L. McGinty and D.C. Wilson (Eds.): ICCBR 2009, LNAI 5650, pp. 509–523, 2009. © Springer-Verlag Berlin Heidelberg 2009 Determining Root Causes of Drilling Problems by Combining Cases and General Knowledge Samad Valipour Shokouhi 1 , Agnar Aamodt 2 , Pål Skalle 1 , and Frode Sørmo 3 1 Department of Petroleum Technology (IPT) 2 Department of Computer and Information Science (IDI) Norwegian University of Science and Technology (NTNU) NO-7491, Trondheim, Norway 3 Verdande Technology AS Stiklestadveien 1- Trondheim, Norway valipour@ntnu.no, agnar.aamodt@idt.ntnu.no, pal.skalle@ntnu.no, frode@verdandetechnology.com Abstract. Oil well drilling is a complex process which frequently leads to op- erational problems. In order to deal with some of these problems, knowledge intensive case based reasoning (KiCBR) has clearly shown potential. An impor- tant problem in oil well drilling is hole cleaning, in which a high number of ob- served parameters and other features are involved. This paper presents how to determine the root causes of poor hole cleaning episodes by means of KiCBR. This will help the drilling crew to apply a proper remedy. The effect of general domain knowledge was demonstrated in a comparative study, in which im- proved results in terms of similarity assessment and explanation capability were achieved. Keywords: Case-based, knowledge intensive, oil well drilling. 1 Introduction Drilling of oil wells is an expensive offshore operation, costing typically 200 000 US$ per day. Any loss of time caused by unwanted events is costly. During drilling all material drilled out need to be removed, i.e. transported to the surface, a process which is referred to as hole cleaning. Often some of the material remains in the well, and hole cleaning is still among the most important problems to deal with during drilling. It is also one of the most studied phenomena within the petroleum industry. Insufficient hole cleaning can in extreme cases lead to loss of the well or a part of it, i.e. stop of the drilling process and blocking of the hole. Due to the number of parameters influencing hole cleaning and the complex mechanisms involved, the phenomenon has not yet been fully understood [1]. Case-based reasoning (CBR) is an approach to problem solving and decision mak- ing where a new problem is solved by finding one or more similar previously solved problems, called cases, and re-using them in the new problem situation. Application- oriented research in the area of case based reasoning has moved mature research