Network Representation and Analysis of a Large-Scale Wastewater Treatment Plant

Abstract

We mapped a large-scale wastewater treatment plant onto a complex network and we investigated how the structural properties of the graph evolve in time as the facility is operated. The Viikinmäki plant is mapped onto a dependence network in which the nodes are online process measurements and interconnectivity between nodes encodes pairwise correlations between the corresponding time series, as estimated over moving-windows. In this initial study, the construction of a graph of Viikinmäki is presented and results are discussed with the goal of understanding its usability as model for process interactions and encoder of latent structures.