Uncertainty modeling in selection of gas/liquid contactors using Bayesian networks

This paper presents efforts at modeling the decision making process involved in selection of gas-liquid contactors that are commonly used for industrial gas treating. The gas-liquid contactors enable one or more components of a gas phase to be absorbed into a liquid phase. Selection of an appropriate contactor is important for functionality, treating efficiency, as well as process economy. However, it is currently difficult to arrive at complete and objective recommendations as to which contactor is appropriate due to the large number of available contactors, the many and inherently uncertain parameters in this selection task, dependencies between these parameters, and vendor biases. Hence, an expert system becomes an essential tool that helps users overcome these difficulties and supports the user in a systematic selection process that takes into consideration all necessary parameters. To model the inherently uncertain parameters in this selection task, a Bayesian approach is adopted as a primary methodology that provides a normative framework. This knowledge modeling phrase is important for later development of an expert decision support tool.