Model selection for risk analysis of wastewater networks

Abstract

In this paper, we test various models of wastewater infrastructure for risk analysis and compare their performance. While many representations are available, existing studies do not consider selection of the appropriate model for risk analysis. In this paper, we define two characteristics of wastewater models: the network granularity and the fidelity of the governing equations. We consider different combinations of these characteristics to determine 6 network representations that could be used as the foundation for risk analysis. We test the performance of each model as compared to predictions from the most detailed model, the full network with dynamic wave flow equations. We demonstrate the model selection for Seaside, Oregon. We conclude that the full network granularity is needed as compared to a coarse network representation. For the fidelity of the governing equations, connectivity analysis is reasonable if the primary goal is to determine the spatial distribution of hazard impact. To more accurately predict nodal performance measures, the dynamic wave equations are needed as they capture important physical phenomena.