The objective of this thesis is to contribute for a better understanding of the decision making process in industrial plants specifically in situations with impact in the long term performance of the plant. The way decisions are made, and especially the motivations that lead to the selection of a specific course of action, are sometimes unclear and lack on justification. This is particularly critical in cases where inappropriate decisions drive to an increase on the production costs. Industrial

more »

... plants are part of these cases, specifically the ones that are still lacking enhanced monitoring technologies and associated decision support systems. Maintenance has been identified as one of the critical areas regarding impact on performance. This is due to the fact that maintenance costs still represent a considerable slice of the production costs. Thus, understanding the way maintenance procedures are executed, and more important, the methods used to decide when maintenance should be developed and how, have been a concern of decision makers in industrial plants. This thesis proposes a methodology to efficiently transform the existing information on the plant behaviour into knowledge that may be used to support the decision process in maintenance activities. The development of an appropriate knowledge model relating the core aspects of the process enables the extraction of new knowledge based on the past experience. This thesis proposes also a methodology to calculate the risk associated to each maintenance situation and, based on the possible actions and on the impacts they may have in the plant costs performance, suggests the most appropriate course. The suggestion is made aiming the minimization of the life cycle costs. Results have been validated in test cases performed both at simulation and real industrial environments. The results obtained at the tests cases demonstrated the feasibility of the developed methodology as well as its adequateness and applicability in the domain of interest.