Abstract

Purpose – The purpose of this paper is to develop an easy and robust tool to develop a decision support system (DSS) for the inspection staff of oil pipelines. The aim is to predict “the class” of each spillage, with respect to some relevant variables such as, mechanical failure or system malfunction. The management will then be able to define which pipelines to monitor and to choose the most suitable monitoring policies, based on the decision tree analysis outcome. Design/methodology/approach – A non-parametric technique based on rule induction is proposed for the identification of the expected spill cause category of cross-country oil pipelines. In particular, the classification and regression trees approach is used to automatically generate inspection or maintenance decision rules. The analysis that is described is based on an extended database concerning information about spill cause category in cross-country oil pipelines in Western Europe. Findings – The proposed technique represents an interesting added value tool for the management. The proposed methodology extrapolates rules for determining the expected spill cause category of cross-country pipelines, depending on the boundary conditions. Practical implications – The methodology here presented will assist maintenance managers of oil pipeline to better plan maintenance activity. In particular, the procedure makes it possible to determine which parts of a pipeline have to be submitted to a monitoring action or particular protection, with the aim of improving the efficiency and reducing the risk of spillages. Originality/value – Effective planning, coordination, and scheduling of the maintenance function can be, and for many years was, accomplished without computer support. The proposed procedure may be included in an information systems tool (sound Computerized Maintenance Information Management System (CMMIS)), for more efficient and effective maintenance/inspection scheduling activitie