Prediction of Temerloh River water level for prediction of flood using Artificial Neural Network (ANN) method

• Main Author: Muhamad Afiq, Mustafa
• Format: Undergraduates Project Papers
• Language: English
• Published: 2015
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/12252/1/FKASA%20-%20MUHAMAD%20AFIQ%20MUSTAFA%20%28CD9276%29.pdf; http://umpir.ump.edu.my/id/eprint/12252/

The purpose of this project is to research more about the flood occurrence in Temerloh, Pahang. The data mining approaches using artificial neural network (ANN) techniques will be use to conduct this research for flood estimation. ANN model will be use to estimate river water level by taking present river water level data. The research will be trained using back propagation method to estimate the flood water level at Temerloh River. ANN’s trained using backpropagation are also known as “feed forward multilayered networks” trained using the backpropagation algorithm. 14 years of rainfall data is get from Department of irrigation and drainage (DID). Rainfall data of 10 years(2000-2010) will be training data to predict the others 4 years(2010-2014) river water level using python software with 1000-4000 iteration of data. At the end of the project we can make parameter model that can use as a tools to predict accurately water level data and achieve high accuracy of flood forecasting. From the result we can see that in this research the best prediction for water level data at Temerloh River is 3-hr lead-time with 6 input 1 output in 4000 iteration because it produce the best CE with 0.998.The average RMSE also less than 500 mm with only small difference error in percentage.