Keywords: reinforced concrete, damage assessment, expert system, rule-based.

Expert systems (ESs) are receiving great attentions in the construction industry to support decision-making processes in diagnostics, design, repair and rehabilitation of the structures. Although several expert systems have been examined in engineering since the 1970's, their applications in construction industry are rare. This was largely due to the lack of expert system tools available to represent the domain knowledge. Lack of flexibility, applicability, and robustness of the classical models, have forced the scientists to discover the ability of the expert systems in problem solving of civil engineering [1,2,3].

In this paper, the applicability of Expert Systems for diagnosis the deterioration of concrete structures is investigated. At first, the essence of an expert system is explained [4,5]. Rule-based and object oriented (frame-based) expert systems, the two major type of expert systems, are described through some examples in the area of the reinforced concrete deterioration problem. Then, the problem of reinforced concrete deterioration is defined. It is emphasised that this problem is one of ill-defined problems that it needs to some engineering judgement and human expert opinion to be solved. Later, we illustrate a problem solving approach (rule-based expert system with forward-chaining inference engine) by means of an example for recognition the crack type in reinforced concrete member.

In this research, we developed an ES prototype for diagnosis the deterioration of concrete structures. Proposed expert system's knowledge-base posses 26 complex "If-Then" rules that has been extracted from references, papers and the negotiation with human experts (domain experts) in the area of our involved problem. This expert system focuses on cracking distress in reinforced concrete elements. It is shown that ES can better represent this problem than the classical methods. Moreover, the inference engine module can reason the real situation similar to the human behaviour. This helps us to solve the problem more robust without eliminating any variable or constraint. Moreover, since the structure of the system is logical oriented, the implementation of the system is easier and more user friendly. Finally, a case study has been presented to examine and evaluate the proposed expert system. The system demonstrates a straightforward method for diagnosing the cause of reinforced concrete elements cracking.

There are some potential future works for this research. The presented system provides diagnosis of common problems specially cracks cause in reinforced concrete. It is of great desire that proposed ES would be expanded so that it can suggest the strategy of crack repair. However, the variables and operators in this system were crisp. Development of a fuzzy expert system for this domain is a potential future work.

1

Salvaneschi, P. Cadei, M. and Lazzari, M. "Applying AI to Structural Safety Monitoring and Evaluation", IEEE Expert, August, 1996, pp. 24-34. doi:10.1109/64.511774

2

Strecker, P.P., "Corrosion Damage Concrete: assessment and repair", Construction Industry Research and Information Association, Butterworths, London, 1987.

3

Adeli, H., "Artificial Intelligence in Structural Engineering", Engineering Analysis, Vol. 3, No. 3, 1986, pp. 154-60. doi:10.1016/0264-682X(86)90053-5

4

Durkin J., "Expert Systems: Design and Development", Macmillan Publishing Company, 1994.

5

Russell S.J., and Norving J., "Artificial Intelligence: A Modern Approach", Prentice-Hall, 1995.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £82 +P&P)