Wee, S (1993) A prototype of an expert system for pavement maintenance and rehabilitation strategy in the state of Ohio (espresso). Unpublished PhD thesis, The Ohio State University, USA.
Abstract
A prototype of an Expert System for Pavement maintenance and REhabilitation in the State of Ohio (ESPRESSO) is developed, tested, and evaluated. ESPRESSO introduces an approach that employs an expert system to enhance a pavement management system (PMS) so as to provide consistent and reasonable strategies for maintenance and rehabilitation (M&R). The development of ESPRESSO consists of three phases: knowledge acquisition, knowledge representation, and knowledge base. The first phase for developing ESPRESSO was to acquire knowledge from a literature review and experts. For this purpose a knowledge acquisition protocol was specially designed and systematically used to facilitate this task. The knowledge acquisition protocol consists of four phases: the preliminary, intermediate, advanced, and organizational phases. The second phase was to develop the knowledge structure and representation based on knowledge acquisition. M&R strategy selection models were developed using decision trees. The decision trees consist of three levels: the evaluation of pavement condition, the M&R strategy for pavement condition, and the M&R strategy for pavement structure. The first level evaluates the pavement condition based on pavement condition rating (PCR) and structural deduct, obtained from a visual inspection of pavement surface condition by experienced engineers and technicians. The second level classifies the pavement condition into three categories: major rehabilitation, minor rehabilitation, and maintenance. In the third level, the M&R strategy for the pavement structure is selected based on the condition of the pavement structure using fuzzy logic concepts. Here, fuzzy modus ponens deduction techniques using angular fuzzy set models are used to find repair strategies for the pavement structure. The last phase was to construct the knowledge base for finding the M&R strategy. For this purpose an expert system shell, named LEVEL5 OBJECT, was chosen. ESPRESSO was tested by knowledge engineers and evaluated by independent experts in the Ohio Department of Transportation (ODOT). ESPRESSO recommended M&R strategies can be used by experienced engineers to check their recommendations. ESPRESSO can be used in the training of less experienced engineers. Recommendations for improvement of ESPRESSO and further research are given.
Item Type: | Thesis (Doctoral) |
---|---|
Thesis advisor: | Hadipriono, F C |
Uncontrolled Keywords: | fuzzy set; pavement; expert system; inspection; rehabilitation; training; fuzzy logic |
Date Deposited: | 16 Apr 2025 19:22 |
Last Modified: | 16 Apr 2025 19:22 |