridge, 1937 (4,2001937 (4200 ft.) ft.)

Thousand Islands Bridge, 1939 (800 ft.) Deer Isle Bridge, 1939 (1,088 ft.) c. 1945 Bronx-Whitestone Bridge, 1939 (2,300 ft.)

Note stay cables, stiffening truss (1947) Stiffening truss added 1947 . . . removed mid-2000s

Tacoma Narrows Bridge, July 1940 2,800 ft., third longest span in the world Opening day, September 1940

Tacoma Narrows Bridge, November 1940

Literally narrow, shallow, and long (lightweight), no truss (flexible), no stays (unconstrained) What made the Niagara Bridge work for railway traffic?

Weight, Stiffness, and Stays Lessons Learned (1941 report)

• wind is the enemy of suspension bridges • flexible decks are vulnerable to the wind

“Progress, far from consisting in change, depends on retentiveness. . . . “Those who cannot remember the past are condemned to repeat it.”

--George Santayana