1974_美国_HAR7402.pdf

8 NATIONAL TRANSPORTATION SAFETY BOARD

HIGHWAY ACCIDENT REPORT

Automobile Crash Off The Silliman Evans Bridge, 1-24/65 Nashville, Tennessee July 27, 1973

ASF. EDT AUG 1 NATIONAL TRANSPORTATATIA SAFETY BOARD

NATIONAL TRANSPORTATION SAFETY BOARD Washington, D.C. 20591

REPORT NUMBER: NTSB-HAR-74-2

NATIONAL TRANSPORTATION SAFETY BOARD Washington, D.C. 20591

SS-H-28

HIGHWAY ACCIDENT REPORT

Automobile Crash Off The Silliman Evans Bridge, I-24/65 Nashville, Tennessee July 27, 1973

Adopted: June 13, 1974

ERRATUM

Please make the following change to subject report:

Page 28, second paragraph, line 6 - change "60" to "600".

August 1, 1974

Report Number: NTSB-HAR-74-2

SS-H-28

HIGHWAY ACCIDENT REPORT

Automobile Crash Off The Silliman Evans Bridge, I-24/65 Nashville, Tennessee July 27, 1973

Adopted: June 13, 1974

NATIONAL TRANSPORTATION SAFETY BOARD Washington, D.C. 20591 Report Number: NTSB-HAR-74-2

TECHNICAL REPORT STANDARD TITLE PAGE

FOREWORD

The accident described in this report has been designated as a major accident by the National Transportation Safety Board under the criteria established in the Safety Board's regulations.

This report is based on facts obtained from an investigation performed by the Safety Board. Cooperation in the investigation was received from the Metropolitan (Nashville and Davidson County, Tennessee) Police Department, the Tennessee Department of Transportation and the Federal Highway Administration (FHWA).

The conclusions, the determination of probable cause, and the recommendations herein are those of the Safety Board.

TABLE OF CONTENTS

Page

SYNOPSIS FACTS. $\ldots$ 13

The Accident. 3

Accident Site .................................................. 3

Marks on the Roadway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Environmental Factors 16

Accident History of the Silliman Evans Bridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Automobile. 17

Automobile Occupants. 17

Postcrash Activities. 18

Applicable Standards and Laws 19

ANALYSIS 20

Crash Events. 20

Identification and Surveillance of Accident Locations . . 24

CONCLUSIONS. 25

PROBABLE CAUSE 28

RECOMMENDATIONS. 29

appendices

Appendix A: Crash Damage to the Automobile. 31

Appendix B: The Box-beam Barrier. 32

Appendix C: Safety Board Letter of Recommendation to Gov. Winfield Dunn of Tennessee . . . . . . 34

Appendix D: Summary of Accident Experience on the Silliman Evans Bridge $\ldots$ 39

NATIONAL TRANSPORTATION SAFETY BOARD WASHINGTON, D.C. 20591

HIGWAY ACCIDENT REPORT

Adopted: June 13, 1974

Automobile Crash off the Silliman Evans Bridge, I-24/65, Nashville, Tennessee, July 27, 1973

SYNOPSIS

Shortly before 5:45 a.m., on July 27, 1973, in a light ground fog or haze, a car crashed through the barrier system on the edge of the Silliman Evans Bridge, I-24/65, in Nashville, Tenn. There were nine persons in the automobile; seven passengers and the driver died. None of the occupants was ejected.

The automobile was traveling on the off-ramp from eastbound I-40 onto northbound I-24/65. When it reached the vicinity of the juncture of the two roadways, the car deviated from the pathway provided and struck the curb face of a raised concrete island. Impact with the island caused the driver to lose control of the automobile, which then traveled north-east across the two lanes of I-24/65 and struck and displaced a boxbeam guardrail located on a concrete curb in front of the bridge rail. As the right front wheel of the car struck and mounted the curb, the automobile moved forward with dynamic uplift, overrode the box-beam, plunged through the bridge rail, and fell 65 feet to the ground.

The pavement on the bridge was dry. There was no evidence of excessive speed, and no other traffic was present at the time of the crash. Alcohol was not involved.

The National Transportation Safety Board determines that the probable cause of this accident was the failure of the driver to maintain her vehicle in the pathway provided.

The following are contributing factors which may have confused the driver: (a) increasing curvature of the ramp; (b) narrowing of the ramp from two lanes to one; (c) a misleading traffic-control sign; (d) misleading pavement markings; and (e) an inadequately delineated and unnecessary section of a concrete island which narrowed the pathway to less than the width of a normal traffic lane--all in the last 600 feet of the ramp.

This confusion probably contributed to the striking of the concrete island, which caused loss of control of the automobile and its subsequent impact with the bridge rail barrier system.

The fall of the automobile to the ground and the resultant fatalities were caused by the failure of the barrier systems to retain or redirect the errant vehicle. Contributing to this failure were: (a) an unnecessary 9-inch-high curb, which provided the impacting vehicle with dynamic uplift; (b) improper installation of a box-beam median-type barrier system rather than a more rigid system; (c) redesign of the box-beam mounting method, which did not secure the beam to its base; (d) inadequate maintenance of the box-beam; and (e) an outside bridge rail not adequate to withstand the vehicle impact.

The probable cause of the failure to detect and correct these hazards was the absence of a systematic program for the identification, surveillance, and correction of potential, or existing, hazardous conditions on the highway.

-3-

FACTS

The Accident

At about 5:45 a.m., on Friday, July 27, 1973, an automobile entered the I-40 off-ramp in Nashville, Tenn. The off-ramp, which curved to the left, connected eastbound I-40 with northbound I-24/65 on the Silliman Evans Bridge. The automobile was a 2-door hardtop sedan with nine occupants. No other traffic was present.

As the automobile approached the vicinity of the juncture of the two roadways, it struck the curb face of an 8-inch-high concrete island which separated the ramp from the traffic lanes of I-24/65. The right wheels of the car overrode the island, and the vehicle headed northeast toward the far side (east edge) of the bridge. (See Figures 1 and 2.)

The east side of the two northbound lanes of I-24/65 was bordered by a 9-inch-high concrete curb, on which was mounted a box-beam (strong beam/ weak post) guardrail system. Three feet east of this guardrail stood a 3-rail steel bridge rail.

Tiremarks indicate that after the automobile impacted the concrete island, the driver lost control of the vehicle. The car crossed the north-bound lanes of I-24/65 and struck the box-beam guardrail. First the right front wheel and then the other three wheels mounted the curb, and the automobile, aided by dynamic uplift, overrode the box-beam guardrail, plunged through the bridge rail, and fell 65 feet from the deck of the bridge to the ground. While airborne, the car pitched forward on its lateral axis. It landed on its roof at a horizontal angle of 8° and 90 feet from the point of impact with the bridge rail, 9 feet outside of the edge of the bridge deck.

During the crash sequence, none of the car's nine occupants was ejected, and no fire broke out. The driver and seven passengers died. Only a 5-year-old child survived; he suffered moderate injuries. Other than the child, there were no witnesses to the accident.

Accident Site

'Interstate 65, a north-south route through Tennessee, crosses the Cumberland River via the Silliman Evans Bridge. Interstate 24, though an east-west route, is concurrent with I-65 through the accident site. Interstate 40, also an east-west route, interchanges with I-24/65 just south of the bridge.

The I-40-off-ramp. The I-40 off-ramp consisted of a compound curve with five radii and a spiral. The design of the ramp satisfied the minimum standards of the American Association of State Highway and Transportation Officials (formerly the AASHO). The spiral at the juncture of the off-ramp and I-24/65 was 400 feet long and was preceded by a curve with a radius of 881 feet.

• 4 -

〔图〕65 FOOTDROPFATAL CRASH OF PASSENGER CARSILLIMAN EVANS BRIDGENASHVILLE, TENNESSEE1-241-24July 27, 19731-651-65Not to scal

Figure 1. Accident site and crash sequence.

〔图〕ISLANDhe Silliman Evans Bridge (picture taken facing south).mPACTRA0 LS00POINTOFME RctneRAFFICS1.60\n

Figure 2. Aerial view of the Silliman Evans Bridge (picture taken facing south).

• 6 -

At its origin on I-40, the ramp was marked for two lanes of traffic. At a point 490 feet before the ramp joined I-24/65, the two traffic lanes began to narrow to one lane. In this narrowing process, the roadway width narrowed from 30 feet to 14 feet. (See Figure 3.) At the point of juncture, the width of the single traffic lane was 10 feet 6 inches between the pavement markings. The lane widened to 12 feet on the bridge as it became the third lane of northbound I-24/65.

As the off-ramp progressed 273 feet into the 881-foot-radius curve, the superelevation uniformly decreased from 0.0950 to 0.0358 feet/foot. Through most of its length, the deck of the off-ramp was an upgrade. Near the juncture with I-24/65, however, it changed to a downgrade.

The bridge rail systems on the ramp were identical to those on the bridge structure. The physical division between the ramp and I-24/65 was continued another 111 feet beyond the point of juncture by a raised concrete island, which decreased in width from 13 feet to 4 feet. The island had 8-inch-high curb faces. The curb face of the island on the ramp side exhibited evidence of numerous vehicle contacts, i.e. tiremarks and chipped concrete.

The average daily traffic count (ADT) on the off-ramp was 7,178 vehicles.

The bridge (I-24/65). As they approached the juncture with the I-40 off-ramp, the two northbound lanes of I-24/65 curved to the right on an ascending grade. The curve became tangent 76 feet south of the south end of the concrete island. The ramp and I-24/65 reached equal elevation near the nose (north end) of the island.

The entire Silliman Evans Bridge consisted of two separate structures, one for northbound traffic and the other for southbound traffic. Each structure had three traffic lanes on a portland-cement concrete deck. The three 12-foot-wide lanes were bounded on both sides by 3-foot-wide shoulders which were delineated by white edge lines. The outside edge of each shoulder was bounded by a 9-inch-high by 10-inch-wide concrete curb, 2 feet 6 inches from the bridge rail.

The bridge rail as originally installed consisted of steel vertical posts, 2 feet 11 inches high, and three horizontal steel rails. According to the manufacturer, the bridge was designed to meet the 1957 AASHO specifications. 1/

〔图〕f the concrete island which separated the I-40 off-ramprthbound lanes of 1-24/65 (right). Note the positionnd the island

Figure 3. North end (nose) of the concrete island which separated the I-40 off-ramp (left) from the northbound lanes of I-24/65 (right). Note the position of the edge line and the island.

• 8 -

The 1965 AASHO "Specifications for Bridge Design" substantially increased the load requirements for bridge rails. Neither the 1957 nor the 1965 specifications required dynamic testing or any standard configuration for rail geometry.

In 1967, the outside bridge rail system was supplemented by a modified box-beam (median-type) guardrail. The guardrail system was designed and installed after the bridge rail had failed to retain impacting vehicles. The plans were signed in May 1967. The box-beam guardrail was installed with Federal funds and was approved by the Federal Highway Administration (FHWA). There was no evidence of any maintenance inspections by the FHWA after the guardrail was installed.

The box-beam barrier system consisted of an 8-inch by 6-inch steel rail, 27 inches above the bridge deck, on support posts which were anchored by three bolts and which were 8 feet apart. (See Figures 4 and 5.) 2/ The anchor bolts had been repeatedly pulled out or dislodged when the beam was struck by errant vehicles. As a result, the post supports were modified 12 to 18 months after the guardrail was installed. To expedite replacement, saddle-type mounts without bolts were fitted over the curb, and the support posts were attached to them. This eliminated anchoring the base supports in the concrete. (See Figure 6.) Thus, at the time of the accident, the replacement supports struck by the car were merely resting on the curb without actually being attached to the curb or the bridge structure. This type of replacement guardrail existed on both the northbound and southbound structures. One 500-foot-long guardrail section on the southbound bridge contained only three short sections, 24 feet long each, which were bolted to the concrete curb as originally designed.

At the time of the accident the State of Tennessee was planning to reconstruct the deteriorating decks of both bridges, to add a traffic lane, and to redesign the barrier rail system. This work is anticipated to require 3 years to complete. The northbound structure was scheduled to be closed for reconstruction in December 1973. In response to Safety Board recommendations issued on August 30, 1973, Tennessee made improvements on both bridge structure, (See Appendix C.)

Crash damage. The curb, box-beam guardrail, and bridge rail were the only components of the bridge damaged during the crash sequence.

The box-beam guardrail was dislodged for 153 feet; 65 feet of guardrail deflected into or beyond the alinement of the bridge rail. (See Figures 7 and 8.) The inside face of the box-beam at its furthermost point was deflected

2/ A box-beam of median barrier design was adopted in 1965 for use on medians as a standard by the State of Tennessee. Tennessee Standard Drawing, dated September 1965, showed the box-beam mounted, on either a curb or above finished ground, on steel posts at 6-foot centers.

〔图〕3/4" Cinch AnchorTraffic Side of Rail(3-unit standardwith z" WasherExisting Curbform a Smooth Profileand SlabTop of Rail

Figure 4. Original design of the box-beam barrier and the bridge rail on the Silliman Evans Bridge.

-10-

〔图〕插图

a. Intact.

〔图〕t<一\n

b. Pulled away from curb.

Figure 5. Post anchorages originally used for the box-beam barrier on the Silliman Evans Bridge.

• 11 -

〔图〕插图

Figure 6. Section of box-beam guardrail on the southbound portion of the Silliman Evans Bridge.

Note the unanchored posts leaning away from the roadway.

• 12 -

〔图〕RAILBOX- BEAMyy\n

Figure 7. View of Silliman Evans Bridge from ground looking up. Note sections of box-beam hammion deflocted 2 foot outside of bri: [mail]

• 13 -

〔图〕RAILBOX-BEAMCURB\n

Figure 8. The bridge deck shortly after the crash. Note the tiremarks and the position (annoure) of unanchowed handbeam current recto.

• 14 -

2 feet beyond the edge of the deck of the bridge. The vertical posts that supported the box-beam were disconnected from the section of guardrail which was deflected outside the deck edge. The total displacement exceeded 5 feet.

A section of the bridge rail (both posts and rails) was separated from the bridge deck. Six posts were broken from their anchorages, and the anchorages (base plates) remained intact and secured to the deck.

The ground below the edge of the bridge deck in the vicinity of impact was littered with damaged components from both the box-beam and the bridge rail, some of which showed evidence of corrosion. Several displaced bridge rail posts were still attached to the base plates and were broken at the height of the lower rail.

An examination of the debris from the box-beam system revealed posts intact with both the original and replacement base plates. (See Figure 9.) The guardrail posts showed no sign of bending and little other damage. The blades which were attached to the top of the posts and designed to fit inside the beam itself showed little or no bending.

The 153-foot-long section of curb involved in this crash exhibited areas of spalled concrete, particularly at points where guardrail posts had rested on the unanchored saddle-mount base plates. In those same areas, the spalled concrete surfaces showed discoloration from aging, and exposed reinforcing steel was corroded.

The box-beam system and bridge rail were repaired on the day of the accident. The deformed section of box-beam was replaced, as were all missing posts, except two. The two missing posts left two 16-foot sections of box-beam, unsupported. The replacement posts were installed on saddle-mount base plates. Other posts already installed with saddle mounts were uprighted into place.

The missing section of bridge rail was replaced in its original design configuration.

Traffic-control Devices and Markings

The posted speed limit on the bridge and adjacent interstate highways was 55mph. The I-40 ramp, however, had a 50-mph advisory speed limit sign.

• 15 -

〔图〕插图 〔图〕C\n 〔图〕插图

Figure 9. Debris below the bridge after crash: (A) original box-beam post and anchorage; (B) broken bridge rail posts and anchorages; (C) revised box-beam post, with saddle-mount anchorage plate; and (D) blade on support post for the connection with the box-beam rail.

• 16 -

At the location on the ramp where the pavement markings indicated the transition from two lanes to one, a black-on-white "MERGE LEFT INTO SINGLE LANE" message sign was posted. An advisory "MERGING TRAFFIC" sign was located on the south end of the concrete island at the juncture of the bridge rails of I-24/65 and the I-40 ramp, 110 feet from north end of the island.

The transition of the two lanes to one lane on the ramp was marked by longitudinal solid white lines with diagonal white hash marks. The right edge line on the ramp at the island terminus, i.e., the north tip of the island, was distorted into the traffic lane. (See Figure 3.) The traffic lanes on the bridge were separated by broken white lines.

The only delineation on the island itself consisted of faded yellow paint on the curb face.

Marks on the Roadway

There were no marks on the roadway to indicate that the automobile braked heavily at any point during the crash sequence. Several tiremarks were found along the path which the car followed from the point of impact with the concrete island to the face of the bridge curb; the marks on the roadway were identified as scuffmarks caused by the lateral movement of rotating tires. The tiremarks were about 140 feet long. The arc-shaped path of the scuffmarks followed two short radii of 204 and 132 feet. The scuffmarks began just to the right, i.e., east, of the nose of the concrete island and continued to within 6 inches of the bridge curb. (See Figure 8.)

Environmental Factors

The accident occurred in daylight. The road surface was dry; the temperature was 71°F. The humidity was 70 percent, and there was no wind.

Some ground fog was reported by persons at the accident scene. Their estimates of the density of the fog ranged from light to moderate. Visibility at the Nashville Airport, about 7 miles from the accident site, was reported by the National Weather Service to be 4 miles with ground fog from 1:40 to 7:50 a.m.

Accident History of the Silliman Evans Bridge

Since the opening of the Silliman Evans Bridge in December 1964, there have been 11 fatal accidents, involving the death of 25 persons. The number of deaths directly attributable to vehicle that plunged from the bridge deck is ascertainable, however, only from January 1969 to the day of the accident. During that period, seven fatal accidents occurred. Three of the seven accidents resulted in 11 fatalities and involved vehicles exiting the bridge deck through or over the box-beam guardrail; one other, which resulted in a single facility, involved vehicle impact with the

• 17 -

concrete island. (See Appendix D.)

There was no evidence that either the police or the Tennessee Bureau of Highways made any effort to conduct an in-depth multidisciplinary investigation into any of the fatal accidents on the bridge. Nor was there evidence to demonstrate that the Maintenance Division of the Bureau of Highways had a program to doucement and analyze crash-damaged highway appurtenance. The primary sources of the Bureau's information regarding crash damage of highway appurtenances were police accident reports and random reporting from highway personnel. Although several "hits" were reported, no record showing the history of crash damage or repair to the barrier system on the Silliman Evans Bridge was made available. Without such data, the precrash position and condition of the box-beam guardrail system at the point of impact in this accident could not be determined by the Safety Board.

The Automobile

The automobile, a 1971 Mercury 2-door, 5-passenger hardtop sedan, was equipped with a V-8 engine, automatic transmission, power steering, power brakes, front bucket seats, and 2 shoulder harnesses and 5 seatbelts. Its tires (Firestone JR78-15 radials) were 2 weeks old and in excellent condition. The odometer read 39,835 miles. The vehicle was registered in Arkansas to Mrs. Georgia Glover of Pine Bluff, Ark. Mrs. Glover was a passenger at the time of the crash.

The automobile had been tuned up and was mechanically inspected 1 week before the accident and displayed a valid Arkansas motor vehicle inspection sticker, dated November 1972. A postcrash examination did not reveal any evidence which would indicate precrash failure of the steering or braking systems. The crash damage sustained by the automobile prevented a detailed inspection of other mechanical components. The automobile was destroyed in the crash. (See Appendix A for a detailed description and analysis of crash damage pertinent to the analysis in this report.)

Automobile Occupants

The automobile was occupied by three adults and six children. The driver and the owner of the automobile, with a 7-year-old child between them, occupied the front seat. The driver should not have been crowded in a manner that would have interfered with her driving. The other adult and five children, ranging from 3 to 9 years, were in the back seat.

The driver of the automobile was a 58-year-old resident of Pine Bluff, Ark. She held a valid Arkansas operator's license, which showed no restrictions. Arkansas had no record of traffic violations or accidents for this driver.

• 18 -

Although the amount of time which the driver operated the car before the crash could not be ascertained, the driver had not driven the automobile throughout the trip. The two adult passengers possessed valid Arkansas operator's licenses. The owner of the automobile was driving when the car left Pine Bluff at 10:30 p.m., on the day before the accident.

The physician in attendance at the scene and later at the hospital indicated that there was no reason to suspect the cause of death as any other than crash trauma. There is no evidence to suggest that the driver suffered from any physical condition which could have caused incapacitation at the time of the crash. A blood alcohol count was negative. No autopsy was performed. The driver was not wearing the available seatbelt or shoulder harness.

The two male and six female passengers were unrestrained in their positions within the automobile. The survivor stated that at the moment of impact, only he and the driver were awake. All of the other occupants were asleep.

Postcrash Activities

An employee of a steel plant located within 500 feet of the accident site notified the Metropolitan Police Department about 1 minute after the crash. Police arrived at 5:48 a.m. Upon arrival, the police did not see or hear any signs of life in the overturned automobile. A tow truck soon arrived and was used to right the vehicle onto its wheels. When the roof was raised from its crushed position, the only sign of life was the muffled crying of a child trapped in the wreckage.

At 6:10 a.m., the first of several ambulances arrived. A check for vital life signs in the occupants of the car who could be reached proved negative. The crying child and other occupants could not be extricated from the wreckage, and a portable power saw unit was sent for. The saw was brought to the scene 20 minutes later. During the delay, a doctor from the Nashville General Hospital arrived. He immediately checked the occupants of the car whom he could reach; only the crying 5-year-old child exhibited any sign of life. The doctor believed that the crash trauma suffered by the occupants was sufficient to have killed them all instantly.

When the power saw arrived at 6:30 a.m., it was used to cut both "A" pillars of the car at the roof line. The roof was then peeled rearward, and all of the car occupants were removed. The lone survivor and the bodies of the eight other occupants were transported to the Nashville General Hospital.

• 19 -

Applicable Standards and Laws

Federal Highway Safety Program Standards. Federal Highway Safety Program Standard (FHSPS) No. 9, "Identification and Surveillance of Accident Locations," requires each State to have a program for identifying and maintaining surveillance over locations with high accident rates, and for taking appropriate steps to reduce accidents at these locations. 3/ Paragraph IA2c of the standard requires the production of an inventory of design and operating features with which high accident frequencies or severities are associated. Paragraph IA4 requires the evaluation of the effectiveness of safety improvements on any specific section of a road and street system. The program manual (Chapter IV, paragraph 3c) recommends the use of multidisciplinary teams to accomplish these requirements.

FHSPS No. 12, "Highway Design, Construction and Maintenance," requires each State to have a program to improve the design, construction, and maintenance of its highways. The program must provide for adequate design standards, removal of highway hazards, proper maintenance procedures, safe construction sites, and efficient postcrash operations. Paragraph II of the standard requires that hazards within the highway right-of-way be identified and corrected. Paragraph IJ4 requires that bridge railings be designed to minimize severity of impact to retain the vehicle, to re-direct the vehicle so that it will move parallel to the roadway, and to minimize danger to traffic below.

FHSPS No. 18, "Accident Investigation and Reporting," requires each State to have a program for uniform comprehensive motor vehicle traffic accident investigation. The program must provide for the gathering of detailed data on traffic accidents and for entering of the data into a traffic records system for use in planning, evaluating, and furthering highway safety program goals. Paragraph IVD2 of this standard requires in-depth accident investigations by teams consisting of two or more persons who represent different disciplines, e.g., police traffic engineer, automotive engineer or medical or social scientist. Each member should gather data for the purpose of determining probable cause of traffic accidents, associated injuries, and fatalities.

United States Code. Title 23, Chapter 1 (Federal-Aid Highways), Section 116 (Maintenance) of the United States Code provides that it is the duty of each State highway department to maintain, or cause to be maintained, any highway project constructed with Federal funds. Under paragraph (c) of Section 116, the U. S. Secretary of Transportation is required, if a project is not being properly maintained, to call the fact to the attention of the State highway department. Failure by the State to comply within 90 days may cause the Secretary of Transportation to withhold approval of all further projects of all types within the entire State until such compliance is achieved.

3/ The Federal Highway Safety Program Standards are contained in 23 CFR 1204.

• 20 -

ANALYSIS

Crash Events

Entrance onto the off-ramp. As it exited to the left from I-40, the automobile traveled on a gently curving ramp which changed into a series of curves of increasingly shorter radius. In the last 600 feet on the ramp before the automobile impacted the curb face of the concrete island, the driver was required to respond to operational features such as:

• the transition from two lanes to one lane on the right side of the ramp;

• roadway geometries which changed from a curve of 1400 feet radius to a curve of 881 feet radius to a spiral;

• decreasing superelevation;

• a "MERGING TRAFFIC" warning sign which preceded the end of the ramp, although the ramp lane continued across the bridge as a third lane of I-24/65;

• limited visibility of northbound traffic approaching the bridge on I-24/65;

• and a narrow pathway, less than usual width of an inter-state lane, at the end of the ramp, bounded on the left by a curb and barrier system and on the right by a raised unmarked concrete island.

As the automobile moved into the spiral curve about 125 feet from the point of initial impact with the island and about 55 feet from the "MERGING TRAFFIC" sign, the driver would not have been able to determine what pathway would be available to her after she left the ramp. Consequently, considering the geometries of the ramp in combination with the warning sign, it is likely that the driver reduced vehicle speed and was looking rearward and to the right in anticipation of a mandatory merging maneuver with the "fast" lanes of I-24/65. No merge was necessary.

The Federal Highway Administration's Manual on Uniform Traffic Control Devices (MUTCD) defines a "MERGING TRAFFIC" sign as "intended for use to warn motorists that merging movements may be encountered in advance of a point where two (2) roadways converge and no turning conflict occurs." The traffic pattern and the physical environment on the ramp and I-24/65 did not require a "MERGING TRAFFIC" sign. If the State highway department was concerned that drivers entering I-24/65 from the I-40 off-ramp would attempt to change lanes on the bridge, then the third lane, i.e., the extreme left lane which began at the interchange between the off-ramp and the two northbound lanes of I-24/65, could have been separated from the other lanes by double striping, as provided for in the MUTCD.

• 21

Since the features in the last 600 feet of the off-ramp did not give the driver sufficient opportunity to anticipate what maneuvers would be necessary at the upcoming interchange, the "MERGING TRAFFIC" sign became her primary source of information. The sign conveyed a message of possible conflict with traffic approaching the interchange on I-24/65. As a result, the driver possibly reacted with excessive visual emphasis to the rear and right in order to check traffic she expected to encounter when she exited the off-ramp. The MUTCD does not include any sign which would advise motorists that a ramp lane does not end at its juncture with a major route but instead continues through as a traffic lane.

If the driver was preoccupied with the problem of merging traffic, her vigilance toward the reducing radius of the ramp and the centrifugal forces acting upon the vehicle may have been diminished. The advisory speed of 50 mph on the off-ramp seems excessive for motorists who are unfamiliar with the roadway. Even though the theoretical design of the ramp permits a 50-mph speed, the many combined operational factors present in the last 600 feet indicate that in reality vehicle speed should be less than 50 mph. Since the accident, the advisory speed has been reduced to 40 mph, as the result of the Safety Board recommendation mentioned above.

Impact with the concrete island. The automobile struck the island curbing at a shallow angle, which was estimated to be less than 5°. The right wheels mounted the island, tracked across its face for 40 feet in a northeast direction, and ran off the nose of the island.

The yellow paint on the curb face of the concrete island was faded and in some areas not visible at all. The use of yellow paint on the curb face is questionable, since traffic is required to move to the left of the obstruction. White paint would more appropriately fit the existing criterion in the MUTCD. Delineators, visible both day and night, positioned at close intervals and at the eye level of a driver, would have more clearly identified the location and geometries of the island. Drivers, especially short persons, would have difficulty in identifying the relative location of a low object, such as the island, to the right of the vehicle. Since the accident, the concrete island has been painted, as the result of the Safety Board recommendation mentioned above.

Traffic-control devices alone, however, cannot make highways safe. Highway design must also consider the ways motorists actually operate or expect to operate their vehicles. The pathway provided on a highway should not force a driver to make unnatural maneuvers.

In 1969, the Safety Board recommended to the Federal Highway Administration that all existing and new standards for drivers, vehicles, and highways be compatible with each other in all aspects. 4/

-‘22-

When the automobile mounted the island, the impact damaged the right front wheel, caused the right rear tire to lose air pressure, and possibly jolted the unbelted driver from her position behind the wheel. A passenger in an automobile involved in a previous accident at the concrete island was ejected upon impact, and the driver lost control of the vehicle. The road-way design could have used some other means of delineation much less hazardous than the last 40 feet of the island.

Dynamic testing as early as 1953 revealed that a curb with a face similar to the curb face of the concrete island served reasonably well as a barrier, but was not consistent in its effect and caused relatively major vehicle damage during low-speed impacts. 5/ At speeds of more than 20 mph and at approach angles of 5°. 10°, and 15°, the test vehicles suffered damage to the rims of the wheels and mounted the curb with little loss of velocity. The test report noted that most of the collisions imparted a sharp jolt to the steering mechanism. The test drivers reported only moderate difficulty in maintaining control; however, the driver of the automobile involved in this accident was not a test driver and was not expecting the impact.

Impact with the guardrail and the bridge rail. After leaving the island, the automobile crossed the northbound lanes toward the curbing, boxbeam guardrail, and bridge rail. The arc-shaped path of the scuffmarks followed two short radii of 204 and 132 feet. Calculations indicate that the upper limit of the speed of the automobile in the first radius was 43 mph, and in the second, 35 mph. 6/

Upon reaching the curb, the right front of the automobile struck the box-beam barrier system. The beam, primarily resting on posts supported by saddle-mount anchorages, offered little resistance to the vehicle and yielded laterally. The displacement of the box-beam exposed the curb to impact.

Curbs of the type used on the edges of this bridge have in the past been installed to control drainage, delineate proper vehicle paths, and prevent vehicles from leaving the roadway at low seppd, or at least deter encroachments. In this instance, however, the curb, in concert with the box-beam guardrail, precipitated the fatal accident by assisting in dynamic uplift of the automobile. The results of the dynamic tests on curbs in California mentioned above demonstrated that dynamic uplift could be anticipated at moderate speed and relatively shallow angles of approach.

5/ "Dynamic Testing of Various Designs of Highway Bridge Curbing," State of California Department of Public Works, Sacramento, California, December 1953.

6/All calculations mentioned in this report can be found in public docket SS-H-28 at Safety Board headquarters in Washington, D.C.

· 23 -

The right front wheel of the automobile struck and mounted the exposed curb. Based on available physical evidence, the Safety Board estimates that the angle of the automobile at impact with the box-beam was between 28° and 32°. As it continued forward, the vehicle began to move upward and started to rotate counterclockwise. That movement caused the rear end of the automobile to drag on the curb and pushed a section of box-beam downward and laterally. As the front of the vehicle started to rise, the path of the automobile was realined within 8° of the alinement of the curb.

The box-beam installed on the curb was very similar to the standard box-beam median barrier used in the State of Tennessee. One major exception was an increase of the post spacing from 6 feet to 8 feet. The increased spacing could substantially increase the lateral deflection capability of the box-beam when impacted by a passenger vehicle.

The anchorage of the posts, i.e., a base plate bolted with inserts drilled into the concrete curb, was insufficient to develop the carrying load for which the posts were designed. This resulted in anchorage failures within the impact area and allowed greater deflection of the rail than had been anticipated. Further modification of the anchorage by the redesign of the base plate to an unanchored saddle-mount compounded the already significant deficiency in the lateral restraint capability of the system. The installation of the saddle mounts to expedite maintenance efforts degraded the structural integrity of the system and created an unsafe environment. The box-beam, in combination with the bridge rail, allowed penetration of the automobile and only absorbed about 17 percent of its velocity.

A search of the literature available in 1966 and 1967 on bridge rails and box-beam barriers did not reveal any basis for using a barrier with the deflective qualities of the box-beam involved in this accident to contain vehicles on bridges. The distance available from the inside face of the box-beam guardrail to the edge of the bridge deck was 3.4 feet. In 1967, the New York State Department of Transportation published the results of tests which had been performed on median barriers similar to the box-beam guardrails used on the Silliman Evans Bridge. (See Appendix B.) The barriers in the New York tests were mounted on posts spaced 6 feet apart; the tests included a rigid bridge rail that allowed no vehicle penetration. The results of the tests clearly indicated that the deflection of the box-beam barrier significantly exceeded 3.4 feet. The results were available to the Tennessee Bureau of Highways at least 1 month before the plans for the box-beam installation on the Silliman Evans Bridge were signed.

In April 1967, a paper which described the 6-year developmental program in New York on box-beam barriers of various configurations and uses was presented to the 49th Annual Tennessee Highway Conference, which was attended by 141 employees of the Tennessee Bureau of Highways. 7/ The conference was

• 24 -

held 1 month before the plans for the installation of the box-beam barrier system on the Silliman Evans Bridge were signed.

The placement of the box-beam on a high curb was inconsistent with the basic concept of the strong beam/weak post guardrail system, which provides for smooth lateral movement of the vehicle and beam after initial contact. A high curb can initiate upward movement of an impacting vehicle and subsequently produce substantial incompatibility between the barrier and the vehicle. One likely result of such incompatibility is vaulting or overriding of the guardrail by the vehicle.

The automobile left the bridge deck at a velocity calculated to be about 30 mph. To compute the exact velocity of the automobile prior to its impact with the concrete island is not possible, but a review of the dynamics of the vehicle before it left the bridge, the tire marks on the bridge deck, the minor damage inflicted on the vehicle by the railing systems, and the calculated 30-mph velocity of the vehicle as it left the bridge deck would suggest that the automobile was traveling on the ramp at a velocity of no more than the posted advisory speed of 50 mph.

Identification and Surveillance of Accident Locations

The best measure of the safety provided by the components of a highway transportation system is well-doucmented accident experience. In this case, very little detailed information was available from the usual, and in most instances the only, source of accident data-the police.

The routine police accident report is not a scientific research-oriented document, but is rather a limited record of data usually related to law enforcement and driver regulations. These reports, therefore, are of limited value to highway officials. However, even a cursory examination of the previous "bridge plunge" accidents described in the police accident reports should have indicated to responsible persons that the ineffectiveness of the barrier system presented the possibility of another catastrophe on the bridge. The failure to apply countermeasures, compounded by the continuing re-stallation of sections that experienced "hits" suggests that the police reports were either not examined or were disregarded. Thus, the need exists for an ongoing systematic program which will focus on identifying problem locations, for investigations to determine cause(s), for establishment of corrective measures, for assurance of implementation, and for evaluation of effectiveness of any new installation. This kind of comprehensive effort is required of each State through the provisions of FHSPS No. 9.

Because the State did not systematically record box-beam and outside bridge rail damage from previous impacts, no data were available for review. Even though an examination of photographs of the bridge after the fatal accident of November 1972, suggested vehicle vaulting, without further documentation it cannot be concluded that the revised anchorage system of the barrier system contributed to the vaulting of vehicles from the bridge or

• 25 -

that the rail system (barrier guardrail and outside bridge rail) failed because of inherent faults. A U, S, House of Representatives Public Works Committee staff report of hearings into highway safety, design, and operations spoke directly to this point:

"A significantly lagging activity, in the judgment of the staff, is the function of operational surveillance for streets and highways under State jurisdiction. Much more participation is recommended for high level professionals who benefit by seeing their facilities in operation first-hand. Office-based engineers and others who confine field inspections to daytime and good weather intervals, often miss seeing significant needs. The deficiencies in design, maintenance, traffic control and other practices can usually be identified readily. Then the proper redirection of effort can be accomplished, often with only minor additional expenditure and with specific benefits to safety." 8/

The events of this accident illustrate the need for a program using both onsite preventive surveillance and multidisciplinary accident investigation teams. Such a program could have identified the critical problems involved in this accident before the accident occurred. Preventive surveillance can be accomplished by State highway personnel through documenting highway damage, reviewing the data, and making onsite evaluations. Multidisciplinary teams, however, require some outside personnel with a degree of independence, because their function involves oversight of the State highway department. This does not mean that highway department personnel should be excluded from the multidisciplinary teams. Their assistance is very much required. But the primary responsibility for the findings and recommendations of multidisciplinary team investigations might best be handled by a member of the team who is not on the staff of the organization that is legally responsible for the highway operations.

CONCLUSIONS

1. The driver was without known physical disabilities which could have contributed to the accident.

1. There is no evidence to indicate that adverse vehicle conditions contributed to the accident.

1. There is no evidence to demonstrate that adverse weather conditions contributed to the accident.

8/ "Highway Safety, design and Operations," (93-14), Hearings before the Subcommittee on Investigations and Review of the Committee on Public Works, House of Representatives, June 12, 13, 14, 1973.

• 26 -

1. The automobile was not traveling at a speed higher than that advised for traffic on the I-40 ramp.

1. The last 40 feet of the raised concrete island served no useful purpose. If the island had been properly delineated by reflectors installed on posts at a height level with a driver's eye, and if the curb face had been adequately painted, the structure would have been more clearly recognizable as a hazard.

1. Traffic-control devices, no matter how well designed, will not adequately warn all motorists of the not always visible hazards created by certain highway designs.

1. The I-40 ramp apparently did not meet this driver's expecta-tions, nor did it convey to her the built-in low tolerance for error.

1. The spacing of the posts on the box-beam guardrail was incorrect for the designed limited lateral displacement.

1. The failure to anchor properly the posts which supported the box-beam resulted in the loss of the guardrail's integrity and increased the guardrail's lateral displacement capability.

1. If the posts had been adequately anchored, the 8-foot spacing would still have permitted a critical deflection of the rail.

1. The box-beam guardrail post anchorage was redesigned to expedite maintenance efforts; it is not clear whether the new design was analyzed for safety effects.

1. The yielding box-beam guardrail was not suited to the bridge environment, because lateral movement beyond a very limited distance was not acceptable.

1. The installation of a weak post/strong beam guardrail system on a narrow curb defeats the basic concept of smooth lateral movement for an impacting vehicle.

1. Highway officials of the State of Tennessee did not make use of the literature available on bridge rails and box-beam barriers, none of which revealed a logical basis for using a box-beam system on the Silliman Evans Bridge.

1. The FHWA, in approving the 1967 installation of the box-beam system, failed to anlayze adequately its application, and thus failed to prevent its inappropriate use.

• 27 -

1. If the FHWA had inspected the box-beam system or assured that the State of Tennessee conducted such an inspection, the bridge could have been readily identified as hazardous.

1. Other locations on the bridge where "saddle mount" anchorages had been installed were even more hazardous than the section impacted in this accident.

1. If the automobile had been retained on the bridge, more occupants may have survived.

1. There is a need for developing and implementing standards to control the impact performance of all classes and sizes of vehicles with traffic barrier systems.

1. The absence of a systematic program to review individual accidents such as those which involved the box-beam guardrail resulted in the failure of the State of Tennessee to identify and correct the extremely hazardous condition on the bridge.

1. The highway operations of the State of Tennessee involved in this accident indicate that the State was not fully complying with FHSPS No. 9, 12, and 18.

• 28 -

PROBABLE CAUSE

The National Transportation Safety Board determines that the probable cause of this accident was the failure of the driver to maintain her vehicle in the pathway provided.

The following are contributing factors which may have confused the driver: (a) increasing curvature of the ramp; (b) narrowing of the ramp from two lanes to one; (c) a misleading traffic-control sign; (d) misleading pavement markings; and (e) an inadequately delineated and unnecessary section of a concrete island which narrowed the pathway to less than the width of a normal traffic lane--all in the last 60 feet of the ramp.

This confusion probably contributed to the striking of the concrete island, which caused loss of control of the automobile and its subsequent impact with the bridge rail barrier system.

The fall of the automobile to the ground and the resultant fatalities were caused by the failure of the barrier systems to retain or redirect the errant vehicle. Contributing to this failure were: (a) an unnecessary 9-inch-high curb, which provided the impacting vehicle with dynamic uplift; (b) improper installation of a box-beam median-type barrier system rather than a more rigid system; (c) redesign of the box-beam mounting method, which did not secure the beam to its base; (d) inadequate maintenance of the box-beam; and (e) an outside bridge rail not adequate to withstand the vehicle impact.

The probable cause of the failure to detect and correct these hazards was the absence of a systematic program for the identification, surveillance, and correction of potential, or existing, hazardous conditions on the highway.

• 29 -

RECOMMENDATIONS

As a result of the investigation of this accident, the Safety Board on August 30, 1973, submitted three recommendations to the Governor of the State of Tennessee. (See Appendix C for a copy of the recommendation letter and the response of the State of Tennessee.) The Safety Board further recommends that:

1. The Federal Highway Administration:

(a) Establish national performance standards, including dynamic testing procedures, for bridge rail systems. Such standards should extend performance criteria to include impact by heavy vehicles and should improve performance characteristics for impacts by all classes of vehicles. The establishment of these standards should be of high priority and compliance should be mandatory for all new bridge rail systems used on public roadways. (Recommendation No. H-74-18).

(b) Assure that the State of Tennessee is maintaining Federal-aid highway projects in accordance with Federal law (U.S.C. Title 23, Section 116). (Recommendation No. H-74-19)

2. The State of Tennessee:

(a) Implement all the provisions of Federal Highway Safety Program Standard No. 9," Identification and Surveillance of Accident Locations," and No. 12," Highway Design, Construction and Maintenance," with special emphasis on the provisions relating to each State's having a systematically organized program to identify and correct potential or existing hazardous conditions within the highway right-of-way. (Recommendation No. H-74-20).

(b) Implement all the provisions of Federal Highway Safety Program Standard No. 18," Accident Investigation and Reporting," with special emphasis on the provisions which require in-depth accident investigations by teams of two or more persons who represent a range of technical disciplines. (Recommendation No. H-74-21).

• 30 -

BY THE NATIONAL TRANSPORTATION SAFETY BOARD

/s/ JOHN H. REED Chairman

/s/ FRANCIS H. McADAMS Member

/s/ LOUIS M. THAYER Member

/s/ ISABEL A. BURGESS Member

William R. Haley, Member, was absent, and did not participate in the adoption of this report.

June 13, 1974

• 31

APPENDIX A

CRASH DAMAGE TO THE AUTOMÓBILE

The principal damage incurred by the automobile was the crushing of the roof, hood, and trunk lid, which occurred when the vehicle hit the ground below the bridge. Other damage, such as buckling of the sides of the vehicle, was induced by crushing forces which followed impact with the ground.

Damage to components of the vehicle before the plunge from the bridge ranged from minor to moderate. Impact with the curb of the concrete island damaged the right front wheel rim and caused the right rear tire to deflate. Impact with the curb at the bridge edge was more severe and further damaged the rim of the right front wheel. The fender area just forward of the wheel showed signs of contact with one of the curbs.

The right front of the vehicle received minor damage when it hit the box-beam. The automobile bumper was deformed rearward. The chrome trim on the fender immediately above the damaged bumper showed no signs of damage, which suggests the right side of the vehicle was somewhat above normal height.

Contact with the curb face of the concrete island probably caused little damage to the right front wheel and tire. The impact with the curb at the edge of the bridge, however, probably caused the rim to bend inward and pinch the tire. That pinching cut a hole in the tire sidewall.

The left front tire deflated when it hit the bridge curb; the wheel sustained no damage. This tire also exhibited scuff marks on the outside corner and sidewall, which suggests lateral movement of the tire.

The left rear wheel and tire made no contact with the concrete island and probably little or no contact with the bridge curb. The tire was intact and inflated to 28 psi. The wheel was undamaged.

The rear bumper was bent up and rearward. That configuration suggests that the bumper contacted an object as the front of the vehicle lifted from the surface of the bridge. A variety of scratches and tears on components located on the underside of the rear portion of the automobile supports this conclusion.

• 32 -

APPENDIX B

THE BOX-BEAM BARRIER

After 6 years of research using mathematical analysis and dynamic testing, the State of New York Department of Transportation in 1967 announced a new highway barrier--the box-beam. The beam, in different configurations, was found to be applicable for use as guardrails, median barriers, and bridge rails.

The stated philosophy of the New York program was:

"It would be desirable to have one all purpose highway barrier for guide rail, median barrier or bridge railing. However, a primary function of any barrier is to minimize the loss of human life by keeping deceleration as low as possible during a collision. This can best be accomplished by a system that yields as it deflects since in doing so it absorbs energy and cushions impact. Unfortunately, large deflections are not always permissible. . . and on bridges, a rail deflection must be kept at a very low value. Therefore, a relatively rigid bridge rail was developed." 9/

The rail was impacted at 55 mph and at an angle of 25°. There was no vehicle penetration. The 3500-pound automobile was then redirected at a velocity of 37 mph and an angle of 10. Damage to the automobile was severe; average impact decelerations ranged from 3 to 12.3 g.

The box-beam median barrier was designed to deflect laterally so that lateral forces on the impacting vehicle would be less than those imposed by a rigid rail. At the same time, the barrier was designed to limit the deflection to 2 feet, so that it could be used in narrow medians 4 feet wide. The box-beam rail is intended to be strong enough to resist lateral forces in bending rather than tension. The support posts are designed to bend and concurrently disconnect at the interface with the beam when impacted. This ease of breakaway prevents vehicle snagging and pocketing. The posts outside the area of immediate impact carry the lateral forces transmitted by the beam. The posts are designed to resist laterally one-tenth of the design loading. In the longitudinal direction, they have 75 percent less strength than in the lateral direction, so that they will bend down when struck by a vehicle bumper or wheel. For the post to resist the designed lateral forces, it is esstantial that the base of the post be restrained sufficiently not to allow premature failure. Designed failure should occur from yielding of the post material rather than from failure of the post's bottom connection. The tendency for the rail to lie down when hit by a car is resisted by the vertical stiffness of the rail, deflections in the vehicle structure, and by the undisturbed posts outside the impact area.

• 33 -

APPENDIX B

For safe dynamic interaction between a barrier and a vehicle, it is essential that the relationship between the vehicle's center of gravity and the rail height be compatible. Any element or function that disturbs that critical relationship can result in a catastrophe, such as a vehicle overturning or vaulting.

In discussing the mathematical model for vehicle trajectory, the New York report stated that:

"the horizontal motions provide nearly all the information that is required to determine the trajectory of a car during the time it is in contact with the barrier provided there is no significant vertical motion. Appreciable roll or pitch of the vehicle during impact is undesirable and if the barrier provides horizontal trajectory it is unlikely that the vehicle will pitch or roll in a vertical plane. Therefore, only horizontal motion was used in the models."

The test performed on the box-beam barrier by New York showed lateral deflections, both dynamically and residually, that exceeded the desired 2 feet. The report stated that changes to the beam splice and post connections were made to reduce the deflection to the desired limit. The final design was for a steel beam 8 by 6 by 17-1/4 inches mounted on 315.7 posts spaced 6 feet apart. The 1971 National Cooperative Highway Research Program Report 118 specifies that the same box-beam design has an expected deflection of 4 feet. The current New York specifications have been modified to allow for a deflection of 3 feet.

• 34 -UNITED STATES OF AMERICA

NATIONAL TRANSPORTATION SAFETY BOARD WASHINGTON, D.C.

APPENDIX C

ISSUED: August 30, 1973

Adopted by the NATIONAL TRANSPORTATION SAFETY BOARD at its office in Washington, D. C. on the 17th day of August 1973

FORWARDED TO:

Honorable Winfield Dunn Governor State of Tennessee Nashville, Tennessee

SAFETY RECOMMENDATION H-73-31 through 33

On July 27, 1973, a passenger car carrying nine persons crashed through a section of guardrail and bridge rail on the northbound portion of the Silliman Evans Bridge on I-65, in Nashville, Tennessee. The vehicle fell 65 feet to the ground. Eight of the occupants were killed and the other was severely injured.

Since September 1971, three trucks and four cars have plunged off the bridge. In these seven crashes, ll persons have been killed. Of the three survivors, one has been in a coma since January 1973.

The National Transportation Safety Board, under authority of the Department of Transportation Act (49 USC 1654(b)(1)), is investigating this accident to determine the probable cause and to make recommendations to those agencies which can take the necessary action to prevent similar crashes in the future.

Preliminary investigation by the Safety Board of the July 27th accident has shown that:

At the point where the ramp from I-40 (eastbound) merges into I-65 (northbound), the two roadways are separated by a raised concrete island. The 9-inch-high island does not have sufficient retaining capability to prevent a vehicle traveling at the advisory speed from mounting the structure and crossing into the adjacent roadway. Furthermore, even a shallow-angle impact with the island can initiate loss of vehicle control.

°The width of the roadway is narrowed abruptly by the location of the island. The boundary-making stripe on the right side of the lane is painted several inches to the left of the island; otherwise the stripe, if it were continued in an arc of the same radius, would run over the curb and onto the top of the island.

• 35 -

Appendix C

°The island is, in effect, an extension of the curb beyond the end of the highly visible railing. The island, however, is much lower than the railing and lacks delineators. A driver's task in negotiating this curve is complicated by the lack of conspicuity of the island, by the narrowing of the roadway to one lane, by a compound curve of decreasing radius, and by a decreasing superelevation of the roadway. It appears that the last several feet of the island serve no useful purpose. There are numerous marks on the top surface of the island which indicate that many vehicles have been driven over it.

°The lateral displacement which is permitted by box-beam guardrails, although appropriate in a median barrier, is not appropriate in a bridge rail which requires heavy resistance to lateral displacement to keep vehicles from passing beyond the edge of the bridge. The saddle mountings used as replacement supports for the box-beam appear to provide only vertical support and little resistance to lateral impact.

°The mounting of a flexible median-type box beam above a curb violates the design concept of that type of rail, in that the curbing may initiate upward movement of an impacting vehicle and thus prevent the containment or redirection of vehicles which strike the rail.

The National Transportation Safety Board therefore recommends that pending a full report by the Board, the State of Tennessee take the following action as soon as practicable to remove or reduce existing hazards on the Silliman Evans Bridge:

1. Modify or remove the last 30 feet of the concrete island at the merge point between the ramp from I-40 (eastbound) and the roadway of I-65 (northbound). (Recommendation No. H-73-31)

1. Provide delineators along the right edge of the ramp lane of I-40 to a point where the single lane becomes tangent on the bridge. $\frac{1}{2}$ (Recommendation No. H-73-32)

1. Initiate immediate action to either replace or greatly strengthen the existing box-beam guardrail to provide a barrier system which does not encourage upward vaulting of vehicles, which provides firm resistance to lateral forces, and which is fastened to the bridge rather than resting upon it. (Recommendation No. H-73-33).

• 36 -

Appendix C

McAdams, Thayer, Burgess, and Haley, Members, concurred in the above recommendations. Reed, Chairman was absent, not voting.

Wallic R. Haly, Acting for

By: John H. Reed Chairman

—

$$-37-$$

Annendix C

TENNESSEE DEPARTMENT OF TRANSPORTATION

NASHVILLE 37219.

WINFIELD DUNN GOVERNOR

November 21, 1973

ROBERT F. SMITH COMMISSIONER

BUREAU OF HIGWAYS BUREAU BUREAU OF PLANNING BUREAU OF BUSINESS MANAGEMENT BUREAU OF AERONAUTICS

RE: Silliman Evans Bridge Recommendations

Mr. John H. ReedChairmanNational Transportation Safety BoardDepartment of TransportationWashington, D. C., 20591

Dear Mr. Reed:

In response to your October 30 letter to Governor Dunn relative to the I-24/I-65 Bridge in Nashville, I would like to advise that the following action has been taken:

Recommendation No. H-73-31 and H-73-32: Delineators have been added as suggested. The advisory speed limit posted in advance of this merge point has been reduced and a concentrated effort has been made to stripe the roadway to position the traffic in such a manner as to reduce, to the extent possible, the risk involved with this movement.

The concrete island has been painted for maximum visibility but has not been removed. This structure will be closed for major reconstruction by the end of this year and the new plans provide for considerable improvement for this merge operation including elimination of the island.

Recommendation No. H-73-33: The box beam railing for the South-bound structure, which will remain in service for at least another year, has been converted from an energy-absorbing flexible system to a non-yielding rigid barrier as suggested.

Since the North-bound structure will soon be closed, no effort was made to convert the railing, however, the present box beam has been structurally fastened to the bridge deck throughout its entire length,

• 38 -

Appendix C

Mr. John H. Reed November 21, 1973 Page 2

We believe that this should give you the information you seek. If we may be of further service, please advise.

Yours very truly,

Henry W. Derthick Engineer of Structures

HWD:1jg

cc: Commissioner R. F. Smith Mr. Michael D. Huggins Mr. Lewis Evans

• 39 -

APPENDIX D

SUMMARY OF ACCIDENT EXPERIENCE ON THE SILLIMAN EVANS BRIDGE

The four crashes described below are somewhat similar to this accident; one caused minor injury, one took the life of one man, one took the lives of two men, and one has placed the single occupant of a vehicle in a coma.

• In December 1971, a pickup truck was northbound on the bridge and, in the language of the police accident report," . . . left the roadway thru the guardrail and bounced over interstate fence and landed upside down on top... "The driver escaped with minor injury.

• In December 1971, a northbound automobile, traveling in the same direction as the vehicle involved in this accident, struck the same concrete island. A passenger was ejected and killed.

• In November 1972, a southbound automobile skidded on the rain-soaked deck of the bridge, and impacted the curb and box-beam guardrail on the east side of the southbound section. The vehicle apparently vaulted (evidenced by the fact that only the top rail of the bridge rail was contacted), and dropped a reported 85 feet to the ground below. (See Figure 10.) Both occupants of the vehicle were killed.

• In January 1973, a southbound automobile which was reportedly traveling at a high rate of speed skidded out of control on dry pavement and crashed through the box-beam guardrail and bridge rail. The vehicle came to rest on the ground below and caught fire. The driver, a seriously injured 19-year-old female, was rescued before the vehicle burned.

• 40 -

〔图〕插图

Figure 10. View of damage to the bridge rail after vehicle left the bridge deck in November, 1972. Arrow identifies missing post and top rail.