Uniform Scaling of Performance Measures for Bridge Management Systems
Transportation asset management uses data and objective analysis to improve decision making, with the objective of providing the required level of service in the most cost effective manner. Asset management systems are tools to quantify relevant performance measures, and forecast future outcomes as they are affected by agency policies, resource levels, and exogenous factors.
Owners of transportation infrastructure are expected to satisfy a wide range of stakeholder expectations for performance outcomes. Under limited funding, these expectations may conflict with each other. For example, achieving a higher safety objective may require drawing resources away from capacity projects, thus having a negative effect on mobility. Efforts to analyze the tradeoffs among performance measures are made difficult by the dissimilar measurement systems used; for example, condition is often measured in terms of element condition states and health index, while safety is in terms of accident rate and mobility is in terms of delay time and distance. In a multi-objective framework, it is necessary to have a way to standardize the scaling of each relevant performance measure, to make it possible for decision makers to make valid comparisons among dissimilar objectives.
NCHRP Project 12-67, published as report 590, developed a utility function framework to enable the optimization of bridge program management decisions considering multiple objectives. It described several methodologies for scaling performance measures to make it possible to compare and weight them in utility functions. The project demonstrated the methods in an informal way, but did not attempt a formal specification or estimation of scaling functions, instead leaving it to each individual agency.
AASHTO, state governments, and international agencies have begun to develop bridge management systems that rely on the methods in NCHRP Report 590. All of these systems will need objective scaling functions in order to realize the potential for multi-objective analysis. While the weighting of competing objectives is likely to vary from one agency to another, the uniform single-objective scaling of performance measures is believed likely to be similar across agencies and across asset types, even extending to asset types beyond bridges. Therefore a significant amount of effort may be spared by developing such scaling functions in cooperative research rather than in each agency individually.
The objective of this research project is to develop a recommended set of scaling functions, including quantitative parameters, for performance measures that would reasonably be used in bridge management systems to implement the utility function methodology of NCHRP Report 590.
In most cases a valid scaled performance measure is a unitless quantity ranging from 0 to 100, where equal intervals have equal value anywhere along the scale.
NCHRP Report 590 described the benefits of multi-objective analysis, in terms of assuring that bridge program decisions fully reflect the importance of all aspects of transportation system performance. This is especially important for considerations of risk, safety, and mobility that are not well represented in traditional life cycle cost analysis.
Implementation of multi-objective analysis requires the development of utility scaling functions, which Report 590 left to each individual agency. The effort to develop these scaling functions may be considerable, and would have to be repeated separately for each agency. It is believed, however, that the results would be very similar from one agency to another. Many agencies would lack the resources to develop their own functions, or to apply the necessary careful study design to ensure that the scaling functions are objective and credible to a wide range of stakeholders typical of asset management.
Performing this study once and for all, would save each state from having to do it individually, thus reducing the total cost by potentially 90-95%, taking advantage of the power of cooperative research.
Relationship to existing body of knowledge
Existing research studies, including NCHRP Report 590 and the FHWA Long-Term Bridge Program, have identified lists of relevant performance measures. Some of these measures are well documented, as in the AASHTO Guide Manual for Bridge Element Inspection and the FHWA National Bridge Inventory Coding Guide. Other measures of performance, especially risk, may need to be gathered and synthesized from current research in order to specify a list to explore further.
The level of service approach to performance assessment may provide a means of standardizing the descriptions of performance measures without tying these measures to specific measurement technology. This approach is commonly used in Maintenance Quality Assurance programs and described in the AASHTO Guide for Asset Management Volume 2: Focus on Implementation. Levels of service correspond to definitions of acceptable or unacceptable performance, and are further refined by identifying the criteria where preventive maintenance, risk mitigation, or life extension actions are applicable.
Currently the only research report that specifically addresses scaling functions for bridge management performance measures, is NCHRP Report 590. This report describes a set of applicable methods but does not try to quantify them. The proposed research will include the data collection activity necessary to quantify and validate scaling functions.
It is envisioned that the research will involve at least the following tasks:
- Gather and synthesize the performance measures relevant to bridge management systems, including domestic and international practice. The research should consider measures that may become possible with the completion of the FHWA long-term performance program, and measures that may be broadly applicable across asset management (not strictly limited to bridges). This task includes a literature review which extends to updating the NCHRP Report 590 literature review as concerns scaling functions.
- Identify a sample panel of 50-100 bridge owners and other stakeholders and experts, representing the full range of such agencies by geography, level of government, size, climate, and organizational environment. The panel could exceed 50 members when considering the various organizational perspectives (engineering, planning, regional, etc.) in each agency. The sample panel will review the list of performance measures to be considered in the research, and will provide the expert elicitation data required in the next task, by electronic means.
- For a set of applicable scaling methods, including but not limited to the methods discussed in NCHRP Report 590, gather and synthesize expert opinion required for development of scaling functions. This includes development of one or more survey instruments and execution of a survey. One or more webinars and/or in-person gatherings of the sample panel will be necessary to ensure panelist understanding and maximize the quality of the data.
- Use statistical estimation methods to develop scaling functions for the applicable methods. Measure the level of consistency and reliability of the various methods for each performance measure. Quantify the variance of scaling parameters by level of government, geography, climate, size, or other relevant agency characteristics.
- Recommend and document a scaling function, including quantitative parameters, for each performance measure. Describe the variance of parameters according to agency characteristics. Recommend procedures to customize the scaling functions for each agency.
- Demonstrate the application of the scaling function in an example multi-objective optimization context. This could be done using Pontis 5.2, or another bridge management system or model that is able to effectively demonstrate the method in action.
- Prepare a Final Report.
This research is envisioned to feed directly into existing efforts to develop multi-objective optimization models in bridge management systems. It will provide default models and methods each agency can use as it customizes BMS models for its own use. Continued FHWA and AASHTO support for bridge management systems, and in general for asset management and performance management, will help to ensure successful and widespread implementation.
Estimated funding requirements
This research is estimated to cost $300,000 and require 24 months to complete, including the Final Report.