Flood Forecasting System

Team

Project Owner

Dr. Upul Jayasinghe, upuljm@eng.pdn.ac.lk

Scrum Master

Imesh Udara Ekanayake, imeshuek@eng.pdn.ac.lk

Team members

E/17/006, ALAHAKOON A.M.H.H, e17006@eng.pdn.ac.lk

E/17/176, KUMARA W.M.E.S.K, e17176@eng.pdn.ac.lk

E/17/338, SRIMAL R.M.L.C, e17338@eng.pdn.ac.lk


Table of Contents

Introduction

Solution

Solution Architecture

Links


Introduction

    Floods are the most destructive form of natural hazards in both local and global context.
    This is true in terms of both loss of life and property damage. Early flood forecasting can
    be used to identify potential areas of flooding in order to develop mitigatory planning and
    evacuation programs to remove people from such areas during flooding
    and also to implement suitable preventive measures to avoid damage to properties.

    In this project, our main objective is to build a flood forecasting system for Mi Oya river
    basin(Sri Lanka).Mi Oya Basin is heavily affected by seasonal flooding and droughts.

Flood afected areas

    As per the available data, floods in the Mi Oya basin are unleashed due to river overflow
    and reservoir spilling. Out of the several reservoirs located in the basin, Tabbowa and
    Inginimitiya are crucial in worsening the flood impacts as these two reservoirs are
    frequently spilling under adverse weather conditions. As such, the prevalence of a
    real-time flood forecasting model with the incorporation of the reservoir operations
    for the entire basin is essential to alleviate the flood induced impacts while
    preserving the optimum volume of water in the major reservoirs in the basin.

Solution

    The proposed solution is to implement a simple data-driven or machine learning model to
    identify a direct mapping between the inputs(e.g., precipitation(P), temperature(T),
    potential evapotranspiration(PET), etc.) and outputs(In-flow level) without detailed
    consideration of the internal structure of the physical process. Also the system is
    consist of a web interface. The interface consists of three major modules as community
    view,control panel, and report. The warning messages are displayed with a map depicting
    the inundation extent in the community view, which is the interface for the public. Further,
    the community members are allowed to register their mobile numbers to receive warnings
    as SMS when such warnings are broadcasting at the disaster management center. A summary
    of past floods can be generated from the Report module for a required time period. Control
    panel provides access to the simulation module, gate operation module, flood warning
    dissemination module, access control module, etc. Only privileged users have access to the
    control panel. The figure below outlines the highlevel structure of the web application.

Solution Architecture

Requirement Analysis

Functional Requirements Nonfunctional Requirements
Forecast Floods and show critical regions in maps Reliability
Warn general public by SMS alerts Scalability
Generate flood reports User experience
Control reservoir operations Security

Usercase Diagram

drawing

Highlevel System Overview

drawing

Data and Control Flow

drawing

Web Application Overview

drawing


Project Repository

Project Page

Department of Computer Engineering

University of Peradeniya