An application of artificial neural networks to the prediction of surface ozone concentrations in the United Kingdom

From the results of investigations at full-scale and pilot plants, it has been found that with ʹusualʹ parameters theASMI is only valid for a small range of SRT as well as ofoxic reactor section (V(OX)N). To make ASM1 applicable lo a wider range of SRT the very slowly biodegradable particulate organic matter has to be taken into consideration. This fraction can be interpreted as part of the ʹheterotrophic biomassʹ X(B,H) which requires modification of the values for the ʹheterotrophic decay rateʹ b(H) and for the fraction of X(l) in the influent. In order to distinguish between ʹactive heterotrophic biomassʹ (X(B,H) and the very slowly biodegradable substrate it seems to be useful to inkoduce the latter mass fraction as a new model component (X(V)) into the ASMI. This extended simulation model has given reasonable results for different activated sludge systems.