Abstract—A knowledge-based expert system with the acronym RASPE is developed as an application tool to help decision makers in construction companies make informed decisions about managing risks in pipeline construction projects. Choosing to use expert systems from all available artificial intelligence techniques is due to the fact that an expert system is more suited to representing a domain’s knowledge and the reasoning behind domain-specific decisions. The knowledge-based expert system can capture the knowledge in the form of conditional rules which represent various project scenarios and potential risk mitigation/response actions. The built knowledge in RASPE is utilized through the underlying inference engine that allows the firing of rules relevant to a project scenario into consideration. Paper provides an overview of the knowledge acquisition process and goes about describing the knowledge structure which is divided up into four major modules. The paper shows one module in full detail for illustration purposes and concludes with insightful remarks. Keywords—Expert System, Knowledge Management, Pipeline Projects, Risk Mismanagement. I. INTRODUCTION HIS paper describes a computerized expert system that can help construction companies especially the ones working in pipeline infrastructure projects make informed decisions about risk management in their projects. The research has been tailored to the Egyptian context. Many pipeline construction projects in the Middle East fail to be completed as per the established targets. A frequently- reported failure is substantial project delays [1]. Pipeline infrastructure projects like other types of construction projects have their share of uncertainties and hardships. It is no exaggeration saying that no project is absolutely risk-free [2]. In real industry practice, many project/construction managers in these countries mostly make their decisions based on intuition, judgment, and experience rather than through a formal and systematic risk management process [2], [3]. It can be useful to aid in such process by providing high- quality domain knowledge to support decision-making. Expert systems can particularly support decision-making because they contain high-level knowledge about a subject area often called a domain [4]. Expert systems are computer applications N. Zabel is with the Institute of Engineering and Technology, fifth Assembly, New Cairo, Egypt. In addition he is with WorleyParson Engineering Consultation, Saudi Arabia (phone: +966-567665507; e-mail naelatnet76@yahoo.com). M. Georgy is with the School of Property, Construction and Project Management, RMIT University, Australia. In addition he is with Faculty of Engineering Cairo University, Egypt (e-mail maged.georgy@rmit.edu.au). M. Ibrahim is with the Faculty of Engineering, Cairo University, Egypt (e- mail moheeb.elsaid@eng.cu.edu.eg). which embody some non-algorithmic expertise for solving certain types of problems [5]. The authors developed a tool named Risk Advisory Smart System for Pipeline Projects in Egypt (RASPE) which can be utilized as smart advisor to help construction companies take suitable and effective actions prior to project starting to avoid or mitigate the unfavorable consequences of risks. RASPE was developed using Visual Studio. II. RASPE RISK MANAGEMENT KNOWLEDGE A. Knowledge Acquisition Almost everyone agrees that this task is the bottleneck in the expert system construction. Since the expert system relies heavily on knowledge, the quality of the acquired knowledge will often be the contributing factor to its success. Knowledge acquisition can be likened to a distillation process where the essential facts and rules must be isolated from information that contains many impurities. Instead of being embedded in various handbooks and manuals, the knowledge in many domains is available only to experts and is never been written down in a structured form. In such cases, the task of building an expert system must begin by acquiring or capturing the expert knowledge [6]. Generally speaking the knowledge acquisition can be made via manual, semi-automatic or automatic means. RASPE knowledge was acquired manually, as it was rather difficult to find great number of cases to apply the automatic knowledge acquisition approach. Besides, risk response/mitigation actions are a subjective issue. Each expert can consider these actions from his/her own perspective. The manual acquisition gives the experts more flexibility to express his/her knowledge and communicate to the researchers while they capture the knowledge in useable form. The knowledge acquisition proceeded via two stages as detailed hereinafter. 1. Analysis of Public Domain Knowledge The public domain knowledge was captured from a variety of published theses and academic references in the area of construction risk management. Also, several international contracts for mega infrastructure projects in Egypt, and which had a foreign party, were reviewed for knowledge worthy of inclusion in the system’s knowledge base. The public knowledge was essential to get the first author –who acted as the knowledge engineer in the knowledge acquisition process– familiar with the domain’s context and terminology. While capturing the knowledge through the relevant clauses in the RASPE – Risk Advisory Smart System for Pipeline Projects in Egypt Nael Y. Zabel, Maged E. Georgy, Moheeb E. Ibrahim T World Academy of Science, Engineering and Technology International Journal of Civil and Environmental Engineering Vol:9, No:3, 2015 731 International Scholarly and Scientific Research & Innovation 9(3) 2015 scholar.waset.org/1307-6892/10000608 International Science Index, Civil and Environmental Engineering Vol:9, No:3, 2015 waset.org/Publication/10000608