MODELING NUTRIENT DYNAMICS IN SEQUENCING BATCH REACTOR

The use of artificial neural networks (ANN) for modeling complex proce sses is an attractive approach that has been successfully applied in v arious fields. However, in many cases the use of an ANN alone may be i nadequate and inaccurate when data are insufficient, because the ANN b lack-box model relies completely on the data. As a result, a hybrid mo del consisting of a simplified process model (SPM) and a neural networ k (residual model) is used in the present study for developing a dynam ic model of sequencing batch reactor systems. The implemented SPM mode l consists of only five discrete rate equations and an ANN is added to the SPM in a parallel connection. Both the SPM and the ANN receive in fluent chemical oxygen demand (COD), total kjeldahl nitrogen (TKN), PO 43- and NH4+ data and timer output signals (for phase control) as inpu ts. The SPM output provides a preliminary prediction of the dynamic be havior of the PO43- and NOx- concentrations. The outputs of the traine d ANN compensate for the output errors of the SPM model. The hybrid mo del output of the final predictions of the process states is obtained by summing the outputs from both the SPM and ANN. Successful applicati on of such a hybrid model is demonstrated.