Method, device, computing equipment, and storage medium for predicting drift velocity

A method for predicting a drift velocity, comprising:

• obtaining environmental characteristic parameters at a location of a floating object to be predicted, wherein the environmental characteristic parameters comprise wave characteristic parameters;
• obtaining a wave-induced drift velocity by inputting the wave characteristic parameters into a deep learning model for the wave-induced drift velocity;
• wherein the deep learning model for the wave-induced drift velocity is trained based on first sample drift data, wherein the first sample drift data comprises a sample drift velocity observation value of sample floating objects and corresponding sample water surface flow characteristic parameters, sample wind characteristic parameters, and sample wave characteristic parameters;
• wherein a step of training the deep learning model for the wave-induced drift velocity based on the first sample drift data comprises:
• taking the sample drift velocity observation values as an approximation of a sum of a flow-induced drift velocity and a wind-induced drift velocity, fitting a first flow-induced drift coefficient and a first wind-induced drift coefficient, and simultaneously determining the first flow-induced drift coefficient and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample flow-induced drift velocity based on the sample water surface flow characteristic parameters and the first flow-induced drift coefficient, and calculating a corresponding estimated value of a sample wind-induced drift velocity based on the sample wind characteristic parameters and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample wave-induced drift velocity based on the sample drift velocity observation value, the corresponding estimated value of the sample flow-induced drift velocity, and the corresponding estimated value of the sample wind-induced drift velocity;
• training the deep learning model for the wave-induced drift velocity using the sample wave characteristic parameters and the corresponding estimated value of the sample wave-induced drift velocity.

A method for training a deep learning model for a wave-induced drift velocity, comprising:

• obtaining multiple groups of first sample drift data, wherein the first sample drift data comprises sample drift velocity observation values of sample floating objects and corresponding sample environmental characteristic parameters, wherein the sample environmental characteristic parameters comprise sample water surface flow characteristic parameters, sample wind characteristic parameters, and sample wave characteristic parameters;
• training the deep learning model for the wave-induced drift velocity based on the first sample drift data;
• wherein a step of training the deep learning model for the wave-induced drift velocity based on the first sample drift data comprises:
• taking the sample drift velocity observation values as an approximation of a sum of a flow-induced drift velocity and a wind-induced drift velocity, fitting a first flow-induced drift coefficient and a first wind-induced drift coefficient, and determining the first flow-induced drift coefficient and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample flow-induced drift velocity based on the sample water surface flow characteristic parameters and the first flow-induced drift coefficient, and calculating a corresponding estimated value of a sample wind-induced drift velocity based on the sample wind characteristic parameters and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample wave-induced drift velocity based on the sample drift velocity observation value, the corresponding estimated value of the sample flow-induced drift velocity, and the corresponding estimated value of the sample wind-induced drift velocity;
• training the deep learning model for the wave-induced drift velocity using the sample wave characteristic parameters and the corresponding estimated value of the sample wave-induced drift velocity.

A computing equipment, comprising:

• a memory and a processor, wherein the memory stores a computer program, when the computer program is executed by the processor, causing the processor:
• obtaining environmental characteristic parameters at a location of a floating object to be predicted, wherein the environmental characteristic parameters comprise wave characteristic parameters;
• obtaining a wave-induced drift velocity by inputting the wave characteristic parameters into a deep learning model for the wave-induced drift velocity;
• wherein the deep learning model for the wave-induced drift velocity is trained based on first sample drift data, wherein the first sample drift data comprises a sample drift velocity observation value of sample floating objects and corresponding sample water surface flow characteristic parameters, sample wind characteristic parameters, and sample wave characteristic parameters;
• wherein a step of training the deep learning model for the wave-induced drift velocity based on the first sample drift data comprises:
• taking the sample drift velocity observation values as an approximation of a sum of a flow-induced drift velocity and a wind-induced drift velocity, fitting a first flow-induced drift coefficient and a first wind-induced drift coefficient, and simultaneously determining the first flow-induced drift coefficient and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample flow-induced drift velocity based on the sample water surface flow characteristic parameters and the first flow-induced drift coefficient, and calculating a corresponding estimated value of a sample wind-induced drift velocity based on the sample wind characteristic parameters and the first wind-induced drift coefficient;
• calculating a corresponding estimated value of a sample wave-induced drift velocity based on the sample drift velocity observation value, the corresponding estimated value of the sample flow-induced drift velocity, and the corresponding estimated value of the sample wind-induced drift velocity;
• training the deep learning model for the wave-induced drift velocity using the sample wave characteristic parameters and the corresponding estimated value of the sample wave-induced drift velocity.