Modelling of a water treatment plant. A multi‐model representation

Abstract

SUMMARYThis paper examines the problem of modelling a non-linear dynamic system by a Takagi±Sugeno model. Although this type of model is well adapted to the representation of a system with non-linear behaviour, its use does however raise real problems in the identi®cation of its structure and parameters, mainly because of the enormous combinatorial increase in model complexity as a function of the number of inputs, and the coupling between the premises and consequence parts. We propose a simpli®ed model structure and an algorithm for the identi®cation of this structure and its parameters. The method is illustrated with reference to data from a wastewater treatment plant of the Greater Nancy Urban Authority.

Predictive monitoring and diagnosis of periodic air pollution in a subway station

“…In other words, improvements in the monitoring process can be achieved by obtaining better knowledge of the system by answering the following questions: which variables characterize the process, what are their internal interactions and what degrees of confidence can be attributed to these measurements? All of these questions are related to the characterization of the system, which involves several fundamental stages: the description of the system, the listing of the variables that characterize the system behavior, the establishment of models between the variables, the identification of the parameters which intervene in these models, the simplification of the models to make them compatible with real-time use, and the validation of these models [6]. Multivariate monitoring methods, including principal component analysis (PCA) and partial least squares (PLS), have been applied to monitor the air quality in many environmental systems [1,[7][8][9] [1,3,[7][8][9].…”

Modeling Carbon and Nitrogen Removal in an Industrial Wastewater Treatment Plant Using an Adaptive Network‐Based Fuzzy Inference System

“…If the resulting fuzzy system performed satisfactorily with respect to the validation data, then the computation was completed. The performance of the resulting model was determined by evaluating measured and estimated results via the correlation coefficient R [11,18,25], root mean square error (RMSE) and average percentage error (APE) [13,16,18] …”

“…AI methods involve the use of artificial neural networks (ANNs), genetic algorithms (GA), fuzzy logic, rule-based systems, knowledgebased systems and their associated applications called , hybrid architectures' such as adaptive network-based fuzzy inference system (ANFIS) [16]. ANFIS, proposed by Jang [17], is based on the first-order Sugeno fuzzy model.…”

ANFIS Modelling of Carbon and Nitrogen Removal in Domestic Wastewater Treatment Plant

“…Literature reveals that several models were developed for wastewater treatment process, for instance, activated sludge models (ASMs) family (ASM1, ASM2, ASM2D, and ASM3) which have proven to be indispensable tools within last decades, made a remarkable impact in achieving reliable treatment plant design, better understanding of system mechanisms and more importantly serves as guidance for research [1], especially the state of art model (ASM1) for carbon and nitrogen removal. Based on this mathematical model concept various models either practically or via simulation were developed [2], [3], [4], [5], [6], [7], [8]. Nevertheless, most of these developed models were either applied to completely mixed tank or contact stabilization and industrial wastewater treatment plant.…”