阅读说明：本技术 一种针对比例阻尼结构的复模态辨识方法 (Complex mode identification method for proportional damping structure ) 是由 曲春绪 伊廷华 李宏男 于 2019-10-24 设计创作，主要内容包括：本发明属于工程结构监测数据分析技术领域,涉及带有比例阻尼结构模态识别方法中的复模态辨识方法。首先通过短时傅里叶变换及单源点检测获取构成实振型的单源点集合,再通过层次距离方法求解实振型；通过自然激励技术和希尔伯特变换获取脉冲响应及其希尔伯特变换,建立模态响应与响应的关系,求解复模态信息。该发明通过明确地表达式给出了带有比例阻尼结构隐藏的复模态的求解步骤,从本质上揭示了结构振动规律。(The invention belongs to the technical field of engineering structure monitoring data analysis, and relates to a complex modal identification method in a modal identification method with a proportional damping structure. Firstly, acquiring a single-source point set forming a real vibration mode through short-time Fourier transform and single-source point detection, and solving the real vibration mode through a hierarchical distance method; and acquiring the impulse response and the Hilbert transform thereof through a natural excitation technology and the Hilbert transform, establishing a relation between modal response and response, and solving complex modal information. The invention provides a complex modal solving step with proportional damping structure hiding through an explicit expression, and essentially discloses a structural vibration law.)

1. A complex mode identification method for a proportional damping structure is characterized by comprising the following steps:

the method comprises the following steps: tap matrix identification

(1) Acquiring the acceleration response y (k) of the structure at time k [ y ]₁(k),y₂(k),…,y_l(k)]^T(ii) a The time domain acceleration response is transformed to a time-frequency domain by adopting short-time Fourier transform, and the expression of the time domain acceleration response is changed into Y (K, omega) ([ Y)₁(K,ω),Y₂(K,ω),…,Y_l(K,ω)]Wherein l is the number of sensors, K represents the Kth time interval, and omega represents the circle frequency;

(2) the single source point reflects single-order modal information, the detection of the single source point is based on the fact that the real part and the imaginary part of the time-frequency coefficient have the same direction, and the unit point is detected by adopting the following formula:

detected single-source point position markIs denoted by (t)_K,ω_K,i) The value is:

Y(K,ω_K,i)＝[Y₁(K,ω_K,i),Y₂(K,ω_K,i),...,Y_l(K,ω_K,i)]^T

wherein, the symbol "K, i" represents the ith order frequency in the Kth time period;

(3) determining the clustering number through the number of peaks with obvious power spectral density of acceleration response, and using a mature hierarchical clustering method to perform single source point Y (K, omega)_K,i) Classifying, calculating the clustering center of each class, and obtaining the real vibration matrix phi^R；

Step two: complex oscillation solving type

(4) The acceleration response y (k) is converted into an impulse response signal y by utilizing the mature natural excitation technology_d(k) For the impulse response signal y_d(k) Performing Hilbert transform to obtain

(5) The following equation is given below:

wherein phi^IRepresenting the imaginary part of the complex mode and the real mode as phi^R±jΦ^IAnd satisfies phi^I＝Φ^RGamma, j represents an imaginary unit satisfying j²＝-1；q^RAnd q is^IRepresents a modal coordinate, and satisfies the following equation:

y_d(k)＝[Φ^R+jΦ^I][q^R+jq^I]^T+[Φ^R-jΦ^I][q^R-jq^I]^T

(6) solving the pseudo-inverse of the constructed equation to obtain q^RAnd q is^IIs expressed, q in the expression^RAnd q is^IExpressed by the unknown parameter γ:

wherein, the symbolRepresents a pseudo-inverse;

(7) q with the parameter gamma^RAnd q is^IThe expression is substituted into the following formula, and the unknown parameter gamma is obtained by solving:

[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

＝[q^R(k+2)+jq^I(k+2)]./[q^R(k+1)+jq^I(k+1)]

wherein, the symbol "/" represents a dot division, that is, each row of elements in the vector are divided respectively, and k represents the kth moment;

(8) the real vibration type matrix phi obtained according to the step (3)^RObtaining the imaginary part phi of the mode matrix from the solved gamma in the step (7)^I＝Φ^Rγ；

(9) The gamma solved in the step (7) is brought into the q given in the step (6)^RAnd q is^IIn the expression, get q^RAnd q is^I；

(10) The complex frequency is found by:

ω^R+jω^I＝[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

wherein, ω is^RAnd ω^IReal and imaginary parts of complex frequencies;

(11) the damping ratio is found by:

wherein the content of the first and second substances,

2. The method of complex mode identification for a proportional damping structure of claim 1, wherein Δ β is 2 °.

Technical Field

The invention belongs to the technical field of engineering structure monitoring data analysis, and relates to a complex modal identification method in a modal identification method with a proportional damping structure.

Background

Structural health monitoring is an important means for guaranteeing structural safety, and modal parameters reflect structural dynamic characteristics and can be used for evaluating structural performance, so that identification of structural modal parameters by using structural monitoring data is of great importance.

The modal parameters of the structure comprise frequency, vibration mode and damping ratio, the actual engineering structure is mostly assumed to have proportional damping, the existing modal parameter identification method identifies the mode of the structure, the obtained modal parameters are often real modal parameters, the real condition is that the mode is a complex mode, the conjugate imaginary parts are mutually offset, the real modal artifact is displayed, the implicit complex modal information is identified, and the key for revealing the dynamic characteristic essence of the structure is realized.

There are many common modal parameter identification methods in engineering, Juang and Pappa propose a characteristic system implementation algorithm in 1985 to identify modal parameters by using an impulse response signal; overscheee and Moor proposed a stochastic subspace approach in 2012 to identify modal parameters using white noise excitation responses; qu equals 2019, proposes to reduce environmental noise in a frequency domain by using a concept of a transfer function, and converts the transfer function into an impulse response function for modal identification; yao equals 2018 and proposes to use a frame of blind source separation to identify the modal parameters; antonnes is equal to 2018, and a blind source separation method is adopted to identify complex modes by analyzing signals; bajri' c and

in 2018, a damping matrix expression formed by complex eigenvectors and eigenvalues of a non-classical damping structure is given. However, the engineering structure is in a stable state for a long time, which shows the characteristics of the proportional damping structure, and it is difficult to obtain the real complex modal information of the structure by the above method, so that it is difficult to accurately grasp the dynamic characteristics of the structure. Therefore, how to perform complex mode identification for the proportional damping structure is necessary.

Disclosure of Invention

The invention aims to provide a complex modal identification method for a proportional damping structure, and solves the problem of hidden complex modal identification in the process of identifying the modal with the proportional damping structure.

The technical scheme of the invention is as follows: a complex modal identification method for a proportional damping structure is deduced, and a structural response signal under environmental excitation is subjected to short-time Fourier transform to obtain a real vibration mode through single-source point detection and a mature hierarchical clustering method. The method comprises the steps of converting an environment excitation response signal into an impulse response signal through a mature natural excitation technology, carrying out Hilbert transform on the impulse response signal, establishing a functional relation between modal response and impulse response and Hilbert transform of the impulse response, solving a relation coefficient between a real vibration mode and a complex vibration mode, introducing the coefficient into the modal response, solving a complex frequency through a ratio of the modal responses at two adjacent moments, and solving a damping ratio through the complex frequency, so that three modal parameters except the complex vibration mode, the complex frequency and the damping ratio are identified.

A complex mode identification method for a proportional damping structure comprises the following steps:

the method comprises the following steps: tap matrix identification

(1) Acquiring the acceleration response y (k) of the structure at time k [ y ]₁(k),y₂(k),…,y_l(k)]^T(ii) a The time domain acceleration response is transformed to a time-frequency domain by adopting short-time Fourier transform, and the expression of the time domain acceleration response is changed into Y (K, omega) ([ Y)₁(K,ω),Y₂(K,ω),…,Y_l(K,ω)]Wherein l is the number of sensors, K represents the Kth time interval, and omega represents the circle frequency;

(2) the single source point reflects single-order modal information, the detection of the single source point is based on the fact that the real part and the imaginary part of the time-frequency coefficient have the same direction, and the unit point is detected by adopting the following formula:

where Re {. and Im {. denotes the real and imaginary parts of the extracted data, respectively, Δ β denotes a threshold for single source point detection, which may be set to Δ β — 2 °;

the detected single-source position is marked as (t)_K,ω_K,i) The value is:

Y(K,ω_K,i)＝[Y₁(K,ω_K,i),Y₂(K,ω_K,i),...,Y_l(K,ω_K,i)]^T

wherein, the symbol "K, i" represents the ith order frequency in the Kth time period;

(3) determining the clustering number through the number of peaks with obvious power spectral density of acceleration response, and using a mature hierarchical clustering method to perform single source point Y (K, omega)_K,i) Classifying, calculating the clustering center of each class, and obtaining the real vibration matrix phi^R；

Step two: complex oscillation solving type

(4) Converting the response signal y (k) into an impulse response signal y by utilizing mature natural excitation technology_d(k) For the impulse response signal y_d(k) Performing Hilbert transform to obtain

(5) The following equation is given below:

wherein phi^IRepresenting the imaginary part of the complex mode and the real mode as phi^R±jΦ^IAnd satisfies phi^I＝Φ^RGamma, j represents an imaginary unit satisfying j²＝-1；q^RAnd q is^IRepresents a modal coordinate, and satisfies the following equation:

y_d(k)＝[Φ^R+jΦ^I][q^R+jq^I]^T+[Φ^R-jΦ^I][q^R-jq^I]^T

(6) solving the pseudo-inverse of the constructed equation to obtain q^RAnd q is^IIs expressed, q in the expression^RAnd q is^IExpressed by the unknown parameter γ:

wherein, the symbolRepresents a pseudo-inverse;

(7) q with the parameter gamma^RAnd q is^IThe expression is substituted into the following formula, and the unknown parameter gamma is obtained by solving:

[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

＝[q^R(k+2)+jq^I(k+2)]./[q^R(k+1)+jq^I(k+1)]

wherein, the symbol "/" represents a dot division, that is, each row of elements in the vector are divided respectively, and k represents the kth moment;

(8) the real vibration type matrix phi obtained according to the step (3)^RObtaining the imaginary part phi of the mode matrix from the solved gamma in the step (7)^I＝Φ^Rγ；

(9) The gamma solved in the step (7) is brought into the q given in the step (6)^RAnd q is^IIn the expression, get q^RAnd q is^I；

(10) The complex frequency is found by:

ω^R+jω^I＝[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

wherein, ω is^RAnd ω^IReal and imaginary parts of complex frequencies;

(11) the damping ratio is found by:

wherein the content of the first and second substances,and

are respectively omega^RAnd ω^II.e. the real and imaginary parts of the ith order complex frequency.

To this end, the complex modal parameter ω^R±jω^I，Φ^R±jΦ^I，ζ_iAll are found.

The invention has the beneficial effects that: the method can be used for identifying the complex modal information hidden in the proportional damping structure by an analytical means, has clear and simple flow, does not need iteration, solves the complex modal of the proportional damping structure, and can better reveal the essential rule of the structural vibration.

Detailed Description

The following further illustrates embodiments of the present invention in conjunction with the technical solutions.

Taking a 3-layer frame structure as an example, the mass of each layer is 1 × 10³kg、2×10³kg、1×10³kg, stiffness matrix and mass damping matrix are as follows:

the excitation form is white noise excitation, the noise level is 20% of the variance of the actual signal, and the response signal is the displacement of each layer of the structure. The specific implementation mode of the method is as follows:

the method comprises the following steps: tap matrix identification

(1) The acceleration response at the time k is acquired, and y (k) is [, y [ ]₁(k),y₂(k),…,y_l(k)]^T(ii) a Converting the time domain acceleration response into a time-frequency domain by adopting short-time Fourier transform, wherein the expression of the time domain acceleration response is changed into Y (K, omega), wherein l is the number of the sensors, K represents the Kth time period, and omega represents the circular frequency;

(2) performing single source point detection according to a formula:

wherein Re {. and Im {. can respectively represent the real part and the imaginary part of the extracted data, and the detected single source points are as follows:

Y(K,ω_K,i)＝[Y₁(K,ω_K,i),Y₂(K,ω_K,i),...,Y_l(K,ω_K,i)]^T

wherein, the symbol "K, i" represents the ith order frequency in the Kth time period;

(3) determining the clustering number to be 3 according to the number of peaks with obvious power spectral density of acceleration response, and using a hierarchical clustering method to single source point Y (K, omega)_K,i) Classifying, and calculating the clustering center of each class to obtain a real vibration matrix

Second, solution of complex oscillation

(4) The acceleration response signal y (k) is converted into an impulse response signal y by utilizing a mature natural excitation technology_d(k) For pulse signal y_d(k) Performing Hilbert transform to obtain

(5) Expressing q by an unknown quantity gamma^RAnd q is^I：

Wherein, the symbol

Represents the pseudo inverse, phi^IRepresents an imaginary part of the complex mode and satisfies phi^I＝Φ^Rγ；；

(6) Q is to be^RAnd q is^IThe expression is substituted into the following expression, and the solution yields an unknown parameter γ ═ diag ([18.0156,17.6482, 17.1870)])：

[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

＝[q^R(k+2)+jq^I(k+2)]./[q^R(k+1)+jq^I(k+1)]

Wherein, "diag" represents a diagonal matrix, the symbol "/" represents a dot division, that is, each row of elements in a vector are divided respectively, j represents an imaginary unit, and j is satisfied²-1, k denotes time k;

(7) the real vibration type matrix phi obtained according to the step (3)^RObtaining the imaginary part phi of the mode matrix from the solved gamma in the step (6)^I＝Φ^Rγ；

(8) The gamma solved in the step (6) is brought into the q given in the step (5)^RAnd q is^IIn the expression, get q^RAnd q is^I；

(9) The complex frequency is found by:

ω^R+jω^I＝[q^R(k+1)+jq^I(k+1)]./[q^R(k)+jq^I(k)]

wherein, ω is^RAnd ω^IThe real and imaginary parts of the complex frequency.

(10) The damping ratio is found by:

wherein the content of the first and second substances,

and

are respectively omega^RAnd ω^II.e. the real and imaginary parts of the ith order complex frequency.