Prediction of Suspended Sediment Concentration in Kinta River Using Soft Computing Techniques
The prediction of suspended sediment concentration in hyperconcentrated rivers is crucial in modeling and designing hydraulic structures such as dams and water intake inlets. In this study, suspended sediment concentration in Kinta River is predicted using soft computing technique, specifically radi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project |
| Language: | English |
| Published: |
Universiti Teknologi Petronas
2013
|
| Online Access: | http://utpedia.utp.edu.my/13426/1/24.pdf http://utpedia.utp.edu.my/13426/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The prediction of suspended sediment concentration in hyperconcentrated rivers is crucial in modeling and designing hydraulic structures such as dams and water intake inlets. In this study, suspended sediment concentration in Kinta River is predicted using soft computing technique, specifically radial basis function. Suspended sediment concentration and stream discharge from the year of 1992 to 1995 and data from the year of 2009 are used as input. The data are divided into three sections, namely training, testing and validation. 824 data are allocated for training, 313 data are allocated for testing purpose and 342 data are allocated for validation purpose. All data are normalized to reduce error. The determination of input neuron is based on correlation analysis. The number of hidden neurons is determined by the application of trial and error method. As for the output, only one output neuron is required which is the predicted value of suspended sediment concentration. The results obtained from the radial basis function model are evaluated to identify the performance of radial basis function model. Performance of the prediction is measured using statistical parameters namely root mean square error (RMSE), mean square error (MSE), Coefficient of efficiency (CE) and coefficient of determination ( ). Radial basis function model performed well producing the value of (0.9856 & 0.9884) for training and testing stages, respectively. However the performance of RBF model in the prediction of suspended sediment concentration for the year 2009 is poor, with the value of of 0.6934. Recommendations to improve the prediction accuracy are by incorporating a wider data span and by including other hydrology parameters that may impact the changes in the value of suspended sediment concentration |
|---|