一种海浪谱预报的新模式

A New Method to Predict Ocean Wave Spectra

为有效解决海上航行船舶难以获取精确的现实海浪信息的问题,提出一种新的基于雷达测波信息的海浪谱预报方法。着眼于小波神经网络对非线性问题的处理能力较强,通过对现实海浪进行筛选比较,决定采用基于小波神经网络的时间序列预报法进行海浪谱预报。频谱预报采用分解子波——时间序列预报——合成频谱的模式,方向谱预报采用方向分布函数预报模式。试验数据表明:预报结果具有较高的精度,实现该预报能够帮助远洋航行舰艇提早掌握海洋环境信息,为实施舰载机起降等对海浪信息精度要求较高的海上作业提供技术支撑。

A new method to get ocean wave spectra based on X-band radar observation is presented which answers for the need of oceangoing vessels to obtain accurate wave information. The time series prediction method based on wavelet neural network is chosen for processing the radar data on the basis of wave characteristics analysis. Frequency spectra are predicted through wavelet decomposition, time series prediction and spectra synthesis. Directional spectra are determined through directional distribution prediction. Experimental results indicate that this method is of high precision, which can help oceangoing vessels to know ocean environment in advance to support offshore operations that require wave information of high-precision, like the aircraft taking off and landing.