Search

CN-122019933-A - Dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link drive

CN122019933ACN 122019933 ACN122019933 ACN 122019933ACN-122019933-A

Abstract

The invention provides a dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link driving, which is applied to safety evaluation of ocean engineering and wave energy power generation systems. Firstly, constructing B spline-lognormal edge distribution taking wind speed as a covariant through data preprocessing, describing dynamic changes of wave height and wave period parameters along with wind speed, secondly, selecting an optimal Copula function, coupling the Copula parameters with wind speed through Logistic mapping, building a dynamic dependency structure model, then dividing low, medium and high wind speed intervals, verifying fitting precision of the model in each interval and extreme wave energy overrun probability prediction performance, and finally, integrating the edge distribution and the dynamic Copula to construct a wind speed driven wave element joint probability prediction model. Compared with the traditional static model, the method can accurately capture the regulation and control effect of wind speed on the wave height-wave period dependency relationship, remarkably improve the prediction accuracy of extreme wave energy events, and provide reliable support for ocean engineering safety and wave energy power station operation.

Inventors

  • HAN LU
  • MA YIDAN

Assignees

  • 西南石油大学

Dates

Publication Date
20260512
Application Date
20260201

Claims (6)

  1. 1. A dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link drive is characterized in that the method synchronously drives wave element edge distribution and dependency structure through unified wind speed covariate (U) to realize complete dynamic depiction of wave height-wave period joint distribution, and the method comprises the following steps: s1, collecting marine meteorological and wave observation data comprising wave height (H), wave period (T) and corresponding wind speed (U), and preprocessing to obtain an effective sample data set; S2, constructing a dynamic edge distribution model of wind speed covariate drive and parameter cooperative response, namely respectively establishing lognormal distribution for wave height (H) and wave period (T), and establishing a cooperative response function relation of position parameter (mu) and scale parameter (sigma) and wind speed (U), wherein: a) For accurately capturing the nonlinear dynamics of the position parameters, fitting the position parameters (mu) into nonlinear functions of the wind speed (U) by adopting a cubic B spline function; b) To ensure the positive nature of the scale parameter and reflect the variation trend thereof, fitting the scale parameter (sigma) to a function of wind speed (U) by adopting a logarithmic linear function; S3, based on a fitting goodness criterion, screening an optimal basic Copula function describing the wave height and wave period dependency relationship from a preset Copula function family; S4, constructing a dynamic Copula dependent structure which is synchronously driven by wind speed with edge distribution, namely mapping the dependent parameter (theta) of the basic Copula function selected in the step S3 into a function of wind speed (U) through a Logistic function, so that the dynamic evolution of the dependent structure and the dynamic evolution of the edge distribution are unified to the same wind speed covariate, and the dependent parameter value is ensured to be always in an effective definition domain; And S5, integrating and predicting, namely integrating the dynamic edge distribution model obtained in the step S2 and the dynamic Copula dependent structure obtained in the step S4, constructing a full-link dynamic Copula joint probability distribution model based on the Sklar theorem, calculating the probability that the wave energy power density exceeds a preset threshold value in different wind speed intervals based on the model, and realizing the probability prediction of the extreme wave energy event.
  2. 2. The method according to claim 1, wherein the fitting of the position parameter (μ) with the cubic B-spline function in step S2 comprises determining at least 3 internal nodes based on the quantiles of the wind speed samples to form a node vector And linear combination is carried out by utilizing a B spline basis function, and the fitting expression is as follows: wherein U is a wind speed covariate, 、 The position parameters of the wave height and the wave period are respectively, As the number of the basic functions, 、 For the coefficients of the B-spline, Is a cubic B-spline basis function.
  3. 3. The method according to claim 1, wherein in the step S2, the fitting expression for fitting the scale parameter (σ) using a log-linear function is: wherein: 、 The scale parameters of wave height and wave period are respectively, 、 、 、 Is a parameter to be estimated.
  4. 4. The method according to claim 1, wherein in the step S4, when the basic Copula function is gaussian Copula, the dynamic expression of the mapping of the dependency parameter ρ (U) by the Logistic function is: wherein: as a function of the Logistic function, 、 For parameters to be estimated, ρ (U) ε (-1, 1) is ensured.
  5. 5. The method of claim 1, wherein the predetermined threshold is defined as the 90 th or 95 th percentile of the training set wave energy power density value, and wherein the wave energy power density J is calculated as: wherein: The density of the seawater is that, The acceleration of the gravity is that, Is the wave height (m), Is the wave period(s).
  6. 6. The method of claim 1, further comprising the step of predicting performance by using Root Mean Square Error (RMSE), mean Absolute Error (MAE) and Kolmogorov-Smirnov test to comprehensively evaluate consistency and accuracy between the extreme wave energy overrun probability predicted by the model and the measured frequency of the test set, wherein the calculation formulas are: Wherein n is the number of effective wind speed intervals, The overrun probability is predicted for the model of the ith interval, And (3) measuring the overrun frequency for the test set of the ith interval.

Description

Dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link drive Technical Field The invention relates to the technical field of ocean engineering and renewable energy sources, in particular to a method for carrying out probability prediction on extreme wave energy events, and especially relates to a dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link driving. Background The development and utilization of wave energy face serious challenges, namely, the power output of the wave energy has strong randomness and intermittence, especially the extreme wave energy event caused by extreme sea conditions, and the event has low occurrence probability, but causes overload impact on the ocean engineering structure and threatens the stability of a power grid, so that the accurate prediction of the occurrence probability of the extreme wave energy event is realized, and the wave energy is a key for improving the safety of ocean engineering and the economy of a wave energy system; The building of a model capable of accurately describing the joint probability distribution between two key wave elements, namely wave height (H) and wave period (T), is a core for calculating the probability of extreme wave working conditions, and the Copula theory is widely applied, but the prior art scheme has the following three remarkable limitations: 1. static model limitations traditional Copula models generally employ a parameter-fixed edge distribution and a static Copula-dependent structure. The model completely ignores 'wind speed (U)' as a core physical driving force for wave generation and evolution, has a strong dynamic regulation and control effect on wave height-wave period joint distribution, and can not capture the dynamic property in a real marine environment, so that serious deviation occurs to the probability prediction of polar joint events; 2. In recent years, although research attempts are made to introduce dynamic Copula, the dynamic property is only given to a single link, namely, only edge distribution parameters are changed along with environmental factors, or only Copula dependent parameters are changed, the scheme of 'local dynamic' cannot construct a 'full link' dynamic evolution model frame which is synchronous by wind speed covariates and systematically drives edge distribution and dependent structures, and the complete physical process of wave energy joint probability distribution along with wind speed real-time change cannot be completely and coherently depicted; 3. The existing method generally does not conduct differentiated fine modeling and verification on the significant heterogeneity shown by the wave high-wave period combined distribution in different wind speed intervals (such as low, medium and high wind speeds), the concentrated distribution at the low wind speed and the dispersion and multimodal distribution characteristics at the high wind speed are described by the same set of fixed parameter models, fitting distortion is necessarily caused, and the prediction reliability of the models in the extreme wave energy main source interval at the high wind speed is seriously insufficient; Therefore, a new method for predicting extreme wave energy, which can deeply couple wind speed influence, realize full-dynamic characterization of edges and dependent structures, and perform refined modeling and verification for different wind speed intervals, is needed to be provided, so as to solve the fundamental problems of insufficient prediction precision and poor adaptability to extreme working conditions in the prior art. Disclosure of Invention The invention aims to provide a dynamic Copula extreme wave energy prediction method based on wind speed covariate full-link driving, aiming at the defects of the existing static Copula model and incomplete dynamic Copula model in extreme wave energy prediction; The method has the core creative idea that a full-link model framework which takes wind speed as a unique physical covariant and synchronously drives the edge distribution of wave elements and the dynamic evolution of a dependent structure is constructed. To achieve this objective, the present invention proposes a set of coordinated parameterized mapping mechanisms: (1) Fitting by adopting a cubic B spline function, and accurately capturing complex nonlinear characteristics of the position parameter along with the change of wind speed according to the excellent local control capability and flexibility of the fitting; (2) Fitting by adopting a logarithmic linear function to ensure that the parameter is constant positive and effectively describe the trend of the parameter along with the change of wind speed; (3) Mapping by adopting a Logistic function to ensure that the dependent parameters are always in the effective definition domain and can sensitively reflect the regulation and control of wind speed on the dependent intensity; According