Many structures maintained by transportation agencies are constructed over waterways that are susceptible to various degrees of scour and stream instability. If not prevented or arrested, scour and instability can ultimately lead to catastrophic failure, with possible loss of human life and high cost of facility replacement. Relatively few experts exist who can use their judgment and experience to effectively address scour and stream stability problems, and, yet, the solution to these problems invariably relies on such individuals. The nature and importance of the problem lead to the potential application of knowledge-based expert systems technology. These systems are particularly applicable to ill-structured problems, where a mathematical solution is not available or is impractical, and the problem must be solved using expert skill, judgment, and experience.

One area of particular concern is the need to provide assistance to field investigators in the assessment of existing structure sites for the potential dangers of scour and stream instability. Field investigators must be aware of the types of data required to make decisions and identify sites for more detailed evaluations.

The objective of this research is to develop an operational, microcomputer-oriented, knowledge-based expert system to advise field personnel in evaluating current and potential scour and stream stability problems and in identifying the urgency for appropriate countermeasures or additional detailed analyses. The resulting information shall be suitable for use by others in the performance of detailed engineering analyses if required. The system shall operate on field-worthy equipment and use selected analytical and heuristic models, in addition to its primary function of capturing the expertise of one or more experts in hydraulics, structures, geology, and structure maintenance and inspection. Transportation agencies will use the system to improve the productivity and effectiveness of structure inspection, maintenance, and replacement programs.

Accomplishment of this objective will require at least the following tasks: (1) Identify parameters and site conditions necessary to evaluate current and potential inland stream stability and scour problems. (2) Determine the relationships among parameters and site conditions and their relative importance for use in predicting trends in stream stability and scour potential. (3) Develop a conceptual plan for the expert system. The system shall include at least the following features: (a) Minimize variations in responses among users through the use of standardized graphic images and photographs. (b) Serve as a diagnostic tool for bridge inspectors and maintenance personnel to assist in selecting bridge rating codes for Items 60, 61, 71, and 113 on the Federal Highway Administration Structure Inventory and Appraisal Sheet, and monitoring conditions that indicate the changing level of stability and scour at structure sites. (c) Be consistent with stream classifications found in the BRI-STARS Expert System for Stream Classification. (d) Provide advice to the user on the relative levels of instability of the stream and the categories of countermeasures or other potential courses of action to be considered. (e) Provide judgments sufficient to advise maintenance personnel or bridge inspectors when expertise outside of their purview is needed. (f) Generate reports that present pertinent information and graphics. The reports shall be capable of being produced on a variety of commonly used output devices. (g) Provide a capability for modification of decision logic variables such that local conditions can be more accurately represented. (4) Implement the conceptual plan developed and approved in Task 3. (5) Test and evaluate the expert system. (6) Refine the expert system and incorporate changes. (7) Prepare a final report documenting the research effort and provide all expert-system software including a well-documented version of the source code on diskette. A user's manual and program documentation manual will be appended to the final report. Any and all software developed will remain the property of the NCHRP and cannot be licensed and sold solely by the developing individual.

Phase II: To achieve the full benefits of the computer-oriented, knowledge-based expert system that was developed in the original research described above, the project panel amended the original scope of work to validate the program in the field and to broaden the expert input. Additional evaluations involving case studies in different parts of the country were also added to refine the system and further improve the systems decision-making skills.

Accomplishing these goals will require at least the following tasks: (1) Critique of the existing system by scour experts and end-users. (2) Refine the program logic to include additional observations and sites factors developed through Task 1. (3) Add user-interface improvements, including archival/retrieval of site-specific photographs. The collective outcome of Tasks 1 through 3 will be a new version of the system that will be ready for final testing and deployment to transportation agencies approximately 1 year after Phase II begins. (4) Validate and calibrate the refined program in 5 states, gathering an additional 5 case studies in each state through hands-on participation of the research personnel in each of the selected states. (5) Develop an online tutorial and multimedia training tool based on the field testing and review by experienced inspectors in Task 4. (6) Prepare a final report that summarizes the concepts, logic, and approaches used in the expert system, a report on system verification, a final users guide, and documentation of the tutorial and intended uses of the multimedia training tool.

Status: Phase II was initiated to further test and refine the system. The final report is finished and has been published as NCHRP Report 426. Further information about the current state of the program, CAESAR (Cataloging And Expert Evaluation of Scour Risk And River Stability at Bridge Sites), and a downloadable copy of the program in its present form, are available at the University of Washington's World Wide Web site (https://www.ce.washington.edu/~scour). The downloadable program runs on Windows 95. An amendment was processed in August 1998 to extend the completion date of this project to June 30, 1999, and to increase the funding by $15,000 to facilitate the implementation of this already successful project. The amendment allows for the contractor to: (1) improve program portability, (2) provide DOT user support, and ( 3) maintain the program homepage through June 1999.

Product Availability: As noted above, the software is available at the research agency's website and the final report is available as NCHRP Report 426 CAESAR: An Expert System for Evaluation of Scour and Stream Stability