| In this paper, basic principle of line of correspoinding stages method and finite-memory least square method have been introduced in detail, for each method, tidal level forecast model is established. Moreover separating water level process line forecast model is designed and developed on the base of characteristic of the water level of tidal reach of Yangtze River attachment, which was affected by the upstream tide and the downstream tide, and the principle of line of correspoinding stages method. In this dissertation artificial nerve network BP model is attempted to apply to tidal forecasting of tidal reach and BP forecasting model is developed. In order to analyze and compare advantage artd weakness of four kinds of forecasting model. These four methods are applied to tidal level forecasting for Nanjing tidal station. It could be concluded from the result and basic principle contrast analysis of every kind forecasting model that line of correspoinding stages method model and separate water level process line forecast model and finite-memory least square model are applicable to tidal reach of the Yangtze River attachment tidal forecasting. But BP forecasting model should be ameliorated and reached because of localization of water level process of oscillating.In this paper, the real-time tidal level forecasting system for Nanjing tidal station wasdeveloped on the base of correlation hydrological model, visual development tool VisualBasic6.0, geographical information system secondary development control MapObjects, andcombined the related principles of prototype design method in system development. This level forecasting system is applied and operated easily. This system is made up of geographical information system sub-module, data management sub-module, forecasting sub-module, and project evaluation sub-module. Every sub-module operates independently, cooperates mutually, and exerts respective function furthest.
|
PDF Full Text Request