A method and system for identifying distributed dynamic loads based on time variable separation

Abstract

The application discloses a kind of distribution dynamic load identification method and system based on time variable separation.The method divides distribution load into distribution load function that can be directly separated space-time variable and distribution load function of general form, based on orthogonal polynomial fitting one-dimensional, two-dimensional function theory, respectively in two cases, time domain load is converted into frequency domain load, using Chebyshev orthogonal polynomial is completed in frequency domain to orthogonal polynomial coefficient identification, combined with time domain-frequency domain Fourier transform theory and least square method further completes the distribution time domain dynamic load identification of continuous beam.The application does not need to use a large number of discrete time domain data for modeling, and can directly and clearly represent the identified distribution load;Since only column vector matrix needs to be identified, the identification efficiency is improved based on a large number of reduction of identification calculation data, while the identification accuracy is also guaranteed.

Description

TECHNICAL FIELD

[0001] In the field of vibration analysis, the input part, the output part and the system characteristic part together constitute a complete vibration system. The determination of the third parameter under the premise of knowing two parameters of the vibration system constitutes three important research contents of structural dynamics. The present application belongs to the field of dynamic load identification, and the input load is determined under the condition that the output response and the system characteristics are known. Specifically, it relates to a distributed dynamic load identification method and system based on time variable separation. BACKGROUND

[0002] In engineering applications, the influence of load on the structure cannot be ignored. Traditionally, most loads are considered as concentrated loads for analysis and identification, but in actual situations, the loads acting on the structure often exhibit distributed characteristics. The existence of these distributed dynamic loads makes the identification of dynamic response more complex and difficult. The identification of distributed dynamic loads is an inverse problem, and its solution is of great significance to the safety and reliability of the structure. The existing technology in this field mainly focuses on the dynamic identification of concentrated loads, while the research on distributed loads is relatively less, resulting in immature technical means. Dynamic load identification, as the second type of inverse problem in vibration problems, started late and is difficult, but it is becoming more and more important in actual engineering problems and is a difficult problem that needs to be solved. Since the end of the 20th century, the theory of dynamic load identification has gradually matured, among which the time domain method and the frequency domain method have been widely applied. The time domain method is generally suitable for short-time loads, and the identification accuracy after time discretization is poor; the frequency domain method started early and is mature, and is suitable for identifying long-time excitations, but its accuracy is insufficient when dealing with short-time excitations.

[0003] The main shortcomings of the prior art are: the frequency domain method cannot clearly and directly represent the distributed load as the time domain method does; the time domain method needs to process a large amount of discrete time data, the matrix is too large, and the ill-posed problem of the matrix is easy to occur, and the fitting form of the distributed load also needs to be sufficient in order to complete the identification of the distributed load, i.e. a larger coefficient matrix is needed to complete the identification of the distributed load, both of which require a large amount of data to be calculated, which seriously leads to the problem of low calculation efficiency. Therefore, there is an urgent need for a new method that can clearly and intuitively represent the distributed load and reduce the calculation data while ensuring the identification accuracy and improving the identification efficiency. SUMMARY

[0004] The present application is proposed to solve the problems of the prior art. The present application provides a distributed dynamic load identification method and system based on time variable separation, which greatly reduces the calculation data of identification and improves the identification efficiency while ensuring the identification accuracy.

[0005] TECHNICAL SCHEME

[0006] A distributed dynamic load identification method based on time variable separation, comprising the following steps:

[0007] Step 1: According to the distributed load f(x, t) loaded on the identification object, it is divided into a distributed load which can directly separate time variable and a distributed load of general form. For the distributed load of general form, the time variable and the position variable are separated by the two-dimensional Chebyshev orthogonal polynomial fitting function principle. For the distributed load which can directly separate time variable, the position function is fitted by one-dimensional Chebyshev orthogonal polynomial fitting function principle;

[0008] Step 2: According to the form classification of the distributed load, the position orthogonal polynomial of appropriate order is selected to fit the frequency domain force F(x, ω) for the distributed load which has separated time variable, and the frequency domain displacement response matrix W and the frequency response function H are respectively obtained;

[0009] Step 3: The fitted frequency domain force is substituted into the displacement response and excitation relationship of the continuous system, the integral matrix of the frequency response function and the orthogonal polynomial is set as the dynamic calibration matrix HT, the coefficient matrix GA in the frequency domain is obtained by W=(HT)(GA);

[0010] Step 4: According to the form classification of the distributed load, for the load which can directly separate time variable, the Fourier transform of the time variable function g(t) is carried out under the fixed frequency to obtain the coefficient For the general form of load, the time orthogonal polynomial of appropriate order is selected Similarly, the coefficient matrix is obtained under the fixed frequency according to the Fourier transform q is the order of the time orthogonal polynomial;

[0011] Step 5: According to the relationship between the coefficient matrix GA in the frequency domain and the coefficient matrix And The original time domain coefficient of the distributed load f(x, t) is obtained by combining the least square method, and the identification of f(x, t) is completed.

[0012] A distributed dynamic load identification system based on time variable separation, comprising:

[0013] The distributed load time position separation module is used for dividing the distributed load f(x, t) loaded on the identification object into a distributed load which can directly separate time variable and a distributed load of general form. For the distributed load of general form, the time variable and the position variable are separated by the two-dimensional Chebyshev orthogonal polynomial fitting function principle. For the distributed load which can directly separate time variable, the position function is fitted by one-dimensional Chebyshev orthogonal polynomial fitting function principle;

[0014] A time-frequency conversion module is configured to select a position orthogonal polynomial of a proper order to fit the frequency domain force F(x, ω) according to the form classification of the distributed load, and to separately obtain a frequency domain displacement response matrix W and a frequency response function H;

[0015] A frequency domain coefficient solving module is configured to substitute the fitted frequency domain force into a displacement response and excitation relationship of a continuous system, set an integral matrix of the frequency response function and the orthogonal polynomial as a dynamic calibration matrix HT, obtain a simplified formula W=(HT)(GA), and solve a coefficient matrix GA in the frequency domain;

[0016] A constant frequency coefficient solving module is configured to perform a Fourier transform on a time variable function g(t) to obtain a coefficient matrix For a general form of load, a time orthogonal polynomial of a proper order is selected Similarly, the coefficient matrix is obtained according to the Fourier transform under the constant frequency q is an order of the time orthogonal polynomial;

[0017] An inverse reconstruction module is configured to obtain original time domain coefficients of the distributed load f(x, t) according to a relationship between the coefficient matrix GA in the frequency domain and the coefficient matrix and in the constant frequency, and to complete identification of f(x, t) by combining a least square method.

[0018] The application further provides a computer device, comprising one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs are used to implement steps of the distributed dynamic load identification method based on time variable separation when executed by the processors.

[0019] The application further provides a computer storage medium, which stores a computer program, and the computer program is used to implement steps of the distributed dynamic load identification method based on time variable separation when executed by the processors.

[0020] Beneficial effects: the distribution dynamic load identification method and system based on time variable separation provided by the application combine time domain theory and frequency domain method identification, divide the distribution dynamic load into two types of time variable and general form which can be directly separated, respectively identify the distribution dynamic load in the two cases in the frequency domain by combining orthogonal polynomials, and restore the time domain distribution force by using the Fourier transform theory of time domain-frequency domain. In the frequency domain identification, there is no need to use a large amount of data after time discretization for modeling, and the matrix ill-conditioning problem that may occur in the time domain identification is avoided; in the time domain restoration and final identification, the distribution dynamic load is more clearly and intuitively characterized than the pure frequency domain identification. At the same time, in the identification process, the two-dimensional orthogonal polynomial fitting coefficient matrix of the time domain force is identified in the time domain in the past, the calculation matrix is too large, the response data is too much, and the calculation efficiency is low, the application changes the solution of the two-dimensional matrix into the solution of the column vector matrix, greatly reduces the identification calculation data, improves the identification efficiency, and ensures the identification accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is a flowchart of the distribution dynamic load identification method based on time variable separation of the application;

[0022] Figure 2 is a simply supported beam model under the action of the distribution load in the embodiment of the application;

[0023] Figure 3 is a comparison of the time history identification results in the embodiment of the application;

[0024] Figure 4 is a comparison of the load distribution identification results in the embodiment of the application;

[0025] Figure 5 is the absolute error of the time history identification results in the embodiment of the application;

[0026] Figure 6 is the absolute error of the load distribution identification results in the embodiment of the application. DETAILED DESCRIPTION

[0027] The technical solutions of the application will be further described below with reference to the drawings.

[0028] The application divides the distribution load into a distribution load function which can directly separate time and space variables and a general form distribution load function, based on the one-dimensional and two-dimensional function theory of orthogonal polynomial fitting, respectively converts the time domain load into the frequency domain load in the two cases, uses Chebyshev orthogonal polynomial to complete the coefficient identification of the orthogonal polynomial in the frequency domain, and further completes the distribution time domain dynamic load identification of the continuous beam by combining the Fourier transform theory of time domain-frequency domain and the least square method. The specific technical solutions include:

[0029] 1. Time variable and position variable separation method of distributed dynamic load

[0030] Taking the one-dimensional structure distributed load f(x, t) as an example, it can be divided into a directly separable time variable form and a general form of distributed load function. When encountering a general form function that cannot directly separate the time variable, the two-dimensional Chebyshev orthogonal polynomial fitting function theory can be used to separate the distributed force using orthogonal polynomials.

[0031] f(x, t) = f(x)g(t) (1)

[0032]

[0033] Equation (1) is a directly separable time function of the distributed load, where f(x) is the position function and g(t) is the time function. Equation (2) is a general form of the distributed load function, The normalized Chebyshev orthogonal polynomial, represents the position polynomial, p is the order of the position polynomial, represents the time polynomial, q is the order of the time polynomial, a ij (ij = 1, 2, …, pq) is the coefficient of the two-dimensional orthogonal polynomial.

[0034] Similarly, using the one-dimensional Chebyshev orthogonal polynomial fitting function principle, the f(x) position function in the separable time variable distributed load is fitted:

[0035]

[0036] In equation (3), is the normalized one-dimensional Chebyshev position orthogonal polynomial, a i (i = 1, 2, …, r) is the coefficient of the one-dimensional polynomial. r is the order of the position orthogonal polynomial.

[0037] In addition, the time term in equation (2) is expanded separately, and there is:

[0038]

[0039] Thus, the time variable in the distributed load is separated from the position variable.

[0040] 2. Construction of time domain load-frequency domain load conversion relationship

[0041] For the two types of time domain distributed loads with separated time variables, equations (3) and (4), according to the Fourier time-frequency transform, they are converted into frequency domain forms, and the frequency domain forces are obtained as follows:

[0042]

[0043] Formula (5) is the frequency domain force that can directly separate the time variable, and formula (6) is the general form of the frequency domain force. The matrix forms are:

[0044]

[0045] In equations (7) and (8), the coefficients obtained by Fourier transform at a known excitation frequency are With coefficient are all known complex constants.

[0046] 3. Distributed dynamic load identification method based on time variable separation

[0047] Taking the continuous beam as an example, combined with the transfer relationship between the displacement response w and the dynamic load F in the frequency domain, the frequency domain forces of Equations (7) and (8) are substituted into the relationship between the n-point response and dynamic load of the continuous system in the frequency domain to obtain:

[0048]

[0049] In the above two equations, x is the spatial position distribution of the load, H xk (k=1,2,…,n) is the frequency response function of the corresponding position, w k (k=1,2,…,n) is the frequency domain response of the response point. Let the integral matrix of the frequency response function and the orthogonal polynomial be the dynamic calibration matrix HT, which can be simplified as follows:

[0050] W=(HT)(GA) (11)

[0051] In Equation (11), W is the frequency-domain displacement response matrix, and the matrix GA is the coefficient matrix of the Fourier transform coefficients of the time function and the fitting coefficients of the time-domain load. Given the known response and dynamic calibration matrix HT, the coefficient matrix can be obtained by combining matrix inversion or generalized inversion methods.

[0052] 4. Inverse reconstruction of time-domain distributed dynamic loads

[0053] After obtaining the coefficient matrix GA in the frequency domain, due to the fixed frequency and As the coefficients are known, the coefficients of the time domain expressions of the two types of distributed forces can be obtained separately by combining the least squares method:

[0054]

[0055]

[0056] The coefficient matrix a i (i=1,2,…,r) and a ijThe reduction of the two types of distributed forces can be completed by substituting the equation (3) and the equation (2) respectively.

[0057] Based on the above inventive concept, the method implementation steps of the present application include:

[0058] Step 1: According to the distributed load f(x, t) loaded on the identification object, the frequency domain displacement response matrix W and the frequency response function H are obtained; in the verification of the simulation identification calculation, the distributed load f(x, t) is loaded on the identification object through forward calculation, and the frequency domain displacement response matrix W and the frequency response function H are obtained; in the actual application, the experimental measurement can be directly performed;

[0059] Step 2: According to the form classification of the distributed load, the position orthogonal polynomial of the appropriate order is selected to fit the frequency domain force F(x, ω); wherein for the selection of the position order: in order to ensure the solution of equation (11), the order (r or p) of the position polynomial needs to be less than or equal to the number n of the response measuring points;

[0060] Step 3: The fitted frequency domain force in step 2 is substituted into the displacement response and excitation relationship formula of the continuous system to obtain the simplified formula W=(HT)(GA), and the coefficient matrix GA in the frequency domain is obtained;

[0061] Step 4: According to the classification of the distributed force, the Fourier transform of the function of the time variable under the fixed frequency is directly performed to obtain the coefficient Or using the time orthogonal polynomial of the appropriate order, the coefficient matrix Through the simulation calculation verification, generally, the time orthogonal polynomial needs to be taken to 3 orders or less than 10 orders, so as to ensure the identification accuracy. Too low order will lead to insufficient fitting order and too low identification accuracy; too high order will affect the identification accuracy due to too many unknown coefficients. In the actual application, the cross-validation method can be used to compare and select an order with the best identification effect.

[0062] Step 5: The original coefficients of the distributed load f(x, t) are obtained through equation (12) or equation (13) by combining the coefficient matrices in step 3 and step 4, and the identification of f(x, t) is completed.

[0063] The actual application of the method of the present application will be illustrated by combining with a specific model, and the performance of the method will be verified. The distributed dynamic load is applied to the simply supported beam structure shown in the figure, and the parameters of the simply supported beam are shown in Table 1: Figure 2

[0064] Table 1 Geometric parameters and material properties of the simply supported beam

[0065]

[0066] ​Taking identification of a distribution load in a more complicated general form as an example, a general form of distribution load is applied on the simply supported beam model shown in Figure 2 The time-varying relationship of the load is The loading time is 10s, 10 points on the beam are taken as response measurement points, and 50 identification points on the beam are taken for dynamic load identification.

[0067] Step 1: calculating the distribution load in the frequency domain The frequency domain displacement response matrix W under the excitation and the frequency response function H are obtained;

[0068] Step 2: selecting 8-order position orthogonal polynomials The fitting form of the frequency domain force F(x, ω) is obtained;

[0069] Step 3: for the continuous beam structure, the frequency domain force is substituted into formula (10) to calculate the coefficient matrix in the frequency domain, and the coefficient matrix GA is obtained, and the matrix size is 8x1;

[0070] Step 4: selecting 10-order time orthogonal polynomials The Fourier transform under the fixed frequency is performed, and the column vector matrix of is obtained;

[0071] Step 5: according to formula (13), the obtained matrix GA and are combined to calculate the distribution force coefficient a in formula (2) n (n=1, 2, …, pq) in the matrix form, and the identification of the general form distribution load is completed. Taking the time history identification result obtained by selecting x=0.15m on the beam and the load distribution identification result obtained by selecting the loading time t=5.5s as examples, the identification effects are shown in Figures 3-6 It can be seen that the load error obtained by the method of the present application is very small, and the precision is very high.

[0072] Based on the same technical concept as the method embodiment, the present application also provides a distribution dynamic load identification system based on time variable separation, comprising:

[0073] A distribution load time position separation module is configured to separate the distribution load f(x, t) loaded on the identification object into a form directly separable from the time variable and a general form of distribution load function. For the general form of distribution load, the time variable and the position variable are separated through the principle of two-dimensional Chebyshev orthogonal polynomial fitting function. For the distribution load directly separable from the time variable, the position function is fitted through the principle of one-dimensional Chebyshev orthogonal polynomial fitting function.

[0074] A time-frequency conversion module is configured to select a position orthogonal polynomial of a proper order to fit the frequency domain force F(x, ω) according to the form classification of the distributed load, and to separately obtain a frequency domain displacement response matrix W and a frequency response function H;

[0075] A frequency domain coefficient solving module is configured to substitute the fitted frequency domain force into a displacement response and excitation relationship of a continuous system, set an integral matrix of the frequency response function and the orthogonal polynomial as a dynamic calibration matrix HT, obtain a simplified formula W=(HT)(GA), and solve a coefficient matrix GA in the frequency domain;

[0076] A constant frequency coefficient solving module is configured to perform a Fourier transform on a time variable function g(t) to obtain a coefficient matrix For a general form of load, a time orthogonal polynomial of a proper order is selected Similarly, the coefficient matrix is obtained according to the Fourier transform under the constant frequency q is an order of the time orthogonal polynomial;

[0077] An inverse reconstruction module is configured to obtain original time domain coefficients of the distributed load f(x, t) according to a relationship between the coefficient matrix GA in the frequency domain and the coefficient matrix and and to complete identification of f(x, t) by using a least square method.

[0078] The application further provides a computer device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs are used to implement steps of the distributed dynamic load identification method based on time variable separation as described above.

[0079] The application further provides a computer storage medium, which stores a computer program, and the computer program is used to implement steps of the distributed dynamic load identification method based on time variable separation as described above.

[0080] Those skilled in the art should understand that embodiments of the application can be provided as a method, device, computer device or computer program product. Therefore, the application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the application can adopt a computer program product in the form of one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0081] The application is described with reference to the Figures according to which the methods of the application are illustrated. It will be understood that each flow of the flow diagram, as well as combinations of the flows of the flow diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, create means for implementing the functions specified in the flow diagram Figure 1 flow or multiple flows. These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flow diagram Figure 1 flow or multiple flows. These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flow diagram Figure 1 flow or multiple flows. These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flow diagram

Claims

1. A method for identifying distributed dynamic loads based on time variable separation, characterized in that, The method comprises the following steps: Step 1: according to the distribution load f(x, t) loaded on the identified object, the distribution load is divided into a distribution load directly separable in time variable and a distribution load in general form, for the distribution load in general form, time variable and position variable are separated by a two-dimensional Chebyshev orthogonal polynomial fitting function principle, for the distribution load directly separable in time variable, a position function is fitted by a one-dimensional Chebyshev orthogonal polynomial fitting function principle; Step 2: according to the form classification of the distribution load, for the distribution load separated in time variable, appropriate order position orthogonal polynomials are selected to fit the frequency domain force F(x, ω), and the frequency domain displacement response matrix W and the frequency response function H are respectively solved; Step 3: the fitted frequency domain force is substituted into the displacement response and excitation relationship of the continuous system, the integral matrix of the frequency response function and the orthogonal polynomial is set as a dynamic calibration matrix HT, a simplified formula W=(HT)(GA) is obtained, and the coefficient matrix GA in the frequency domain is solved; Step 4: According to the form of the distributed load, for the load that can directly separate the time variable, the Fourier transform of the time variable function g(t) is carried out under the fixed frequency to obtain the coefficient For the general form of the load, the time orthogonal polynomial of the appropriate order is selected Similarly, the coefficient matrix is obtained under the Fourier transform at the fixed frequency n = 1, 2, … q, q is the order of the time orthogonal polynomial; Step 5: According to the relationship between the coefficient matrix GA under the frequency domain and the coefficient matrix G under the constant frequency, combined with the least square method, the original time domain coefficient of the distributed load f(x, t) is obtained respectively, and the identification of f(x, t) is completed. and the relationship, combined with the least square method, the original time domain coefficient of the distributed load f(x, t) is obtained respectively, and the identification of f(x, t) is completed.

2. The method of claim 1, wherein, In the step 1, for the distribution load in general form, time variable and position variable are separated by a two-dimensional Chebyshev orthogonal polynomial fitting function principle, comprising: According to the two-dimensional Chebyshev orthogonal polynomial fitting function theory, the distribution load is separated by using orthogonal polynomials: wherein is a normalized Chebyshev orthogonal polynomial, represents a location polynomial, i = 1, 2, …, p, p is an order of the location polynomial, represents a time polynomial, j = 1, 2, …, q, q is an order of the time polynomial, a ij is a coefficient of a two-dimensional orthogonal polynomial.

3. The method of claim 2, wherein, For the distribution load directly separable in time variable, a position function is fitted by a one-dimensional Chebyshev orthogonal polynomial fitting function principle, comprising: The distribution load directly separable in time function is expressed as f(x, t)=f(x)g(t), wherein f(x) is a position function and g(t) is a time function, the position function f(x) in the distribution load directly separable in time variable is fitted by using a one-dimensional Chebyshev orthogonal polynomial fitting function principle: wherein is the normalized Chebyshev position-orthogonal polynomial, a i are coefficients of the one-dimensional polynomials, i = 1, 2,..., r, r is the order of the position-orthogonal polynomials.

4. The method of claim 3, wherein, In the step 2, for the distribution load separated in time variable, appropriate order position orthogonal polynomials are selected to fit the frequency domain force F(x, ω), comprising: For the two types of distribution load separated in time variable, they are converted into frequency domain forms according to Fourier time-frequency transformation, and frequency domain force expressions are obtained; Appropriate order position orthogonal polynomials are selected to fit the frequency domain force F(x, ω).

5. The method of claim 4, wherein, For the distribution load directly separable in time variable, the frequency domain force expression after Fourier time-frequency transformation is: In matrix form: For the distribution load in general form, the time term is expanded separately, and the frequency domain force expression after Fourier time-frequency transformation is: Comprising: In matrix form:

6. The method of claim 1, wherein, In step 5, for the distributed load which can be directly separated from the time function, the coefficients a are obtained by using the least square method i , i = 1, 2, …, r, and a i is brought back to the load expression of the one-dimensional Chebyshev orthogonal polynomial expansion to obtain the distributed load identification result.

7. The method of claim 1, wherein, In step 5, for the general form of distributed load, according to The coefficients a of the time-domain expression of such distributed force are obtained by using the least square method ij , a ij The load expression with the two-dimensional Chebyshev orthogonal polynomial expansion is brought back to obtain the distributed load identification result.

8. A time variable separation based distributed dynamic load identification system, characterized in that, A distribution load time position separation module is used for separating the distribution load f(x, t) loaded on the identified object into a distribution load directly separable in time variable and a distribution load in general form, for the distribution load in general form, time variable and position variable are separated by a two-dimensional Chebyshev orthogonal polynomial fitting function principle, for the distribution load directly separable in time variable, a position function is fitted by a one-dimensional Chebyshev orthogonal polynomial fitting function principle; A time-frequency conversion module is used for selecting appropriate order position orthogonal polynomials to fit the frequency domain force F(x, ω) for the distribution load separated in time variable according to the form classification of the distribution load, and the frequency domain displacement response matrix W and the frequency response function H are respectively solved; ​ A frequency domain coefficient solving module is configured to substitute the fitted frequency domain force into a displacement response and excitation relationship of the continuous system, set an integral matrix of the frequency response function and the orthogonal polynomial as a dynamic calibration matrix HT, obtain a simplified formula W=(HT)(GA), and solve a coefficient matrix GA in the frequency domain; The coefficient solving module at constant frequency is used for classifying the form of distributed load, and performing Fourier transform on the time variable function g(t) for the load which can be directly separated from time variable to obtain the coefficient at constant frequency For the general form of load, the time orthogonal polynomial of appropriate order is selected Similarly, the coefficient matrix is obtained at constant frequency according to Fourier transform n=1, 2, … q, q is the order of time orthogonal polynomial An inverse reconstruction module is configured to obtain original time-domain coefficients of the distributed load f(x, t) according to a relationship between the coefficient matrix GA in the frequency domain and the coefficient matrix G in the fixed frequency, in combination with a least square method, to complete identification of f(x, t). and ​ 9. A computer device, comprising: comprise: one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs, when executed by the processors, implement the steps of the time variable separation based distributed dynamic load identification method according to any one of claims 1-7.

10. A computer storage medium having stored thereon a computer program, characterized in that The computer programs, when executed by the processors, implement the steps of the time variable separation based distributed dynamic load identification method according to any one of claims 1-7. The computer programs, when executed by the processors, implement the steps of the time variable separation based distributed dynamic load identification method according to any one of claims 1-7.

Images