Inferencing Economic Impacts of Accidents in the Oil & Gas Supply Chain Using Fuzzy Systems

Currently the sustainability of traditional technologies employed in the oil and gas supply chain constitutes one of the major challenges. New regulation, public awareness and the climate change have bring new concerns for the decision making in safety of oil and gas infrastructure. One of the key elements for the decision making in safety is related to sizing the magnitude of accidental impacts. However, modelling impacts involve large amount of uncertainty due to its inherent complexity.

Taking in account the fuzzy systems as suitable tools to deal with uncertainty, this paper presents a methodology to determine the magnitude of the impacts of accidents in the oil and gas infrastructure. Then, the overall aim of this paper is to present the application of knowledge management technology into business practices, which is related with the estimation of the accidental impacts of the oil and gas infrastructure by using knowledge-based fuzzy inference systems.

The overall content is established by the interaction of Safety of energy infrastructure, artificial intelligence, and business value maximization. This framework establishes as its ultimate goal to create value for stakeholders from the safety management using artificial intelligence methods.

The general approach of this research is a hybrid technique that combines theoretical and empirical methods in order to develop an inference model to estimate economic impacts of failures in the oil and gas supply chain. The parameters of the model are determined using expert knowledge and statistical data.

The content of this research is divided in three major parts: 1. the determination of the range values of impacts, 2. the implementation of the inference system using Matlab, and 3. the simulation test. The first activity consists in defining the mean values of the impacts as well as their classifications. The second activity is related with the construction of the inference model using the fuzzy toolbox of Matlab. And the third activity consists in the simulation of the model developed using a different set of data and coefficient values.

This paper is intended to serve as a relevant tool to create business value from the implementation of safety measures, that comprehensive study aims to introduce a novel technique to assess accidental risks of energy infrastructure using hybrid artificial intelligence methods to handle with the uncertainty and vagueness derived from the estimation of probabilities and consequences of failure.