Translating safety factors to sources of variability and uncertainty; Model-based design evaluation for the Eindhoven WWTP

Abstract

Design guideline safety factors (SF) are based on the accumulated experience of the industry and account for the uncertainty and variability inherent in the wastewater treatment process. Models on the other hand, as implemented in most commercial simulators, do not explicitly account for uncertainty. This paper discusses how to deal with variability and uncertainty in an explicit way when using a dynamic model combined with uncertainty analysis to perform a post-project evaluation. Through the use of time series, the response of the design to the variability of key inputs can be captured in the dynamic model. In order to incorporate uncertainty, selected model parameters can be described with probability density functions (PDF). Scenario analysis can be used to simulate discreet operational conditions or events. Uncertainty can then be propagated to the dynamic model outputs with the use of Monte Carlo simulations. To test this approach, the upgrade of the Eindhoven WWTP (Waterboard De Dommel, The Netherlands) was selected as a case study. The project focused on how variability and uncertainty are incorporated in design guidelines and highlighted key issues linked to the inclusion of the sources of uncertainty in the calibrated model.

Probabilistic design of wastewater treatment plants

“…The effect of influent uncertainty (an important source of variability) which heavily depends on the site-specific conditions (e.g. characteristics of the connected sewershed, climate conditions) has been considered but not to a great extent (Rousseau et al, 2001;Bixio et al, 2002;Benedetti, 2006;Belia et al, 2012). In addition, little effort has been dedicated to the communication of the concept of PONC which may sound abstract to many designers and decision makers.…”

“…The upgrade of the WWTP was deemed necessary mainly due to more stringent (Schilperoort, 2011) In this study the data and information representative of conditions prior to the upgrade of the Eindhoven WWTP were used to make a design using the proposed probabilistic design method in Chapter CHAPTER 3:. The general information and data regarding the desired capacity of the plant, effluent requirements, as well as the basic characteristics of the connected sewershed were acquired from the available reports and studies on the Eindhoven WWTP (Schilperoort, 2011;Belia et al, 2012).…”

“…The ranges of uncertainty depicted in Table 5-2, were derived using the information obtained from the previous studies on the Eindhoven WWTP (i.e. Schilperoort, 2011;Belia et al 2012), ATV (2000) design guideline recommendations, effluent standard (imposed by regulations), and expert…”

“…Contrary to previous decades during which the application of dynamic models was in its infancy and limited to the academic and research domain, these models have become quite popular in the wastewater industry and are now being used to assess the performance of (Rousseau et al, 2001;Bixio et al, 2002;Huo et al, 2006;Sin et al, 2009;Belia et al, 2012;Benedetti et al, 2013). However, in most of the studies of this kind, the main emphasis was laid on the uncertainty of model parameters while the characterization of the influent variability lacked statistical rigor (e.g.…”