Thalles Vitelli Garcez

Multicriteria and Multiobjective Models for Risk, Reliability and Maintenance Decision Analysis

Multiobjective and Multicriteria Problems and Decision Models.- Multiobjective and Multicriteria Decision Processes and Methods.- Basic Concepts on Risk Analysis, Reliability and Maintenance.- Multidimensional Risk Analysis.- Preventive Maintenance Decisions.- Decision Making in Condition-Based Maintenance.- Decision on Maintenance Outsourcing.- Spare Parts Planning Decisions.- Decision on Redundancy Allocation.- Design Selection Decisions.- Decisions on Priority Assignment for Maintenance Planning.- Other Risk, Reliability and Maintenance Decision Problems.

Multidimensional Risk Assessment of Manhole Events as a Decision Tool for Ranking the Vaults of an Underground Electricity Distribution System

Even though an underground electricity distribution system is safer than an overhead system, several accidents have occurred in them. Assessing the risk of hundreds or even thousands of underground vaults is a hard task. Furthermore, given the large variability in external and internal environments and, hence, there being a wide range of possible consequences when an accident occurs, an approach to risk assessment under a multidimensional view is required. Moreover, in terms of decision making, the aggregation of the decision makers preferences in modeling, by multiple-criteria decision-making methods, is more complete, comprehensive, and, in particular, includes considering the decision makers desires. Therefore, this study puts forward a multidimensional assessment of the risks from underground vaults by generating a decision tool, which ranks the vaults in a risk hierarchy. Multiattribute utility theory was used to achieve this ranking. An application was generated to demonstrate the applicability of the model, under the following aspects of consequences: those that are human, financial, and operational; and disruptions to local vehicular traffic. The use of information arising from analysis of the differences between risks enabled the decision maker to make an in-depth analysis of the range of possibilities over which alternatives may be chosen in order to implement preventive actions.

A risk measurement tool for an underground electricity distribution system considering the consequences and uncertainties of manhole events

This paper explores a risk measure of underground vaults that considers the consequences of arc faults. The increasing use of underground systems, together with the aging of networks, the lack of maintenance and interference from other (third party) underground systems nearby have caused many accidents in urban areas, thus endangering human life. The involvement of a large number (hundreds or thousands) of underground vaults with different characteristics, the lack of historical data on modes of failure, the rarity of the occurrence of some faults, the magnitude of their consequences and the involvement of a complex environment surrounding the hazard zone make risk management even more complex and uncertain. Furthermore, given that the (monetary, time, staff, etc.) resources of an electrical power company are limited and scarce, it is necessary to use decision-making tools that aggregate the consequences and the uncertainties to assess the risks jointly with the preference structure of the company, thus solving the problem more realistically. Therefore, this paper puts forward the use of an additional risk analysis for manhole events in underground electrical distribution networks with a view to its being used as a decision aid tool in risk management. As an illustration of the use of the risk measurement tool proposed, a numerical application is presented. The result rather than showing a ranking of underground vaults, gives a measure of the risk used that can show the decision-maker (DM) how much better one group of alternatives (formed by alternatives with quite similar risk values) is than other groups, based on the DM’s attitude to risk and grounded on the axiomatic structure of utility theory.

Decisions on Priority Assignment for Maintenance Planning

This chapter presents multicriteria (MCDM/A) models to classify and assign maintenance priorities in order to allow maintenance planning to be more effective. From traditional maintenance planning techniques such as RCM (Reliability Centered Maintenance), TPM (Total Productive Maintenance) and others, a common aspect of these techniques is the definition of maintenance priorities, based on a criticality classification for RCM, for example. As maintenance planning has to satisfy multiple objectives, such as availability, maintainability, detectability, safety and reliability besides cost, the maintenance manager is a decision maker (DM) who has to establish tradeoff amongst multiple criteria. This chapter presents an MCDM/A model integrated with the RCM structure using Utility Theory principles to include states of nature and the DM’s behavior to risk (prone, neutral and averse) in a decision model based on Multi-attribute utility theory (MAUT). To illustrate situations when a DM has a non-compensatory rationality, and requires an outranking method, a decision model based on ELECTRE TRI is applied. In addition, TPM aspects are discussed in order to emphasize potential MCDM/A problems that may be approached.

Decision on Maintenance Outsourcing

This chapter presents key aspects of multicriteria (MCDM/A) approaches for decisions on maintenance outsourcing regarding maintenance contract, which includes contract selection (e.g. repair contract) and supplier selection. Contract design is a multi-objective task that leads the maintenance manager (or decision maker - DM) to decide amongst a combination of contracts and suppliers’ bids for the service. Given the multiple objective nature of this kind of problem, this chapter presents models that include maintainability, dependability, quality of repair and other aspects besides cost. The decision models presented consider methods such as Multi-attribute utility theory (MAUT) to address compensatory preferences and ELECTRE for preferences that require an outranking method. The DM’s behavior to risk (prone, neutral and averse) is considered by using Utility Theory and Decision theory foundations in order to include the state of nature in decision models. Thus, most of the problems are related to supplier and contract selection, which may be modeled into a single problem when considering all combinations of contracts and suppliers as alternatives, including the possibility of in-house maintenance being undertaken by a maintenance service supplier. Depending on the organization and in how strategic its maintenance function may be, decisions in maintenance outsourcing may be approached in different stages. Thus, a key performance indicator (KPI) for such problems are defined depending on the type of organization, its capabilities and the number of maintenance activities, while the tradeoff amongst strategic objectives is balanced in order to assure the system’s availability.

Preventive Maintenance Decisions

Technological advances in equipment and the increase in process automation has led to the maintenance function having a role in business competitiveness. The contribution of preventive maintenance is discussed, as an important part of this function, with some emphasis on methods for planning replacement, in the sense of time interval of preventive maintenance. The classical optimization approach is used to illustrate the original preventive maintenance problem, thereby enabling insights and discussion of the main features that require the use of MCDM/A approaches for these decisions, and thus considering the multidimensional consequence space. A structured framework to build a multicriteria decision model for supporting the selection of time interval is presented. Two different MCDM/A methods are applied depending on the decision maker’s (DM) preferences. The first illustrates the application of Multi-attribute Utility Theory (MAUT) as an example of compensatory method and; the second details the application of a non-compensatory PROMETHEE method, which considers outranking relations.

Multidimensional Risk Analysis

Accidents involve critical consequences that require an appropriate and efficient form of risk management. A multidimensional risk analysis allows a broader view. MCDM/A approaches enable more consistent decision-making, taking into account the DM’s rationality (compensatory or non-compensatory), DM’s behavior regarding risk (prone, neutral or averse) and the uncertainties inherent in the risk context. This chapter presents numerical applications illustrating the use of multicriteria models in two different contexts: a natural gas pipeline and an underground electricity distribution system. Two different MCDM/A approaches are considered: MAUT (Multiattribute Utility Theory) and the ELECTRE TRI outranking method. In the numerical applications, MCDM/A approach steps for building decision models are presented: identifying hazard scenarios, estimating the set of payoffs, eliciting the MAU function (Multi-attribute Utility function), computing the probability function of consequences and estimating multidimensional risk. Loss functions are introduced in the models to calculate the probability distribution functions over the multiple criteria such as impact on humans, and environmental and financial losses. Therefore, Decision Theory concepts are applied to estimate risk in industrial plants and modes of transportation. Finally, other decision problems related to multidimensional risk analysis, using MCDM/A, are considered in different contexts, such as: power electricity systems, natural hazards, risk analysis on counter-terrorism, nuclear power plant.

Design Selection Decisions

The design selection problem in the RRM context considers long-term performance, and represents higher additional costs if unforeseen features that should have been included during the project design phase had to be implemented afterwards. Design decision involves multiple aspects and may be more critical depending on the kind of item, such as consumer appliances, industrial equipment or projects that have to consider safety aspects (airplanes or facilities). Reliability has an essential role for design selection although other aspects have to be considered such as maintainability and risk depending on the specific design problem. Therefore, a multidimensional approach is usually required. In this chapter, all these aspects are discussed in order to illustrate the importance of a broader perspective when facing design decision problems. The fundamental requirements are to consider reliability, maintainability and risk aspects so as to establish features in the design project, including the definition of material, redundancies, control systems and safety barriers. To illustrate these decisions, aspects such as reliability (e.g. MTBF), maintainability (e.g. MTTR), safety, cost, service life, efficiency, are discussed as criteria for these problems. Multi-attribute utility theory (MAUT) is applied in this chapter to illustrate how reliability, maintainability and risk aspects are included in an MCDM/A model for design selection incorporating states of nature. The decision regarding the selection of which features to include in a design project may be considered as an MCDM/A portfolio problem. Finally, an introductory view is given of how the redesign problem arises in the maintenance context with multicriteria approaches.

Multiobjective and Multicriteria Problems and Decision Models

The decision-making process for any organization may be a key factor for its success. Many decision problems have more than one objective that need to be dealt with simultaneously. This chapter introduces decision problems with multiple objectives, with a description of the basic elements needed to build decision models and focuses on multicriteria methods (MCDM; MCDA; MCDM/A), in which the DM’s preference structure is considered. An overview for classification of MCDM/A methods is given, including a discussion on the DM’s compensatory and non-compensatory rationality and on multi-objective and multicriteria approaches. The concepts and basic elements of MCDM/A methods are presented, including preference structures in a multi-attribute context, and intra-criterion and intra-criteria evaluation. The basic elements of a decision process for building decision models and the actors in this process are also presented. Differences between the descriptive, normative, prescriptive and constructivism decision approaches are discussed, considering the decision process. Although these concepts are presented in a general sense, this description deals mainly with the main context that this book explores: Risk, Reliability and Maintenance (RRM). Decision problems in a RRM context may affect the strategic results of any organization, as well as, human life (e.g. safety) and the environment. Therefore, an explanation is given as to why and how a MCDM/A arises in the RRM context. In particular, some peculiarities of service producing systems for MCDM/A models are presented, as well as for goods producing systems.

Decision on Redundancy Allocation

Redundancy allocation is a decision that involves assessing and choosing where to locate additional components or subassemblies, above the minimum required for an existing system to operate, in order to promote the system’s reliability. Specifically, the field of multi-objective redundancy allocation has received several contributions since the 1970s and the combinatorial complexity of these problems has mainly encouraged researchers to develop search algorithms focused on the Pareto front definition, the most frequent approach in this literature. Finding a set of non-dominated solutions based on heuristics is a step that demands much computational effort to solve the problem. Despite these difficulties, the DM’s preferences should be evaluated in order to recommend a solution that represents the best compromise among the criteria considered, such as reliability, cost and weight. This chapter covers redundancy allocation problems from a multicriteria perspective. Therefore, basic concepts related to the typical criteria and tradeoff in redundancy allocation problems are presented and a brief review of the literature on MCDM/A redundancy allocation is given. To illustrate the MCDM/A approach for redundancy allocation, a decision model considering a standby system based on Multi-attribute Utility Theory (MAUT) is presented including the DM’s behavior to risk (prone, neutral and risk averse). The problem approached in this chapter involves a question about how to select a suitable maintenance strategy in order to evaluate the tradeoff between a system’s availability and cost, including experts’ prior knowledge to deal with the uncertainty of failure and repair rate parameters.