基于BP神经网络模型研究背景流对于ADCP波浪估计的影响

Effect of background current on ADCP wave estimation based on BP neural network molel

ADCP利用接近水面位置处的海流进行波浪估计,其中浪致流是波浪估计所需要的要素,但是当背景流较强时,会影响波浪估计的准确度,现利用WAVESON软件对背景流幅值大于0.75m/s时进行背景流修正。研究表明,当背景流存在时,影响波浪估计的因素有背景流幅值、背景流流向、波浪方向和水深,现有的背景流修正算法是不精确的。文章拟将BP神经网络模型应用于背景流存在条件下,分析波浪估计误差与各影响因素之间的复杂关系,提升背景流存在条件下波浪估计的准确性。

ADCP wave estimation is carried out by using current velocity closed to water surface,and the only wave induced velocity is needed during wave estimation.When background current is strong,wave estimation accuracy is affected by background current.The software WAVESMON suggests that background current correction should be done if background current amplitude is larger than 0.75m/s.The research shows that when background current exists,the factors of background current amplitude,background current direction,wave direction and water depth can affect wave estimation.And the present background current correction algorithm is not accurate.This paper applies BP neural network model in wave estimation when background current exists.And the relationship between wave estimation error and various factors are analyzed,which can improve wave estimation accuracy.