Risk Assessment of Marine LNG Operations
摘要
Various methodologies for the risk assessment of LNG transfer operations at the ship-shore interface of gas terminals were presented. These include a qualitative risk matrix approach, a multiple attribute utility model and a fuzzy inference system. The use of multiple attribute utility theory in risk assessment of LNG operations allows the ranking of risk alternatives based on a unified utility measure. A maximum risk alternative is selected to minimize the overall expected utility. This methodology allows modeling of the decision maker’s attitude towards risk, i.e., risk aversion/neutral and/or risk taker. Available software tools allow ranking of risk alternatives and sensitivity analyses to be carried out to assess the sensitivities of the risk model’s recommendations to various modeling variables. An approach for the assessment of multiple attribute risk using fuzzy set theory was also presented. The developed methodology is an alternative to qualitative risk assessment matrices currently used in many industries and by ship classification societies. A three dimensional risk envelope or surface is generated and used for computation of risk values as replacement to the traditional risk matrix.The use of fuzzy sets and a fuzzy inference engine is suited for handling imprecision often associated with accident likelihood and consequence data. The total number of rules needed to construct the fuzzy inference engine is the product of the number of rows and the number of columns for the corresponding qualitative risk matrix. The proposed approach improves upon existing qualitative methods and allows the ranking of risk alternatives based on a unified measure. A fuzzy risk index was adopted for aggregation of multiple consequences into a unified measure. Both the Mamdani and Sugeno type inference methods were adopted. Results show that while the Mamdani method is intuitive and well suited to human input, the Sugeno method is computationally more efficient and guarant