Cable-stayed bridge cable wind-induced abnormal vibration monitoring and early warning method

Abstract

The invention discloses a monitoring and early warning method for wind-induced abnormal vibration of a cable-stayed bridge cable, and the method comprises the steps: calculating an energy concentration coefficient R according to a vibration intensity peak value P1st corresponding to a first large peak value and a vibration intensity peak value P2nd corresponding to a second large peak value; adding the energy concentration coefficient R into the array {R} to obtain an array {R} of bmin, and calculating the acceleration root mean square of the vibration acceleration data of the inhaul cable bmin according to the vibration acceleration data of the inhaul cable bmin; if k continuous energy concentration coefficients R in the array {R} of bmin exceed the early warning lower limit value and the acceleration root mean square exceeds the acceleration RMS upper limit threshold value, an early warning signal is sent out. The method has universality and effectiveness in identifying the second large vibration intensity peak value of different stay cable spectrograms, a bridge management and maintenance unit can conveniently know the abnormal vibration condition of the stay cable in time and maintain the stay cable, the service life of the stay cable is prolonged, and a guarantee is provided for safe operation of a bridge.

Description

Technical Field

[0001] The invention relates to a method for monitoring and early warning abnormal vibration of cables of a cable-stayed bridge caused by wind, and belongs to the technical field of bridge monitoring. Background Art

[0002] For cable-stayed bridges, cables are one of the important components of the bridge. Changes in the stiffness of the bridge structure, changes in boundary constraints, and emergencies such as overload and ship collision will cause changes in the working state of the cables. Therefore, for operating cable-stayed bridges, it is extremely important to accurately and efficiently obtain the vibration state of the cable system, which helps bridge managers to timely grasp the working status of the entire bridge.

[0003] Wind-induced vibration of cables is mainly caused by periodic vortex shedding when the airflow generated by the wind bypasses the cables. When the wind speed reaches a certain value, the frequency of vortex shedding is close to the natural frequency of the cables, causing vortex-induced resonance of the cables. Cable vortex vibration usually manifests as a limited amplitude vibration with a small amplitude that is not easily detected, and is mainly manifested as frequency-locked vibration.

[0004] Therefore, the study and prevention of frequency locking phenomenon in abnormal vibration of wind-induced cables has important engineering practical significance. Frequency locking vibration is an abnormal vibration. Unlike normal periodic vibration, it is less affected by external excitation, but generates continuous vibration energy through the nonlinear factors of the system, which increases the vibration amplitude of the cable and generates large alternating stress, which may cause structural damage and safety accidents.

[0005] Although existing technologies have explored and studied the frequency-locked vibration monitoring and early warning of the cable-stayed cable, the monitoring method is often used for post-analysis, which makes it difficult to achieve online real-time automatic monitoring and lacks timeliness and effectiveness. At present, semi-automatic monitoring methods can be realized based on data analysis codes and energy concentration coefficient calculation formulas, but the second largest vibration intensity peak automatically found in the spectrum analysis based on the energy concentration coefficient calculation formula is not necessarily accurate, and manual regular data verification is also required, which is cumbersome and labor-intensive.

[0006] Therefore, those skilled in the art are in urgent need of improving the wind-induced abnormal vibration monitoring and early warning method for bridge cables. Summary of the invention

[0007] Purpose: In order to overcome the deficiencies in the prior art, the present invention provides a method for monitoring and early warning of wind-induced abnormal vibration of cables of cable-stayed bridges, which can monitor and analyze abnormal vibration of cables in near real time and automatically, and has a good automatic early warning effect of abnormal vibration frequency locking phenomenon of cables, so as to understand the operation status of bridges in real time and ensure the safety of structures.

[0008] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is:

[0009] A method for monitoring and early warning of abnormal wind-induced vibration of cables of a cable-stayed bridge, specifically comprising the following steps:

[0010] Step 1: Obtain the vibration acceleration data of the cable as, where a represents the sampling period and s represents seconds.

[0011] Step 2: Perform frequency domain transformation on the cable as vibration acceleration data to obtain the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable, and form the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable (P x ,f) variable group,P x is the power spectrum energy value set of the short-time vibration acceleration of the cable, f is the frequency set of the short-time vibration acceleration of the cable, for (P x f) The variable group uses the peak method to identify the frequency corresponding to the peak of the spectrum of the as vibration acceleration data and sort it to obtain the frequency sequence corresponding to the peak of the spectrum of the as vibration acceleration data, and find the vibration intensity peak value P corresponding to the first largest peak. 1st .

[0012] Step 3, read (P x , f) variable group, remove the signal noise component, and then perform smooth curve processing to obtain a peak value with a smooth curve (P x , f) variable group, for the variable group with obvious peak and smooth curve (P x , f) The peak method is used to identify the variable group with obvious peaks and smooth curves (P x , f) The peak values ​​of the variable groups correspond to the frequencies and are sorted to obtain the peak values ​​with obvious peak values ​​and smooth curves (P x , f) The frequency sequence corresponding to the peak value of the variable group.

[0013] Step 4: According to the frequency sequence corresponding to the peak value of the spectrum of the vibration acceleration data as well as the peak value with a smooth curve (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, for the updated (P x , f) The variable group uses the peak method to identify the updated (P x , f) The frequency corresponding to the peak value of the variable group, and find the vibration intensity peak value P corresponding to the second largest peak value 2nd .

[0014] Step 5: According to the vibration intensity peak value P corresponding to the first maximum peak value 1st The vibration intensity peak value P corresponding to the second largest peak 2nd , calculate the energy concentration factor R.

[0015] Step 6, add the energy concentration coefficient R to the array {R}, repeat steps 1 to 5, obtain the array {R} of bmin, b represents the time period, min represents minutes, and calculate the acceleration root mean square of the cable bmin vibration acceleration data based on the bmin cable vibration acceleration data.

[0016] Step 7: If the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper limit threshold, a warning signal is issued.

[0017] As a preferred solution, it also includes: Step 8, if the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value for k consecutive times or the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value, return to step 1.

[0018] As a preferred scheme, it also includes: Step 9, if the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit value for k consecutive times and the root mean square of acceleration exceeds the acceleration RMS upper limit threshold and the wind speed and wind direction meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, it is determined to be a vortex vibration event.

[0019] As a preferred solution, it also includes: Step 10, if the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit value for k consecutive times and the root mean square of acceleration exceeds the RMS upper limit threshold of acceleration and the wind speed and wind direction do not meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, then combine the monitoring system data and manual observation and analysis to determine whether the resonant frequency locking phenomenon is caused by other external load sources.

[0020] As a preferred solution, bmin=2min, and the acceleration RMS upper threshold is set to 100cm / s 2 .

[0021] As a preferred solution, k is equal to 6, and the warning lower limit value is equal to 0.1.

[0022] As a preferred solution, the energy concentration factor R is expressed as follows:

[0023] .

[0024] As a preferred solution, the vortex vibration initiation wind speed threshold is the vortex vibration initiation wind speed at the bridge site. , the calculation formula is as follows:

[0025]

[0026] Where: B is the outer diameter of the cable; S t is the Strouhal number, It is the main frequency of the cable history.

[0027] The vortex vibration initiation wind direction threshold is set to take the bridge axis as 0° and the wind direction angle as being in the range of 75°-105°.

[0028] As a preferred solution, in step 4, according to the frequency sequence corresponding to the peak value of the vibration acceleration data spectrum and the peak value and smooth curve (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, including:

[0029] Step 4-1: Select the peak value with smooth curve (P x , f) The first frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data with a difference within the difference limit and the smallest sequence number is selected and added to the matched frequency sequence.

[0030] Step 4-2, remove the selected frequency from the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data to obtain the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal.

[0031] Step 4-2: select the peak value with smooth curve (P x , f) The second frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal with a difference within the difference limit and the smallest serial number is selected and added to the matched frequency sequence.

[0032] Step 4-3, and so on, traverse the peak values ​​with smooth curves (P x , f) The remaining frequencies in the frequency sequence corresponding to the peak value of the variable group are added to the matched frequency sequence to obtain a complete matched frequency sequence.

[0033] Step 4-4: Calculate the difference between adjacent frequencies in pairs in the complete matched frequency sequence. When the difference is less than the verification threshold, delete the back-end value of the adjacent frequency to obtain the verified frequency sequence.

[0034] Step 4-5: Add the verified frequency sequence and the power spectrum energy values ​​corresponding to all frequencies to the updated (P x , f) variable group, and get the updated (P x , f) variable groups.

[0035] As a preferred solution, the difference limit is equal to ±0.05f 1,1 , where f 1,1 Indicates the first frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data. The verification threshold is equal to f 3,1 , where f 3,1 Indicates the first frequency in the complete matched frequency sequence.

[0036] Beneficial effects: The present invention provides a method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridges, which is universal and effective in identifying the second largest vibration intensity peak of different cable spectrum graphs, making it easier for bridge maintenance units to timely understand the abnormal vibration of cables and maintain them, thereby increasing their service life and providing protection for safe bridge operation. Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The implementation effect of the cable wind-induced abnormal vibration monitoring and early warning method based on wind field characteristics and dynamic response is basically consistent with the abnormal vibration of the cable observed by on-site maintenance personnel. It has the advantages of timeliness, effectiveness and accuracy, and is convenient for engineering application.

[0038] 2. The present invention is novel in concept. It identifies the frequency locking phenomenon in abnormal vibration by monitoring multiple monitoring indicators and obtains the vibration-prone mode. Subsequently, it can realize the integration of cable vibration early warning and vibration reduction by coordinating vibration reduction measures, maintain the working safety of the cable, and ensure the safety of bridge operation.

[0039] 3. The entire system is automatically controlled by a computer, which is more energy-efficient and more responsive, fully demonstrating the superiority of intelligent monitoring and maintenance of the bridge over manual inspection and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a structural diagram of the wind-induced abnormal vibration monitoring and early warning system for cable-stayed bridge cables.

[0041] Figure 2 This is a schematic diagram of the architecture of the wind-induced abnormal vibration monitoring and early warning system for cable-stayed bridge cables.

[0042] Figure 3 The figure is a flow chart of the vortex-induced vibration monitoring and early warning method for cable-stayed bridge cables.

[0043] Figure 4 Schematic diagram of the time-varying power spectrum for monitoring and early warning of wind-induced abnormal vibration of cable-stayed bridge cables.

[0044] Figure 5 This is a schematic diagram of the power spectrum of a typical cable wind-induced vibration event in the cable-stayed bridge wind-induced abnormal vibration monitoring and early warning system, where: Figure 5 Figure a is a normal vibration diagram. Figure 5 Figure b is a schematic diagram of abnormal vibration.

[0045] Figure 6 Schematic diagram of the all-day energy concentration factor of the wind-induced abnormal vibration monitoring and early warning system for cable-stayed bridge cables.

[0046] Figure 7 Schematic diagram of the short-time energy concentration factor of the wind-induced abnormal vibration monitoring and early warning system for cable-stayed bridge cables.

[0047] The numbers in the figure are as follows:

[0048] 1. Bridge; 2. Abutment; 3. Bridge tower; 4. Cable; 5. Wind speed and direction sensor; 6. Cable vibration sensor; 7. Cable damper; 8. Line; 9. Computer; 10. Data acquisition device; 11. Memory; 12. Signal processor. DETAILED DESCRIPTION

[0049] The following is a clear and complete description of the technical solutions in the examples of the present invention in conjunction with the accompanying drawings in the examples of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present invention.

[0050] The present invention will be further described below in conjunction with specific embodiments.

[0051] Embodiment 1:

[0052] This embodiment introduces a cable-stayed bridge cable wind-induced abnormal vibration monitoring and early warning system. Figure 1 to Figure 2 As shown, it includes: a bridge 1, a pedestal 2 is provided under the bridge tower 3 of the bridge 1, cables 4 are provided above the bridge, and a plurality of cables 4 are inclined between the bridge tower 3 and the bridge 1; it also includes a monitoring module and an automatic control module; the monitoring module includes a wind direction and speed sensor 5, a cable vibration sensor 6 and a computer 9, the wind direction and speed sensor 5 is provided at the bridge site, and the cable vibration sensor 6 is bundled at the typical cable cross section; the wind direction and speed sensor and the cable vibration sensor are both connected to the computer 9.

[0053] Preferably, the wind direction and speed sensor and the cable vibration sensor are connected to the data collector 10 of the computer 9 through the line 8. The data collector 10 collects data and stores it in the memory 11. The signal processor 12 of the computer monitors and analyzes the data. When early warning information appears, an audible and visual alarm is issued to remind the maintenance personnel to pay attention to the vibration of the cable.

[0054] Preferably, one wind direction and wind speed sensor is provided on each of the bridge towers on both sides of the bridge and in the middle of the span. In this case, the bridge system has a total of three wind direction and wind speed sensors to monitor the wind direction and wind speed of the side wind.

[0055] Preferably, the sampling frequency of the wind direction and speed sensor is 1 Hz, and the sampling frequency of the cable vibration sensor is 50 Hz.

[0056] Preferably, the plurality of cables 4 are provided with cable dampers 7 .

[0057] Embodiment 2:

[0058] The more practical monitoring and early warning method in existing projects is to take the order frequency f of the easy vibration i ±0.5f1 (i=1,2,3...n) corresponding to the maximum energy value to determine whether it emits abnormal vibration. However, abnormal vibration of a structure is generally a certain order frequency such as 1.524, but this frequency fluctuates, sometimes 1.499, sometimes 1.555, so in the XY correspondence in the power spectrum diagram, it is not possible to select 1.524 as the maximum value of Y every time. Instead, the maximum energy Y in the previous and next area is used as the maximum energy value of this vibration. Among them: f1 is the first order frequency of the cable, f i is the i-th order frequency of the cable.

[0059] Therefore, although the more practical monitoring and early warning methods in existing projects are simple and fast, and are universal for most conventional cables, there are still the following four problems:

[0060] 1) It is difficult to accurately identify the fundamental frequencies of different cables. Manual analysis of a single cable has a good effect, but it is difficult to achieve automated analysis of cable groups.

[0061] 2) For sensors with abnormal signals, their spectrum will also be abnormal. When this method is used for analysis, it is difficult to identify spectrum abnormalities, let alone identify the energy concentration coefficient.

[0062] 3) This method does not take into account the impact of factors such as icing, installation of vibration dampers, and temperature on the sensor, which may cause significant changes in the cable frequency.

[0063] 4) Due to the cumulative effect of errors of each order, this method is more effective for identifying low-order vibrations, but the frequency calculated for high-order vibrations may not be accurate. Therefore, it is not conducive to the identification of medium and high-order frequencies of vortex vibrations, and it is difficult to identify and warn of abnormal wind-induced vibration phenomena such as vortex vibrations.

[0064] This embodiment introduces a method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables to overcome the shortcomings of existing methods, which specifically includes the following steps:

[0065] Step 1: Obtain the vibration acceleration data of the cable as, where a represents the sampling period and s represents seconds.

[0066] In one embodiment, the AS takes 20 seconds.

[0067] In one embodiment, the original vibration acceleration data can be converted into an excel format stored on a daily basis into an as or a mat format for a month by reading the code, and stored in the system background for easy access at any time.

[0068] Step 2: Perform frequency domain transformation on the cable as vibration acceleration data to obtain the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable, and form the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable (P x ,f) variable group,P x is the power spectrum energy value set of the short-time vibration acceleration of the cable, f is the frequency set of the short-time vibration acceleration of the cable, for (P x f) The variable group uses the peak method to identify the frequency corresponding to the peak of the spectrum of the as vibration acceleration data and sort it to obtain the frequency sequence corresponding to the peak of the spectrum of the as vibration acceleration data, and find the vibration intensity peak value P corresponding to the first largest peak. 1st .

[0069] In one embodiment, the frequency domain transform adopts FFT transform.

[0070] In one embodiment, the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data is f 1,1 , f 1,2 , f 1,3 , ...f 1,i ..., f 1,n , i∈[1,n], n represents the order.

[0071] Step 3, read (P x , f) variable group, remove the signal noise component, and then perform smooth curve processing to obtain a peak value with a smooth curve (P x , f) variable group, for the variable group with obvious peak and smooth curve (P x , f) The peak method is used to identify the variable group with obvious peaks and smooth curves (P x , f) The peak values ​​of the variable groups correspond to the frequencies and are sorted to obtain the peak values ​​with obvious peak values ​​and smooth curves (P x , f) The frequency sequence corresponding to the peak value of the variable group.

[0072] In one embodiment, the signal noise component is removed by using a frequency domain decomposition method and a finite difference method.

[0073] In one embodiment, the peak value is obvious and the curve is smooth (P x , f) The frequency sequence corresponding to the peak value of the variable group is: f 2,1 , f 2,2 , f 2,3 ...f 2,j ..., f2,m , where j∈[1,m], m represents the sequence number, m <n。

[0074] Step 4: According to the frequency sequence corresponding to the peak value of the spectrum of the vibration acceleration data as well as the peak value with a smooth curve (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, for the updated (P x , f) The variable group uses the peak method to identify the updated (P x , f) The frequency corresponding to the peak value of the variable group, and find the vibration intensity peak value P corresponding to the second largest peak value 2nd .

[0075] In one embodiment, the vibration intensity peak value P corresponding to the second largest peak value of the present invention is 2nd The acquisition method overcomes the influence of burr peaks near the peak where the vibration intensity peak corresponding to the first largest peak is located, and realizes the function of automatic recognition.

[0076] Step 5: According to the vibration intensity peak value P corresponding to the first maximum peak value 1st The vibration intensity peak value P corresponding to the second largest peak 2nd , calculate the energy concentration factor R.

[0077] Step 6, add the energy concentration coefficient R to the array {R}, repeat steps 1 to 5, obtain the array {R} of bmin, b represents the time period, min represents minutes, and calculate the acceleration root mean square of the cable bmin vibration acceleration data based on the bmin cable vibration acceleration data.

[0078] One embodiment further includes: linearizing the array {R}.

[0079] In one embodiment, bmin is 2 minutes.

[0080] Step 7: If the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper limit threshold, a warning signal is issued.

[0081] In one embodiment, k is equal to 6.

[0082] In one embodiment, the warning lower limit value is equal to 0.1.

[0083] In one embodiment, the early warning signal uses an audible and visual alarm signal to remind maintenance personnel to check the vibration condition.

[0084] Preferably, the method further includes: step 8, if the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value for k consecutive times or the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value, then return to step 1.

[0085] Preferably, it also includes: Step 9, if the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit value for k consecutive times and the root mean square of acceleration exceeds the acceleration RMS upper limit threshold and the wind speed and wind direction meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, it is determined to be a vortex vibration event.

[0086] Preferably, it also includes: Step 10, if the energy concentration coefficient R in the array {R} of bmin exceeds the lower warning limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper threshold and the wind speed and wind direction do not meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, then combine the monitoring system data and manual observation and analysis to determine whether the resonant frequency locking phenomenon is caused by other external load sources (such as vehicles passing at the bridge head). After the energy concentration coefficient warning, the wind speed and direction data of the wind speed and direction sensor can be combined to analyze the vibration type.

[0087] Preferably, the peak method is the simplest and most widely used frequency domain mode identification method in the civil engineering field. The basic principle is that the peak value of the frequency response function of the system can represent its own oscillation frequency. The mode vibration shape is determined by the peak value. Each peak value represents an oscillation component, and the frequency response function is mainly based on the following formula:

[0088]

[0089] Where: is the independent variable, i is the modal order, is the residue matrix, the poles Contains frequency and damping information, namely: + , is the i-th order characteristic frequency, is the i-th order characteristic damping, j is an imaginary unit, for and The adjoint matrix of .

[0090] Since the modal vibration shape is difficult to estimate finitely, some oscillation modes may be lost during modal identification. Therefore, the average regularized power spectrum is introduced to integrate the response data. Compared with the autopower spectrum density of a single channel, the peak information is more accurate and the curve is clear.

[0091]

[0092] Where: k is the data number, m is the discrete frequency number, n is the discrete frequency number for Fourier transform, is the data length, and PSD is the power spectral density matrix.

[0093] Preferably, the frequency domain decomposition method and the finite difference method are combined to reduce noise based on the following formula:

[0094] 1) Fourier transform: In digital signal processing, discrete Fourier transform (DFT) or fast Fourier transform (FFT) is usually used to process discrete signals. First, the noisy signal x(t) is Fourier transformed to obtain its representation X(f) in the frequency domain (where j is an imaginary unit):

[0095]

[0096] 2) Frequency domain filtering: Based on the finite difference method, the high-order derivatives of the signal are estimated to construct a more complex smoothing filter. In this method, the central difference method is used. For the function At the point The central difference at is the step size), and estimates the second-order derivative:

[0097]

[0098]

[0099] Then, in the frequency domain, a filter H(f) is designed to remove the noise. This filter can be a low-pass, high-pass, band-pass, or band-stop filter, depending on the characteristics of the noise and signal. The weights of the smoothing filter are adjusted based on the estimated value of the second-order derivative to better adapt to the characteristics of the signal.

[0100]

[0101] 3) Inverse Fourier transform: Perform an inverse Fourier transform on the filtered frequency domain signal Y(f) to obtain the denoised time domain signal y(t):

[0102]

[0103] Preferably, in one embodiment, considering that the frequency of wind-induced cable vibration events in coastal areas is high and cable diseases are accumulated, the upper threshold of the RMS acceleration corresponding to the 2-min vibration acceleration of the cable is set to 100 cm / s 2 .

[0104] In one embodiment, the root mean square acceleration of the cable bmin vibration acceleration data is a statistical parameter characterizing the total energy of random vibration, and the calculation formula is:

[0105]

[0106] In the formula, A a It represents the root mean square value of acceleration in centimeters per square second (cm / s 2 );a i Indicates the vibration acceleration of the cable in centimeters per square second (cm / s 2 ); N represents the number of bmin acceleration sampling points.

[0107] Preferably, the energy concentration coefficient can reflect the outstanding situation of the amplitude of the vibration power spectrum of the cable at a certain moment. The smaller the coefficient, the greater the main frequency vibration energy, and its calculation formula is:

[0108]

[0109] Where P 1st It represents the maximum amplitude of the power spectrum density of the structural vibration response, in square millimeters per cubic second (mm 2 / s 3 );P 2nd It represents the second largest amplitude in the power spectrum density of the structural vibration response, in square millimeters per second (mm 2 / s 3 ).

[0110] Further, in one embodiment, R as It represents the energy concentration factor R of the AS time-course data output, a=20. It can be optimized to a shorter time period as the spectrum analysis technology improves.

[0111] Preferably, the vortex vibration initiation wind speed threshold adopts the vortex vibration initiation wind speed at the bridge site. , the calculation formula is as follows:

[0112]

[0113] Where: B is the outer diameter of the cable; S t is the Strouhal number. For a real circular cylinder, S t Take 0.2; It is the main frequency of the cable history.

[0114] Preferably, when vortex vibration of the cable-stayed cable occurs, it is mainly affected by the cross-bridge wind. Therefore, the vortex vibration initiation wind direction threshold is set to take the bridge axis as 0° and the wind direction angle within the range of 75°-105°.

[0115] Preferably, in step 3, the frequency domain decomposition method is first used to perform spectrum analysis on the signal to identify and separate the noise component. Then, the finite difference method is used to discretize the signal and extract the characteristic information in the signal. Finally, noise suppression or filtering is performed through the characteristic information to remove the noise component, and peak recognition is achieved by calling the matlab built-in function findpeaks. The burr peak in the singular value curve can be well removed by front and rear mean filtering, and the burr peak part of the curve becomes relatively smooth, and the true peak part of the curve will not change before and after filtering.

[0116] Preferably, all peak-to-trough difference indicators are classified based on K-means clustering method to distinguish true peaks from false peaks, but the effect is only good for the first few orders and higher energy frequency bands.

[0117] Preferably, in step 4, according to the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data and the peak value and smooth curve (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, including:

[0118] Step 4-1: Select the peak value with smooth curve (P x , f) The first frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data with a difference within the difference limit and the smallest sequence number is selected and added to the matched frequency sequence.

[0119] Step 4-2, remove the selected frequency from the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data to obtain the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal.

[0120] Step 4-2: select the peak value with smooth curve (P x , f) The second frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal with a difference within the difference limit and the smallest serial number is selected and added to the matched frequency sequence.

[0121] Step 4-3, and so on, traverse the peak values ​​with smooth curves (P x , f) The remaining frequencies in the frequency sequence corresponding to the peak value of the variable group are added to the matched frequency sequence to obtain a complete matched frequency sequence.

[0122] Step 4-4: Calculate the difference between adjacent frequencies in pairs in the complete matched frequency sequence. When the difference is less than the verification threshold, delete the back-end value of the adjacent frequency to obtain the verified frequency sequence.

[0123] Step 4-5: Add the verified frequency sequence and the power spectrum energy values ​​corresponding to all frequencies to the updated (P x , f) variable group, and get the updated (P x , f) variable groups.

[0124] Step 4 is a specific step for removing differences and retaining similarities. When the residual component allows, the intersection of the two peak frequency arrays is obtained. At the same time, the removal of similarities and retention of differences is only processed in the first analysis. The processing results are directly saved in the background and directly called later. The purpose is to remove the false peaks of the power spectrum and retain its true peaks.

[0125] In one embodiment, the difference limit is equal to ±0.05f 1,1 , where f 1,1 Indicates the first frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data.

[0126] In one embodiment, the verification threshold is equal to f 3,1 , where f 3,1 Indicates the first frequency in the complete matched frequency sequence.

[0127] In one embodiment, the peak value of the spectrum of the vibration acceleration data corresponds to the frequency sequence n=5, the peak value is obvious and the curve is smooth (P x , f) Take the frequency sequence m=3 corresponding to the peak value of the variable group as an example.

[0128] Step 4-1, f 2,1 , f 2,2 , f 2,3 The first frequency f 2,1 Sequentially with f 1,1 , f 1,2 , f 1,3 , f 1,4 , f 1,5 Perform difference calculation and select the frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data within the difference limit, such as f 1,1 , f 1,2 Select the smallest number and get f 1,1 And add the matched frequency sequence and record it as f 3,1 .

[0129] Step 4-2: f 1,1 From 1,1 , f1,2 , f 1,3 , f 1,4 , f 1,5 Remove from the equation, and we get f 1,2 , f 1,3 , f 1,4 , f 1,5 .

[0130] Step 4-2: f 2,1 , f 2,2 , f 2,3 The second frequency f 2,2 Sequentially with f 1,2 , f 1,3 , f 1,4 , f 1,5 Calculate the difference of all frequencies in and select the frequency f whose difference is within the different limit 1,3 , f 1,4 Select the smallest number and get f 1,3 And add the matched frequency sequence and record it as f 3,2 , where f 1,2 The corresponding differences are not within the dissimilarity limits.

[0131] Step 4-3, and so on, f 2,3 Sequentially with f 1,2 , f 1,4 , f 1,5 Calculate the difference of all frequencies in and select the frequency f whose difference is within the different limit 1,5 , we get f 1,5 And add the matched frequency sequence and record it as f 3,3 , where f 1,2 , f 1,4 The corresponding differences are not within the dissimilarity limits.

[0132] Step 4-4: Put the complete matched frequency sequence into f 3,1 , f 3,2 , f 3,3 The difference between two adjacent frequencies is calculated in sequence. When the difference is less than the verification threshold, the back-end value of the adjacent frequency is deleted, such as f 3,2 , it is also possible that the difference is not less than the verification threshold, and the verified frequency sequence f is obtained 3,1 , f 3,3 or 3,1 , f 3,2 , f 3,3 .

[0133] Step 4-5: Add the verified frequency sequence and the power spectrum energy values ​​corresponding to all frequencies to the updated (P x , f) variable group, and get the updated (P x , f) variable groups.

[0134] Embodiment 3:

[0135] This embodiment introduces an application example of the method of the present invention. A bridge in a coastal area is located at the mouth of the Yongjiang River, spanning two areas. The total length of the bridge is 2482m, of which the main bridge is 567m long, and the span is arranged as 74.5m+258m+102m+83m+49.5m. The bridge is 29.5m wide, with 6 lanes in both directions and a design speed of 60km / h; the main tower is 148.4m high, with a total of 102 upstream and downstream cables. A total of 30 sets of cable force sensors are arranged upstream and downstream of the entire bridge. Typical cable vibration wind speeds are shown in Table 1:

[0136] Table 1 Wind speed of the cable-stayed cable (m / s)

[0137]

[0138] When the vortex vibration event occurs, the vibration modes of each cable are single, and a "locked frequency phenomenon" occurs. The main vibration frequencies are mainly 2nd, 4th, and 8th orders. The vibration energy reaches 10 3 -10 5 The maximum and second-largest values ​​of the power spectrum are quite different, and the cable vibration is obvious. Figure 4 shown.

[0139] Table 2 Energy changes during vibration

[0140]

[0141] In addition to observing whether the frequency locking state occurs, it is also necessary to focus on whether the energy concentration coefficient exceeds the limit. Table 2 lists the maximum and second maximum values ​​and energy concentration coefficients of the power spectra of the six cables before and after the vibration event. It can be seen from the vortex vibration early warning subsystem that before the event, the energy concentration coefficients of the six cables did not exceed the threshold of 0.1; during the event, except for the cable SL-P-08, the energy concentration coefficients of the remaining five cables were all lower than 0.1, exceeding the lower limit of the vortex vibration. The five cables can be warned of vortex vibration in time, and the bridge maintenance unit can remind the drivers and passengers to limit the speed and drive slowly through the bridge, and take maintenance measures when necessary.

[0142] Take the cable SL-P-13 as an example. Figure 5 , Figure 6 Displays the energy concentration factor throughout the day and in short periods of time. Figure 7 The short-term energy concentration coefficient of the wind-induced abnormal vibration monitoring and early warning system of the cable-stayed bridge is shown. Combined with Table 3, it can be seen that the number of vortex vibrations and their start time, end time and duration can help the bridge maintenance unit to understand the cable vibration in real time.

[0143] Table 3 Vortex vibration frequency and wind conditions of cable SL-P-13

[0144]

[0145] Based on the historical dominant vibration frequency of vortex-induced vibration of typical cables, targeted vibration reduction measures are set up in subsequent maintenance to effectively reduce the acceleration amplitude of abnormal vibration events, greatly reduce the duration of each vortex-induced vibration event, extend the working life of the cables, and ensure the normal operation of the bridge.

[0146] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for monitoring and early warning of abnormal wind-induced vibration of cables of a cable-stayed bridge, characterized in that: The specific steps include: Step 1, obtain the vibration acceleration data of the cable as, where a represents the sampling period and s represents seconds; Step 2: Perform frequency domain transformation on the cable as vibration acceleration data to obtain the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable, and form the frequency and power spectrum energy value corresponding to the short-term vibration acceleration of the cable (P x ,f) variable group,P x is the power spectrum energy value set of the short-time vibration acceleration of the cable, f is the frequency set of the short-time vibration acceleration of the cable, for (P x f) The variable group uses the peak method to identify the frequency corresponding to the peak of the spectrum of the as vibration acceleration data and sort it to obtain the frequency sequence corresponding to the peak of the spectrum of the as vibration acceleration data, and find the vibration intensity peak value P corresponding to the first largest peak. 1st ; Step 3, read (P x , f) variable group, remove the signal noise component, and then perform smooth curve processing to obtain a peak value with a smooth curve (P x , f) variable group, for the variable group with obvious peak and smooth curve (P x , f) The peak method is used to identify the variable group with obvious peaks and smooth curves (P x , f) The peak values ​​of the variable groups correspond to the frequencies and are sorted to obtain the peak values ​​with obvious peak values ​​and smooth curves (P x ,f) frequency sequence corresponding to the peak value of the variable group; Step 4: According to the frequency sequence corresponding to the peak value of the spectrum of the vibration acceleration data as well as the peak value with a smooth curve (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, for the updated (P x , f) The variable group uses the peak method to identify the updated (P x , f) The frequency corresponding to the peak value of the variable group, and find the vibration intensity peak value P corresponding to the second largest peak value 2nd ; Step 5: According to the vibration intensity peak value P corresponding to the first maximum peak value 1st The vibration intensity peak value P corresponding to the second largest peak 2nd , calculate the energy concentration factor R; Step 6, adding the energy concentration coefficient R to the array {R}, repeating steps 1 to 5, obtaining the array {R} of bmin, where b represents the time period and min represents minutes, and calculating the acceleration root mean square of the cable bmin vibration acceleration data according to the bmin cable vibration acceleration data; Step 7: If the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper limit threshold, a warning signal is issued.

2. The method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge according to claim 1, characterized in that: Also includes: Step 8: If the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value for k consecutive times or the energy concentration coefficient R in the array {R} of bmin does not exceed the warning lower limit value, return to step 1.

3. The method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge according to claim 1, characterized in that: Also includes: Step 9: If the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper limit threshold and the wind speed and wind direction meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, it is determined to be a vortex vibration event.

4. The method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to claim 1, characterized in that: Also includes: Step 10: If the energy concentration coefficient R in the array {R} of bmin exceeds the warning lower limit for k consecutive times and the acceleration root mean square exceeds the acceleration RMS upper limit threshold and the wind speed and wind direction do not meet the vortex vibration initiation wind speed threshold and vortex vibration initiation wind direction threshold, then combine the monitoring system data and manual observation and analysis to determine whether the resonant frequency locking phenomenon is caused by other external load sources.

5. A method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to any one of claims 1 to 4, characterized in that: bmin=2min, the acceleration RMS upper limit threshold is set to 100cm / s 2 .

6. A method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to any one of claims 1 to 4, characterized in that: The k is equal to 6, and the warning lower limit value is equal to 0.

1.

7. A method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to any one of claims 1 to 4, characterized in that: The energy concentration factor R is expressed as follows: 。 8. A method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to any one of claims 1 to 4, characterized in that: The vortex vibration initiation wind speed threshold is the vortex vibration initiation wind speed at the bridge site. , the calculation formula is as follows: ； Where: B is the outer diameter of the cable; S t is the Strouhal number, It is the main frequency of the cable history; The vortex vibration initiation wind direction threshold is set to take the bridge axis as 0° and the wind direction angle as being in the range of 75°-105°.

9. A method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to any one of claims 1 to 4, characterized in that: In step 4, the peak value of the spectrum of the vibration acceleration data corresponds to the frequency sequence and the peak value is obvious and the curve is smooth (P x , f) The peak value of the variable group corresponds to the frequency sequence, and the updated (P x , f) variable group, including: Step 4-1: Select the peak value with smooth curve (P x , f) The first frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data with a difference within the difference limit and the smallest sequence number is selected and added to the matched frequency sequence; Step 4-2, removing the selected frequency from the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data, to obtain the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal; Step 4-2: select the peak value with smooth curve (P x , f) The second frequency in the frequency sequence corresponding to the peak value of the variable group is calculated in turn with all the frequencies in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal, and a frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data after removal with a difference within the difference limit and the smallest sequence number is selected and added to the matched frequency sequence; Step 4-3, and so on, traverse the peak values ​​with obvious and smooth curves (P x , f) The remaining frequencies in the frequency sequence corresponding to the peak value of the variable group are added to the matched frequency sequence to obtain a complete matched frequency sequence. Step 4-4, calculate the difference between adjacent frequencies of all frequencies in the complete matched frequency sequence in sequence, and when the difference is less than the verification threshold, delete the back-end value of the adjacent frequency to obtain the verified frequency sequence; Step 4-5: Add the verified frequency sequence and the power spectrum energy values ​​corresponding to all frequencies to the updated (P x , f) variable group, and get the updated (P x , f) variable groups.

10. The method for monitoring and early warning of abnormal wind-induced vibration of cable-stayed bridge cables according to claim 9, characterized in that: The difference limit is equal to ±0.05f 1,1 , where f 1,1 Indicates the first frequency in the frequency sequence corresponding to the peak value of the spectrum of the as vibration acceleration data; the verification threshold is equal to f 3,1 , where f 3,1 Indicates the first frequency in the complete matched frequency sequence.

Images