Maximum wave height forecasting method based on wave energy spectrum and deep learning model

By using a method based on wave energy spectrum and deep learning model, the problem of insufficient accuracy of traditional maximum wave height forecast under high sea states is solved, and a more accurate and adaptable forecast is achieved, which is applicable to marine engineering and shipping safety.

CN121682103APending Publication Date: 2026-03-17SHANGHAI MARINE METEOROLOGICAL CENTER
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-11
Publication Date
2026-03-17

AI Technical Summary

Technical Problem

Existing maximum wave height prediction methods lack sufficient accuracy under high sea states. Traditional methods suffer from insufficient information utilization, limited theoretical assumptions, and poor adaptability to extreme sea states, making it difficult to meet the needs of accurate early warning.

Method used

A method based on wave energy spectrum and deep learning model is adopted. Historical data is acquired and standardized preprocessed to construct a deep learning model to learn the nonlinear mapping relationship from wave energy spectrum to maximum wave height. Combined with nonlinear transformation and progressive weight function, the ability to predict extreme sea states is improved.

Benefits of technology

It significantly improves the accuracy of maximum wave height forecast, especially in extreme sea conditions where the significant wave height is greater than 2.0 meters, reduces forecast errors, and ensures that the forecast results conform to physical laws and are applicable to existing wave forecasting systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121682103A_ABST
    Figure CN121682103A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of marine environment forecasting, and discloses a maximum wave height forecasting method based on a wave energy spectrum and a deep learning model, and the method comprises the steps: obtaining historical actual measurement maximum wave height and historical actual measurement or forecast wave energy spectrum density data, and carrying out the standardization preprocessing of the historical actual measurement maximum wave height and historical actual measurement or forecast wave energy spectrum density data; energy spectrum zero moment is calculated through numerical integration, and the significant wave height is obtained according to a physical formula; constructing a deep learning model, taking the standardized energy spectrum as input, and taking a nonlinear transformation value of a ratio of an actually measured maximum wave height to an effective wave height obtained by a formula as a training target training model, so that the model learns nonlinear mapping from the energy spectrum to a target variable; and inputting a standardized energy spectrum to be forecasted into the model to obtain a transformation value forecasting result, obtaining a wave height ratio forecasting value through inverse transformation, and multiplying the wave height ratio forecasting value with the effective wave height to output a final maximum wave height forecasting value. The energy spectrum information is fully utilized, the forecasting precision is improved, the extreme sea condition forecasting performance is particularly improved, and the method can be widely applied to the fields of ocean engineering, shipping safety and disaster early warning.
Need to check novelty before this filing date? Find Prior Art