Use this URL to cite or link to this record in EThOS:
Title: Design and modelling of pollutant removal in stormwater constructed wetlands
Author: Kiiza, Christopher
ISNI:       0000 0004 7962 0885
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
Growth in urban population, urbanisation and economic development have increased the demand for water, especially in water-scarce regions. Stormwater treatment has the potential to reduce water demand. Furthermore, the use of constructed wetlands (CWs) in the treatment of stormwater also has the benefit that CWs can lower peak flow discharges and hence lessen floods; as well as improve the aesthetics of urban landscapes. In this research study, 8 pilot-scale vertical flow constructed wetlands (VFCWs) were configured to examine the influence of design and operational variables on the performance of tidal-flow VFCWs. The rationale of the research was that tidal-flow operational strategy draws atmospheric oxygen in the VFCWs, thereby increasing the concentration of dissolved oxygen in the wetland system. Moreover, a combination of dissolved oxygen and a fixed retention time of 24 hours enhances the removal of nutrients N and P. Therefore; the research was conducted in two major parts. The first part consisted of outdoor and laboratory experiments, which were carried out over a continuous period of 2 years at Cardiff School of Engineering, Cardiff University. The physical models of the VFCWs were configured from a series of media compositions and were fed with loads of influent stormwater to simulate various storm events over different catchment sizes. The performance of the VFCWs regarding total suspended solids, nutrients (N and P), and heavy metals were monitored during the experimental period. The data obtained were analysed using descriptive and inferential statistics. The second part of the study involved exploring the experimental data to develop artificial neural network models (ANNs) to predict pollutant removal in the VFCWs, an essential aspect of the design process. Accordingly, the outputs from the research show that the different designs of VFCWs significantly reduce priority pollutants in stormwater; and that pollutant removal is related to the design and operational variables. Additionally, exploratory data analysis by principal components analysis (PCA) is relatively effective at reducing the dimensionality of input variables. Subsequently, the ANN models developed produced satisfactorily accurate generalisations of TN and TP removal, as measured by the different statistical indices. Generally, the good agreement between the predicted and experimental data suggests that ANNs can adequately predict TN and TP removal up to 4 months in advance. Furthermore, the ANNs had fewer inputs, indicating that monitoring costs and time can be reduced.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available