DOI
https://doi.org/10.1002/clen.200700155

Summary
The dynamic model for mathematical modeling of chemical oxygen demand (COD) concentrations and monitoring large-scale wastewater treatment plant performance was presented in this article. Analysis and modeling of wastewater processes enables to deepen knowledge in plant design, wastewater treatment practices and other related processes. Two different approaches, that is, analytical and the one using artificial neural networks (ANN) were described with the procedure to evaluate the model performance. The approaches were applied to Ankara central wastewater treatment plant. It was demonstrated that after choosing the appropriate ANN architecture, the ANN-based model gives better results and better describes the operation conditions that analytical model. The authors state that such approach can be transferred to predict other pollutant concentrations.

Relevance for Complex Systems Knowledge
The wastewater treatment modelling is an example of a complex system as it requires knowledge and skills in physical, chemical and biological fields. Its complexity is also defined by broadness of the data range and diversity, evaluation of the output parameters. Such problem is often used in teaching engineering courses due to various aspects that must be considered in order to solve this problem. The authors present wastewater models which incorporate artificial neural networks in mathematical and mechanical model of prediction wastewater treatment. Obviously, the development of this model includes evaluation of model accuracy.

Point of Strength
The incorporation of artificial intelligence methods (artificial neural networks) to solve complex problems is provided in this article.