The Analysis of Chemical Engineering Process Plants and their Models Represented by Networks

Abstract

Chemical process plants can be interpreted as networks where energy mass and momentum are exchanged. Therefore, they can be represented as such using graph theory. This work presents a decomposition methodology that allows to conceptually deconstruct chemical process plants and their models to be represented by networks. By the selection of minimum constituent units of the plant, it is possible to describe the interdependencies among those minimum units. This procedure allows the correct representation of process plants as networks along with a detailed analysis of them. Using network metrics such as centrality measures, it was possible to analyze, as an example, the HDA process plant under two selected plantwide control strategies, allowing the identification of its more important units and controllers. This permitted an insight of how plantwide control strategies responded to perturbations from a qualitative point of view. It was also possible to establish the main coincidences and discrepancies between the selected plantwide control strategies.