Sediment load prediction for rivers in Malaysia using evolutionary polynomial regression technique
Mining has become one of the most challenging phenomena that can cause serious sedimentation process and aggregation problems in rivers in Malaysia. The process will increase the amount of sediment load in rivers that can result in them becoming shallow and subject to flood problems during heavy sto...
| Main Authors: | , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/23175/ http://umpir.ump.edu.my/id/eprint/23175/1/ICWR-2018_paper_93.pdf |
| Summary: | Mining has become one of the most challenging phenomena that can cause serious sedimentation process and aggregation problems in rivers in Malaysia. The process will increase the amount of sediment load in rivers that can result in them becoming shallow and subject to flood problems during heavy storms. In such a problem, management cost to monitor the amount of sediment in rivers can be really high. A new model using Evolutionary Polynomial Technique has been developed to assist researchers to predict sediment load in sandy rivers, in view of Malaysia. The model utilises 273 data of twelve rivers in Malaysia. Based on these data, it has been found that 173 is suitable for use in the model, that later gave 100% results within the range of the required difference ratio of 0.5 to 2.0. A new set of data that contains 82 data is used for validation and the results also give a good prediction of sediment loadings. The new model predictions are also compared with the present sediment model using other techniques, such as the regression technique by Ariffin in 2004, a modified Graf by Chan in 2005 and a multiple regression technique by Sinnakaudan in 2006. As a conclusion, the model developed using the Evolutionary Polynomial Technique indicates performance which gave the best results compared to the other three models; with a 100% accuracy within the stipulated range, followed by the multiple regression technique with 90.24%, the regression technique with 69.51% and finally the modified Graf with 23.17%. |
|---|