Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.507834
Title: Modelling activated sludge wastewater treatment plants using artificial intelligence techniques (fuzzy logic and neural networks)
Author: Rustum, Rabee
Awarding Body: Heriot-Watt University
Current Institution: Heriot-Watt University
Date of Award: 2009
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
Activated sludge process (ASP) is the most commonly used biological wastewater treatment system. Mathematical modelling of this process is important for improving its treatment efficiency and thus the quality of the effluent released into the receiving water body. This is because the models can help the operator to predict the performance of the plant in order to take cost-effective and timely remedial actions that would ensure consistent treatment efficiency and meeting discharge consents. However, due to the highly complex and non-linear characteristics of this biological system, traditional mathematical modelling of this treatment process has remained a challenge. This thesis presents the applications of Artificial Intelligence (AI) techniques for modelling the ASP. These include the Kohonen Self Organising Map (KSOM), backpropagation artificial neural networks (BPANN), and adaptive network based fuzzy inference system (ANFIS). A comparison between these techniques has been made and the possibility of the hybrids between them was also investigated and tested. The study demonstrated that AI techniques offer viable, flexible and effective modelling methodology alternative for the activated sludge system. The KSOM was found to be an attractive tool for data preparation because it can easily accommodate missing data and outliers and because of its power in extracting salient features from raw data. As a consequence of the latter, the KSOM offers an excellent tool for the visualisation of high dimensional data. In addition, the KSOM was used to develop a software sensor to predict biological oxygen demand. This soft-sensor represents a significant advance in real-time BOD operational control by offering a very fast estimation of this important wastewater parameter when compared to the traditional 5-days bio-essay BOD test procedure. Furthermore, hybrids of KSOM-ANN and KSOM-ANFIS were shown to result much more improved model performance than using the respective modelling paradigms on their own.
Supervisor: Adeloye, Adebayo Sponsor: Damascus University
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.507834  DOI: Not available
Share: