Bridge Damage and Repair Costs from Hurricane Katrina

Hurricane Katrina caused significant damage to the transportation system in the Gulf Coast region. The overall cost to repair or replace the bridges damaged during the hurricane is estimated at over $1 billion. This paper describes the observed damage patterns to bridges, including damage attributed to storm surge, wind, impact from debris, scour, and water inundation, as well as examples of repair measures used to quickly restore functionality to the bridges and transportation system. Using the data from the 44 bridges that were damaged, relationships between storm surge elevation, damage level, and repair costs are developed. The analysis reveals that, in general, regions with higher storm surge had more damage, although there were several instances where this was not the case, primarily due to damage resulting from debris impact. It is also shown that a highly nonlinear relationship exists between the normalized repair cost and the damage state. The paper concludes with a brief discussion on the efficacy of using typical seismic design details for mitigating the effects of hurricane loads, and potential design considerations for bridge structures in vulnerable coastal regions.     

Pavement Structures Damage Caused by Hurricane Katrina Flooding

In September of 2005, Hurricane Katrina devastated New Orleans and caused sustained flooding. Limited pre- and postflooding tests indicated that the pavement structures tested were adversely impacted by the flood water. Consequently, the Louisiana Dept. of Transportation and Development hired an independent contractor to structurally test approximately 383 km (238?mi) of the region’s federally aided urban highway system both inside and outside of the flooding area. Falling weight deflectometer (FWD) tests were performed every 161 m (0.1?mi) over each selected roadway, along with other field tests. The FWD data were imported into a geographical information system and plotted against a USGS geo-referenced map. Comparative analyses were made possible through the use of extensive flood maps made available through NOAA and FEMA. This arrangement made it possible to classify spatially and graphically all test points on the basis of flooding versus nonflooding, short flooding duration versus longer flooding duration, shallow flooding versus deep flooding, and thin pavements versus thick pavements. Three pavement types, asphaltic concrete, Portland cement concrete, and composite, were considered in this analysis. The statistical inference about the difference in the means of compared data groups was conducted with 95% confidence.     

Observations of Structural Damage Caused by Hurricane Katrina on the Mississippi Gulf Coast

The loads associated with Hurricane Katrina led to the destruction or severe damage of approximately 130,000 homes and over 200 deaths in the state of Mississippi. This paper discusses the results of a field inspection of structural damage along the state’s Gulf Coast area caused by this hurricane. It was found that reinforced concrete, steel frame, and heavy timber structures generally performed well, with minimal structural damage. Precast concrete, light frame wood, and bridge structures generally performed poorly. Nonstructural components of all building types, in particular facades and interior partitions subjected to storm surge, were typically destroyed. For various structures, the primary cause of failure was found to be insufficient connection strength. A comparison of Katrina’s storm surge and wind loads is made to those specified in current design standards. It was found that Katrina’s forces exceeded those specified in design standards in many parts of the state.     

Roof Damage in New Homes Caused by Hurricane Charley

Hurricane Charley was the first Category 4 hurricane to strike Florida after 1992. This paper presents results of a study to investigate the performance of 425 of 747 roofs of new homes in Punta Gorda Isles, a subdivision of Punta Gorda that was directly in the path of Hurricane Charley shortly after it made landfall. The homes examined were larger, concrete/clay tiled roof homes having irregular floor plans and complex roof configurations not explicitly addressed by prevailing wind load codes. Roof damage was evaluated using images from aerial photographs taken at an elevation of approximately 762?m?(2,500?ft.) Specialized software was used to quantify damage. Damage was classified based on tile loss area. The study showed that the vast majority of the roofs were either undamaged or sustained minor damage. Fewer than 14% were classified as damaged. The most common observed tile loss was along ridges, corners, or in the hip zone where negative uplift pressures are recognized to be the highest. Given the modest observed damage, prevailing methods for estimating wind loads for irregular buildings specified in codes may be adequate. Problems encountered may be best resolved through new details for attaching tiles on ridges, corners, and hip zone.     

Hydrodynamic Investigation of Coastal Bridge Collapse during Hurricane Katrina

A number of U.S. coastal bridges have been destroyed by hurricanes, including three highway bridges in Mississippi and Louisiana during Hurricane Katrina (2005). This paper addresses three fundamental questions on the coastal bridge failures: (1) what were the hydrodynamic conditions near the failed bridge during the hurricane; (2) what was the cause of the bridge collapse; and (3) what was the magnitude of the hydrodynamic loading on the bridge under the extreme hurricane conditions. Guided by field observations of winds, waves, and water levels, two numerical models for storm surges and water waves are coupled to hindcast the hydrodynamic conditions. Fairly good agreement between the modeled and measured high watermarks and offshore wave heights is found, allowing an estimate of the surge and wave conditions near the bridges in nested domains with higher resolutions. The output of the coupled wave-surge models is utilized to determine the static buoyant force and wave forces on the bridge superstructure based on empirical equations derived from small-scale hydraulic tests for elevated decks used in the coastal and offshore industry. It is inferred that the bridge failure was caused by the wind waves accompanied by the storm surge, which raised the water level to an elevation where surface waves generated by strong winds over a relatively short fetch were able to strike the bridge superstructure. The storm waves produced both an uplift force and a horizontal force on the bridge decks. The magnitude of wave uplift force from individual waves exceeded the weight of the simple span bridge decks and the horizontal force overcame the resistance provided by the connections of the bridge decks to the pilings. The methodology for determining the hydrodynamic forcing on bridge decks can be used to produce a preliminary assessment of the vulnerability of existing coastal bridges in hurricane-prone areas.     

R-Factor Parameterized Bridge Damage Fragility Curves

Damage fragilities describe the probability that a bridge will incur certain (discrete) damage states conditioned on the intensity of the earthquake it may experience. Reinforced concrete box girder highway overpass bridges are prevalent among the total inventory of bridges in California. For this class of bridges, a method for computing damage fragilities for three damage states (concrete cover spalling, longitudinal bar buckling, and column failure) based on the bridge force reduction factors (R-factors) is derived in this paper. Bridge damage fragilities are described by equations relating the median intensity and uncertainty of ground motion to discrete damage states of column concrete cover spalling, column bar buckling, and column failure using the bridge R-factor as the principal parameter describing the bridge structure. Such damage fragility equations are furnished for earthquake intensities measured using pseudospectral acceleration (Sa) and cumulative absolute displacement (CAD) in both the bridge longitudinal and transverse directions.     

New Orleans Levee System Performance during Hurricane Katrina: London Avenue and Orleans Canal South

Hurricane Katrina was one of the worst natural disasters in U.S. history. The effects of the hurricane were particularly devastating in the city of New Orleans. Most of the damage was due to the failure of the levee system that surrounds the city to protect it from flooding. This paper presents the results of centrifuge models conducted at Rensselaer Polytechnic Institute and the U.S. Army Corps of Engineers simulating the behavior of the levees at London Avenue North and South that failed during Hurricane Katrina. Those levees failed without being overtopped by the storm surge. Also included are the results of a centrifuge model of one levee section at Orleans Canal South, which did not fail during the hurricane. The key factor of the failure mechanism of the London Avenue levees was the formation of a gap between the flooded side of the levee and the sheetpile. This gap triggered a reduction of the strength at the foundation of the protected side of the levee. The results are fully consistent with field observations.     

Development of FRP Short-Span Deployable Bridge—Experimental Results

For military and civilian applications, there exists a need for lightweight, inexpensive, short-span bridges that can be easily transported and erected with minimal equipment. Owing to its favorable properties, fiber-reinforced polymer (FRP) has been shown to be feasible for the construction of such bridges. Investigations into the behavior of a short-span bridge structural concept, adapted to the material properties of commercially available glass FRP (GFRP) pultruded products, are presented. A 4.8-m span prototype was built from GFRP sections, bonded throughout to form a tapered box beam, with a width of 1.2?m and a height at midspan of approximately 0.5?m. The box beam represents a single trackway of a double-trackway bridge, whose trackways could be connected by light structural elements. The quasi-static and dynamic behavior of the prototype box beam was investigated in ambient laboratory and field conditions to assess the design and construction techniques used, with a view to designing a full-scale 10-m GFRP bridge. Laboratory testing of the prototype box beam used single and pairs of patch loads to simulate wheel loading. These tests confirmed that the box beam had sufficient stiffness and strength to function effectively as a single trackway of a small span bridge. Field testing of the structure was undertaken using a Bison vehicle (13,000?kg), driven at varying speeds over the structure to establish its response to realistic vehicle loads and the effects of their movement across the span.     

Single-Span Prestressed Girder Bridge: LRFD Design and Comparison

This report summarizes the comparative design of a single-span AASHTO Type III girder bridge under the AASHTO Standard Specification for Highway Bridges, 16th Edition, and the AASHTO LRFD Bridge Design Specification. The writers address the differences in design philosophy, calculation procedures, and the resulting design. Foundation design and related geotechnical considerations are not considered. The LRFD design was similar in most respects to the Standard Specification design. The significant differences were: (1) increased shear reinforcement; (2) increased reinforcement in the deck overhang; and (3) increased reinforcement in the wing wall. The comparisons would likely change if the bridge were designed purely according to LRFD Specifications rather than as a comparative design. Design procedures under the LRFD Specification tend to be more calculation-intensive. However, the added complexity of the LRFD Specification is counterbalanced by the consistency of the design philosophy and its ability to consider a variety of bridges.     

Experimental Validation of a 10-m-Span Composite UHPFRC–Carbon Fibers–Timber Bridge Concept

This paper presents the experimental study of a new structure for a 10-m-span bridge deck, which takes into account the range of possibilities offered by new and high-strength materials along with the advantages of a traditional environmental friendly material. This 10-m-span element is formed by wooden beams braced at their ends on supports, a thin (7-cm-thick) upper slab made of precast ultrahigh performance fiber-reinforced concrete (UHPFRC), and fiber-reinforced polymer at the lower chord of these beams. The test program has been aimed at identifying the major critical aspects involved in producing an initial estimate of safety margins as well as validations of the design process and its underlying assumptions. Under the first loading configuration derived from live traffic loads, both the transverse and local bending of the thin UHPFRC slab were activated and confirmed by means of a three-dimensional finite-element computation. The second loading configuration corresponds to pure global longitudinal bending, with the bearing capacity being monitored up to the theoretical ultimate limit state loading and then beyond, up to experimental failure. Critical mechanisms and safety factors have also been identified. Though concept feasibility has been demonstrated, some aspects still need to be further optimized in order to obtain greater ductility and safer control over failure modes and occurrences.     

New Orleans and Hurricane Katrina. I: Introduction, Overview, and the East Flank

The failure of the New Orleans regional flood protection systems, and the resultant catastrophic flooding of much of New Orleans during Hurricane Katrina, represents the most costly failure of an engineered system in U.S. history. This paper presents an overview of the principal events that unfolded during this catastrophic hurricane, and then a more detailed look at the early stages of the event as the storm first drove onshore and then began to pass to the east of the main populated areas. The emphasis in this paper is on geotechnical lessons and it also includes broader lessons with regard to the design, implementation, operation, and maintenance of major flood protection systems. This paper focuses principally on the early stages of this disaster, including the initial inundation of Plaquemines Parish along the lower reaches of the Mississippi River as Katrina made landfall, and the subsequent additional early levee breaches and erosion along the eastern flanks of the regional flood protection systems fronting Lake Borgne that resulted in the flooding of the two large protected basins of New Orleans East and St. Bernard Parish. Significant lessons learned include (1) the need for realistic assessment of risk exposure as an element of flood protection policy; (2) the importance of considering erodibility of embankment and foundation soils in levee design and construction; (3) the importance of considering all potential failure modes; and (4) the problems inherent in the construction of major regional systems over extended periods of multiple decades. These are important lessons, as they are applicable to other regional flood protection systems in other areas of the United States, and throughout much of the world.     

Analytic Model of Long-Span Self-Shored Arch Bridge

An innovative self-shoring staged construction method was developed to build the world’s longest reinforced composite concrete arch bridge across the Yangtze River at Wanxian, in Chongqing, China. The method uses a steel tube truss frame constructed by the conventional cantilever launching technique. This steel frame with concrete-filled tubes performs the dual role of arch falsework and arch main reinforcement for the final reinforced concrete arch bridge. An optimized schedule for concrete placement was proposed to control the stresses, deflections, and stability of the arch rib during construction. The time dependent effects of concrete, the nonlinear stress-strain relationship of steel and concrete, as well as the geometric nonlinearility were considered. Control information at various stages of construction can be provided using the model developed. A program was developed to conduct parametric studies for selection of the final construction scheme and to direct the construction progress by monitoring and comparing actual and predicted stress and deflection.     

Temporal Nature of Fatigue Damage in Highway Bridges

To develop a valid method of estimating fatigue life of a highway bridge, it is first necessary to have a reasonable understanding of the manner in which fatigue damage occurs over time. This paper presents the findings from an extensive highway bridge monitoring program focused on monitoring vehicle-induced strain cycles; the cause of fatigue damage. The purpose of the study was to identify the temporal character (if any) of fatigue damage accumulation. A sample of 24 bridges in the State of Ohio was monitored; each bridge in the sample for 1?year. The data collected during the study captured hourly, daily, and monthly variation in fatigue damage accumulation. The raw data were resolved to a fatigue damage metric representation. The damage metric was analyzed with the intent of determining the temporal nature of fatigue damage accumulation. The analyses led to the development of postulates regarding fatigue damage in highway bridges. The postulates stated in this study should provide a sound basis for development of a fatigue-life estimation procedure using site-specific strain cycle histograms collected over an abbreviated time window.     

Study of Passive Deck-Flaps Flutter Control System on Full Bridge Model. I: Theory

A passive aerodynamic control method for suppression of the wind-induced instabilities of a very long span bridge is presented in this paper. The control system consists of additional control flaps attached to the edges of the bridge deck. Control flap rotations are governed by prestressed springs and additional cables spanned between the control flaps and an auxiliary transverse beam supported by the main cables of the bridge. The rotational movement of the flaps is used to modify the aerodynamic forces acting on the deck and provides aerodynamic forces on the flaps used to stabilize the bridge. A time-domain formulation of self-excited forces for the whole three-dimensional suspension bridge model is obtained through a rational function approximation of the generalized Theodorsen function and implemented in the FEM formulation. This paper lays the theoretical groundwork for the one that follows.     

Study of Passive Deck-Flaps Flutter Control System on Full Bridge Model. II: Results

Using the analytical format set forth in the first of these two companion papers, numerical simulations of a passive aerodynamic control method for suppression of the wind-induced instabilities of a very long span bridge are presented in this paper. At first, the aerodynamic stability of the uncontrolled system is discussed using a sectional model and full bridge model, respectively. Next, the efficiency of the proposed asymmetric and symmetric cable connection control systems is studied. Both systems offer similar maximum improvements in critical wind speed. The asymmetric cable connection system turns out to be very sensitive to horizontal motions, which strongly degrade its effectiveness. On the contrary, the symmetric cable connection system is not sensitive to horizontal motions, but its efficiency is limited by divergence.     

Retrofit of a Three-Span Slab Bridge with Fiber Reinforced Polymer Systems—Testing and Rating

A 45-year old, three-span reinforced concrete slab bridge with insufficient capacity was retrofitted with 76.2- and 127-mm wide bonded carbon fiber-reinforced polymer (FRP) plates, 102-mm wide bonded carbon FRP plates with mechanical anchors at the ends, and bonded carbon FRP fabrics. The use of four systems in one bridge provided a unique opportunity to evaluate field installation issues and to examine the long-term performance of each system under identical traffic and environmental conditions. Using controlled truckload tests, the response of the bridge before retrofitting, shortly after retrofitting, and after one year of service was measured. The stiffness of the FRP systems was small in comparison to the stiffness of the bridge deck, and accordingly the measured deflections did not change noticeably after retrofitting. The measured strains suggest participation of the FRP systems, and more importantly, the strength of the retrofitted bridge was increased. A detailed 3D finite-element model of the original and retrofitted bridge was developed and calibrated based on the measured deflections. The model was used to predict more accurately the demands for computing the rating factors. The addition of FRP plates and fabrics led to a 22% increase in the rating factor and corresponding load limits. During a one-year period, traffic loading and environmental exposure did not apparently affect the performance of the FRP systems. The increased capacity and acceptable performance of the FRP systems enabled the engineers to remove the load limits in order to resume normal traffic. Future tests are necessary to monitor the long-term behavior of the FRP systems.     

In-Service Diagnostics of a Highway Bridge from a Progressive Damage Case Study

Development of diagnostic and prognostic routines for application to in-service measurements from highway bridges necessitates analysis of experimental measurements from in-service highway bridges under natural or prescribed induced damage. This is generally limited to the unique opportunity of investigating end-of-service life bridges prior to reconstruction and consequently only a limited library of such case studies exist. This paper documents a field test of an end-of-service bridge span with prescribed progressive damage to a bearing as well as several diaphragm connections. Thirty dual-axis accelerometers were distributed across the bridge span with data acquisition and transmission facilitated by a real-time lossless wireless sensor network. A highway department service truck applied traffic excitation to the structure through routine passes on a consistent lane of traffic. Output-only system identification was applied to the baseline time history response to develop a state-space model of the bridge dynamics used for forward prediction in the form of a Kalman filter. Simple statistical evaluation of the prediction error in the model demonstrates the variance can be used to localize and generally quantify the degree of damage in the structure. The case study additionally illustrates the potential importance of monitoring lateral acceleration along the girders to permit identification of damage to elements, such as the diaphragms, that contributing primarily to the lateral and torsional response of primary structural members.     

Damage Identification on the Tilff Bridge by Vibration Monitoring Using Optical Fiber Strain Sensors

Vibration testing is a well-known practice for damage identification of civil engineering structures. The real modal parameters of a structure can be determined from the data obtained by tests using system identification methods. By comparing these measured modal parameters with the modal parameters of a numerical model of the same structure in undamaged condition, damage detection, localization, and quantification is possible. This paper presents a real-life application of this technique to assess the structural health of the 50-year old bridge of Tilff, a prestressed three-cell box-girder concrete bridge with variable height. A complete ambient vibration survey comprising both vertical accelerations and axial strains has been carried out. The in situ use of optical fiber strain sensors for the direct measurement of modal strains is an original contribution of this work. It is a big step forward in the exploration of modal curvatures for damage identification because the accuracy in calculating the modal curvatures is substantially improved by directly measuring modal strains rather than deriving the modal curvatures from acceleration measurements. From the ambient vibrations, natural frequencies, damping factors, modal displacements and modal curvatures are extracted by the stochastic subspace identification method. These modal parameters are used for damage identification which is performed by the updating of a finite element model of the intact structure. The obtained results are then compared to the inspections performed on the bridge.