A Closer Look at the Dangers of Uncalibrated Simulators
Sedran et al. (2006) determined the required aerobic reactor volume to achieve 1 mgN/L effluent ammonia in a nitrifying plant, using three commercial simulators (BioWin, GPS-X and Plan-It- Stoat). The paper reports significant differences between the required reactor volumes calculated by different packages using several activated sludge models in each. The authors were unable to find an explanation for these differences. In one scenario of their investigation, Sedran et al. (2006) aimed to ensure that the same models were used, with the same parameters and the same influent data, and still obtained different results from the three packages. This result implies that modeling is an unreliable tool to support wastewater treatment design. The aim of this paper is to clarify some of the questions raised, and provide advice to users of simulation packages to obtain sound process predictions using their simulators. Three simulation companies participated in this study, and all three simulators produced similar or the same results under controlled conditions. The authors of this paper intend to demonstrate that computer modeling can produce consistent
results, while at the same time there are model and implementation differences between the packages as taken “out of the box”. Simulation providers should strive to provide users with general, easy to follow descriptions of the logic of their models, as well as detailed instructions for specific model details.