Application of the differential neural network observer to the kinetic parameters identification of the anthracene degradation in contaminated model soil

In this work a new technique dealing with differential neural network observer (DNNO), which is related with differential neural networks (DNN) approach, is applied to estimate the anthracene dynamics decomposition and to identify the kinetic parameters in a contaminated model soil treatment by simple ozonation. To obtain the experimental data set, the model soil (sand) is combined with an initial anthracene concentration of 3.24 mg/g and treated by ozone (with the ozone initial concentration 16 mg/L) during 90 min in a reactor by the “fluid bed” principle. The anthracene degradation degree was controlled by UV-vis spectrophotometry and HPLC techniques. Based on the HPLC data, the obtained results confirm that anthracene may be decomposed completely in the solid phase by simple ozonation during 20 min and by-products of ozonation are started to be destroyed after 30 min of treatment. In the ozonation process the ozone concentration in the gas phase at the reactor outlet is registered by an ozone detector. The variation of this parameter is used to obtain the summary characteristic curve of the anthracene ozonation (ozonogram). Then, using the experimental decomposition dynamics of anthracene and the ozonogram, the proposed DNNO is trained to reconstruct the anthracene decomposition and to estimate the anthracene ozonation constant using the DNN technique and a modified Least Square method.