数据融合下的短期潮位预报测试及分析

针对短期验潮数据分析难度大、预报精度低的问题,设计了一种基于数据融合技术的短期潮位预报方法,利用扩展卡尔曼滤波(EKF),将通用的潮汐模型计算水位融合到调和预报模型之中,生成精度更高的融合预报值。数据测试表明,对于3天的验潮数据,EKF方法至少对向后5天的预报有效,融合值较两种源数据的平均优化度分别为33%和60%;对于7天的验潮数据,EKF方法在向后20天的预报中几乎都产生了优化效果。该方法适宜在资料缺乏的海区使用,为短期潮位预报提供了一种新的解决方案。

Short-term tidal level forecast test and analysis based on data fusion

Aiming at the problems of difficult analysis and low forecasting accuracy of short-term tide data,this paper designs a short-term tidal level forecasting method based on data fusion technology by using extended Kalman filter (EKF) to fuse the general tidal model calculation water level into the harmonic prediction model,and generates the fusion forecast data with higher accuracy than both source data.Data test shows,for the 3-day tide data,the EKF method is effective for at least 5-day forecast,and the accuracy of fusion value is 33% and 60% improved compared with two source data,for the 7-day tide Data,the EKF method effective would reach 20-day forecast.This method is suitable for use in sea areas where data is lacking,and provides a new solution for short-term tide level forecasting.