An automatic identification method for single-mode and multi-mode vortex-induced vibration of a stay cable

Abstract

The present invention discloses a method for automatically identifying single-mode and multi-mode vortex-induced vibrations of a cable, which includes two stages: the first stage is the automatic screening of potential vortex-induced vibrations, which takes the frequency domain characteristics of vortex-induced vibrations as the theoretical basis, and preliminarily determines the possibility of vortex-induced vibrations according to the number of power spectrum peaks of the cable vibration signal; the second stage is the automatic confirmation of vortex-induced vibrations, which takes the time domain characteristics of vortex-induced vibrations as the theoretical basis, and determines whether vortex-induced vibrations actually occur by the stability of the amplitude of each modal vibration signal, thereby realizing automatic identification of vortex-induced vibrations. The present invention solves the shortcoming that the existing vortex-induced vibration identification method can only identify single-mode vortex vibrations, and improves the vortex-induced vibration identification from only being able to identify single-mode vortex vibrations to being able to identify vortex-induced vibrations involving any order of modes. At the same time, the method automates the entire identification process, which is suitable for long-term automatic identification and online monitoring of structures.

Description

Technical Field

[0001] The invention belongs to the field of bridge health monitoring, and relates to a method for automatically identifying vortex-induced vibration of a cable under environmental excitation, and specifically to a unified method for automatically identifying single-mode and multi-mode vortex-induced vibration of a cable. Background Art

[0002] Vortex-induced vibration is a regular vibration. Vortex-induced vibration of cable-stayed cables can cause cable fatigue and other problems. Therefore, identifying and tracking vortex-induced vibration of cable-stayed cables is an important part of cable-stayed cable monitoring. The traditional view is that vortex-induced vibration is a single-mode vibration, that is, a regular vibration controlled by one mode. However, as the span of cable-stayed bridges becomes larger and larger and the cables become longer and longer, researchers have found that super-long cables often experience multi-mode vortex-induced vibration during service, that is, regular vibration involving multiple main vibration modes.

[0003] The existing vortex-induced vibration identification method can only identify single-mode vortex-induced vibration. Its mechanical principle is that when a single-mode vortex-induced vibration occurs in a cable, its vibration response is controlled by the single-mode vibration. The frequency domain characteristic of the single-mode vortex-induced vibration is that its power spectrum density diagram is a single peak, and the time domain characteristic is that the amplitude of its time domain vibration signal is stable, such as Figure 1 .

[0004] However, the above characteristics are not applicable in multi-modal vortex-induced vibration. Figure 2 The frequency domain and time domain characteristics of typical dual-mode vortex-induced vibration (i.e., regular vibration controlled by two modes) are given. Figure 1 In comparison, the frequency domain and time domain characteristics of dual-mode vortex-induced vibration are quite different from those of single-mode. Therefore, the existing single-mode vortex-induced vibration identification method is unable to identify dual (multi)-mode vortex-induced vibration.

[0005] Single-mode vortex-induced vibration and multi-mode vortex-induced vibration may occur during the service of the cable. Therefore, a unified identification method that can identify both single-mode vortex-induced vibration and multi-mode vortex-induced vibration is needed. At the same time, in order to achieve continuous, online identification and online monitoring, the identification method needs to be automated. Summary of the invention

[0006] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a method for automatically identifying single- and multi-modal vortex-induced vibrations of inclined cables, so as to overcome the shortcomings of the existing methods for identifying vortex-induced vibrations of inclined cables that cannot identify multi-modal vortex-induced vibrations. The method has the characteristics of full process automation and the ability to identify vortex-induced vibrations of any order.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions to achieve the above problems:

[0008] A method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable comprises the following steps:

[0009] Step 1: Perform spectrum analysis on the vibration acceleration signal of the cable and calculate the energy similarity ratio vector;

[0010] Step 2: compare the nth index value in the energy similarity ratio vector with the threshold value. If the nth index value is greater than the threshold value, it is considered that potential nth-order vortex-induced vibration occurs, otherwise it is considered to be random vibration.

[0011] Step 3: Determine the instantaneous frequency, use the adaptive frequency modulation mode tracking algorithm to perform mode decomposition, and automatically extract the single mode signal;

[0012] Step 4, performing Hilbert transform on each single-mode signal obtained by decomposition and establishing an analytical signal, and calculating an amplitude similarity ratio vector;

[0013] Step 5: Compare the nth index value in the amplitude similarity ratio vector with the amplitude similarity ratio threshold. If the nth index is greater than the threshold, it is considered that the nth-order vortex-induced vibration actually occurs.

[0014] The present invention also includes the following technical features:

[0015] Specifically, the step one includes: first, performing spectrum analysis on the vibration acceleration signal of the inclined cable, using the peak extraction method to automatically extract the peaks and corresponding frequencies in the spectrum diagram, and arranging these peaks in descending order according to the power spectrum density values; then, calculating the energy similarity ratio of the two adjacent spectrum peaks after the arrangement, and forming an energy similarity ratio vector, that is, automatically forming a frequency domain feature index vector.

[0016] Specifically, in step 1, the peak values ​​are arranged in descending order according to the power spectrum density values ​​as follows:

[0017] A=[A1,A2,…,A i ,…,A n ] (1)

[0018] In the formula, A1>A2>A3>…>A n ; A i is the ith power spectrum density value after arranging in descending order of power spectrum density, and n is the number of peaks.

[0019] Specifically, in step 1, the energy similarity ratio vector calculation formula is:

[0020]

[0021] In the formula, is the energy similarity ratio corresponding to the i-th peak.

[0022] Specifically, the step three includes:

[0023] Step 3.1, signal time-frequency transformation: use synchronous squeezed short-time Fourier transform to transform the signal into a time-frequency signal to calculate the frequency spectrum characteristics of the vibration signal;

[0024] Step 3.2, spectrum result extraction: extract the spectrum result at a certain moment from the time-frequency result;

[0025] Step 3.3, frequency sorting: Use the peak extraction method to automatically extract all peaks in the frequency domain and their corresponding frequencies, sort the frequencies by peak value, extract the frequency corresponding to the maximum peak as the instantaneous frequency at this moment, and based on the same process, automatically obtain the instantaneous frequency of the mode at other times;

[0026] Step 3.4, single-mode signal calculation: After obtaining the initial instantaneous frequency, the adaptive frequency modulation mode tracking algorithm is used to automatically extract the single-mode signal corresponding to this order frequency, and the process is repeated until all single-mode signals are obtained.

[0027] Specifically, in step 3.1, the expression of the synchronous squeezed short-time Fourier transform is:

[0028]

[0029] Where FSST(t,ξ) represents the frequency value after the synchronous compression short-time Fourier transform at time t and the frequency ξ after synchronous compression; δ is the synchronous compression operator; ξ is the time-frequency spectrum frequency after synchronous compression; STFT(t,ω) is the short-time Fourier transform; ω0(t,ω) is the instantaneous frequency corresponding to time t.

[0030] Specifically, in step 4, the signal z is analyzed i (t):

[0031]

[0032] In the formula, z i (t) is the analytical signal of the i-th eigenmode component; s i (t) is the original signal of the i-th eigenmode component; is the signal after Hilbert transformation of the i-th eigenmode component; j is the imaginary unit.

[0033] Specifically, in step 4, each analytical signal z i (t) is projected onto the complex domain. The instantaneous modulus of the analytical signal represents the instantaneous amplitude of the signal. The time domain feature vector is automatically established by calculating the similarity ratio of the instantaneous amplitude at each moment:

[0034]

[0035] In the formula, A i(t) is all instantaneous amplitudes of the i-th single-mode signal in this period; t is the corresponding time in this period; is the amplitude similarity ratio corresponding to the i-th unimodal signal; m=n-1 is the number of decomposed unimodal signals.

[0036] Compared with the prior art, the present invention has the following technical effects:

[0037] The unified method for automatic identification of single and multi-modal vortex-induced vibrations proposed in the present invention can realize automatic identification of vortex-induced vibrations controlled by any order of modal control. By sorting the spectrum peaks of the acceleration signal and calculating the energy similarity ratio index vector, the number of modes involved in the vortex-induced vibration can be determined. The adaptive frequency modulation modal tracking algorithm is used to realize automatic decomposition of multi-modal signals. The automatic identification of multi-modal vortex-induced vibrations is realized through Hilbert transform and time domain characteristic index vector calculation. The automation of the entire identification process is realized through the automatic extraction of peaks, automatic determination of the initial frequency of signal decomposition, and automatic signal decomposition.

[0038] The present invention can realize the vortex-induced vibration identification of any number of modal controls of the inclined cable, and has wider applicability than the traditional method. The present invention realizes automation of the whole process of signal decomposition, index calculation and vortex-induced vibration identification, and realizes automatic identification of vortex-induced vibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Signal characteristics of single-mode vortex-induced vibration (a) frequency domain characteristics of single-mode vortex-induced vibration (b) time domain characteristics of single-mode vortex-induced vibration.

[0040] Figure 2 The signal characteristics of dual-mode vortex-induced vibration (a) dual-mode vortex-induced vibration frequency domain characteristics (b) dual-mode vortex-induced vibration time domain characteristics.

[0041] Figure 3 Identify the first stage for vortex vibration.

[0042] Figure 4 Automatic determination of initial frequency and signal decomposition process.

[0043] Figure 5 Identify the second stage for vortex vibration.

[0044] Figure 6 It is the whole process of automatic identification of single and multi-mode vortex-induced vibration of inclined cables.

[0045] Figure 7 It is the time history curve of the acceleration data of the inclined cable during this period.

[0046] Figure 8 This is a diagram of vortex vibration results of signal recognition during this period using the present invention.

[0047] Fig. 9This is the vibration signal characteristic diagram at time 1 (2022-07-04 14:02).

[0048] Fig.10 This is the vibration signal characteristic diagram at time 2 (2022-07-04 14:05).

[0049] Fig.11 This is the vibration signal characteristic diagram at time 3 (2022-07-04 14:09).

[0050] Fig.12 This is the vibration signal characteristic diagram at time 4 (2022-07-04 14:15).

[0051] Fig.13 This is the vibration signal characteristic diagram at time 5 (2022-07-04 14:25).

[0052] Fig.14 This is the vibration signal characteristic diagram at time 6 (2022-07-04 14:37). DETAILED DESCRIPTION

[0053] The present invention provides a method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable, comprising:

[0054] The first stage of vortex-induced vibration identification:

[0055] Data preprocessing: First, use low-pass filtering to filter the original signal to remove high-frequency signals and reduce the impact of high-frequency noise on subsequent recognition.

[0056] Step 1: Automatically establish a frequency domain characteristic index vector: First, perform spectrum analysis on the vibration acceleration signal of the inclined cable, use the peak extraction method to automatically extract the peaks and corresponding frequencies in the spectrum diagram, and arrange these peaks in descending order according to the power spectrum density value; then, calculate the energy similarity ratio of the two adjacent spectrum peaks after arrangement, and form an energy similarity ratio vector, that is, automatically form a frequency domain characteristic index vector;

[0057] The peaks are arranged in descending order by power spectral density value:

[0058] A=[A1,A2,…,A i ,…,A n ] (1)

[0059] In the formula, A1>A2>A3>…>A n ; A i is the ith power spectrum density value after arranging in descending order of power spectrum density, and n is the number of peaks.

[0060] The energy similarity ratio vector calculation formula is:

[0061]

[0062] In the formula, is the energy similarity ratio corresponding to the i-th peak.

[0063] Step 2: Automatically screen potential vortex-induced vibrations: According to the n-th order vortex-induced vibration to be identified, compare the n-th index value in the energy similarity ratio vector with the threshold value. If the n-th index value is greater than the threshold value, it is considered that potential n-th order vortex-induced vibration occurs, otherwise it is considered to be random vibration.

[0064] If E(i)<0.9 for all i=1,2,...,n-1, the cable is judged to be in an ambient vibration state; if E(n)=max(E) and E(n)≥0.9, the cable may be subjected to vortex-induced vibration caused by the combined action of n-order modes.

[0065] The flow chart of the first stage of vortex vibration identification is shown in Figure 3 ,This process is fully automated from the index calculation to the preliminary judgment of ,vortex-induced vibration.

[0066] The second stage of vortex vibration identification:

[0067] In the second stage, time domain indicators are used to further determine whether the inclined cable actually undergoes n-th order vortex-induced vibration.

[0068] Step 3: determine the initial instantaneous frequency, use the adaptive frequency modulation mode tracking algorithm to perform mode decomposition, and automatically extract the single mode signal;

[0069] Step 3.1, signal time-frequency transformation: Use synchronous squeezed short-time Fourier transform to perform time-frequency transformation on the signal to calculate the spectral characteristics of the vibration signal; improve the energy concentration of the frequency domain characteristic signal component to facilitate better extraction of the initial frequency.

[0070] The expression of the synchronous squeezed short-time Fourier transform is:

[0071]

[0072] Where FSST(t,ξ) represents the frequency value after the synchronous compression short-time Fourier transform at time t and the frequency ξ after synchronous compression; δ is the synchronous compression operator; ξ is the time-frequency spectrum frequency after synchronous compression; STFT(t,ω) is the short-time Fourier transform, that is, Where x(t) is the time-varying signal, g(t) is the window function, * represents conjugate, t is the corresponding time, ω is the signal angular frequency, τ is the window center position, and j is the complex unit; Where ω0(t,ω) is the instantaneous frequency corresponding to time t; is the partial derivative of the short-time Fourier transform with respect to time t; j is a complex unit.

[0073] Step 3.2, spectrum result extraction: extract the spectrum result at a certain moment from the time-frequency result;

[0074] Step 3.3, frequency sorting: Use the peak extraction method to automatically extract all peaks in the frequency domain and their corresponding frequencies, sort the frequencies according to the peak value (power spectrum energy level), extract the frequency corresponding to the maximum peak (highest energy value) as the instantaneous frequency at this moment, and based on the same process, automatically obtain the instantaneous frequency of the mode at other times;

[0075] Step 3.4, single-mode signal calculation: After obtaining the initial instantaneous frequency, the single-mode vibration signal corresponding to the order frequency is automatically extracted using the adaptive frequency modulation mode tracking algorithm, and the process is repeated until all single-mode signals are obtained;

[0076] The above process flow is as follows Figure 4 .

[0077] Step 4, perform Hilbert transform on each single-mode signal obtained by decomposition and establish an analytical signal, and finally calculate the amplitude similarity ratio vector;

[0078] Analyze signal z i (t):

[0079]

[0080] In the formula, z i (t) is the analytical signal of the i-th eigenmode component; s i (t) is the original signal of the i-th eigenmode component; is the signal after Hilbert transformation of the i-th eigenmode component; j is the imaginary unit.

[0081] Each analytical signal z i (t) is projected onto the complex domain. The instantaneous modulus of the analytical signal represents the instantaneous amplitude of the signal. The time domain feature vector is automatically established by calculating the similarity ratio of the instantaneous amplitude at each moment:

[0082]

[0083] In the formula, A i (t) is all instantaneous amplitudes of the i-th single-mode signal in this period; t is the corresponding time in this period; is the amplitude similarity ratio corresponding to the i-th unimodal signal; m=n-1 is the number of decomposed unimodal signals.

[0084] Step 5: Compare the n-th index value in the amplitude similarity ratio vector with the amplitude similarity ratio threshold. If the n-th index is greater than the threshold, it is considered that the n-th order vortex-induced vibration actually occurs.

[0085] Taking the acceleration data during the vortex-induced vibration and random vibration of the stay cable as samples, the threshold r of the time-domain index vector S is statistically calculated. The threshold r can be custom-set according to the strictness of identification. The larger the threshold r is set, the more accurate the identification result is, but the fewer the number of identified vortex vibrations will be. On the contrary, the smaller the threshold r is set, the less accurate the identification result is, but the number of identified vortex vibrations will increase. If S(i) < r for all i = 1, 2,..., n - 1, it is determined that the stay cable is in the ambient vibration state. If S(n) ≥ r, it is determined that the stay cable has vortex-induced vibration with the n-th order mode interaction.

[0086] The flow chart of the second stage of vortex vibration identification is shown in Figure 5 .

[0087] Based on the above two-stage identification, a unified method for automatic identification of single and multi-modal vortex-induced vibration of stay cables is established. The whole process is as shown in Figure 6 .

[0088] The following gives specific embodiments of the present invention. It should be noted that the present invention is not limited to the following specific embodiments, and all equivalent transformations made on the basis of the technical solutions of this application fall within the protection scope of the present invention.

[0089] Embodiment:

[0090] Taking a certain cross-sea bridge as an example to illustrate the practicability of the present invention. The bridge is a double-cable-plane harp-shaped arrangement of a double-tower steel-concrete composite beam stay cable bridge. The transverse distance between the two cable planes is 1.0 m, and the bridge span arrangement is 85 + 127.5 + 268 + 127.5 + 85 = 693 m. A total of 16 acceleration sensors are installed on the stay cables of its health monitoring system, and the sampling frequency is 50 Hz. Single-modal vortex vibration and double-modal vortex vibration are the most likely vortex vibrations to occur in actual stay cables. Therefore, only single-modal vortex vibration and double-modal vortex vibration are identified in this case, and other multi-modal vortex vibrations can refer to the process of this case. The threshold r of the amplitude similarity index is set to 0.5.

[0091] Select the acceleration data of stay cable No. 9 from July 4, 2022, [14:01, 14:41) for 40 minutes as the calculation sample to verify the accuracy of the identification result of the stay cable vortex vibration automatic identification algorithm.

[0092] Figure 7 It represents the time history curve graph of the stay cable acceleration data during this period.

[0093] Figure 8Indicates the result of vortex vibration identification of the signal in this period using this method. The identification result in the figure is 1 represents single-mode vortex-induced vibration, 2 represents dual-mode vortex-induced vibration, and the marked moments 1 to 6 in the figure correspond to the development and stability stage of single-mode vortex-induced vibration, the development, stability and disappearance stage of dual-mode vortex-induced vibration, and the disappearance stage of single-mode vortex-induced vibration.

[0094] Fig. 9 Represents the vibration signal characteristics at time 1 (2022-07-04 14:02), Fig. 9 (a) Original data in time domain (b) Frequency domain (c) 2nd modal component in time domain (d) 2nd modal component in complex domain (e) 4th modal component in time domain (f) 4th modal component in complex domain; It can be seen that this is the development stage of single-mode vortex-induced vibration. From the perspective of frequency domain, it can be found that the power spectrum density diagram is a single peak at this moment. From the perspective of time domain and complex domain, the signal amplitude increases steadily at this moment, and the projection of the component in the complex domain presents an obvious hollow circle, with an amplitude similarity ratio of 0.83. The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this method accurately identifies the development stage of single-mode vortex-induced vibration of the inclined cable.

[0095] Fig.10 Represents the vibration signal characteristics at time 2 (2022-07-04 14:05), Fig.10 (a) Original data in time domain (b) Frequency domain (c) 2nd modal component in time domain (d) 2nd modal component in complex domain (e) 4th modal component in time domain (f) 4th modal component in complex domain; It can be seen that this is the stable stage of single-mode vortex-induced vibration. From the perspective of frequency domain, it can be found that the power spectrum density diagram is a single peak at this moment. From the perspective of time domain and complex domain, the signal amplitude is stable at this moment, and the projection of the component in the complex domain presents an obvious hollow circle, and its amplitude similarity ratio is 0.86; The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this method accurately identifies the stable stage of single-mode vortex-induced vibration of the inclined cable.

[0096] Fig.11 Represents the vibration signal characteristics at time 3 (2022-07-04 14:09), Fig.11 (a) Original data in time domain (b) Frequency domain (c) 2nd-order modal component in time domain (d) 2nd-order modal component in complex domain (e) 4th-order modal component in time domain (f) 4th-order modal component in complex domain; It can be seen that this is the stage of development of dual-mode vortex-induced vibration. From the perspective of frequency domain, it can be found that two peaks appear in the power spectrum corresponding to the 2nd and 4th-order modes respectively. From the perspective of time domain and complex domain, the amplitude of the 2nd-order modal component is stable and remains unchanged, while the amplitude of the 4th-order modal component is steadily increasing. The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this method accurately identifies the development stage of dual-mode vortex-induced vibration of the cable.

[0097] Fig.12 Represents the vibration signal characteristics at time 4 (2022-07-04 14:15), Fig.12 (a) original data time domain (b) frequency domain (c) 2nd-order modal component time domain (d) 2nd-order modal component complex domain (e) 4th-order modal component time domain (f) 4th-order modal component complex domain; it can be seen that this is the stable stage of dual-mode vortex-induced vibration. From the perspective of the frequency domain, it can be found that there are two significant peaks in the power spectrum and the peak values ​​are close. From the perspective of the time domain and complex domain, it can be found that the amplitudes of the 2nd and 4th-order modal components tend to be stable, and the amplitude similarity ratios of the two are 0.90 and 0.72 respectively. The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this patent accurately identifies the stable stage of dual-mode vortex-induced vibration of the inclined cable.

[0098] Fig.13 Represents the vibration signal characteristics at time 5 (2022-07-04 14:25), Fig.13 (a) original data time domain (b) frequency domain (c) 2nd-order modal component time domain (d) 2nd-order modal component complex domain (e) 4th-order modal component time domain (f) 4th-order modal component complex domain; it can be seen that this is the decay stage of the dual-modal vortex-induced vibration. From the perspective of the frequency domain, it can be found that there are still two peaks in the power spectrum, but the peak corresponding to the 4th-order mode is significantly reduced. From the perspective of the time domain and complex domain, the amplitude of the 4th-order modal component gradually decreases during this time period, while the amplitude of the 2nd-order modal component remains stable. The signal gradually changes from dual-modal vortex-induced vibration to single-modal vortex-induced vibration. The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this patent accurately identifies the decay stage of the dual-modal vortex-induced vibration of the inclined cable.

[0099] Fig.14 Represents the vibration signal characteristics at time 6 (2022-07-04 14:37), Fig.14 (a) original data time domain (b) frequency domain (c) 2nd-order modal component time domain (d) 2nd-order modal component complex domain (e) 4th-order modal component time domain (f) 4th-order modal component complex domain; it can be seen that this is the decay stage of single-mode vortex-induced vibration. In the frequency domain, in addition to the main peak corresponding to the 2nd-order mode, peaks at other modal positions can also be observed. In the time domain, the amplitude of the 2nd-order modal component gradually decreases, and its various modes gradually participate in the vibration, and the vibration of the cable gradually turns into random vibration. The time-frequency domain characteristics of the vibration signal are consistent with the identification results, indicating that this patent accurately identifies the decay stage of single-mode vortex-induced vibration of the cable.

[0100] The above cases prove that this patent can automatically identify the process of single-mode and dual-mode vortex-induced vibration of the cable from development to stability and finally to disappearance. The above process only demonstrates the recognition of single-mode and dual-mode vortex-induced vibration by this method, and other multi-mode vortex-induced vibrations can be identified in the same way.

[0101] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, a variety of simple modifications can be made to the technical solution of the present invention, and these simple modifications all belong to the protection scope of the present invention.

[0102] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

[0103] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable, characterized in that: The following steps are involved: Step 1: Perform spectrum analysis on the vibration acceleration signal of the cable and calculate the energy similarity ratio vector; Step 2: compare the nth index value in the energy similarity ratio vector with the threshold value. If the nth index value is greater than the threshold value, it is considered that potential nth-order vortex-induced vibration occurs, otherwise it is considered to be random vibration. Step 3: Determine the instantaneous frequency, use the adaptive frequency modulation mode tracking algorithm to perform mode decomposition, and automatically extract the single mode signal; Step 4, performing Hilbert transform on each single-mode signal obtained by decomposition and establishing an analytical signal, and calculating an amplitude similarity ratio vector; Step 5: Compare the nth index value in the amplitude similarity ratio vector with the amplitude similarity ratio threshold. If the nth index is greater than the threshold, it is considered that the nth-order vortex-induced vibration actually occurs.

2. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 1, characterized in that: The step one comprises: first, performing spectrum analysis on the vibration acceleration signal of the inclined cable, using the peak extraction method to automatically extract the peaks and corresponding frequencies in the spectrum diagram, and arranging these peaks in descending order according to the power spectrum density values; then, calculating the energy similarity ratio of two adjacent spectrum peaks after the arrangement, and forming an energy similarity ratio vector, that is, automatically forming a frequency domain feature index vector.

3. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 2, characterized in that: In step 1, the peak values ​​are arranged in descending order according to the power spectrum density values ​​as follows: A=[A1,A2,…,A i ,…,A n ] (1) In the formula, A1>A2>A3>…>A n ; A i is the ith power spectrum density value after arranging in descending order of power spectrum density, and n is the number of peaks.

4. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 3, characterized in that: In the step 1, the energy similarity ratio vector calculation formula is: In the formula, is the energy similarity ratio corresponding to the i-th peak.

5. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 1, characterized in that: The step three comprises: Step 3.1, signal time-frequency transformation: use synchronous squeezed short-time Fourier transform to transform the signal into a time-frequency signal to calculate the frequency spectrum characteristics of the vibration signal; Step 3.2, spectrum result extraction: extract the spectrum result at a certain moment from the time-frequency result; Step 3.3, frequency sorting: Use the peak extraction method to automatically extract all peaks in the frequency domain and their corresponding frequencies, sort the frequencies by peak value, extract the frequency corresponding to the maximum peak as the instantaneous frequency at this moment, and based on the same process, automatically obtain the instantaneous frequency of the mode at other times; Step 3.4, single-mode signal calculation: After obtaining the initial instantaneous frequency, the adaptive frequency modulation mode tracking algorithm is used to automatically extract the single-mode signal corresponding to this order frequency, and the process is repeated until all single-mode signals are obtained.

6. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 5, characterized in that: In step 3.1, the expression of the synchronously squeezed short-time Fourier transform is: Where FSST(t,ξ) represents the frequency value after the synchronous compression short-time Fourier transform at time t and the frequency ξ after synchronous compression; δ is the synchronous compression operator; ξ is the time-frequency spectrum frequency after synchronous compression; STFT(t,ω) is the short-time Fourier transform; ω0(t,ω) is the instantaneous frequency corresponding to time t.

7. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 6, characterized in that: In the step 4, the signal z is analyzed i (t): In the formula, z i (t) is the analytical signal of the i-th eigenmode component; s i (t) is the original signal of the i-th eigenmode component; is the signal after Hilbert transformation of the i-th eigenmode component; j is the imaginary unit.

8. The method for automatically identifying single-mode and multi-mode vortex-induced vibration of a stay cable according to claim 7, characterized in that: In the step 4, each analytical signal z i (t) is projected onto the complex domain. The instantaneous modulus of the analytical signal represents the instantaneous amplitude of the signal. The time domain feature vector is automatically established by calculating the similarity ratio of the instantaneous amplitude at each moment: In the formula, A i (t) is all instantaneous amplitudes of the i-th single-mode signal in this period; t is the corresponding time in this period; is the amplitude similarity ratio corresponding to the i-th unimodal signal; m=n-1 is the number of decomposed unimodal signals.

Images