Please use this identifier to cite or link to this item: http://localhost:8081/xmlui/handle/123456789/10410
Title: LONG TERM OCEAN WAVE FORECASTING
Authors: Krishna, Bal
Keywords: OCEAN WAVE;REAL TIME FORECASTING;SMB MODEL;WATER RESOURCES DEVELOPMENT AND MANAGEMENT
Issue Date: 2010
Abstract: Knowledge of magnitude, as well as behavior of wind waves is essential in the entire planning, design, construction, operation and maintenance related activities carried out in the whole ocean area including harbour, coastal and offshore regions. The requirement for each activity however may vary. For example design exercises require prediction of wave heights over a period of 100 years, as against planning and construction related works which call for forecasting of wave heights as well as that of their behavior for a short period of a few hours or so. The most important factors for establishing design wave for a coast are the long-term storm data and storm wave modelling. In this study, the authors have applied various deep-water parametric storm wave prediction models like SMB(Sverdrup-Munk-Bretschneider), Wilson and CEM(Coastal Engineering manual Model,U.S.Army,2006) including artificial neural network to a 115-year period of historical severe storms. The wave heights for each storm were hindcasted and the wave heights which were significant to the respective coasts were extrapolated to long term using Gumbel, Weibull and Log-Normal distributions. Assuming, no theoretical justification is available as to which distribution is to be used (Burcharth and Liu, 1994). The average of these three distributions for specified return periods were taken as a design wave height. The results of the offshore design wave height (m) predicted for both the locations, are represented in the following Table. Table: Predicted Extreme Wave Height (m) for Various Return Return Period 25 50 75 100 150 Off Mumbai 12.6 14.2 15 15.7 16.2 Off Pondicherry 18.9 21.2 22.4 23.5 24.7 The studies on the applicability of ANN to the problem of wave prediction indicated that the appropriate trained network could provide satisfactory results.
URI: http://hdl.handle.net/123456789/10410
Other Identifiers: M.Tech
Research Supervisor/ Guide: Kansal, M. L.
Agarwal, J. D.
metadata.dc.type: M.Tech Dessertation
Appears in Collections:MASTERS' DISSERTATIONS (WRDM)

Files in This Item:
File Description SizeFormat 
WRDMG20668.pdf6.2 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.