Skip to content
News & media

Unisa online - CSET water research flagship seminar


Prof Francois Ilunga (Chair of the Department of Civil and Chemical Engineering)

In their struggle to address the issues of water scarcity, the Water research flagship in the College of Science, Engineering and Technology hosted a seminar themed Hydrology and water resources modelling on 31 May 2012. The seminar was held on the Unisa Florida Campus in the Gencor auditorium.

In his presentation, Prof Francois Ilunga (Chair of the Department of Civil and Chemical Engineering) spoke on the topic of “Application of Analytic Hierarchy Process (AHP) for Artificial Neural Network Model Selection in Streamflow Prediction.” He also highlighted that in the planning and management of water resources, models are needed to assist in the prediction of streamflows.

AHP is used for the selection of the best predicting rainfall-runoff models between Artificial Neural Network (ANN) techniques, namely; the M5P Model Tree and the multilayered perception (MLP) ANN. The techniques were used to predict discharge at the Luvuvhu River Mhinga gauging station of the Limpopo province in South Africa.


Ednah Onyari (Lecturer in the Department of Civil and Chemical Engineering)

Based on a consistent judgment, the AHP process is perceived to be a powerful tool for model preference. AHP can assist decision makers in the choice of suitable streamflow prediction models since streamflows are used for design, operation and management purposes. 

Mrs Onyari (Lecturer in the Department of Civil and Chemical Engineering) presented on the topic of “Integrated Sediment Management in a Reservoir using a DPSIR Framework.” She told the audience that reservoirs lose their storage capacities and affect water availability due to excessive sedimentation.

Moreover, she spoke about the impact complexity of the assessment and management of sediment in such reservoirs, considering various factors such as the environment, society and economy.

“The two presentations showed a huge potential to accommodate information from water experts in hydrological modelling. This aspect lacks in most hydrological modelling exercises”, said Prof Ilunga.



Other Unisa online News | Latest | Archive