Please use this identifier to cite or link to this item:
Full metadata record
DC FieldValueLanguage
dc.thesis.supervisorRanasinghe, D.N. (Dr.)-
dc.contributor.authorLiyanage, H.K.E.en_US
dc.description.abstractIt is considered that predicting floods for Rathnapura in a acceptable ac- curacy is a challenging task with conventional hydrological methods. Build- ing a bridge between operational hydrology and computer science, in here we try to solve that problem with Artificial Neural Networks, a well known approach for modeling highly complex non linear patterns which cannot be captured by an algorithmic way. But we had to do our research with con- strained availability of data, thus past 13 flood events from 2001 to 2008 were taken into consideration. Input is taken from the daily precipitations of Hapugasthanna, Wellandura, Lellopitiya and Rathnapura gauging stations. Output is taken from the hourly water heights of Rathnapura gauging sta- tion. Two prediction models were developed based on daily precipitations and rainfall intensity. Feed forward neural network was used to do the predic- tions, trained with back propagation algorithm. Our results show that daily precipitation based model works with significant accuracy with correlation coefficient over 0.8 while the intensity based model is less in ability in gener- alizing, hence training it with adequate data is expected to over perform the daily precipitation based model.en_US
dc.titleFlood Prediction for Rathnapuraen_US
Appears in Collections:SCS Individual Project - Final Thesis (2008)

Files in This Item:
File Description SizeFormat 
  Restricted Access
418.86 kBAdobe PDFView/Open Request a copy

Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.