Incremental time-varying structure operational modal parameter real-time identification method and system

Abstract

The application relates to a kind of incremental time-varying structure operating modal parameter real-time identification method and system. By determining the modal shape matrix and modal response matrix at initial time i-1 according to multi-channel vibration response data and the time length of sliding window, it is judged whether the sum of current time i and the time length L of sliding window is less than or equal to the end time of sliding window. If yes, the modal coordinate response matrix at initial time i-1 is substituted into the calculation of the sliding window at the i-th time, and the incremental independent component analysis algorithm is used to determine the mixing matrix and source signal at the i-th time. According to the mixing matrix and source signal, the modal shape matrix and modal response matrix corresponding to the i-th time are determined, and the modal shape matrix and modal response matrix at the i+1-th time, the i+2-th time,..., the i+L-th time are determined in turn. The application can reduce the time of identifying time-varying structure operating modal parameters and improve the accuracy of the identification result.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of operational modal parameter identification, in particular to an incremental real-time identification method and system for time-varying structure operational modal parameters. BACKGROUND

[0002] A mode is a vibration characteristic of a structure itself. By identifying the parameters (such as modal natural frequency, mode shape, damping ratio, etc.) of each mode through experimental modal analysis, the dynamic characteristics of the structure can be understood, and then damage identification of the structure and fault detection of equipment can be performed. Modal parameter identification refers to exciting (i.e. inputting) a vibration system, obtaining input and output signal data of the system by measurement, processing and analyzing the input and output signal data, and identifying the structural modal parameters (frequency, damping ratio, mode shape, etc.) of the system according to different identification models and methods. Operational modal analysis (OMA) can identify the modal parameters of a structure only from vibration response signals collected by vibration sensors.

[0003] Most engineering structures in reality have time-varying characteristics. The physical characteristics (mass, stiffness, damping, etc.) of a structure change over time, and therefore the modal parameters of the structure also change over time. For example, a train passing over a bridge, launching a satellite, and rotating machinery all reflect the time-varying characteristics of a structure. Vibration response signals of a structure are usually not obtained all at once, but are sampled slowly over time, and therefore a method for identifying the operational modal parameters of a time-varying structure needs to be proposed according to the time-varying characteristics of the structure.

[0004] The sliding window method based on the "time freezing" theory is a method for identifying the operational modal parameters of a time-varying system, and the sliding window method has been applied to some algorithms. The traditional sliding window method has the disadvantages of high time and space complexity, low accuracy, and unsuitability for embedding into portable devices. SUMMARY

[0005] The present application aims to provide an incremental real-time identification method and system for time-varying structure operational modal parameters, which can reduce the time for identifying the operational modal parameters of a time-varying structure and improve the accuracy of the identification results.

[0006] To achieve the above-mentioned purpose, the present application provides the following solutions.

[0007] An incremental real-time identification method for time-varying structure operational modal parameters comprises the following steps:

[0008] Step 101: obtaining multi-channel non-stationary vibration response signals of a linear time-varying structure under environmental excitation within a set time length;

[0009] Step 102: obtaining a time length L of a sliding window of the vibration response data;

[0010] Step 103: determining a modal shape matrix and a modal response matrix at an initial time i-1 according to the multi-channel vibration response data and the time length of the sliding window;

[0011] Step 104: judging whether a sum of a current time i and the time length L of the sliding window is less than or equal to an end time of the sliding window, to obtain a judgment result;

[0012] Step 105: if the judgment result is yes, substituting the modal coordinate response matrix at the initial time i-1 into calculation of the sliding window at the i-th time, and respectively determining a mixing matrix and a source signal at the i-th time by using an incremental independent component analysis algorithm;

[0013] Step 106: determining a modal shape matrix corresponding to the i-th time according to the mixing matrix;

[0014] Step 107: determining a modal response matrix corresponding to the i-th time according to the source signal;

[0015] Step 108: sequentially determining modal shape matrices and modal response matrices at the i+1-th time, the i+2-th time, …, and the i+L-th time according to the modal shape matrix and the modal response matrix corresponding to the i-th time, wherein i+L=T END , T END is the end time of the sliding window;

[0016] Step 109: if the judgment result is no, directly outputting the modal shape matrix and the modal response matrix;

[0017] The modal shape matrix and the modal response matrix are incremental time-varying structure operational modal parameters.

[0018] Optionally, the modal coordinate response matrix is:

[0019]

[0020] wherein, is a modal shape matrix composed of modal shape vectors of n order modes of a structure in a window at the i-th time, is a modal response matrix composed of modal responses of n order modes of the structure in the window at the i-th time, is a modal shape matrix composed of modal shape vectors of n order modes of a structure in a window at the i-th time, is a modal response matrix composed of modal responses of n order modes of the structure in the window at the i-th time, L is a time length of the sliding window, and T END is the end time of the sliding window.

[0021] Optionally, the determining the modal response matrix corresponding to the i-th time instant according to the source signal comprises:

[0022] The modal response matrix corresponding to the i-th time instant is obtained by fast Fourier transform according to the source signal, and the modal response matrix represents an inherent frequency.

[0023] Optionally, the method further comprises quantitatively evaluating the accuracy of the operational modal parameter identification by using a modal confidence parameter formula.

[0024] Optionally, the modal confidence parameter formula is:

[0025]

[0026] wherein φ i is the i-th modal shape to be identified, is the i-th modal shape to be identified, and are the transpose of φ i and , respectively, is the inner product of two vectors, is the similarity degree of φ i and , and the value is closer to 1, the higher the accuracy of the operational modal parameter identification.

[0027] An incremental real-time identification system of time-varying structure operational modal parameters comprises:

[0028] A multi-channel non-stationary vibration response signal acquisition module is configured to acquire multi-channel non-stationary vibration response signals of a linear time-varying structure under environmental excitation within a set time length;

[0029] A sliding window time length acquisition module is configured to acquire the time length L of the sliding window of the vibration response data;

[0030] An initial time instant modal shape matrix and modal response matrix determination module is configured to determine the modal shape matrix and the modal response matrix at an initial time instant i-1 according to the multi-channel vibration response data and the time length of the sliding window, and the initial time instant i=1;

[0031] A judgment module is configured to judge whether the sum of the current time instant i and the time length L of the sliding window is less than or equal to the end time of the sliding window to obtain a judgment result;

[0032] The mixing matrix and source signal determination module is used to, when the judgment result is yes, substitute the modal coordinate response matrix at the initial time i-1 into the calculation of the sliding window at time i, and use the incremental independent component analysis algorithm to determine the mixing matrix and source signal at time i respectively.

[0033] The module for determining the mode shape matrix at time i is used to determine the mode shape matrix corresponding to time i based on the hybrid matrix.

[0034] The module for determining the modal response matrix at time i is used to determine the modal response matrix at time i based on the source signal.

[0035] The module for determining the modal shape matrix and modal response matrix at other times is used to determine the modal shape matrix and modal response matrix at times i+1, i+2, ..., i+L, sequentially, based on the modal shape matrix and modal response matrix corresponding to time i, where i+L = T. END T END This refers to the end time of the sliding window.

[0036] The modal shape matrix and modal response matrix output module is used to directly output the modal shape matrix and modal response matrix when the judgment result is negative.

[0037] Optionally, the modal coordinate response matrix is:

[0038]

[0039] in, For the mode shape vectors of the structure in the nth order mode within the window at time i. The resulting modal shape matrix, For the i-th time window, the structure is the nth order modal response. The modal response matrix is ​​formed, where L is the time length of the sliding window, and T is the modal response matrix. END This is the end time for the sliding window.

[0040] Optionally, the modal response matrix determination module at time i specifically includes:

[0041] The modal response matrix determination unit at time i is used to obtain the modal response matrix corresponding to time i by performing a fast Fourier transform on the source signal. The modal response matrix represents the natural frequency.

[0042] Optionally, it also includes an accuracy determination module for working modal parameter identification, used to quantitatively evaluate the accuracy of the working modal parameter identification using a modal confidence parameter formula.

[0043] Optionally, the modal confidence parameter formula is:

[0044]

[0045] wherein φ i is the i th identified modal shape, is the i th real modal shape, and are the transpose of φ i and respectively, is the inner product of two vectors, is the similarity degree of φ i and , and the value is closer to 1, the higher the accuracy of the operational modal parameter identification.

[0046] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0047] The present application provides a kind of incremental time-varying structure operational modal parameter real-time identification method, by obtaining linear time-varying structure in the set length under environmental excitation in multiple channel non-stationary vibration response signal;Obtain the time length L of sliding window window of vibration response data;According to the time length of multiple channel vibration response data and sliding window window, determine the modal shape matrix and modal response matrix at initial time i-1;Whether the sum of current time i and the time length L of sliding window window is less than or equal to the end time of sliding window window is judged, to obtain the judgment result;If the judgment result is yes, then the modal coordinate response matrix of initial time i-1 is substituted into the calculation of the i th sliding window window, and the incremental independent component analysis algorithm is used to determine the mixing matrix and source signal at the i th time respectively;According to the mixing matrix, the modal shape matrix corresponding to the i th time is determined;According to the source signal, the modal response matrix corresponding to the i th time is determined;According to the modal shape matrix and the modal response matrix corresponding to the i th time, the modal shape matrix and the modal response matrix of i+1 time, i+2 time …, i+L time are determined in turn.Compared with sliding window independent component analysis algorithm, all the present application uses incremental sliding window model parameter independent component analysis ICA algorithm to reduce time and memory overhead, with very low time and space complexity and very high identification accuracy, so as to be more easily embedded in portable device, realize online real-time monitoring device condition. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below only illustrate some of the embodiments of the present application, and all other embodiments obtained by those of ordinary skill in the art without any creative work on the premise of the accompanying drawings also belong to the protection scope of the present application.

[0049] Figure 1 Flow chart of the incremental time-varying structure working modal parameter real-time identification method of the present application;

[0050] Figure 2 Structure diagram of the incremental time-varying structure working modal parameter real-time identification system of the present application;

[0051] Figure 3 Schematic diagram of the linear time-varying structure for identifying working modal parameters;

[0052] Figure 4 Schematic diagram of the modeling of the linear time-varying structure in MATLAB / Simulink;

[0053] Figure 5 Natural frequency of the working modal parameter identification based on the sliding window independent component analysis algorithm in the 0-2000s time period;

[0054] Figure 6 Confidence coefficient MAC value change diagram of the working modal parameter identification based on the sliding window independent component analysis algorithm in the 0-2000s time period;

[0055] Figure 7 Natural frequency of the working modal parameter identification based on the incremental sliding window model parameter independent component analysis algorithm in the 0-1950s time period;

[0056] Figure 8 Confidence coefficient MAC value change diagram of the working modal parameter identification based on the incremental sliding window model parameter independent component analysis algorithm in the 0-1950s time period. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without any creative work on the premise of the accompanying drawings also belong to the protection scope of the present application.

[0058] The present application aims to provide an incremental time-varying structure working modal parameter real-time identification method and system, which can reduce the time for identifying the working modal parameters of the time-varying structure and improve the accuracy of the identification results.

[0059] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0060] The traditional independent component analysis algorithm is to preprocess the data (whitening), the data preprocessing uses the PCA algorithm to extract the characteristic value, performs characteristic value decomposition, so that each dimension of the vibration signal is linearly independent, and then through continuous iteration, the vibration signal has greater Gaussianity, so as to realize the separation of the source signal. However, this method must calculate the mixing matrix, i.e. the modal shape, each time. Therefore, in the sliding window independent component analysis calculation process, the mixing matrix needs to be calculated once each time, which leads to high time and space complexity of the sliding window independent component analysis method. The present application provides an incremental time-varying structure working modal parameter real-time identification method and system, which can reduce the complexity of the calculation time and improve the accuracy of the calculation result.

[0061] As shown in Figure 1 , the present application provides an incremental time-varying structure working modal parameter real-time identification method, which comprises the following steps:

[0062] Step 101: acquiring a multi-channel non-stationary vibration response signal of a linear time-varying structure under environmental excitation within a set time length Specifically, a plurality of vibration response sensors are arranged on the linear time-varying structure, and a non-stationary vibration signal matrix of the linear time-varying structure under environmental excitation within a period of time is acquired

[0063] The working modal parameters of the time-varying structure change with time, according to the dynamics theory of the structure, within a time length t∈[T BEGIN ,T END ], for an n-degree-of-freedom linear time-varying vibration structure system, the motion equation in the physical coordinate system is:

[0064]

[0065] wherein, and respectively represent the mass matrix, the damping matrix and the stiffness matrix varying with time t∈[T BEGIN ,T END ]. At the same time, they change with time due to the influence of the structure. represents an excitation vector of the external load, and respectively represent the acceleration response signal, the velocity response signal and the displacement response signal.

[0066] Based on the theory of "time freeze," a time-varying discrete multi-degree-of-freedom system can freeze within a very small time interval τ∈[t]. begin ,t end Within t∈[T], its mass, damping, and stiffness can be considered time-invariant; therefore, within the complete t∈[T]... BEGIN ,T END During the time period, the dynamic equations of the structure in the physical coordinate system can be expressed as:

[0067]

[0068] Here, S'(τ) represents the time-invariant structure at time t = τ, and S' represents a set of time-varying structures consisting of K linear time-invariant structures.

[0069] Step 102: Obtain the time length L of the sliding window for the vibration response data.

[0070] Initially, set the sliding window length L. The vibration response signal is for the first window with length L. Let L be the vibration response signal of the i-th window with length L. In each window... Within this system, the system is considered to be short-time invariant.

[0071] Step 103: Based on the multi-channel vibration response data and the time length of the sliding window, determine the modal shape matrix and modal response matrix at the initial time i-1, where the initial time i=1. The modal shape matrix and modal response matrix are incremental time-varying structural working modal parameters.

[0072] For a lightly damped structure, the response data can be divided into a finite number of parts. In the τ-th part, with a certain window length, the modal coordinate response of the linear system is:

[0073]

[0074] Where Φ(τ) and Q(τ,t) represent the mode shape matrix and modal response matrix of the τ-th window, respectively.

[0075] When the natural frequency of each modal of the structure is ω i When all are unequal, the mode shapes satisfy normalized orthogonality, and the modal responses are uncorrelated, as follows:

[0076]

[0077]

[0078] Suppose that in a very short time interval τ∈[t] begin ,t endWithin [t], the system is considered to be short-time invariant, that is, time t∈[t] BEGIN ,t END ] is divided into finite segments τ∈[t begin ,t end ], in each time period τ∈[t begin ,t end Within a given time period, the system is considered to be short-term and time-invariant. Therefore, the working modal parameter identification algorithm with a time-invariant structure is used to identify the working modal parameters for that time period. The window slides to the right to calculate the working modal parameters for the next time period, and so on. Finally, each time period is arranged in chronological order to form the modal parameters with a time-varying structure.

[0079]

[0080] Wherein, the window length of the response data is L, n represents the number of sensors, and T represents the sampling time.

[0081] For a structure with n degrees of freedom and relatively small damping, the time-varying vibration response data can be obtained by selecting a certain window length L = t. begin -t end Divided into a finite number of parts, in the i-th window (τ∈[t] begin ,t end The modal coordinate response of the linear system is:

[0082]

[0083] in, For the mode shape vectors of the structure in the nth order mode within the window at time i. The resulting modal shape matrix, For the i-th time window, the structure is the nth order modal response. The modal response matrix is ​​formed, where L is the time length of the sliding window, and T is the modal response matrix. END The end time of the sliding window. The natural frequency ω of each modal of the i-th window of the structure. i When all are unequal, the mode shapes satisfy normalized orthogonality, and the modal responses are uncorrelated, as follows:

[0084]

[0085]

[0086] Where, m l This represents the mass of the l-th modal.

[0087] Step 104: judging whether the sum of the current time i and the time length L of the sliding window is less than or equal to the end time of the sliding window, to obtain a judgment result;

[0088] Step 105: if the judgment result is yes, then the modal coordinate response matrix of the initial time i-1 is substituted into the calculation of the i-th time sliding window, and the mixed matrix and the source signal of the i-th time are determined respectively by using the incremental independent component analysis algorithm.

[0089] The data window of the i-th time is known And the mixed matrix of the i-1-th time An independent component analysis model is established in the data window to solve the model of the working modal parameters of the window.

[0090] The mixed matrix obtained by the independent component analysis algorithm Then The modal shape matrix corresponding to the i-th time Thus, the working modal parameters of the i-th time are obtained.

[0091] Step 106: determining the modal shape matrix corresponding to the i-th time according to the mixed matrix.

[0092] Step 107: determining the modal response matrix corresponding to the i-th time according to the source signal, specifically including:

[0093] The modal response matrix corresponding to the i-th time is obtained by fast Fourier transform according to the source signal, and the modal response matrix represents the natural frequency.

[0094] The source signal is the n-order modal response, The mixed matrix is the modal shape vector The source signal and the modal response matrix correspond to each other, and the mixed matrix and the modal shape matrix correspond to each other.

[0095] Step 108: determining the modal shape matrix and the modal response matrix of the i+1-th time, the i+2-th time, …, the i+L-th time in turn according to the modal shape matrix and the modal response matrix corresponding to the i-th time, wherein i+L=T END , T END is the end time of the sliding window.

[0096] The window slides to the right, the mixed matrix solved in the last time period is used as the input, and the working modal parameters in the next time period are calculated, until all the windows are identified, and each time period is arranged in time sequence, so as to track the time-varying characteristics of the structure, and form the modal parameters of the time-varying structure.

[0097] Step 109: if the result of the judgment is no, then directly output the modal shape matrix and the modal response matrix.

[0098] The working modal parameter identification method provided by the application can further comprise step 110: using a modal confidence parameter formula to quantitatively evaluate the accuracy of the working modal parameter identification. The modal confidence parameter formula is:

[0099]

[0100] wherein, φ i is the i th modal shape to be identified, is the i th real modal shape, and respectively are the transpose of φ i and , is the inner product of two vectors, is the similarity degree of φ i and , the value is closer to 1, the higher the accuracy of the working modal parameter identification.

[0101] The application provides a kind of incremental time-varying structure working modal parameter real-time identification method, which is a kind of linear structure working modal parameter identification method based on incremental sliding window model parameter independent component analysis algorithm, can carry out real-time online parameter identification to structure with time-varying characteristics, identifies the working modal parameter (modal shape, modal frequency) of system, real-time effective monitoring system dynamic change characteristics, can be used for equipment fault diagnosis, health monitoring and system structure analysis and optimization.The method can identify working modal parameter (only by measured response signal can identify the characteristics of system), and is proved from mathematical theory analysis and experiment, endows the method with physical explanation, compared with traditional test modal parameter identification technology needing to simultaneously measure excitation and response signal has greater advantage.The main idea of the method is, combined with short time invariant theory and incremental sliding window model parameter independent component analysis algorithm, by reusing the hybrid matrix of last time period, the working modal parameter (including the inherent frequency and modal shape of each order mode) of each time is estimated, then the working modal parameter obtained at each time is connected, thereby realizing time-varying linear structure working modal parameter identification.Compared with the linear time-varying structure working modal parameter identification method based on traditional sliding window independent component analysis, the application has lower time and space complexity, and is beneficial to embedded into portable hardware equipment.

[0102] The steps of the method are mainly to set the sliding window length, establish an independent component analysis model in each window, solve the structural operating modal parameter model by using the mixing matrix obtained in the previous window, and finally fit the modal parameters at each time, so that the time-varying characteristics of the structure are tracked, and the modal parameters of the linear time-varying structure are obtained. Compared with the sliding window ICA algorithm which needs to calculate all the at each time, the independent component analysis algorithm based on incremental sliding window model parameters can reduce the time and memory overhead, has very low time and space complexity, and is more easily embedded in portable devices to realize online real-time monitoring of equipment conditions. The independent component analysis algorithm based on incremental sliding window model parameters is a working modal parameter identification method, which only needs to obtain the non-stationary vibration response signal of the structure to achieve online real-time identification of the time-varying transient working modal parameters (instantaneous working modal shape and instantaneous working modal natural frequency) of the linear time-varying structure. Compared with the traditional sliding window modal parameter identification technology, it has great advantages.

[0103] The sliding window method was first proposed by Rubinger in 1974, and the sliding window method has been widely used in many fields. The present application combines the incremental modal parameter, the sliding window method and the independent component analysis algorithm and applies it to the linear time-varying structure operating modal parameter identification. By combining the short-time invariance theory and different independent component analysis algorithms, the independent component analysis algorithm based on incremental sliding window model parameters is used to estimate the working modal parameters (including the natural frequency and modal shape of each order mode) at each time in each window, and then the working modal parameters obtained at each time are connected to realize the working modal parameter identification of the time-varying linear structure. Compared with the working modal parameter identification method of the sliding window independent component analysis algorithm, the independent component analysis algorithm based on incremental sliding window model parameters has lower time and space complexity in solving the working modal parameters in each window. Therefore, the independent component analysis algorithm based on incremental sliding window model parameters not only improves the identification accuracy, but also reduces the time and space complexity of the algorithm, and better realizes real-time and effective detection of the working modal parameters of the structure. This helps to use the method in hardware embedding, equipment fault diagnosis, health monitoring and system structure analysis and optimization.

[0104] The present application also provides the following scheme:

[0105] As shown in Figure 2 An incremental time-varying structure operating modal parameter real-time identification system includes:

[0106] A multi-channel non-stationary vibration response signal acquisition module 201 is configured to acquire multi-channel non-stationary vibration response signals of a linear time-varying structure under environmental excitation within a set time length

[0107] The sliding window time length acquisition module 202 is used to acquire the time length L of the sliding window of the vibration response data.

[0108] The initial time modal shape matrix and modal response matrix determination module 203 is used to determine the modal shape matrix and modal response matrix at the initial time i-1 based on the multi-channel vibration response data and the time length of the sliding window, where the initial time i=1.

[0109] The judgment module 204 is used to determine whether the sum of the current time i and the time length L of the sliding window is less than or equal to the end time of the sliding window, and obtain the judgment result.

[0110] The mixing matrix and source signal determination module 205 is used to, when the judgment result is yes, substitute the modal coordinate response matrix at the initial time i-1 into the calculation of the sliding window at time i, and use the incremental independent component analysis algorithm to determine the mixing matrix and source signal at time i respectively.

[0111] The mode shape matrix determination module 206 at time i is used to determine the mode shape matrix corresponding to time i based on the hybrid matrix.

[0112] The modal response matrix determination module 207 at time i is used to determine the modal response matrix corresponding to time i based on the source signal.

[0113] The module 208 for determining the modal shape matrix and modal response matrix at other times is used to determine the modal shape matrix and modal response matrix at times i+1, i+2, ..., i+L, sequentially, based on the modal shape matrix and modal response matrix corresponding to time i, where i+L = T. END T END This is the end time for the sliding window.

[0114] The modal shape matrix and modal response matrix output module 209 is used to directly output the modal shape matrix and modal response matrix when the judgment result is negative.

[0115] The modal coordinate response matrix is:

[0116]

[0117] in, For the mode shape vectors of the structure in the nth order mode within the window at time i. The resulting modal shape matrix, For the i-th time window, the structure is the nth order modal response. The modal response matrix is composed, L is the time length of the sliding window, T END is the end time of the sliding window.

[0118] The i-th moment modal response matrix determination module 207 specifically comprises:

[0119] The i-th moment modal response matrix determination unit is configured to obtain the modal response matrix corresponding to the i-th moment by fast Fourier transform according to the source signal, and the modal response matrix represents the natural frequency.

[0120] The working modal parameter identification accuracy determination module is configured to quantitatively evaluate the accuracy of the working modal parameter identification by using a modal confidence parameter formula.

[0121] The modal confidence parameter formula is:

[0122]

[0123] Wherein, φ i is the i-th modal vibration mode to be identified, is the i-th modal vibration mode in reality, and are the transposition of φ i and , respectively. is the inner product of two vectors, is the similarity degree of φ i and . The value is closer to 1, and the higher the accuracy of the working modal parameter identification is.

[0124] The application further provides a device fault diagnosis method, which comprises:

[0125] Step 301: performing working modal parameter identification according to an incremental time-varying structure working modal parameter real-time identification method (steps 101-109) to obtain working modal parameters; the modal parameters comprise a modal vibration mode matrix and a modal response matrix.

[0126] Step 302: obtaining working modal parameters under normal operation of a measured device.

[0127] Step 303: comparing the working modal parameters with the working modal parameters under normal operation of the measured device to determine whether a fault occurs in the device and the location of the fault.

[0128] The application further provides a device fault diagnosis system, which comprises:

[0129] The working modal parameter identification module 401 is configured to identify working modal parameters according to the incremental time-varying structure working modal parameter real-time identification method, and obtain the working modal parameters; the modal parameters include a modal modal shape matrix and a modal response matrix.

[0130] The working modal parameter acquisition module 402 is configured to acquire working modal parameters of the device under test in a normal operation state.

[0131] The fault detection module 403 is configured to compare the working modal parameters with the working modal parameters of the device under test in the normal operation state, and determine whether a fault occurs in the device and a location of the fault.

[0132] In this embodiment, the incremental sliding window model parameter independent component analysis algorithm working modal parameter identification and application adopts a mass slow time-varying three-degree-of-freedom structure to simulate a time-varying structure, and a linear time-varying structure needs to be identified, as shown in FIG. 1. Figure 3

[0133] The sampling frequency is 40 Hz, the sampling interval is 0.025 s, the sampling time is t = 2000 s, the initial conditions of three-stage modal displacements of the system are all zero, the stiffness is set as k1(t) = k2(t) = k3(t) = 1000 N / m, 0 ≤ t ≤ 2000 s, the damping is set as c1(t) = c2(t) = c3(t) = 0.01 N.s / m, 0 ≤ t ≤ 2000 s, and the mass is set as m2(t) = m3(t) = 1 kg, 0 ≤ t ≤ 2000 s. The above parameters are time-invariant, and m1(t) is slow time-varying.

[0134] The dynamic equation of the mass time-varying at time t is represented as

[0135]

[0136] The excitation F1(t) received by m1(t) is a Gaussian white noise, and the slow time-varying data after 50 s is taken for research. Figure 4 The modeling schematic diagram of the linear time-varying structure in MATLAB / Simulink is shown in FIG. 2, and the linear time-varying structure herein refers to a three-degree-of-freedom spring oscillator system.

[0137] The response data obtained after the white noise excitation is applied to the three-degree-of-freedom structure is identified by using the incremental sliding window model parameter independent component analysis algorithm working modal parameter identification method, and the identified modal parameters are compared with the working modal parameters identified by the sliding window-based independent component analysis algorithm and the real modal parameters.

[0138] Figure 5 ​The figure of the change of the confidence coefficient MAC value of the working modal parameter identification based on the sliding window independent component analysis algorithm in the time period of 0-2000s. Figure 6 The figure of the change of the confidence coefficient MAC value of the working modal parameter identification based on the sliding window independent component analysis algorithm in the time period of 0-2000s. Figure 7 The figure of the change of the confidence coefficient MAC value of the working modal parameter identification based on the sliding window independent component analysis algorithm in the time period of 0-2000s. Figure 8 The figure of the change of the confidence coefficient MAC value of the working modal parameter identification based on the sliding window independent component analysis algorithm in the time period of 0-2000s. Figures 5-8 It can be known that the application can reduce the time for identifying the working modal parameter of the time-varying structure and improve the accuracy of the identification result.

[0139] Table 1 is the sliding window independent component analysis method, and Table 2 is the incremental sliding window independent component analysis method.

[0140] Table 1

[0141]

[0142] Table 2

[0143]

[0144]

[0145] In the specification, each embodiment focuses on the difference from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the system disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the relevant part can be referred to the method part.

[0146] The principles and implementation manners of the application are described by using specific examples in the specification, and the above embodiment description is only used to help understand the method of the application and its core idea; meanwhile, for the general technical personnel in the field, the specific implementation manner and application range of the application will have changes according to the idea of the application. In conclusion, the content of the specification should not be understood as the limitation of the application.

Claims

1. A real-time incremental time-varying structural operational modal parameter identification method, characterized in that, The method comprises the steps of: Step 101: acquiring a multi-channel non-stationary vibration response signal of a linear time-varying structure under environmental excitation within a set time length; Step 102: acquiring a time length L of a sliding window of the multi-channel non-stationary vibration response signal; Step 103: determining a modal shape matrix and a modal response matrix at the initial time instant t0according to the multi-channel non-stationary vibration response signal and the time length of the sliding window. i -1; and ; Step 104: judging whether the sum of the current time and the time length L of the sliding window is less than or equal to the end time of the sliding window, to obtain a judgment result. i and the sliding window end time, to obtain a judgment result. Step 105: If the result of the judgment is yes, the initial time i is substituted into the calculation of the modal coordinate response matrix of the first i time sliding window, and the incremental independent component analysis algorithm is used to determine the mixing matrix and the source signal of the first i time, respectively. Step 106: determining the modal shape matrix corresponding to the time instant according to the mixing matrix i Step 106: determining the modal shape matrix corresponding to the time instant according to the mixing matrix Step 107: determining the modal response matrix corresponding to the time instant according to the source signal i Step 107: determining the modal response matrix corresponding to the time instant according to the source signal Step 108: according to the modal shape matrix and the modal response matrix corresponding to the first time i , the modal shape matrix and the modal response matrix of the first time i , the first time i , the first time , the first time , is the end time of the sliding window Step 109: if the judgment result is no, directly outputting the modal shape matrix and the modal response matrix; The modal shape matrix and the modal response matrix are incremental time-varying structure operational modal parameters; The modal coordinate response matrix is: wherein, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the length of the sliding window, T END is the end time of the sliding window.

2. The incremental time-varying structural operational modal parameter real-time identification method according to claim 1, characterized in that, The step of determining the first based on the source signal. i The modal response matrix at time t, specifically includes: According to the source signal, a modal response matrix corresponding to the time moment is obtained by fast Fourier transform, and the modal response matrix represents the natural frequency. i According to the source signal, a modal response matrix corresponding to the time moment is obtained by fast Fourier transform, and the modal response matrix represents the natural frequency.

3. The incremental time-varying structural operational modal parameter real-time identification method according to claim 1, characterized in that, The method further comprises the step of quantitatively evaluating the accuracy of the operational modal parameter identification by using a modal confidence parameter formula.

4. The incremental time-varying structural operational modal parameter real-time identification method according to claim 3, characterized in that, The modal confidence parameter formula is: ； wherein, is the identified first i mode shape, is the true first i mode shape, and are respectively and the transpose of, is the inner product of two vectors, is and the degree of similarity, the value of which is closer to 1, the higher the accuracy of the operational modal parameter identification.

5. A system for real-time identification of incremental time-varying structural operational modal parameters according to any one of claims 1-4, characterized in that, The method comprises the steps of: A multi-channel non-stationary vibration response signal acquisition module is configured to acquire a multi-channel non-stationary vibration response signal of a linear time-varying structure under environmental excitation within a set time length; A sliding window time length acquisition module is configured to acquire a time length L of a sliding window of the multi-channel non-stationary vibration response signal; The initial time modal shape matrix and modal response matrix determination module is configured to determine the modal shape matrix and the modal response matrix at the initial time t0 according to the multi-channel non-stationary vibration response signal and the time length of the sliding window. i -1 of the modal shape matrix and the modal response matrix at the initial time t0. ; A judging module is configured to judge whether the sum of the current time and the time length L of the sliding window is less than or equal to the end time of the sliding window, and obtain a judgment result. i A judging module is configured to judge whether the sum of the current time and the time length L of the sliding window is less than or equal to the end time of the sliding window, and obtain a judgment result The hybrid matrix and source signal determination module is used to determine the initial time when the judgment result is yes. i Substituting the modal response matrix of -1 into the first... i In the calculation of the sliding window at each time step, an incremental independent component analysis algorithm is used to determine the time step. i The mixing matrix and source signal at each time step; The method comprises the following steps: i A time modal shape matrix determination module is configured to determine a modal shape matrix corresponding to the time moment according to the mixed matrix. i A time modal shape matrix determination module is configured to determine a modal shape matrix corresponding to the time moment according to the mixed matrix. The method comprises the following steps: i A time moment modal response matrix determination module is configured to determine a modal response matrix corresponding to the time moment according to the source signal. i A time moment modal response matrix determination module is configured to determine a modal response matrix corresponding to the time moment according to the source signal. Other time modal shape matrix and modal response matrix determination module, for determining the first i The modal shape matrix and the modal response matrix corresponding to the time are determined in turn i The modal shape matrix and the modal response matrix of the first i The modal shape matrix and the modal response matrix of the first The modal shape matrix and the modal response matrix of the first ; A modal shape matrix and a modal response matrix output module is configured to directly output a modal shape matrix and a modal response matrix when the judgment result is no; The modal coordinate response matrix is: wherein, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the modal shape matrix of the structure at the time window, is the length of the sliding window, T END is the end time of the sliding window.

6. The incremental time-varying structural operational modal parameter real-time identification system according to claim 5, characterized in that, The first i The time modal response matrix determination module specifically comprises: The first i The time modal response matrix determination unit is configured to obtain a modal response matrix corresponding to the first time according to the source signal through fast Fourier transform, and the modal response matrix represents an inherent frequency. i The time modal response matrix determination unit is configured to obtain a modal response matrix corresponding to the first time according to the source signal through fast Fourier transform, and the modal response matrix represents an inherent frequency.

7. The incremental time-varying structural operational modal parameter real-time identification system of claim 5, wherein, The method further comprises an accuracy determination module of operational modal parameter identification, which is configured to quantitatively evaluate the accuracy of the operational modal parameter identification by using a modal confidence parameter formula.

8. The incremental time-varying structural operational modal parameter real-time identification system according to claim 7, characterized in that, The modal confidence parameter formula is: ； in, For the identified first i One mode shape, For the true first i One mode shape, and They are respectively and transpose, Let be the dot product of two vectors. for and The degree of similarity, The closer the value is to 1, the higher the accuracy of the working modal parameter identification.

Images