National Bridge Inventory Data Research Articles

Evaluating geospatial accuracy for equity analyses: A case study of the U.S. National Bridge Inventory

This short communication seeks to develop and evaluate a generalizable methodological approach for geospatially matching infrastructure assets to socioeconomic data to better facilitate equity-informed transportation decisions. This study focuses specifically on providing insight into the spatial accuracy of the U.S. Department of Transportation (DOT)'s National Bridge Inventory (NBI) dataset at a national level via a method that is less computationally or manually intensive than past approaches. Phase 1 focuses on assessing the quality of existing NBI locational data using a geospatially-driven method. We find that 98.8% and 95.4% of the NBI records matched the reported state and county tag respectively, according to their reported latitude and longitude. Phase 2 focuses on potential ways to correct locational mismatches. We find that a 30.5 m (100 ft) buffer and Thiessen polygons provide the greatest improvements in matching accuracy in the NBI record over point coordinates. Phase 3 assesses the feasibility of enhancing location granularity through a hand-checked case study. We find that for coarser geospatial levels (state and county), equity-based analyses may be possible, but for more granular geospatial levels (place and census tract), which is where equity-based analyses are often conducted, the data may need greater buffering and hand checks to ensure accuracy. To use more automated processes at these high spatial resolutions, governments and other policy-making entities will likely need to require quality control checks of longitude and latitude infrastructure data. This work provides the first systematic analysis of the robustness of the NBI data for equity-based analyses, as well as possible policies to improve its usage for equity analyses.

Survival Analysis Using Cox Proportional Hazards Regression for Pile Bridge Piles Under Wet Service Conditions

This paper studies the deterioration of bridge substructures utilizing the Long-Term Bridge Performance (LTBP) Program InfoBridgeTM and develops a survival model using Cox proportional hazards regression. The survival analysis is based on the National Bridge Inventory (NBI) dataset. The study calculates the survival rate of reinforced and prestressed concrete piles on bridges under marine conditions over a 29-year span (from 1992 to 2020). The state of Maryland is the primary focus of this study, with data from three neighboring regions, the District of Columbia, Virginia, and Delaware to expand the sample size. The data obtained from the National Bridge Inventory are condensed and filtered to acquire the most relevant information for model development. The Cox proportional hazards regression is applied to the condensed NBI data with six parameters: Age, ADT, ADTT, number of spans, span length, and structural length. Two survival models are generated for the bridge substructures: Reinforced and prestressed concrete piles in Maryland and reinforced and prestressed concrete piles in wet service conditions in the District of Columbia, Maryland, Delaware, and Virginia. Results from the Cox proportional hazards regression are used to construct Markov chains to demonstrate the sequence of the deterioration of bridge substructures. The Markov chains can be used as a tool to assist in the prediction and decision-making for repair, rehabilitation, and replacement of bridge piles. Based on the numerical model, the Pile Assessment Matrix Program (PAM) is developed to facilitate the assessment and maintenance of current bridge structures. The program integrates the NBI database with the inspection and research reports from various states’ department of transportation, to serve as a tool for condition state simulation based on maintenance or rehabilitation strategies.

Bridge Infrastructure Management System: Autoencoder Approach for Predicting Bridge Condition Ratings

Being an essential component of our economy, bridges play a vital role in facilitating transportation of people and goods all over the world. Bridge infrastructure systems help in developing an effective way to monitor and protect structures in all aspects. However, the bridges are exposed to various kinds of damages due to aging, heavy load of traffic, quality of construction, and so on. Hence, determining the condition of such bridges through a proper infrastructure management system is important for understanding the potential loss in the longevity of the structures. This research paper presents the development and evaluation of an autoencoder-random forest (AE-RF) model to predict the condition rating of bridges using the National Bridge Inventory (NBI) database. To demonstrate the proposed model, a case study using bridges present in the US state of Florida has been performed using historical NBI data (2011–2020). Through this research, it was identified that when one of the deep learning models, named autoencoder, is combined with random forest (RF), it results in an efficient model for determining the condition ratings of the bridge components with fewer input parameters. The developed model was about 90% accurate in predicting the bridge’s deck condition by using other rating values as input variables and about 79% accurate without the use of any other rating factors, thereby addressing the existing research gap in determining the condition rating without the use of historic condition rating. On the other hand, the model was 78% and 77% accurate in determining the condition ratings for superstructure and substructure without using historic ratings and evaluation parameters. Hence, the proposed model would be more reliable in the evaluation of condition of existing bridges across the nation.

Investigating Factors Influencing Deck Conditions of Concrete Bridge and Steel Bridge Using an Interpretable Machine Learning Framework

Good deck condition is critical for a bridge to transport vehicles across physical obstacles on the landscape. This study uses the National Bridge Inventory data provided by LTBP InfoBridge to assess the influence of weather, structure, and traffic factors on deck conditions of concrete and steel bridges in the ten states with the highest percentage of bridges with poor deck conditions. For this analysis, an interpretable machine learning framework takes advantage of the high prediction accuracy while maintaining interpretability. A model is built for each structure type producing results that reveal the impacts of the factors on deck conditions for both types of bridges. The effects of various factors on the deck conditions of steel and concrete bridges are analyzed by comparing the similarities and differences among these factors. The results show that the concrete cast-in-place is the only deck structure associated with good deck conditions for steel bridges, while multiple other structure types associate with good deck conditions for concrete bridges. For the wearing surface, the bituminous wearing surface is highly associated with poor deck conditions of steel bridges. The number of snow days is negatively associated with the deck condition of steel bridges, while the concrete bridge could maintain a good performance with extremely long snow days. The findings also confirmed that increasing the truck percentage could deteriorate the deck conditions faster.

Implementation of a Mathematical Model for the Prediction of the Future Condition Rating for Bridge Components

Bridges are continuously exposed to environmental changes and dynamic loading effects caused by moving loads. As a result, bridge deterioration is a critical problem in the U.S.A. The National Bridge Inventory (NBI) database contains historical bridge condition information for bridges in the U.S.A. and currently it is the best available database for describing the historical condition of bridges in the U.S.A. The objective of this paper is to develop a mathematical model that can be used to predict the future condition ratings of each bridge component and, more specifically, to estimate the probability of each bridge component being at any condition rating at any future year. Two different types of future condition rating prediction models, namely the current practice model (CPM) and the deterioration prediction model (DPM), were developed. The CPM is capable of simulating the effects of historical maintenance activities when predicting the future condition rating probabilities, whereas the DPM does not consider the effects of historical maintenance activities when predicting the future condition rating probabilities. Both models were illustrated and validated using most current NBI data. The performance of both models was evaluated on hundreds of bridges in the states of Iowa and Wisconsin. The results indicated that the CPMs tend to converge to condition rating 6 within 15 years, whereas the DPMs tend to converge to condition rating 4 with 15 years. This suggests that conducting current maintenance activities helps to keep the nation’s bridges in at least “satisfactory condition.” However, a lack of performing any maintenance could lead to bridges being structurally deficient within 15 years.

Enhancing Reliability Analysis with Multisource Data: Mitigating Adverse Selection Problems in Bridge Monitoring and Management

Data collected using sensors plays an essential role in active bridge health monitoring. When analyzing a large number of bridges in the U.S., the National Bridge Inventory data as been widely used. Yet, the database does not provide information about live loads, one of the most indeterminate variables for monitoring bridges. Such asymmetric information can lead to an adverse selection problem in making maintenance, rehabilitation, and repair decisions. This study proposes a data-driven reliability analysis to assess probabilities of bridge failure by synthesizing NBI data and Weigh-In-Motion (WIM) data for a large number of bridges in Georgia. On the resistance side, tree ensemble methods are employed to support the hypothesis that the NBI operating load rating represents the distribution of bridge resistance capacities which change over time. On the loading side, the live load distribution is derived from field data collected using WIM sensors. Our results show that the proposed WIM data-enabled reliability analysis substantially enhances information symmetry and provides a reliability index that supports monitoring of bridge conditions, depending on live loads and load-carrying capacities.

Performance of Aging Timber Bridges based on Field Tests and Deterioration Models

This study investigated the performance of timber trestle railroad bridges through field tests on aging open- and ballast-deck bridges. It also developed nonparametric deterioration models based on average time in condition rating (ATICR) and Kaplan–Meier (K-M) estimates to assess the performance of timber highway bridge components using National Bridge Inventory data. This study proposes ATICR-based deterioration models for predicting the condition of timber bridge components. Average service life of highway timber multibeam bridge components under routine maintenance without any reconstruction or rehabilitation was found to be 23 years for the decks, 26 years for the superstructures, and 27 years for the substructures. K-M estimate-based reliability models were developed for probabilistic prediction of bridge safety while accounting for partial information from incomplete bridge condition observations. Based on these models, the median survival year—the number of years for which 50% of the bridge population remains in its current condition without deterioration—was obtained as 10 years for the decks, 14 years for the superstructures, and 12 years for the substructures of timber multibeam bridges. The models also suggested there was an 80% chance the timber bridge components would remain in their current condition without deterioration to a lower condition for a minimum of 3 years. The analysis carried out in this study could help improve the management and maintenance of timber bridges that are currently in operation worldwide.

Risk-based life-cycle analysis of highway bridge networks under budget constraints

This paper describes the life-cycle analysis of the risk to highway bridge networks subjected to overweight traffic loads under maintenance budget constraints. The risk to the network is initially evaluated based on the current condition of its constituent bridges and then the life-cycle analysis is performed to identify the level of risk reduction that is possible given a particular budget allocation. The deterioration rate is extracted from National Bridge Inventory data. The risk is monitored as the deterioration of bridges progresses over the years. Bridges are ranked in descending order starting from the bridge whose failure has the highest impact on network risk. Maintenance is scheduled sequentially on the riskiest bridges until the entire budget allocation is depleted. The methodology is illustrated using as an example the highway network of major interstate and state roads in New York State. The paper compares the results of the proposed life-cycle risk analysis for several budget levels. The results show that network risk can be reduced from its present level by 20% if current expenditure rates are maintained. Risk reduction can reach 37% if the budget for bridge rehabilitation is increased to match that recommended in the ASCE report card for America’s infrastructure.

Binary logistic regression approach for decision making in bridge management

Bridge management professionals need effective tools to help guide the decision making process and maintain quality infrastructure in a region. Using logistic regression for a novel binary response defined by at-risk and not-at-risk bridges based on their existing overall condition scores, the probability of a bridge being at risk is expressed in terms of the primary bridge factors age, load, the type of construction material and structural design and conditions of the deck, superstructure and substructure. These estimated probabilities multiplied by specified consequence values are used to introduce the risk classes and their ranks. Employing the method for training and validating sets of sizes 13 540 and 3385 in 2017 and 13 481 and 3370 in 2018 data in the National Bridge Inventory (NBI) Indiana, a statistically significant model is established containing the age, load and conditions of both superstructure and substructure. Moreover, at-risk bridges are identified from Indiana NBI data in both years and for a subset from Connecticut in 2017. The innovative bridge-ranking tool prioritises bridges for maintenance purposes such as replacing or repairing and hence efficiently guides the management in the decision making process for capital expenditures and perhaps for predicting the missing overall bridge condition.

Bridge condition rating data modeling using deep learning algorithm

This paper presents a deep learning-based bridge condition rating data modeling approach using selected data from the National Bridge Inventory (NBI) database. The objective of this research is to develop a data-driven approach that enables prediction of future conditions of highway bridge components from historical inspection data. The problem is solved by training a Convolutional Neural Network (CNN) model with online available NBI data. One prominent feature of the CNN model is that if well-trained it can represent the high dimensional data in the dataset abstractions for which conventional mathematical models may be difficult to describe. A case study of Maryland and Delaware highway bridges using historical data (1992–2017) sourced from the NBI database has been performed to demonstrate the proposed method. CNN models for three primary components of these highway bridges including the deck, superstructure, and substructure have been established. Optimization of model parameters is achieved through a parametric study. Research findings suggest that the deep learning model offers a promising tool as a data-driven condition forecasting approach for bridge components with a demonstrated prediction accuracy over 85%.

Application of Machine Learning Techniques for the Analysis of National Bridge Inventory and Bridge Element Data

The MAP-21 Act requires information on bridge assets to be at the element level for management operations in the U.S.A. This approach has the objective of improving future predictions of the performance of bridge assets for a more precise evaluation of condition and correct allocation of management funds to keep bridges in a good state of repair. Although bridge element conditions were introduced in the 1990s, the application of such data had never been mandatory for bridge asset management until 2014, therefore, the amount of historical data on bridge element (BE) condition is still limited. On the other hand, National Bridge Inventory (NBI) ratings have been collected since the 1970s and a wide range of data are available. Therefore, it is natural to ask whether BE condition can be predicted using NBI data. In the past, researchers statistically related BE and NBI data, but little has been done to revert NBI to BE. This paper addresses both challenges of mapping BE–NBI condition data using several machine learning techniques. The results of the analysis of these techniques applied to a sample of about 9,000 bridges from northeastern states of the U.S.A. shows that between 79.8% and 100% of the NBI ratings for deck, superstructure, and substructure can be predicted within a rating error of ± 1. The back-mapping operation of NBI time-dependent ratings to BE deterioration profiles for deck, superstructure, and substructure can also be predicted accurately with a probability greater than 50% at the 95% confidence level.

Drivers of Bridge Decommissioning in the United States

Extending the useful life of bridges through better design, construction, and management is a shared effort among the bridge management community. Data in the National Bridge Inventory (NBI) is valuable for understanding the behavior of bridges throughout their service lives. While the primary reason for bridge replacement, intuitively, would be condition, research has indicated that bridge replacement may not necessarily be driven by the condition of the bridge. The estimated median service life for bridges, 53 years, is much lower than the desired service life of 75 years. This paper summarizes the results of an NCHRP Project which identified the main drivers for bridge decommissioning in the United States, based on findings from three complementary analyses using historic NBI data files, select agency records, and data from old and new structure pairs. A common finding in previous studies was that a significant portion (15–30%) of decommissioning could not be associated with any particular reason. Although poor condition is a significant factor, the major driver of bridge decommissioning is functional improvement, and this explains the majority of the unexplained cases. Structures replaced due to functional reasons tend to be replaced at a younger age, leading to a decrease in the overall decommissioning age. While decisions on functional improvement projects are not led by bridge offices, bridge networks are substantially affected by these decisions. Coordinating functional improvement decisions at the agency level and integrating relevant information with decision support tools can improve financial planning and asset management processes.

Erratum for “Hybrid Particle Swarm Optimization and K-Means Analysis for Bridge Clustering Based on National Bridge Inventory Data” by Silvia Galvan-Nunez and Nii Attoh-Okine

Erratum for “Hybrid Particle Swarm Optimization and K-Means Analysis for Bridge Clustering Based on National Bridge Inventory Data” by Silvia Galvan-Nunez and Nii Attoh-Okine

Regression Model Evaluation for Highway Bridge Component Deterioration Using National Bridge Inventory Data

Accurate prediction of bridge component condition over time is critical for determining a reliable maintenance, repair, and rehabilitation (MRR) strategy for highway bridges. Based on bridge inspection data, regression models are the most-widely adopted tools used by researchers and state agencies to predict future bridge condition (FHWA 2007). Various regression models can produce quite different results because of the differences in modeling assumptions. The evaluation of model quality can be challenging and sometimes subjective. In this study, an external validation procedure was developed to quantitatively compare the forecasting power of different regression models for highway bridge component deterioration. Several regression models for highway bridge component rating over time were compared using the proposed procedure and a traditional apparent model evaluation method based on the goodness-of-fit to data. The results obtained by applying the two methods are compared and discussed in this paper.

Hybrid Particle Swarm Optimization and K-Means Analysis for Bridge Clustering Based on National Bridge Inventory Data

AbstractBecause of budget constraints for maintaining highway bridges in the United States, it is necessary to accurately monitor bridge structures so that repair and rehabilitation strategies can be performed. In this paper, an optimization approach based on the hybrid of the metaheuristic particle swarm optimization and the k-means method (KPSO) in data clustering is presented. The goal is to group bridges by similar structural deficiency attributes by minimizing the sum of squares error associated with assigning data points to each cluster and the determination of the most suitable number of clusters. The presented approach was compared to the basic version of particle swarm optimization (PSO) and the traditional clustering method k-means. The algorithms were tested using the National Bridge Inventory (NBI) database. The results show that KPSO provides better results in terms of the objective function as well as showing an opportunity to implement optimization techniques for data analysis in civil infr...

Automatic Logical Inconsistency Detection in the National Bridge Inventory

Studies about the data quality of National Bridge Inventory (NBI) reveal missing, erroneous, and logically conflicting data. Existing data quality programs lack a focus on detecting the logical inconsistencies within NBI and between NBI and external data sources. For example, within NBI, the structural condition ratings of some bridges improve over a period while having no improvement activity or maintenance funds recorded in relevant attributes documented in NBI. An example of logical inconsistencies between NBI and external data sources is that some bridges are not located within 100 meters of any roads extracted from Google Map. Manual detection of such logical errors is tedious and error-prone. This paper proposes a systematical “hypothesis testing” approach for automatically detecting logical inconsistencies within NBI and between NBI and external data sources. Using this framework, the authors detected logical inconsistencies in the NBI data of two sample states for revealing suspicious data items in NBI. The results showed that about 1% of bridges were not located within 100 meters of any actual roads, and few bridges showed improvements in the structural evaluation without any reported maintenance records.

Estimating Inspection Intervals for Bridges Based on Statistical Analysis of National Bridge Inventory Data

This paper describes a statistical approach to estimating an appropriate inspection interval for highway bridges based on historical condition data collected from 20 years of routine inspections. Condition ratings of 4,270 bridges with concrete, steel, and prestressed-concrete superstructures were analyzed in this study to determine the time-in-condition rating (TICR) for superstructure components. The Anderson-Darling test was used to assess five conventional distributions to determine goodness of fit for describing the distribution of the TICR probabilistically. The test results revealed that the Weibull distribution was the best fit for the TICR for superstructure components. Parameters of the Weibull distribution were estimated for concrete, steel, and prestressed concrete in different condition ratings and used to define probability-density functions (PDFs), which were then used to determine the likelihood of a bridge deteriorating from a good condition to a poor condition over some time period. These data can be used for rational decision making regarding the appropriate inspection interval for bridges of a given type based on quantitative historical analysis of data. This study is beneficial because of the simplicity of the models developed and their accessibility for practicing engineers.

The Influential Factors and Association Rules for Bridge Deck Deterioration - Case Study of National Bridge Inventory Data

Bridge is the hub of a road, playing an important role in transportation. With the age of bridges increasing in recent years, the condition of bridges is deteriorating. Bridge deterioration usually results from multiple factors. Studies in exploring the factors leading to bridge deterioration and its state estimation are booming. However, studies in the relationship between bridge deterioration and external environment by using database are still limited. Since 1992, National Bridge Inventory (NBI) has collected basic and historical information of bridges in every state, a very rich database which should include some valuable knowledge. Hence, through clustering and classifying in data mining, this study explores association rules for bridge deterioration, which is displayed in decision tree with clear routes for reference. Using association rules developed by thus study, bridge maintenance personnel can understand the future state of the bridge under their jurisdiction, respond to risks and costs of bridge deterioration, extend the service life of bridges, and protect the safety of road users.

A Web-Supported National Bridge Inventory Management Tool

The U.S. National Bridge Inventory (NBI) contains approximately 600,000 bridges. Its size, numerous items, and electronic format make management and analysis of its data difficult. The objective of this study is to develop a tool that will provide a user-friendly environment as well as advanced services for accessing and analyzing NBI data sets. The tool can import text- or spreadsheet-based NBI data sets, provide tools for managing and analyzing them, and use Web services for pinpointing bridge positions. It is considered useful for managing the NBI data sets and for further examining and analyzing the characteristics and condition of the NBI.

Estimating the Future Condition of Highway Bridge Components Using National Bridge Inventory Data

The accurate prediction of the future condition of bridge components is an important part of any bridge management system. Past bridge inspection data along with information on any repair and/or retrofit can provide a baseline for predicting future conditions of bridge components. As expected, such data are subject to a rather large uncertainty, primarily due to the variation in the inspection process. This uncertainty is also caused sometimes by unrecorded repairs and/or replacements conducted on various parts of the bridge. If not properly considered in the bridge data analysis, the uncertainty may result in an erroneous prediction for future bridge conditions. To develop a reasonable estimate for future bridge conditions, this paper discusses two possible methods. In these methods, discrepancies inherent in bridge condition ratings, that may have been due to unrecorded improvement works, are removed to arrive at more consistent estimates for future bridge conditions. In one method, the adjustment in condition ratings is done based on the notion that unless there is evidence of improvement work, the condition rating cannot be larger than previous ratings. In the other method, the duration between consecutive inspections is used as a means to construct deterioration curves. These methods are applied to rating data collected from 2,601 Illinois bridges in the period of 1976–1998. The results obtained by applying the two methods are presented in the form of bridge deterioration models, compared with one another and discussed in the paper.