Basic Concepts on Risk Analysis, Reliability and Maintenance

Abstract

Abstract:Man's level of dependence on equipment is increasing. This degree of dependence requires high levels of availability, which has been changing the impact that disruption of these systems causes. For many systems, an interruption has consequences that go beyond the dimension of financial loss, thus justifying a multidimensional consequence approach by using multicriteria (MCDM/A) models. Thus, understanding the relationships between and among risk, reliability and maintenance (RRM) is essential in order to offer more comprehensive solutions to the various problems often treated in isolation from each other, and which are the most important problems of the competitive market. This chapter discusses fundamental topics about RRM, including tools for risk analysis and hazard identification, concepts of reliability, maintenance techniques such as RCM and TPM and eliciting expert's knowledge. These topics are presented in order to provide a basis for structuring different MCDM/A problems that are addressed in several chapters. Some fundamental aspects could be used as input to decision models in different forms such as attributes, objectives, criteria, and problem context. Basic Concepts on Risk AnalysisThere are many concepts on risk found in the literature and also different perceptions to it. However, if a decision is being made and risk is involved, then, the risk concept should combine consequences and probabilities, incorporating the DM's preferences over that, as seen in Chap. 2.Actually, a 'decision process' with no DM's preference has no decision being made, as discussed at the end of Chap. 1. Instead of that, that process either: a) has some preference structure incorporated within the model, at random; b) is just arbitrary following a previous decision of someone else.Even so, in most of real cases, the consequences are multidimensional, and therefore, require an MCDM/A approach for building a decision model. The following topics are mainly based on the basic RRM literature and do not incorporate the idea of decision support, as given in Chap. 2. That is, DM's preferences are not necessarily considered in the model.