Publicação

Application of computational intelligence techniques for monitoring and prediction of biological wastewater treatment systems

Detalhes bibliográficos
Resumo:	Computational intelligence models are being increasingly employed for the supervision and control of biological wastewater treatment systems. These models can be described as mathematical methodologies which explain relations between cause (input data) and effects (output data) irrespective to the process and without the need for making assumptions considering the nature of the relations. In this work both Artificial Neural Network and Neural Fuzzy models were used for monitoring and prediction of biological wastewater treatment systems. The proposed approaches were tested for their ability to detect external and internal disturbances in data obtained from the IWA/COST Benchmark Simulation Model. The models were also applied to predict, with one hour is advance, the response of the system to a sequence of two large increases in the influent flow rate. Both models learned well from the training data and exhibited good and fast predictions of the performance of the system submitted to the tested shocks. The results obtained indicate that the Neural Fuzzy model is slightly superior to the Neural Network model being however the correlation coefficients obtained for both models superior to 0.96.
Autores principais:	Dias, A. M. A.
Outros Autores:	Alves, M. M.; Ferreira, Eugénio C.
Assunto:	Adaptive neural fuzzy model Artificial neural network COST Benchmark Monitoring Prediction
Ano:	2007
País:	Portugal
Tipo de documento:	comunicação em conferência
Tipo de acesso:	acesso aberto
Instituição associada:	Universidade do Minho
Idioma:	inglês
Origem:	RepositóriUM - Universidade do Minho

Descrição
Resumo:	Computational intelligence models are being increasingly employed for the supervision and control of biological wastewater treatment systems. These models can be described as mathematical methodologies which explain relations between cause (input data) and effects (output data) irrespective to the process and without the need for making assumptions considering the nature of the relations. In this work both Artificial Neural Network and Neural Fuzzy models were used for monitoring and prediction of biological wastewater treatment systems. The proposed approaches were tested for their ability to detect external and internal disturbances in data obtained from the IWA/COST Benchmark Simulation Model. The models were also applied to predict, with one hour is advance, the response of the system to a sequence of two large increases in the influent flow rate. Both models learned well from the training data and exhibited good and fast predictions of the performance of the system submitted to the tested shocks. The results obtained indicate that the Neural Fuzzy model is slightly superior to the Neural Network model being however the correlation coefficients obtained for both models superior to 0.96.