Application of artificial neural networks to tide forecasting
Zhaochen Sun · Dalian Ligong Daxue xuebao · 2007
In ocean environment,accurate tide forecasting is a kind of significant work for constructions and human activities.Artificial neural networks(ANN) model is presented for forecasting the tidal-level using limited field data of one site or multi-sites.Based on the characteristics of tide time series,a new conception-the periodicity analysis on tide data is proposed and applied to the testing of the performance of ANN model.The numerical results indicate that this method is efficient to solve the time lag problem existing in previous ANN models.Based on the pertinence of tidal-level data among different tidal gauge stations,a new tide forecasting method-site to site is developed,and the numerical results agree well with field data.