High-order random wave representation method, simulation method, system and storage medium for non-Gaussian wind field

Abstract

The present invention discloses a method for representing high-order random waves of a non-Gaussian wind field, comprising the following steps: S1: deriving a frequency-wavenumber spectrum of a random wave based on a target power spectrum density and a theoretical coherence function model γ(ξ,ω); S2: defining a frequency-wavenumber tri-spectrum according to the frequency-wavenumber spectrum, and obtaining a partial tri-coherence function of a high-order random wave; S3: calculating a pure frequency-wavenumber spectrum through the partial tri-coherence function; S4: using a Fourier-Stears integral expansion for the random wave, decomposing an orthogonal increment process, and establishing a new orthogonal increment process according to the pure frequency-wavenumber spectrum and the pure frequency-wavenumber tri-spectrum; S5: substituting the established orthogonal increment into the Fourier-Stears integral expansion of the random wave in step S4, and obtaining a high-order spectrum representation of a non-Gaussian random wave. The present invention also discloses a simulation method, system, and storage medium for a non-Gaussian wind field.

Description

Technical Field

[0001] The present invention belongs to the technical field of non-Gaussian wind field process simulation and wind load testing, and specifically relates to a high-order random wave representation method, a simulation method and a system and a storage medium for a non-Gaussian wind field. Background Art

[0002] In traditional wind resistance theory and design specifications, the average wind profile is described by exponential or logarithmic law, and the fluctuating wind is regarded as a random process of stable Gaussian distribution, that is, the third-order moment (skewness) is 0 and the fourth-order moment (kurtosis) is 3. However, the results of mountain wind measurement and terrain model wind tunnel test show that the average wind profile of mountain wind is complex; the spectral characteristics of fluctuating wind tend to shift to the high frequency band, which is significantly different from the wind spectrum model of existing design specifications; more importantly, the probability distribution of fluctuating wind presents unique non-Gaussian characteristics, that is, the skewness deviates from 0 and the kurtosis is far away from 3. At this time, Gaussian theory is no longer applicable to describe mountain wind with non-Gaussian characteristics. If the traditional stable Gaussian theory is still used for the wind resistance design of mountain transmission tower line system, it will lead to serious safety problems of the structure. Existing studies have shown that the non-Gaussian characteristics of fluctuating wind will accelerate the serious fatigue damage of wind turbine structure and even cause excessive extreme value response of the structure. At the same time, the non-Gaussianity of the load will cause the structural response to be highly non-Gaussian, causing a large structural response, which will affect the durability and safety of the structure.

[0003] However, for linear, slender, flexible structures and long-span structures, such as large wind turbines, long-span bridges, transmission towers, transmission lines, and flexible photovoltaic devices, wind loads are often the main design consideration. The randomness of dynamic wind loads often causes unpredictable vibration responses, which may damage the structural integrity or lead to catastrophic failures, such as the collapse of the Tacoma Narrows Bridge (Billal and Scanlan, 1991). To ensure the safety and performance of long-span structures, it is necessary to understand the wind-induced vibration mechanism and perform dynamic aeroelastic analysis (Davenport, 1961; Chen et al., 2000, 2010). The Monte Carlo simulation method is considered to be the standard for solving the randomness of dynamic wind loads (Gurley et al., 1996; Melaku et al., 2022; Jiang et al., 2023), but this method requires a large number of sample functions generated under the Gaussian assumption (Shinozuka et al., 1972). Actual measurements show that wind fields in mountainous areas often exhibit non-Gaussian characteristics, and using the Gaussian assumption to evaluate aerodynamic effects may lead to incorrect safety assessments. Therefore, accurate simulation of non-Gaussian wind fields is crucial for accurately estimating the dynamic response of long-span structures.

[0004] Currently, a variety of methods have been developed to simulate non-Gaussian random wind fields. One is to use random variables with deterministic basis functions, such as Hermite polynomials (Puig et al., 2002). Another method is "nonlinear translation" or "memoryless nonlinear transformation" (Gioffrè et al., 2000). These methods have played an important role in enhancing the analysis of complex systems, but compatibility issues remain challenging, and iterative improvements may introduce errors and reduce efficiency (Lagaros et al., 2005; Bocchini et al., 2008). To address these limitations, Shields et al. (2017) developed a third-order spectral representation method using bispectral, allowing the simulation of skew non-Gaussian processes. Vandanpu et al. (2024) developed third-order random waves to effectively simulate skew non-Gaussian wind fields of long-span bridges. The above methods do not consider fourth-order characteristics. Li et al. (2023; 2024) introduced the concept of trispectral and extended SRM to simulate fourth-order non-Gaussian processes. However, as the number of simulation points increases, the matrix decomposition becomes unstable and inefficient. Therefore, developing a high-order random wave method to simulate non-Gaussian wind fields is crucial to improve the efficiency of non-Gaussian wind field simulation. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a high-order random wave representation method, simulation method, system and storage medium for non-Gaussian wind fields, so as to realize high-order spectral representation and simulation of non-Gaussian wind fields, and provide a theoretical basis for accurately simulating multi-point non-Gaussian wind fields of large-span structures in actual engineering.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention first proposes a high-order random wave representation method for a non-Gaussian wind field, comprising the following steps:

[0008] S1: Derivation of the frequency-wavenumber spectrum S(ω,κ) of random waves based on the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω);

[0009] S2: Define the frequency-wavenumber triple spectrum T(κ) according to the frequency-wavenumber spectrum S(κ,ω) 1 ,κ 2 ,κ 3 ,ω 1 ,ω 2 ,ω 3 ), and obtain the partial tricoherence function of the high-order random wave

[0010] S3: Through the partial three-coherence function Calculate the pure frequency-wavenumber spectrum S p (κ,ω);

[0011] S4: Fourier-Stears integral expansion is used for random wave u(x, t), the orthogonal incremental process is decomposed, and a new orthogonal incremental process is established based on the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum;

[0012] S5: Substitute the established orthogonal increment into the Fourier-Stears integral expansion of the random wave u(x, t) in step S4 to obtain a high-order spectrum representation of the non-Gaussian random wave.

[0013] Furthermore, in step S1, the frequency-wavenumber spectrum S(ω,κ) of the random wave is derived by multiplying the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω):

[0014]

[0015] Among them: κ represents the spatial wave number; ξ represents the spatial distance; ω represents the angular frequency; i represents the imaginary unit.

[0016] Further, in step S2, the frequency-wavenumber triple spectrum T(κ 1 ,κ 2 ,κ 3 ,ω 1 ,ω 2 ,ω 3 ) is expressed as:

[0017]

[0018] Where: 1 , κ 2 and κ 3 represents the spatial wave number; ω 1 ,ω 2 and ω 3 represents the angular frequency;

[0019] Partial Tricoherence Functions of Higher-Order Random Waves It is expressed as:

[0020]

[0021] in: represents the partial tri-coherence function; S p (κ,ω) represents the pure frequency-wavenumber spectrum; and denote the spatial wave number respectively; and They represent angular frequency respectively; Δk represents the wave number interval; Δω represents the frequency interval.

[0022] Further, in step S3, the pure frequency-wavenumber spectrum S p (κ, ω) is expressed as:

[0023]

[0024] Where: S p (κ l ,ω m ) represents the pure frequency-wavenumber spectrum; S(k l ,ω m ) represents the traditional frequency-wavenumber spectrum; represents the partial tricoherence function of the higher-order random wave; κ l represents the spatial wave number; ω m represents the angular frequency; l represents the lth value of the spatial wave number; m represents the mth value of the frequency.

[0025] Further, in step S4, the random wave u(x, t) is expanded by Fourier-Stears integral to obtain:

[0026]

[0027] After discretization, we get:

[0028]

[0029] Where: l =lΔk,ω m =mΔω;du(κ l ,ω m ) and dv(κ l ,ω m ) represents an orthogonal incremental process.

[0030] Furthermore, for the orthogonal incremental process du(κ l ,ω m ) and dv(κ l ,ω m ) is decomposed to obtain:

[0031] du(κ l ,ω m )=du p (κ l ,ω m )+du I (κ l ,ω m )

[0032] dv(κ l ,ω m )=dv p (κ l ,ω m )+dv I (κ l ,ω m )

[0033] Among them: p (κ l ,ω m ) and dv p (κ l ,ω m ) represents a specific element corresponding to the pure frequency wave number spectrum; du I (κ l ,ω m ) and dv I (κ l ,ω m ) represents the component of power due to cubic coupling of waves;

[0034] According to the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum, a new orthogonal incremental process is established, and the following is obtained:

[0035]

[0036]

[0037] in:

[0038]

[0039] Where: S p (κ l ,ω m ) represents the pure frequency-wavenumber spectrum; S(κ l ,ω m ) represents the traditional frequency-wavenumber spectrum, is the square of the partial tricoherence function; represents the partial tri-coherence function; represents three phases; φ is a random phase angle uniformly distributed in [0,2π]; φ lm represents a random phase angle uniformly distributed in [0,2π]; and They represent random phase angles uniformly distributed in [0,2π]; Re[·] represents the real part; Im[·] represents the imaginary part.

[0040] Further, in step S5, the high-order spectrum of the non-Gaussian random wave is expressed as:

[0041]

[0042] Represented using finite frequency and wavenumber cutoffs:

[0043]

[0044] Where: N 1 and N 2They represent the discrete number of spatial wave numbers and the discrete number of frequencies respectively.

[0045] The present invention also proposes a method for simulating a non-Gaussian wind field, comprising the following steps:

[0046] Step 1: using the high-order random wave representation method of the non-Gaussian wind field as described above to obtain a high-order spectrum representation of the non-Gaussian random wave;

[0047] Step 2: Discretize the high-order spectrum representation of non-Gaussian random waves, incorporate fast Fourier transform into the high-order spectrum representation method of random waves, and obtain the high-order spectrum representation enhanced by fast Fourier transform;

[0048] Step 3: Determine the cutoff frequency ω u and the cutoff wave number κ u , the frequency, wave number, time and space are discretized, and the parameters are substituted into the high-order spectrum representation enhanced by fast Fourier transform to simulate the non-Gaussian wind field.

[0049] The present invention also proposes a non-Gaussian wind field simulation system, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program and implements the non-Gaussian wind field simulation method as described above.

[0050] The present invention further proposes a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the non-Gaussian wind field simulation method as described above is implemented.

[0051] The beneficial effects of the present invention are:

[0052] The simulation method of the non-Gaussian wind field of the present invention innovatively expands the classical second-order random wave representation by introducing the frequency-wavenumber spectrum and the frequency-wavenumber triple spectrum, and obtains the high-order spectrum representation of the non-Gaussian wind field, so that the non-Gaussian wind field can be modeled in the form of high-order random waves; in addition, in order to improve the simulation efficiency, the fast Fourier transform (FFT) is integrated into the high-order spectrum representation of the non-Gaussian random wave to generate a sample function of the non-Gaussian random wave, so that the multi-point fourth-order non-Gaussian wind field can be accurately simulated through the high-order spectrum representation, which provides a theoretical basis for accurately evaluating the wind load of large-span structures in actual engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0054] Figure 1 This is a schematic diagram of a long-span cable-stayed bridge;

[0055] Figure 2is a flow chart of a method for simulating a non-Gaussian wind field according to the present invention;

[0056] Figure 3 is the frequency-wavenumber spectrum of random waves;

[0057] Figure 4 (a) is the pure frequency-wavenumber spectrum of random waves;

[0058] Figure 4 (b) is a comparison diagram of the traditional frequency-wavenumber spectrum and the pure frequency-wavenumber spectrum of random waves;

[0059] Figure 5 (a) is a sample function diagram of the non-Gaussian wind field of a long-span cable-stayed bridge simulated by the high-order spectrum representation of random waves;

[0060] Figure 5 (b) is a sample function diagram of the wind field of a long-span cable-stayed bridge simulated by the second-order spectrum representation of random waves;

[0061] Figure 5 (c) is the difference between the wind field sample function simulated by the high-order spectrum representation of random waves and the traditional second-order spectrum representation;

[0062] Figure 6 This is a schematic diagram for comparing wind field profiles along the bridge main beam at different times;

[0063] Figure 7 It is a schematic diagram of wind speed time history comparison at different spatial positions of the bridge;

[0064] Figure 8 (a) is the probability density function diagram of the simulated wind field sample function;

[0065] Figure 8 (b) is the cumulative density function diagram of the simulated wind field sample function. DETAILED DESCRIPTION

[0066] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0067] In this embodiment, the proposed non-Gaussian wind field simulation method is used to simulate the wind field on a long-span cable-stayed bridge to evaluate the effectiveness and accuracy of the proposed non-Gaussian wind field simulation method in simulating non-Gaussian random waves. Specifically, the cable-stayed bridge is 600 meters long and the main beam is 46 meters high. Figure 1 shown.

[0068] like Figure 2 As shown, the simulation method of the non-Gaussian wind field of this embodiment includes the following steps.

[0069] Step 1: Use the high-order random wave representation method of the non-Gaussian wind field to obtain the high-order spectrum representation of the non-Gaussian random wave.

[0070] Specifically, the high-order random wave representation method of the non-Gaussian wind field includes the following steps.

[0071] S1: The frequency-wavenumber spectrum S(ω,k) of the random wave is derived based on the target power spectral density S(ω) and the theoretical coherence function model γ(ξ,ω).

[0072] Specifically, in this embodiment, the frequency-wavenumber spectrum S(ω, k) of the random wave is derived and expressed by the product of the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ, ω):

[0073] Specifically, in this embodiment, the target power spectrum density uses the Kaimal wind spectrum model as the target spectrum for simulating the wind field of a large-span cable-stayed bridge. The Kaimal wind spectrum model is expressed as follows:

[0074]

[0075] Where: ω represents the angular frequency; z represents the height above the ground; represents the average wind speed at height z; u * is the friction velocity, and:

[0076]

[0077] Where: k represents the von Karman constant, z 0 is the parameter of ground roughness.

[0078] The Davenport coherence function model is used to characterize the spatial correlation within the wind farm. The expression of the coherence function is:

[0079]

[0080] Where: λ is the attenuation factor; ξ represents the spatial distance between two points in the wind field.

[0081] Specifically, in this embodiment, the frequency-wavenumber spectrum S(ω,κ) of the random wave is derived and expressed by the product of the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω):

[0082]

[0083] Among them: κ represents the spatial wave number; ξ represents the spatial distance; ω represents the angular frequency; i represents the imaginary unit.

[0084] In this way, substituting the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω) into the above formula, the frequency-wavenumber spectrum can be obtained as follows:

[0085]

[0086] Define the parameters required in the frequency-wavenumber spectrum: L = 600m, z = 46m, k = 0.4, z 0 =0.03,λ=10,ω u =πrad / s, N ω =128, N κ =128,

[0087] Wherein, L represents the total length of the bridge; z represents the height of the main beam from the ground; represents the average wind speed at height z; 0 is the parameter of ground roughness; λ is the attenuation factor; ω u Indicates the cutoff frequency; N ω represents the discrete number of frequencies; Δω represents the frequency interval; N k represents the discrete number of spatial wave numbers; κ u represents the cutoff wavenumber; Δκ represents the wavenumber interval.

[0088] Substituting the parameters, we can get the frequency-wavenumber spectrum of random waves, such as Figure 3 shown.

[0089] S2: Define the frequency-wavenumber triple spectrum T(κ) according to the frequency-wavenumber spectrum S(κ,ω) 1 ,κ 2 ,κ 3 ,ω 1 ,ω 2 ,ω 3 ), and obtain the partial tricoherence function of the high-order random wave

[0090] In this embodiment, the frequency-wave number triple spectrum T(κ 1 ,k 2 ,k 3 ,ω 1 ,ω 2 ,ω 3 ) is expressed as:

[0091]

[0092] Where: 1 , κ 2 and κ 3 represents the spatial wave number; ω 1 ,ω 2 and ω 3 represents the angular frequency;

[0093] Partial Tricoherence Functions of Higher-Order Random Waves It is expressed as:

[0094]

[0095] in: represents the partial tri-coherence function; S p (k,ω) represents the pure frequency-wavenumber spectrum; and denote the spatial wave number respectively; and They represent angular frequency respectively; Δk represents the wave number interval; Δω represents the frequency interval.

[0096] S3: Through the partial three-coherence function Calculate the pure frequency-wavenumber spectrum S p (k,ω).

[0097] In this embodiment, the pure frequency-wave number spectrum S p (κ,ω) is expressed as:

[0098]

[0099] Where: S p (κ l ,ω m ) represents the pure frequency-wavenumber spectrum; S(κ l ,ω m ) represents the traditional frequency-wavenumber spectrum; represents the partial tricoherence function of high-order random waves; k l represents the spatial wave number; ω m represents the angular frequency; l represents the lth value of the spatial wave number; m represents the mth value of the frequency.

[0100] Specifically, the calculated pure frequency-wave number spectrum cloud diagram is as follows Figure 4 (a) In order to compare the difference between the traditional frequency-wavenumber spectrum and the pure frequency-wavenumber spectrum of random waves, a comparative schematic diagram of the two is given, as shown in Figure 4 (b) as shown.

[0101] S4: The random wave u(x, t) is subjected to Fourier-Stilles integral expansion to decompose the orthogonal incremental process, and a new orthogonal incremental process is established based on the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum.

[0102] Specifically, the random wave u(x, t) is expanded using the Fourier-Stears integral to obtain:

[0103]

[0104] After discretization, we get:

[0105]

[0106] Where: k l =lΔκ,ω m =mΔω;du(k l ,ω m ) and dv(κ l ,ω m ) represents an orthogonal incremental process.

[0107] For the orthogonal incremental process du(k l ,ω m ) and dv(κ l ,ω m ) is decomposed to obtain:

[0108] du(κ l ,ω m )=du p (κ l ,ω m )+du I (κ l ,ω m )

[0109] dv(κ l ,ω m )=dv p (k l ,ω m )+dv I (κ l ,ω m )

[0110] Among them: p (κ l ,ω m ) and dv p (κ l ,ω m ) represents a specific element corresponding to the pure frequency wave number spectrum; du I (κ l ,ω m ) and dv I (κ l ,ω m ) represents the component of power due to cubic coupling of the waves.

[0111] According to the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum, a new orthogonal incremental process is established, and the following is obtained:

[0112]

[0113] in:

[0114]

[0115] Where: S p (k l ,ω m ) represents the pure frequency-wavenumber spectrum; S(k l ,ω m ) represents the traditional frequency-wavenumber spectrum, is the square of the partial tricoherence function; represents the partial tri-coherence function; represents three phases; φ is a random phase angle uniformly distributed in [0,2π]; φ lm represents a random phase angle uniformly distributed in [0,2π]; and They represent random phase angles uniformly distributed in [0,2π]; Re[·] represents the real part; Im[·] represents the imaginary part.

[0116] S5: Substitute the established orthogonal increment into the Fourier-Stears integral expansion of the random wave u(x, t) in step S4 to obtain a high-order spectrum representation of the non-Gaussian random wave.

[0117] Specifically, the high-order spectrum of non-Gaussian random waves is expressed as:

[0118]

[0119] Represented using finite frequency and wavenumber cutoffs:

[0120]

[0121] Where: N 1 and N 2 They respectively represent the discrete number of spatial wave numbers and the discrete number of frequencies.

[0122] Step 2: In order to reduce the computational cost of simulating non-Gaussian random waves, the high-order spectrum representation of non-Gaussian random waves is discretized, and the fast Fourier transform (FFT) is incorporated into the high-order spectrum representation method of random waves to obtain the high-order spectrum representation enhanced by the fast Fourier transform.

[0123] Specifically, the high-order spectrum representation of the non-Gaussian random wave is discretized to obtain:

[0124]

[0125] in: represents the wave number interval, Indicates the frequency interval.

[0126] Substituting into Euler's formula, using symmetry to simplify the sum to only positive exponents, rewriting the high-order spectrum of the random wave to be:

[0127]

[0128] Where: φ (1|2) represents a suitable random phase angle and satisfies the uniform distribution in the interval [0,2π]; when i 2 、j 2 and k 2 When the phase angle is less than 0, φ (2) ; On the contrary; when i 2 、j 2 and k 2 When greater than 0, select the phase angle φ (1) In addition, when using the Fast Fourier Transform (FFT), the interval product must obey the following relationship:

[0129]

[0130] In order to further simplify the high-order spectrum representation of random waves, let:

[0131]

[0132]

[0133] Get the simplified high-order spectral representation of the random wave:

[0134]

[0135] Let: B represents B lm The matrix form of express The matrix shape of the non-Gaussian random wave can be represented by the pseudo-symbolic representation of FFT as follows:

[0136]

[0137] Step 3: Determine the cutoff frequency ω u and the cutoff wave number κ u , the frequency, wave number, time and space are discretized, and the parameters are substituted into the high-order spectrum representation enhanced by fast Fourier transform to simulate the non-Gaussian wind field.

[0138] Specifically, the non-Gaussian wind field of a large-span cable-stayed bridge is simulated according to the non-Gaussian wind field simulation method of this embodiment. The results are as follows: Figure 5 In addition, the second-order spectrum representation method of random waves is used to obtain the Gaussian wind field sample function, as shown in Figure 5 (b). To facilitate direct comparison, both methods used the same random phase angle in the sample function generation process. The difference between the sample functions generated by the two methods is shown in Figure 5(c) It is obvious that the sample function obtained from the higher-order spectral representation method is significantly different from the sample function obtained using the second-order spectral representation.

[0139] In order to further compare the difference between the high-order random wave representation method and the second-order spectrum representation method of the non-Gaussian wind field in this implementation, the following Figure 6 The wind speed profiles at different spatial positions along the longitudinal axis of the main beam of the cable-stayed bridge at a specific time as shown in the figure and Figure 7 The wind speed time history of different spatial points on the main beam shown in FIG. 1 is compared. After comparison, it is obvious that the non-Gaussian sample function generated by the high-order random wave representation method of the non-Gaussian wind field in this embodiment has a significant deviation from the result obtained using the traditional second-order spectrum representation. This difference is particularly obvious at the peak speed.

[0140] This embodiment uses the high-order random wave representation method and the second-order spectrum representation method of the non-Gaussian wind field to simulate 100 sample functions to estimate their probability density function (PDF) and cumulative density function (CDF). Specifically, the simulated probability density function and cumulative density function of the non-Gaussian random wave and Gaussian random wave are as follows: Figure 8 (a) and 8(b). Obviously, the PDF of the simulated non-Gaussian random wave is higher and narrower, while the PDF of the Gaussian random wave is fatter and wider. This proves that the kurtosis value of the non-Gaussian random wave simulated by the higher-order spectral representation is larger.

[0141] The simulation method of the non-Gaussian wind field of this embodiment simulates the non-Gaussian wind field. By introducing the frequency-wave number three spectrum, the classical second-order random wave representation is innovatively expanded to obtain a non-Gaussian wind field for high-order spectrum representation, thereby modeling the non-Gaussian wind field in the form of high-order random waves. In addition, in order to improve the simulation efficiency, the fast Fourier transform is integrated into the high-order spectrum representation method to generate a sample function of a non-Gaussian random wave. Within this theoretical framework, it is proved that the generated non-Gaussian random wave sample function satisfies the first-order, second-order and fourth-order ensemble properties. Finally, the method is applied to a 600-meter long-span cable-stayed bridge to model the non-Gaussian wind field and simulate the wind speed field. The results show that the first-order, second-order and fourth-order statistical characteristics of the generated sample function are consistent with the target characteristics. Therefore, the effectiveness and accuracy of the proposed algorithm have been verified. It provides a theoretical basis for accurately simulating non-Gaussian wind fields and evaluating the wind loads of structures in actual engineering. It should be noted that the simulation method of the non-Gaussian wind field in this embodiment is not limited to large-span bridges, but also applicable to other structures such as high-rise buildings, large-span cable structures, wind turbines and photovoltaic structures.

[0142] This embodiment also proposes a non-Gaussian wind field simulation system, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program and implements the non-Gaussian wind field simulation method as described above.

[0143] This embodiment further proposes a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the non-Gaussian wind field simulation method as described above is implemented.

[0144] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for simulating a non-Gaussian wind field, characterized in that: The steps include: Step 1: Use the high-order random wave representation method of the non-Gaussian wind field to obtain the high-order spectrum representation of the non-Gaussian random wave; The high-order random wave representation method of the non-Gaussian wind field comprises the following steps: S1: Derivation of the frequency-wavenumber spectrum S(ω,k) of random waves based on the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω); S2: Define the frequency-wavenumber triparticle T(κ1,κ2,κ3,ω1,ω2,ω3) according to the frequency-wavenumber spectrum S(k,ω) to obtain the partial tricoherence function of the high-order random wave S3: Through the partial three-coherence function Calculate the pure frequency-wavenumber spectrum S p (2,ω); S4: Fourier-Stears integral expansion is used for random wave u(x, t), the orthogonal incremental process is decomposed, and a new orthogonal incremental process is established based on the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum; S5: Substitute the established orthogonal increment into the Fourier-Stears integral expansion of the random wave u(x, t) in step S4 to obtain a high-order spectrum representation of the non-Gaussian random wave; Step 2: Discretize the high-order spectrum representation of non-Gaussian random waves, incorporate fast Fourier transform into the high-order spectrum representation method of random waves, and obtain the high-order spectrum representation enhanced by fast Fourier transform; Step 3: Determine the cutoff frequency ω u and the cutoff wave number κ u , discretize the frequency, wave number, time and space, substitute the parameters into the high-order spectrum representation enhanced by fast Fourier transform, and simulate the non-Gaussian wind field; In step S1, the frequency-wavenumber spectrum S(ω,κ) of the random wave is derived and expressed by the product of the target power spectrum density S(ω) and the theoretical coherence function model γ(ξ,ω): Among them: κ represents the spatial wave number; ξ represents the spatial distance; ω represents the angular frequency; i represents the imaginary unit; In step S2, the frequency-wavenumber triple spectrum T(κ1,κ2,κ3,ω1,ω2,ω3) is expressed as: Where: κ1, κ2 and κ3 represent spatial wave numbers; ω1, ω2 and ω3 represent angular frequencies; Partial Tricoherence Functions of Higher-Order Random Waves It is expressed as: in: represents the partial tricoherence function; S p (κ,ω) represents the pure frequency-wavenumber spectrum; and denote the spatial wave number respectively; and They represent angular frequency respectively; Δκ represents wave number interval; Δω represents frequency interval; In step S3, the pure frequency-wavenumber spectrum S p (κ,ω) is expressed as: Where: S p (κ l ,ω m ) represents the pure frequency-wavenumber spectrum; S(κ l ,ω m ) represents the traditional frequency-wavenumber spectrum; represents the partial tricoherence function of the higher-order random wave; κ l represents the spatial wave number; ω m represents the angular frequency; l represents the lth value of the spatial wave number; m represents the mth value of the frequency; In step S4, the random wave u(x, t) is expanded using Fourier-Stears integral to obtain: After discretization, we get: Among them: Mr l =lΔκ,ω m =mΔω;du(κ l ,oh m ) and dv(k l ,oh m ) display the process of increasing For the orthogonal incremental process du(κ l ,ω m ) and dv(κ l ,ω m ) is decomposed to obtain: du(k l ,oh m )=du p (k l ,oh m )+du I (k l ,oh m ) dv(k l ,oh m )=dv p (k l ,oh m )+dv I (k l ,oh m ) Among them: p (k l ,ω m ) and dv p (k l ,ω m ) represents a specific element corresponding to the pure frequency wave number spectrum; du I (k l ,ω m ) and dv I (k l ,ω m ) represents the component of power due to cubic coupling of waves; According to the pure frequency-wavenumber spectrum and the pure frequency-wavenumber triple spectrum, a new orthogonal incremental process is established, and the following is obtained: in: Where: S p (κ l ,ω m ) represents the pure frequency-wavenumber spectrum; S(κ l ,ω m ) represents the traditional frequency-wavenumber spectrum, is the square of the partial tricoherence function; represents the partial tri-coherence function; Indicates three phases; φ lm represents a random phase angle uniformly distributed in [0,2π]; and Respectively represent random phase angles uniformly distributed in [0,2π]; Re[·] represents the real part; Im[·] represents the imaginary part; In step S5, the high-order spectrum of the non-Gaussian random wave is expressed as: Represented using finite frequency and wavenumber cutoffs: Where: N1 and N2 represent the discrete number of spatial wave numbers and the discrete number of frequencies respectively.

2. A non-Gaussian wind field simulation system, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program and implements the method for simulating a non-Gaussian wind field as claimed in claim 1.

3. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the method for simulating a non-Gaussian wind field as claimed in claim 1 is implemented.

Images