Predicting Bridge Life-Cycle Deterioration By Integrating Condition States with System Performance -- A NSF Career Development Award

The overall goal of this research is to develop life-cycle prediction models for use in bridge condition assessment. A critical component in achieving this goal is partnering with state highway departments.  Through these efforts data collection methods will be improved and analytical relationships will provide a link between component and system deterioration leading to improved prediction models in bridge management systems.

[ Back to Research or SCA Home ]

Background: At a time when our civil infrastructure systems are in dire need of renewal, efforts are being made to improve maintenance procedures and condition assessment practice.  Innovative techniques which include the implementation of new technologies and bridge management systems give bridge inspectors and engineers the necessary information to determine an appropriate action.  Often, such a decision is dependent on a combination of the quantitative information obtained from various measurements, qualitative information obtained from a bridge visit, and general engineering knowledge about the entire bridge system.  To properly allocate funds, a bridge owner needs a bridge management system which uses historical deterioration trends and predicative relationships. Combining existing management systems such as PONTIS with bridge specific deterioration models (which consider the systemsí structural behavior) will improve a highway departmentís ability to make bridge specific decisions and allocate funds.

[ top ]

Research Focus: This research addresses current issues related to infrastructure renewal.  The proposed work is integrated into a research and education program which revolves around topics related to infrastructure renewal engineering.  As part of the research program, performance-based prediction models for bridge deterioration will be developed.   The prediction models will track component and system parameters (such as component capacity and system load rating) as a function of time and will be based on the mixed quantitative and qualitative information obtained in the field.  By using case histories developed for individual bridges and isolating parameters which relate specific causes to bridge deterioration it will be possible to generalize the models to a set of parameters which consider bridge type, use, and environmental conditions.   The prediction model will consider changes in structural performance due to deterioration and damage.

This research will evaluate the integrity and performance of highway bridges located in the Massachusetts Highway Department (MHD), the New York State Department of Transportation (NYSDOT), and the California Department of Transportation (Caltrans) bridge inventories.  The results of this research will include a new method to consider structural uncertainties applicable to a wide range of structural systems and applications.   Additionally, a link between component condition state and overall system performance will be created.  It is important that this assessment be performed for individual structural systems to determine the risk for a particular bridge structure in addition to a regional evaluation.  Both models will support practices already in place by highway departments while making fundamental contributions to conventional evaluation and analysis techniques.

[ top ]

Current Limitations: Currently, bridge inspection data is primarily qualitative and prediction models are obtained empirically based on statistical trends. Previous inspections can provide global information for the deck, superstructure, and substructure.  Bridge management systems (BMS) have made it possible to track larger amounts of data and have incorporated information at the component level thereby, improving resource allocation.  However, the limitations in the data collected and stored in  BMS (and other storage mediums) which hinder the evaluation of system performance are:

  • There is a lack of  a relationship between condition states to component performance and overall structural performance or behavior.
  • The prediction models are based on statistical relationships of visual condition ratings and do not consider the cause of deterioration or damage as governing parameters.
  • BMS do not facilitate structural performance evaluation due to damaging events.
  • There is no standard method to combine the qualitative condition states with the quantitative measurements obtained from NDE and controlled testing.  The quantitative measures obtained from detailed evaluations are not combined with condition ratings.
[ top ]

Research Objectives:  A life-cycle model is needed to predict bridge performance which considers historical trends, physical phenomena which cause the deterioration or damage, structural mechanisms, and system improvement (i.e. maintenance).  The objectives for the research component of this career development plan are:

  • To relate deterioration trends obtained from case-histories into performance based life-cycle models developed as a result of this research.
  • To create a link between prediction models and qualitative information in BMS with other visual and quantitative data. These enhancements include:  integrating quantitative information such as damage indicators obtained using instrumentation, considering the cause of the deterioration and mapping component deterioration to system deterioration.
  • To work with transportation departments in Massachusetts, New York, and California and to develop and implement a performance-based prediction model which combines qualitative and quantitative condition assessment data.
  • To incorporate the uncertainty associated with future condition states by evaluating the use of different types of uncertainty models for bridge performance prediction.
  • To develop damage models to assess a system performance at an arbitrary point in time.
[ top ]