The Forecasting Alarm System Of Sediment-Cleaning-Project Based On Neural Network

Harbor Sediment has become the key nodus of Port Management allover the world. To fit the requirements of sailing, each year a mass ofregular deposit clean work in sea-route, haven and berth has to be donefrequently, especially in those Silt deposited based harbor. Currentlythose port cleaning companies frame their estimative clean plans uponglancing experience and former data without accurate gist. Accordingly areasonable and effective work schedule of equipment maintenance andusing is hard to establish, neither the maximum exertion of labor forcenor the maximum efficiency of equipment usage. Thus the cost ofsediment cleaning is greatly increased. Simultaneously along with thedevelopment of computer science, plenty of harbors and relativeorganizations has established quite consummate database to storeabundant harbor-sediment-related data. According to the reason above, aiming at the lack of theharbor-sediment-cleaning project forecast, this thesis finds outdisadvantages of forecasting technology in Harbor Sediment by analyzingthe status quo. At the meantime, by researching theory and characteristicof Neural Network method of Data Mining and analyzing its nonlinearapproaching capacity, the writer probed into the feasibility of theapplication of Neural Network method of Data Mining applying topredictive measurement of the harbor-sediment-cleaning project. Andhereon basing on the multi-layer forward neural network theory of BPalgorithm, with the most widely used engineering calculation softwareplatform Matlab 6.5 and its plentiful, practical resource-neural networktool box, more interactive graphic interface and visible simulating tools,analyzed the harbor-sediment-relative data and established a set ofeffective sediment-cleaning-project anticipation system, thereby strivedfor the intention of supervising sediment-cleaning-project practice.Finally, put the completing forecasting alarm system into some subjectsof harbor sediment prediction, the operation results prove that this systemcan forecast harbor sediment well and truly and can also give guidance tothe harbor-sediment-cleaning project.