FLOOD FREQUENCY ANALYSIS OF ADYAR RIVER USING GUMBEL METHOD AND ARTIFICIAL NEURAL NETWORK

Abstract

Floods are the most affected natural disaster which may leads to damages in movable and immovable properties. Floods are caused due to the extreme rainfall such that which develops within short period of time. The major factors causing floods are events such as deforestation, uncontrolled urbanization and global warming. The major flood event that occurred in December 2015 during northeast monsoon season in Adyar river basin has driven attention from the public nationwide due to many peoples lost their life and properties. The flood is very dangerous occurrence which depends on several factors such as rainfall conditions, catchment characteristics such as its size, shape and land use. Flood frequency analysis plays major role in effective design of hydraulic structures such as roads, flood control structures, culvert, bridges, railways, spillways, urban drainage system and waterways. The objective is to assess the flood frequency curve by Gumbel method and the Artificial Neural Networks (ANN) is a soft computing technique which is used to analyze the frequency of flood. To analyze the frequency of flood the various return periods were used to obtained the frequency curve using Gumbel method and ANN model was used for further accuracy in occurrence of flood. The flood frequency curve of analysis shows that the selection ANN models are more accurate and appropriate that Gumbel method. This study show that ANN models can be effective tools for flood frequency analysis for water resource planners, managers and analyzers to take safety measures and save the nation.