A RECOMMENDATION FOR “STRUCTURAL” BRIDGE INSPECTIONS
The sudden collapse of the I-35W Bridge in Minneapolis on Aug. 1, 2007 underscores
the need for increased research attention to, and stronger guidance in, the areas of bridge
inspection, evaluation, and rehabilitation. Over the years the Transportation Research Board has
devoted its attention to the area of bridge inspection, evaluation, and rehabilitation, especially in
the area of load capacity and fatigue evaluation specifications, and published several manuals for
bridge inspection, evaluation, repair, and strengthening.
During the last 20 years there has been a great emphasis placed on developing bridge
management programs with electronic data processing and since the early 1970s very little
improvement in qualifying bridge inspectors and engineers. We now find the bridge inspection
program is primarily used to gather an over-abundance of information on all bridges while there
has been a down-grading of the personnel in the field inspections and follow-up evaluations.
The bridge inspection program of today is a superficial cookbook approach to making an
inspection using individuals with a minimum of qualifications. In addition to the inclusion of
less qualified inspectors and engineers, the certification program has made the lesser qualified
individuals more permanent and has created a condition where more qualified personnel are not
allowed to serve as bridge inspectors, even to the exclusion of professional structural engineers.
In 1967, shortly before Christmas, the Silver Bridge in Point Pleasant, West Virginia,
collapsed during rush hour and spilled cars and trucks into the Ohio River and took a heavy toll
of life. This tragic incident was singly responsible for inciting public awareness of bridge
failures and resulted in today‟s ongoing program of inspecting our highway bridges.
Forty years later, the I-35W Bridge in Minneapolis, collapsed, during rush hour, into the
Mississippi River and human life was lost. Will this tragic event also initiate a new dedication to
public safety within the transportation system or will we concern ourselves with saving face.
Was our inspection effort sufficient in identifying developing problems within the bridge
structure? Did we follow up on our findings with sound engineering evaluations? Did we act
responsibly in maintaining the bridge? These questions must be addressed and answered and
corrective action taken, as needed, within the agencies responsible for the safety of the traveling
public. Or will we seek ways to exonerate the responsible parties and continue to reward them
for inadequate performance?
In the aftermath of the tragic I-35W Bridge collapse we have examined the wreckage and
have reviewed the events leading up to that fateful day of August 1, 2007. Those of us close to
Minnesota‟s bridge inspection program feel a responsibility to share our views and comments
with the NTSB, FHWA and others in an effort to prevent a repeat of the tragedy.
Page 2 of 9
The collapse of the I-35W Bridge was not an act of God, nor was it the cause of any one
person. The I35W Bridge collapsed because we:
1. Failed to detect a design error in sizing the gusset plates,
2. Failed to maintain the bridge and thereby allowed corrosion to reduce these same
undersized gusset plates, and
3. Once we detected the corrosion on these gusset plates, which further reduced the
thickness to 30 percent of what it should have been, we failed to do anything until
the bridge collapsed.
I Bridge Management
The Minnesota Department of Transportation maintains a bridge management program
known as “Pontis.” This program is designed to gather data useful in scheduling maintenance,
setting priority repairs, and making statistical analysis. It is to this database that the bridge
inspector‟s findings are sent, and from this database MnDOT decides which bridges should be
considered for rehabilitation or replacement. Such systems are, therefore, intended to link
inspection findings with maintenance decisions and to predict optimum response time for repairs.
Within MnDOT‟s Bridge Management Program (Pontis) there exists a smart flag for
indicating the inspection has uncovered a critical finding. This smart flag must be hoisted by the
inspector making the inspection and cannot be removed until the critical finding has been
properly address and resolved. The effectiveness of this alarm is dependent on the inspector
recognizing a critical finding exists and then involving the agency in immediate action to resolve
the finding. Should the required corrective action create a significant hardship on the agency the
inspector or reviewing engineer may be reluctant to record the finding as a critical finding.
Pontis is a tool useful in bridge maintenance and long range forecasting bridge program
needs, especially for large populations of bridges. Pontis is not particularly useful at the local
level where the number of bridges to be managed is small. It is not a bridge program that should
be depended upon for predicting a potential failure.
Analysis and evaluations in regard to bridge safety must remain the responsibility of a
qualified bridge engineer reviewing the inspection reports, especially when evaluating the
structural condition of each bridge.
Between the collapse of the Silver Bridge in 1967 and the collapse of the I-35W Bridge
in 2007 other similar events occurred which revealed certain weaknesses in the Nation‟s bridge
inspection program. In 1980 one span of the Sunshine Skyway Bridge in Florida was destroyed
by a ship impact. This event led to new developments in pier protection and ship impacts. In
1983 one span of the Miamus River Bridge in Connecticut fell and resulted in in-depth
examinations of pin and hanger assemblies and non-redundant bridge designs. In 1987 New
York‟s Schoharie Creek Bridge collapsed due to river scouring of the pier footings and resulted
in greater attention to the need for hydraulic evaluations. In 1975 a fatigue crack in one of two
Page 3 of 9
main girders of the Lafayette Bridge in St. Paul emphasized greater design control in welding
details. This near collapse was averted by the timely evaluation and inspection of a reported sag
in one of the bridge spans.
Within this same time period a bridge in Tennessee collapsed during flood conditions
containing high flow and heavy debris accumulations. The past inspection reports, citing pier
scour and lack of sufficient foundation piling to withstand additional scouring, was not properly
reviewed and evaluated by Tennessee‟s engineering staff. These events, if documented in a
thorough and accurate inspection report and reviewed by a bridge engineer qualified to properly
evaluate bridge conditions, could have been prevented.
II BRIDGE INSPECTION PERSONNEL
Routine inspections should be made, not for the specific purpose of developing bridge
rehabilitation projects, but to determine the existing condition of the bridge and to provide a
written record from which repairs, if needed, can be formulated. The bridge inspector should
have a working knowledge of, (1) how the bridge should perform structurally, (2) basic design
criteria, and (3) how various materials change with age, fatigue, stress, weathering, corrosion and
It is imperative that the inspector be capable of writing clear and concise notes. The
inspection report must be complete and accurate. Areas not accessible for close up inspections
must be identified as not inspected, and as a follow-up, special efforts need to be made to
evaluate their capacity.
The engineer responsible for the bridge safety inspection program, needs to appreciate
the importance of being entrusted with public safety and must be capable for making a structural
evaluation of the inspection made.
When reviewing the inspection report the engineer must be familiar with the
characteristics of the structure inspected. It is his or her responsibility to grade the completeness
of the report and to make a judgment or evaluation of the conditions being reported. Where
problems are noted, the location and severity of the deterioration or improper functioning of the
unit needs to be defined and its probable cause determined and, critical findings passed on to the
State Bridge Engineer.
A number of States say Federal Standards for bridge inspections are minimal and lead to
lax reports. “There are many kinds of professional engineers and we want to make sure they
have the proper engineering training and qualifications. Just having any type of engineering
degree won‟t do.
Page 4 of 9
III Bridge Inspections
The 1993 inspection reports of the I-35W Bridge, cited loss of section within (under-
designed) gusset plates and in 1982 the inspection reports noted non-performing expansion
bearings. Neither condition was followed-up by an engineering review and evaluation. The
competence of MnDOT‟s bridge inspection program and the qualifications of its inspectors and
review engineers, therefore, must be critically reviewed and improved.
The instructions on steel truss inspections always emphasize a thorough examination of
tension members for cracks, especially fracture-critical members. What is usually overlooked in
the instructions is the need to closely examine compression members and their connections for
any evidence of a developing compression failure. Loss of one compression member due to
buckling within the member or loss of section through corrosion in the batten plates, lacing bars
or connection plates could result in sudden failure of the entire span. Had bowing of the gusset
plates on the I35W Bridge been reported, the follow-up evaluation by the reviewing engineer
with structural competence would have discovered the plates were under-designed and severely
The thermal effect on bridges has concerned engineers for as long as bridges have been
built. Failure to provide for movements associated with thermal expansion or contraction can
result in a major failure of the entire bridge. The thermal stresses induced by changes in
temperature, combined with non-performing bearings are not accounted for in the original design
calculations. Expansion bearings on the I-35W Bridge were reported as not moving, however,
this appears to be a casual observation as measurements of expansion joints and bearings at
various temperatures were never taken.
Bearings are designed for the static dead load on the structure and the dynamic loads
generated by traffic, wind, tractive forces, temperature variations and earthquakes, among others.
Bearings must be inspected at regular intervals for any signs of deterioration, slippage, lack of
performance, or structural distress. The stability of the entire bridge and the safety of the
traveling public is dependent on these inspections.
Of the many external forces causing movements in bridges, the greatest movement
usually occurs from temperature change. Bridge elements, whether concrete, steel or wood,
expand when heated and contract when cooled. Restricting these movements in any way can
generate tremendous forces. A steel bridge member, not allowed to expand or contract with a
given change in temperature of 45 degrees Fahrenheit, will develop an internal unit stress of
approximately 8,500 psi. If the size of that steel member contains a cross-sectional area of 10
square inches, the force applied internally to the member and its connections will be 85,000 lbs.
Should the member be restricted throughout a temperature change of 90 degrees F, the resulting
force developed would be approximately 17,000 psi or 170 kips for the example member.
Maintaining a well-functioning expansion bearing is extremely important and the destructive
effect of a frozen bearing could be severe.
Page 5 of 9
The sports arena in Hartford collapsed due to corrosion of bolts. Two weeks later the
sports arena in Kansas City collapsed due to the same condition.
Recognizing and understanding the behavior of connections and details is crucial if the
inspector is to properly inspect fracture critical members. This is because connections and
details are often the location of highest stress concentrations.
Within the inspection reports for the I-35W Bridge can be found the notation of section
loss on components of the lower chord. In 1993 the report states there is section loss of 3/16th
inches deep on the L11E gusset plate. Significant section loss should always be reviewed and
evaluated by an engineer. There does not appear to be any record of this being done.
Corrosion, especially corrosion of three-sixteenths of an inch deep on a one-half inch
gusset plate, if undetected or not responded to will eventually have destructive consequences.
Every structural engineer would like to know how the inspectors could seemingly miss a simple
deficiency for fourteen years, and even worse, how, after discovery, this deficiency could go
unevaluated and unfixed. Had the reviewing engineer, or any of the consultant‟s engineers
making in-depth inspections, checked the capacity of the reduced connection plate, they would
have discovered a grossly under-designed gusset plate and closed the bridge immediately.
Another area with extremely high section loss was found at L14E where the one-half inch
gusset plates were measured at 75 percent section loss (1/8th of an inch remaining section).
Failure to detect and evaluate the serious corrosion occurring on the I-35W Bridge cannot be
accepted as a mere oversight. Something must be corrected within the inspection program
and/or the personnel making the inspections and reviews.
MnDOT not only continued to use marginally qualified inspectors and engineers of
questionable structural ability but completely failed to inspect the „hard to reach‟ expansion
bearing assemblies and failed to follow-up on findings of section loss of connection plates first
noted 25 years earlier. Apparently the inspector or the reviewing engineer was not aware of the
seriousness of inoperative expansion bearings.
IV Quality Assurance
Quality assurance (QA) is the verification of the level of quality of the bridge inspection.
This is accomplished by the re-inspection of a sample of bridges by an independent inspection
team. For de-centralized state inspections, the QA program can be performed by the central staff
or their agent, (e.g., consultants.) If the inspections are centralized within the state, then the QA
program should be performed by consultants or a division separate and independent of the
The quality of the inspection and the inspection reports rests primarily with the inspection
team leaders and team members and their knowledge and professionalism in developing a quality
product. A QA/QC program is a means by which random inspections, reviews, and evaluations
Page 6 of 9
are performed in order to provide feedback concerning the quality and uniformity of the state‟s
inspection program. The feedback is then used to improve the training of the bridge inspectors
and the quality of the inspection report. QA/QC should not be limited to a review of the
inspectors, the inspection report, and the bridge maintenance program within a given agency. Of
greatest importance is the qualification and performance of the engineer directly responsibility
for supervising the inspection program and reviewing the results obtained. Low ratings, notes
and sketches identifying structural problems within the bridge require a close follow-up and
evaluation by the reviewing engineer.
MnDOT‟s QA/QC program failed to identify inspection and engineering review
problems on the I-35W Bridge simply because the same personnel making the I-35W Bridge
inspection were also assigned the responsibility to conduct QA/QC reviews.
A review of MnDOT‟s bridge inspection program, both before and after the collapse of
the I-35W Bridge, requires a complete forensic examination by an outside agency not hired by
MnDOT. This examination should also include the role of the FHWA over-seeing MnDOT‟s
V I-35W Bridge Inspection Summary
Annual inspection reports dating from 1982 to 2006 were reviewed for completeness and
accuracy in describing the condition of the bridge. In this respect it was found the reports were
superficial and lacking in detail regarding the problems discovered. Reports which lack
sufficient detail, require additional efforts be made to gather pertinent information needed to
evaluate the problems identified. The record shows “Fracture Critical” inspections were made,
not as an addition to the routine inspection, but as part of the Annual Inspection. The Fracture
Critical inspection not only added more depth to the Annual Inspection but increased the report
size from 4 pages to 50 pages, most of which were details of non fracture critical items. Fracture
Critical Inspection Reports are very important documents and should not be cluttered with non-
During the review of the last 24 Annual Inspection Reports (1994 was missing) we found
inspection items relevant to the Deck Truss Spans were understated and lacked specifics as to the
location and amount of deficiency being reported.
Conditions which may have a direct effect on the cause of the bridge collapse, such as
severe corrosion and inoperative expansion bearings, were first noted in 1982. The succeeding
reports were essentially the same, however, in 1987 the comments of lower chord corrosion and
inoperative roller bearings were dropped from the reports until 1993.
In 1993, the deck truss was reported as having section loss of 18” long and 3/16” deep on
the L11E and L13E connection plates. The inspection report shows the same person made the
Page 7 of 9
inspection and the engineer‟s review. In most of the reports the engineer making the review is
In 1995, the inspection reporting system, known as Pontis, was initiated and the element
Deck Truss was coded as 10% of its length has failed paint with minor section loss. The
Expansion Bearings were coded as 100% in good condition. The only comment on the Deck
Truss was “surface rust inside box.”
In 1996, the Expansion Bearings were coded as 42 of the 125 expansion bearings being in
Condition State 3 (CS 3) which means they may have section loss, loss of support, loss of
alignment, but not necessarily inoperative. Nowhere in the definition does CS 3 directly state the
expansion bearing is not moving as intended. There is no narrative comment associated with the
poorest rating of the bearings nor was any comment made as to which expansion bearings are CS
3 or the actual conditions of these bearings.
The report shows no change in Deck Truss rating or comment.
The next three years, 1997 – 1999, show no change in Deck Truss or Expansion Bearing
ratings or comments.
In 2000, the report shows no change in Deck Truss or Expansion Bearing rating, except
for the comment “Main Truss bearings have moderate corrosion.”
The next two years, 2001 & 2002, show no change in Deck Truss or Expansion Bearing
coding or comments.
In 2003, the Deck Truss Coding is changed to show 32 LF of the deck truss is now CS 5
and the words “section loss” was added to the comments. The Expansion Bearing coding is
changed to 6 of the 125 expansion bearings are CS 3. “Section loss” was added to the comments
referring to the condition of the main truss bearings. Condition State 5 requires an evaluation be
done by an engineer. This evaluation was never made.
In 2006, the Deck Truss Coding is changed to show 0 LF of the deck truss is CS 5 and
the previous 32 LF of corrosion that was in CS 5 is now in CS 4 (minor section loss – no
engineer‟s evaluation is needed). The Pontis item for Section Loss is coded CS 2, 1 each, and
the associated note reads “Section loss: pitting, flaking & surface rust on steel.”
The engineer‟s evaluation of extreme conditions, such as loss of structural section and
frozen bearings, should be written and made part of the bridge record. Should the evaluation of
section loss indicate the bridge‟s service capacity is not affected and the safety of the traveling
public has not been compromised, then and only then should the quantities listed in condition
state 5 revert back to a less deficient states. Merely repainting a member with section loss does
not correct the structural deficiency of that member. The Pontis definitions for the five condition
states of painted steel members tends to place too much emphasis on inspecting the surface
Page 8 of 9
conditions and in effect minimizes the need to examine the structural performance of the
Pontis is a bridge maintenance management program and should be in the summary of
the inspection report. The need to inspect bridges is first, to insure public safety, and second, to
protect the investment in the structure. Pontis tends to direct the inspector to look at the bridge
superficially. Steel members need to be evaluated in the context of load paths and stress levels.
Compression members bowing under heavy overloads, under Pontis, would be rated as condition
state 1 if their paint coat is in prime condition.
Maintaining the protective paint coat is important to the long life of steel members,
however, the condition of the paint does not affect the structural condition of the member.
Evaluating the condition of the paint should be a separate element in Pontis. In the case of
bridge no. 9340 the primary indicator of condition is 2,127 LF of deck truss and what the
percentage of paint deterioration is for condition states 1 through 3.
Only in condition state 4 is the loss of steel section estimated. If the inspector feels the
loss of section is minor and the paint is severely deteriorated then the quantity of all similar areas
are measured in LF and entered in the report under condition state 4. If the section loss is
measurable, the amount of loss should be recorded and the reviewing engineer must determine if
the loss is minor or significant.
In the coding of Moveable Bearings the Pontis‟ Condition State 2 mentions the
supporting concrete may have cracked, and in Condition State 3 Pontis mentions the concrete
may be broken and no longer capable of supporting the bearing loads. The Pontis program is
remiss in not directing the inspector to refer the condition to an engineer for evaluation.
In the case of Bridge 9340 the condition of frozen roller bearings was identified in 1982
but never considered as dangerous.
Even when identified as a condition 3 deficiency, neither the reviewing engineer nor the
writers of the Pontis program identified the condition of inoperative roller bearings as a Critical
VI Final Comments
Since my retirement as the state bridge inspections engineer I have served other agencies
as a consultant engineer in bridge inspections. The inspections are primarily written for the
bridge owner‟s engineer in simple to understand terms (user friendly). The structure is reported
as primary units in conditions as Good, Fair or Poor, and the condition is described in a narrative
statement. Of particular value to the engineer is the comments made under the headings:
“Changed Conditions and Hazardous Conditions.”
Page 9 of 9
NBI ratings are listed on the reverse side of the report along with a summary of the Pontis
elements and condition codes.
When reviewing the Pontis report for condition of the truss we find the truss has xx no. of
LF in condition state 4 where the paint is identified as “areas of rust with minor section loss” and
no area of significant rust with section loss. If the significant area of “rust with section loss” on
gusset plates at L11E had been recorded in condition state 5 the 8 LF in condition state 5 as a
percentage of the total LF of the truss would have been rounded off to zero.
Inspection reports for large and complex structures, such as the I35W Bridge, should not
be written in the standard Pontis format that is acceptable for culverts. Where the various spans
are of different designs the inspection report should be by type. On I35W the steel beam
approach spans were multiple girders with frequent out-of-plane bending cracks. This one
deficiency codes the entire bridge, including the deck truss and voided concrete slab spans, as an
NBI rating of four and the definition of structurally deficient. One size doesn‟t always fit all.
And, finally, just as one inspection format is not acceptable for all structures, neither is
one level of qualified inspector acceptable for all structures. MnDOT has flooded its ranks of
certified bridge inspectors with personnel from all worker classifications with little regard for the
capability of these individuals. By sending the applicants to three weeks of classroom training,
where less than 1 percent fail the final exam, and having them work with other inspectors for a
period of five years, they now become fully qualified by law to inspect all types of bridges
including the I35W Bridge. To enhance the quality of the inspection MnDOT requires
applicants to take a field test to show proficiency in making bridge inspections and like the
classroom very, very few ever fail this field test.
Because the qualification of bridge inspectors is set by state statutes, the certification of
bridge inspectors is unnecessary and only serves to keep other more qualified individuals, such
as professional structural engineers and bridge design engineers, from making bridge safety
inspections, and assures the lesser qualified maintenance worker permanent tenure.
Written by Ronald W. Olds, P.E. – Ret.