AuthorsAnyachebelu, Tochukwu Kene
Subjectswater quality index
artificial neural networks
Subject Categories::G730 Neural Computing
MetadataShow full item record
AbstractSurface water quality is a dynamic quantity to deal with. There are factors which affect the surface water quality. These factors include weather changes, anthropogenic activities and urbanization. There is an underlying problem with water quality monitoring and management around the world especially in developing countries. Pollution outbreak in a Luton town lake in 2012 which killed a lot of fishes is one instance of such devastating outbreak and the underlying effects. This thesis is aimed at utilising measured data values for certain physico chemical parameters in the determination of surface water quality through analysis, indexing and model prediction. The physical parameters measured are temperature, conductivity and turbidity while the chemical parameters are dissolved oxygen, pH and ammonium. These parameters are measured at two locations on the Luton Hoo Lake which is used as our research study site to monitor the water quality and possible sources of contamination. Data with regards to the values of the different parameters is collected with the help of multi parameter probe sensors that were installed at the site in form of a remote monitoring station. Manual sampling of the water and collection of parameter value readings is also used to substantiate the values derived from the remote monitoring stations. A preliminary analysis is carried out using descriptive Statistics and correlation analysis to determine the dependencies between the various parameters measured for water quality monitoring. We evaluate the relationship of the measured parameters to contamination sources and its impact on the water quality as it affects aquatic life. With the correlation analysis, it is discovered that some of the parameters exhibited the expected relationship whereas some could not relatively show any dependence. This led to the need for further analysis to help in the parameter selection since the aim of this work is to use minimal parameters to establish the water quality status. It is determined that dissolved oxygen is the most important parameter compared to other measured parameters. This was actualized through the use of principal component analysis to identified the major components. Multiple linear regression is used to establish the relationship between Dissolved oxygen and the other measured parameters. The two locations being monitored exhibit the same trend from the results of the box plot analysis that was carried out. Dissolved oxygen is inversely proportional to Temperature which confirms the same trend pattern as exhibited by the correlation analysis. Principal component analysis helps in the establishment of the hierarchy of the parameters measured and the level of importance which determines the input parameters for the water quality index. This research work looked at various water quality indices developed by various researchers and discovered that most available indices were limited by the fact that the data collection was done manually. This work adopted the use of a tailored Water quality index with three parameters which are Dissolved oxygen, Conductivity and Turbidity. These parameters were selected based on the results derived from the Principal components analysis. Water quality index is a good way of uniquely rating the overall water quality status of a water body using a single term. In this research study, we utilised available water quality indices that have been developed through expert opinions and modified them to suit our requirements. The results obtained were satisfactory and proved that the use of minimal parameters can give a good indication of the water quality in same way as the use of many parameters. The minimal Water Quality Index is recommended where faster and more economical approach is needed in decision making with regards to water quality monitoring. A hybrid neural network is finally proposed for the prediction of parameters for water quality index which can be tailored based on the location of the surface water and the primary use of the surface water for the end users. The major problem of identifying a low cost way of monitoring surface waters for developing countries is achieved through the use of minimal parameters and the availability of the sensor probes in the market is of great importance to the work.
CitationAnyachebelu, T. (2019) 'Prediction of a water quality index using online sensor data'. PhD Thesis. University of Bedfordshire.
PublisherUniversity of Bedfordshire
TypeThesis or dissertation
Description"A thesis submitted to the University of Bedfordshire in partial fulfilment of the requirements for the degree of Doctor of Philosophy"
The following license files are associated with this item:
- Creative Commons
Except where otherwise noted, this item's license is described as Attribution-NonCommercial-NoDerivatives 4.0 International