(343b) A Model Reduction Approach to Activated Sludge Systems

The Activated Sludge Model No. 3 (ASM3) extended for two-step nitrification and denitrification has proven to be a very accurate model to describe the Activated Sludge Process (ASP) (Kaelin et al. 2009). The division of the nitrification-denitrification reaction in a two-step reaction is essential when trying to describe the bypassing nitrate generation process in the ASP (Katsogiannis et al. 2002). Moreover, the extended ASM3 enables the description of the substrate consumption and the Oxygen uptake with a higher precision than the older versions of the ASM family because of the addition of energy storage effects. This model extension facilitates the calculation of the NO2 concentration as an independent variable. In this work, the extended Activated Sludge Model No. 3 (ASM3) is analyzed and reduced to enable fast and accurate simulations under typical process conditions. An exhaustive analysis of the model results in a number of simplifications, which reduces the model drastically while keeping its description accuracy.

The model proposed in this work is called 8-state model (8SM) according to the resulting number of state variables. Eight differential equations describe the process basic variables (concentrations), namely: (1) substrate, (2) heterotrophic bacteria, (3) ammonia oxidizers, (4) nitrite oxidizers, (5) dissolved oxygen, (6) ammonia, (7) nitrite, and (8) nitrate. In addition, an algebraic equation is added to explain the effects of energy storage on biomass growth and on the substrate as well as oxygen uptake, as described by the extended ASM3. The model reduction is based on the principle that an activated sludge process should stop once all concentrations comply with the regulations. Furthermore, these concentrations should be reached in the shortest time possible. Consequently, lower output concentrations than required are an indicator of a suboptimal operation. Another important assumption is that the bacteria never exhaust their stored energy. Except for the recycle process, the bacteria are always in a medium, which is rich in substrate. Therefore, the stored energy value should be permanently high during the process and never limit bacterial growth.

In the 8-state model, the new set of equations is presented in such a way that both the substrate uptake increment and the oxygen uptake increment caused by the storage are included in the original substrate and oxygen differential equations. By these means that the algebraic equation, which describes the behavior of the stored energy, can be removed without affecting the behavior of the substrate or oxygen concentration provided that the substrate concentration is above zero. The simulations were performed using the in-house solver DaCL, which is a differential algebraic system solver based on orthogonal collocation on finite elements. The results show that the 8-state model mimics the behavior of the extended ASM3 in a broad operation range. The results obtained in this work suggest that the reduced model can also be applied to continuous and to SBR ASP.

References:

Kaelin, D., R. Manser, et al. (2009). "Extension of ASM3 for two-step nitrification and denitrification and its calibration and validation with batch tests and pilot scale data." Water Research 43(6): 1680-1692.

Katsogiannis, A. N., M. Kornaros, et al. (2002). "Enhanced nitrogen removal in SBRs bypassing nitrate generation accomplished by multiple aerobic/anoxic phase pairs." Water Science and Technology 47(11): 53-59.