CN-121997094-A - Sea state prediction method, sea state prediction device, sea state prediction equipment, sea state prediction medium and program product

CN121997094ACN 121997094 ACN121997094 ACN 121997094ACN-121997094-A

Abstract

The invention belongs to the technical field of sea state prediction, and particularly relates to a sea state prediction method, a sea state prediction device, sea state prediction equipment, a sea state prediction medium and a sea state prediction program product. The sea condition prediction method comprises the steps of obtaining sea wave characteristics and wind field characteristics of a target sea area within a preset time range, extracting sea wave time sequence characteristics and wind field time sequence characteristics from the sea wave characteristics and the wind field characteristics within the preset time range by utilizing a bidirectional gating circulation unit, determining sea wave space-time attention characteristics and wind field space-time attention characteristics according to the sea wave time sequence characteristics and the wind field time sequence characteristics, determining wind wave fusion characteristics according to the sea wave space-time attention characteristics and the wind field space-time attention characteristics, extracting extreme perception characteristics from the wind wave fusion characteristics by utilizing an extreme perception module, and predicting sea conditions of the target sea area according to the extreme perception characteristics. And the sea state trend prediction and extreme state intelligent early warning integration is realized.

Inventors

LI HAOJIA
ZHAN XIAOMING
WANG BOTE
WANG QING
ZHENG SHUQIAN
Yu Dieran
SHEN YAOJIE

Assignees

浙江华东智联科技有限公司
中国电建集团华东勘测设计研究院有限公司

Dates

Publication Date: 20260508
Application Date: 20251208

Claims (10)

1. A method for predicting sea conditions, comprising: Acquiring sea wave characteristics and wind field characteristics of a target sea area within a preset duration range; Extracting wave time sequence features and wind field time sequence features from the wave features and the wind field features within a preset duration range by using a bidirectional gating circulating unit; Determining sea wave time-space attention characteristics and wind field time-space attention characteristics according to the sea wave time sequence characteristics and the wind field time sequence characteristics respectively; determining wind-wave fusion characteristics according to the wave space-time attention characteristics and the wind field space-time attention characteristics; extracting extreme sensing characteristics from the wind-wave fusion characteristics by using an extreme sensing module; And predicting the sea state of the target sea area according to the extreme perception characteristics.
2. The method of claim 1, wherein the acquiring sea wave characteristics and wind farm characteristics of the target sea area comprises: acquiring a sea surface echo map of the target sea area within a preset duration range by utilizing a micro wave radar; Determining the wave characteristics of the target sea area at each moment within a preset duration range according to the sea surface echo map; acquiring wind speed fields of a plurality of radial points in front of a fan impeller within a preset time range by using a laser wind-finding radar; And determining the wind field characteristics of the target sea area at each moment within a preset duration range according to the wind speed field.
3. The method according to claim 1, wherein extracting the wave timing feature and the wind farm timing feature from the wave features and the wind farm features, respectively, within a preset time period range using a bi-directional gating cycle unit comprises: time-aligning the sea wave features and the wind field features; Respectively inputting the time-aligned sea wave characteristics and the wind field characteristics into a bidirectional gating circulation unit; And obtaining the sea wave time sequence characteristic and the wind field time sequence characteristic through the bidirectional gating circulating unit.
4. The method of claim 1, wherein the determining the wave spatiotemporal attention feature and the wind farm spatiotemporal attention feature from the wave timing feature and the wind farm timing feature, respectively, comprises: calculating the wave attention weight of the wave time sequence characteristics at each moment; Determining the wave time-space attention characteristic according to the wave time sequence characteristic and the wave attention weight at each moment; Calculating wind field attention weights of the wind field time sequence characteristics at all moments; And determining the wind field time-space attention characteristic according to the wind field time sequence characteristic and the wind field attention weight at each moment.
5. The method of claim 1, wherein extracting the extreme perceptual features from the stormy waves fusion features using an extreme perceptual module comprises: Calculating extreme probability scores of the storm fusion characteristics by using a multi-layer perceptron network; Determining an extreme weight from the extreme probability score; amplifying the salient features of the stormy waves fusion features by using a nonlinear amplification function to obtain fusion amplification features; and determining the extreme perception feature according to the fusion amplification feature, the stormy waves fusion feature and the extreme weight.
6. The method of claim 1, wherein predicting the sea state of the target sea area from the extreme perceptual features comprises: predicting sea wave parameters of the future period of the target sea area according to the extreme perception characteristics by using a double-layer fully-connected network; predicting extreme event probability according to the extreme perception characteristics by using a single-layer fully-connected network and an activation function; Acquiring a preset probability threshold; And determining whether to send out extreme event early warning according to the extreme event probability and the probability threshold.
7. A sea state prediction apparatus, comprising: The first acquisition module is used for acquiring sea wave characteristics and wind field characteristics of a target sea area within a preset duration range; the first execution module is used for extracting wave time sequence characteristics and wind field time sequence characteristics from the wave characteristics and the wind field characteristics within a preset duration range by utilizing a bidirectional gating circulation unit; The first determining module is used for determining the wave time-space attention characteristic and the wind field time-space attention characteristic according to the wave time sequence characteristic and the wind field time sequence characteristic respectively; the second determining module is used for determining wind-wave fusion characteristics according to the wave space-time attention characteristics and the wind field space-time attention characteristics; The second execution module is used for extracting extreme perception features from the wind-wave fusion features by utilizing the extreme perception module; And the first prediction module is used for predicting the sea condition of the target sea area according to the extreme perception characteristics.
8. A computer electronic production device comprising a memory, a processor and a computer program stored on the memory, the processor executing the computer program to carry out the steps of the method of any one of claims 1 to 6.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method of any of claims 1-6.

Description

Sea state prediction method, sea state prediction device, sea state prediction equipment, sea state prediction medium and program product Technical Field The invention belongs to the technical field of sea state prediction, and particularly relates to a sea state prediction method, a sea state prediction device, sea state prediction equipment, a sea state prediction medium and a sea state prediction program product. Background The high-precision acquisition of wave and sea wind factors is a core link for grasping marine environment conditions in the process of offshore wind power construction and operation and maintenance, and the wave and wind load analysis of a fan foundation and a tower structure is closely related to the power output efficiency, fatigue damage assessment and the prediction of the service life of a platform of a wind turbine. The running environment of the offshore wind farm is complex and changeable, and extreme wind wave events (such as sudden strong wind changes, billows, typhoon residual waves and the like) often generate strong impact on a wind turbine tower, a foundation, a submarine cable and an operation and maintenance platform in a short time. The real-time prediction and early warning of the extreme sea conditions are important technical links for guaranteeing the safety of wind power equipment and reducing the unplanned shutdown rate. While traditional numerical modes such as a shallow sea wave mode (SWAN) and a weather research and forecast model (WRF) have physical rigor, the traditional numerical modes are difficult to quickly respond to locally abrupt wind wave power conditions depending on external background field driving, and have the problems of long calculation time consumption and difficulty in real-time updating, while statistical regression models and neural network regression models (such as LSTM (short-term set) and GRU (generalized-weighted unit)) can deal with part of time sequence characteristics, extreme values are easily underestimated under the conditions of scarce extreme samples and obvious nonlinear coupling of wind waves, so that early warning failure or delay is caused. Disclosure of Invention The sea condition prediction method comprises the steps of obtaining sea wave characteristics and wind field characteristics of a target sea area within a preset time range, extracting sea wave time sequence characteristics and wind field time sequence characteristics from the sea wave characteristics and the wind field characteristics within the preset time range by utilizing a bidirectional gating circulation unit, determining sea wave space-time attention characteristics and wind field space-time attention characteristics according to the sea wave time sequence characteristics and the wind field time sequence characteristics, determining wind wave fusion characteristics according to the sea wave space-time attention characteristics and the wind field space-time attention characteristics, extracting extreme perception characteristics from the wind wave fusion characteristics by utilizing an extreme perception module, and predicting sea conditions of the target sea area according to the extreme perception characteristics. The method solves the problem that the early warning is invalid or delayed due to the fact that extreme samples are scarce and wind wave nonlinear coupling is obvious in the prior art, achieves sea state trend prediction and extreme state intelligent early warning integration, and is applicable to real-time operation monitoring and risk prevention and control systems of offshore wind farms. In order to solve the technical problems, the application provides five aspects. The application provides a sea condition prediction method, which comprises the steps of obtaining sea wave characteristics and wind field characteristics of a target sea area within a preset time range, extracting sea wave time sequence characteristics and wind field time sequence characteristics from the sea wave characteristics and the wind field characteristics within the preset time range by utilizing a bidirectional gating circulation unit, determining sea wave time-space attention characteristics and wind field time-space attention characteristics according to the sea wave time sequence characteristics and the wind field time sequence characteristics, determining wind wave fusion characteristics according to the sea wave time-space attention characteristics and the wind field time-space attention characteristics, extracting extreme perception characteristics from the wind wave fusion characteristics by utilizing an extreme perception module, and predicting sea conditions of the target sea area according to the extreme perception characteristics. In some embodiments, the acquiring the sea wave characteristics and the wind field characteristics of the target sea area comprises acquiring a sea surface echo map of the target sea area within a preset duration range by utilizi