Safety based decision support systems for marine structures

Analyzing risk is an essential and very powerful tool to deal with uncertainties. New technologies, new developments and new methodologies always include uncertainties and thereby risks. There are many different methods to deal with risks. Qualitative and quantitative methods are both used for decades with success in many different industries. Quantitative studies require more statistical data then qualitative ones. Ability to use both tangible and intangible data in analyses enables to perform risk analysis at any stage of the project. Hazardous Identification (HAZID) and Hazardous and Operability (HAZOP) studies are powerful and widely used analysis methods. In these studies a group of experts make an effort to define and assess possible risks. Work groups may be problematic if group domination by one or more participants happens. And generally because of the consensus about the subject, it is hard to include fuzziness. Risk assessment is a very important figure in analysis of new systems. Assessment has three main stages; risk identification, risk analysis and risk evaluation. Identification starts with questioning "when, what, who, how, where", after finding answers to these questions analysis stage starts. Answers to questions at identification stage helps to understand the boundaries of the system. Analysis focuses on "how much, how often, how critical" type questions. These questions help analyst to understand the nature, period and level of risk. After analyzing the risk evaluation period starts; this point is the one where the action to the risk is selected, taking no action or transferring risk such as an insurance company might also be an alternative. There are several techniques defined by ISO to deal with risk management. About 30 techniques are listed in the standard. Whatever the method is used, knowledge of experts, their understanding of problem and their position to view the risk will directly affect the result. In order to get proper assessment results, problem has to be thought from all sides of aspect. The success is methodology depended, however the main component of all analysis methods is human. Whether or not be aware of it, decision making is something done in every daily life. Decisions made, draw the path of life, decisions shapes the life and every decision has it's own consequences that has to be faced. Historically, human decisions theories have focused on outcome prediction. Modern decision making is based on ; understanding of decision making process, thoughts and application of technology tools to support process by human beings. Analytical Hierarchy Process (AHP) proved that using Eigen vectors to solve decision problems is possible. With this study Saaty opened slightly the door to the new studies. Difficulties about modeling real problems in a hierarchical structure are limiting the usage of AHP; therefore a better way of modeling in network structure with dependencies and feedback is presented by Saaty. This methodology is called Analytical Network Process (ANP). ANP is a successful and powerful tool to model complex decision problems. However modeling and solving ANP requires a lot of patience and effort. Fuzzy Multi-Criteria Decision Making using Fuzzy Set Theory has been recommended as an alternative method for overcome the complexity of ANP. Fuzzy sets contain uncertainty and they are easy to apply to all kinds of problems. There are also some xx challenging difficulties with fuzzy sets. Definition of fuzzy sets, membership functions require experience. Changing defined rules is not so easy. Due to these problems in multi-criteria decision making, hybrid methods such as Fuzzy-AHP, Fuzzy-ANP, etc. are being used commonly. There are generated hybrid solution methodologies, thus applier, depending of the nature of the problem, may choose one of them and directly apply the solution method. In this study, LNG bunkering operation was selected as case study. LNG is one of the most probable alternatives to current fuel oils. Bunkering operation is the most challenging part of LNG usage on board. HAZID/HAZOP studies for bunkering, classification societies bunkering guidelines were used to define the hazards. Hazards are grouped under clusters by using Fault Tree Analysis (FTA). Experts participating in the study were asked to fill in the questionnaire, according to pre-defined ANP network structure. Risk values for alternatives are calculated using likelihood and consequences. Final results obtained by using group decision making techniques such as aggregation of individual priorities (AIP), expert weights also calculated by using a separate ANP network. As result LNG leaks have been found the most critical risk alternative for all calculation methodologies.