Penelusuran Banjir (Stage Hydrograph) Menggunakan Jaringan Saraf Tiruan (Studi Kasus : DAS Siak)

Abstract

The purpose of this research is to predict the height of water level of Siak River Sub DAS Siak Hulu Pantai Cermin Station in 2012 by using water level data recorded by Sub DAS Tapung Kiri Tandun Station in the same year. This is to search for a more convenient and accurate method in flood routing from all methods that have been applied in the hope that this research can be considered as an alternative method.
This research is conducted by using backpropagation algorithm artificial neural network approach with single input and single output as network model configuration. The approach itself uses Matlab 7.8.0.347 (R2009a) as supporting program.
This research shows that the results of the training, test, and validation of artificial neural network model have a fairly good level of correlation with the value of R 0.58713, 0.64818 and 0.65933 though not yet provide maximum results. The level of correlation between prediction result with actual data is 0,518.
Keywords: flood routing, water level, artificial neural network, backpropagation, tapung kiri, siak hulu