A fully automatic cable frequency extraction, order determination and cable force identification method and its application

Abstract

The present invention relates to the fields of bridge engineering and signal processing technology, and in particular to a method and application for fully automatic cable frequency extraction, order determination, and cable force identification. The present invention first uses a sensor to obtain time-course signals such as acceleration and displacement of cable vibration, and uses Fourier transform and other methods to obtain the spectrum of the vibration signal; then uses a memory-type sieve peak algorithm to search for peaks in the spectrum, that is, the modal frequency of the cable; hash voting is performed on all acquired frequency peaks to obtain the nominal fundamental frequency, and then the order corresponding to all picked frequencies is determined; after obtaining all frequencies and orders, the cable-beam vibration theory can be used to obtain parameters such as cable force. The present invention combines the memory-type sieve peak algorithm and the hash voting statistical algorithm to achieve a fully automatic and intervention-free cable force identification process, and has strong engineering applicability.

Description

Technical Field

[0001] The present invention relates to the fields of bridge engineering and signal processing technology, and in particular to a method and application for fully automatic cable frequency extraction, order determination and cable force identification. Background Art

[0002] Cables are a widely used civil engineering structural component, offering advantages such as strong tensile strength, high strength, and lightweight. In bridge engineering, cables are widely used in cable-supported bridges, including cable-stayed bridges, suspension bridges, and suspender arch bridges. As the primary load-bearing component, the cable tension in the suspender or cable directly affects the overall internal force distribution of the structure, determining the overall condition and safety of the bridge. Therefore, cable tension is a crucial indicator in bridge inspection and long-term monitoring.

[0003] Existing methods for identifying cable tension primarily include the pressure gauge method, pressure sensor testing, magnetic flux method, and vibration method. The vibration method identifies the cable's vibration modal information based on its vibration signal and then uses an appropriate dynamic model to calculate the cable tension. This method is the most widely used due to its advantages, such as easy sensor installation, no impact on cable tension, no calibration requirements, and excellent durability. For longer cables, single-point vibration measurement can be used to obtain the cable's vibration frequency, providing a more accurate estimate of the cable tension. The present invention is a long cable tension measurement algorithm based on the frequency method.

[0004] The key to identifying cable forces using the vibration method lies in frequency extraction and accurate determination of the modal order. Furthermore, given the large number of bridge cables, the detection and monitoring process requires processing large amounts of data. Automating this data processing, without human intervention, is highly effective and has important practical implications. Furthermore, due to the influence of ambient noise, the identification of some characteristic frequencies can be erratic, with peaks sometimes missing (particularly the fundamental frequency). Automated frequency extraction and accurate order determination are challenging tasks.

[0005] When the number of cables and operating conditions is small, manual operation can be used to filter out peaks in the spectrum (i.e., characteristic frequencies of various orders) and estimate the cable's fundamental frequency by taking the difference between two large peaks. However, modern large-scale cable-stayed bridges contain a large number of cables, and the advent of non-contact instruments such as microwave radar allows the vibration time history of a large number of cables across the entire cable surface to be obtained in a single measurement. This results in a huge amount of data to be processed, making manual operation extremely time-consuming and labor-intensive, as well as subjective and uncertain in accuracy. Therefore, it is necessary to propose an automatic peak screening algorithm and an automatic fundamental frequency extraction algorithm to improve the efficiency and accuracy of cable force identification. Summary of the Invention

[0006] In order to solve the above problems, the purpose of the present invention is to provide a fully automatic cable frequency extraction, order determination and cable force identification method and application, aiming to solve the above problems of existing cable force identification technology and improve the cable force detection efficiency and accuracy of a large number of inclined cables under a large number of working conditions. The present invention first uses a sensor to obtain the acceleration, displacement and other time-course signals of the cable vibration, and uses Fourier transform and other methods to obtain the spectrum of the vibration signal; then uses a memory-type sieve peak algorithm to search for the peak in the spectrum, that is, the modal frequency of the cable; hash voting is performed on all the obtained frequency peaks to obtain the nominal fundamental frequency, and then the order corresponding to all picked frequencies is determined; after obtaining all frequencies and orders, the cable-beam vibration theory can be used to obtain parameters such as cable force. The present invention combines the memory-type sieve peak algorithm and the hash voting statistical algorithm to achieve a fully automatic and non-interventional cable force identification process, and has strong engineering applicability.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] The first object of the present invention is to provide a fully automatic cable frequency extraction, order determination and cable force identification method, comprising the following steps:

[0009] (S1) collecting vibration time history data of the cable;

[0010] (S2) performing spectrum analysis on the vibration time history data measured in step (S1) to obtain a displacement spectrum;

[0011] (S3) identifying the characteristic frequency in the displacement spectrum obtained in step (S2) by a memory-type sieve peak algorithm to obtain the modal frequency of the cable;

[0012] (S4) extracting the nominal fundamental frequency of the cable from the modal frequency of the cable in step (S3) by a voting statistical algorithm of a hash map;

[0013] (S5) determining the orders of all the cable modal frequencies extracted in step (S3) according to the nominal fundamental frequency of the cable obtained in step (S4) by a modal frequency order determination algorithm;

[0014] (S6) extracting the cable force based on the cable modal frequency obtained in step (S3) and the order of the cable modal frequency obtained in step (S5) in combination with the cable-beam vibration theory.

[0015] In one embodiment of the present invention, in step (S1), the vibration time history data of the cable is one of cable acceleration data or cable vibration displacement data.

[0016] In one embodiment of the present invention, the cable acceleration data is measured by a wired or wireless acceleration sensor.

[0017] In one embodiment of the present invention, the cable vibration displacement data is obtained by observing using microwave radar or video technology.

[0018] In one embodiment of the present invention, in step (S3), the memory-type sieve peak algorithm specifically includes the following steps:

[0019] (S301) Setting the size S of the filter window;

[0020] (S302) defining a Boolean array B whose initial values ​​are all true;

[0021] (S303) traverse each peak in the displacement spectrum sequence. For the current index i, if B[i] is false, skip the current value and go to the next value in the sequence.

[0022] (S304) For the current value, compare it with the next S / 2 points, and update the B array;

[0023] (S305) For the current value, compare it with the previous S / 2 points;

[0024] (S306) After comparison, if the current value is greater than the surrounding S values, then the point value is the peak point;

[0025] (S307) Repeat the process from step (S303) to (S306) until the traversal is completed and all peak points are output.

[0026] In one embodiment of the present invention, the screening window size S is initialized to 1.5 times the design value of the cable fundamental frequency.

[0027] In one embodiment of the present invention, in step (S4), the hash map voting algorithm specifically includes the following steps:

[0028] (S401) Calculate the difference sequence of the characteristic frequency sequence in step (S3), that is, the frequency difference sequence; and simultaneously calculate the amplitude weight corresponding to each frequency difference, whose value is the smaller value of the spectrum amplitudes corresponding to two adjacent frequencies. The corresponding formula is:

[0029] AW i =min(A i ,A i-1 );

[0030] in,

[0031] i-index number of the characteristic frequency sequence;

[0032] A-amplitude of characteristic frequency;

[0033] AW-amplitude weight;

[0034] (S402) traverse each element in the frequency difference sequence, multiply the element by an integer factor to obtain a hash value; for two integers with adjacent hash values, calculate their distance weights respectively, using the formula:

[0035] PW floor (val) = ceil(val) - val;

[0036] PW ceil (val) = val - floor (val);

[0037] in,

[0038] PW-distance weight;

[0039] floor - round down;

[0040] ceil - round up;

[0041] val - the value of an element in the sequence;

[0042] (S403) voting for two integers with adjacent hash values, with the number of votes being the product of the amplitude weight and the distance weight;

[0043] (S404) looping through steps (S402) and (S403) until the traversal is complete and the hash voting results are counted to find the hash value with the most votes;

[0044] (S405) traverse each element in the differential sequence again, multiply the element by the integer factor in step (S402) to obtain a hash value, and compare the hash value with the hash value with the most votes in step (S404);

[0045] (S406) If the difference in step (S405) does not exceed 1, the corresponding difference can participate in the weighted average calculation of the fundamental frequency, and its weight is the product of the amplitude weight and the distance weight;

[0046] (S407) Steps (S405) and (S406) are repeated until the traversal is completed. At the same time, a weighted average is performed on the characteristic frequency difference, and the result is the nominal fundamental frequency.

[0047] In one embodiment of the present invention, the hash value is obtained by multiplying the difference by 200, and the number of votes is the result of the amplitude weight and the distance weight, which is generally not an integer.

[0048] In one embodiment of the present invention, in step (S5), the modal frequency order determination algorithm specifically includes the following steps:

[0049] (S501) using the forward pointer to traverse the characteristic frequency sequence obtained in step (S3);

[0050] (S502) Dividing the current frequency by the nominal fundamental frequency, if it is approximately an integer, the first valid modal order is obtained; otherwise, returning to step (S501) to continue traversal;

[0051] (S503) traversing the remaining frequency sequence using the back pointer;

[0052] (S504) Subtracting the frequencies corresponding to the two front and rear pointers. If the difference divided by the nominal fundamental frequency is approximately an integer, a valid modal order is obtained, and the rear pointer is used to update the front pointer.

[0053] (S505) Steps (S503) and (S504) are repeated repeatedly until the traversal is completed, and all identified modal orders are output.

[0054] The second object of the present invention is to provide a fully automatic cable frequency extraction, order determination and cable force identification method for application in cable force identification of cable-stayed bridges, arch bridges and suspension bridges in load testing, periodic inspection and long-term monitoring.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] (1) The memory-type sieve peak algorithm proposed in the present invention has low complexity and better efficiency than existing algorithms, and is suitable for processing the vibration responses of multiple cables simultaneously;

[0057] (2) The hash mapping voting algorithm proposed in this paper achieves a stable estimation of the nominal fundamental frequency, is less affected by the missing of some modal frequencies and environmental noise, and greatly improves the robustness of the algorithm;

[0058] (3) The present invention combines the memory-type sieve peak algorithm and the hash voting statistical algorithm to realize the fully automatic and intervention-free cable force identification process, and has strong engineering applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of a fully automatic cable frequency extraction, order determination and cable force identification method according to Example 1;

[0060] Figure 2 is a spectrum containing a missing mode with identified characteristic frequencies of all orders in Example 1;

[0061] Figure 3 Detailed flow chart of the memory-type peak screening algorithm in Example 1;

[0062] Figure 4 This is a detailed flow chart of the hash mapping voting statistics algorithm in Example 1;

[0063] Figure 5 Detailed flow chart of the modal frequency order determination algorithm in Example 1;

[0064] Figure 6 This is a schematic diagram of order-frequency linear fitting. DETAILED DESCRIPTION

[0065] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] In the following examples, unless otherwise specified, the detection means and methods used are conventional detection means and methods in the art.

[0067] Example 1

[0068] This embodiment provides a fully automatic cable frequency extraction, order determination and cable force identification method.

[0069] Figure 1 The overall flow chart of the method is shown below. The fully automatic cable frequency extraction, order determination and cable force identification method of the present invention is further described below. The main operation is in step (S4), which is as follows:

[0070] Step (S1), using microwave radar to measure the vibration displacement time history of a certain cable;

[0071] Step (S2) uses Fast Fourier Transform (FFT) on the original data obtained in step (S1) to obtain Figure 2 The spectrum shown;

[0072] Step (S3) uses a memory-type peak screening algorithm to automatically identify the peak (characteristic frequency) in the spectrum. The specific process is as follows: Figure 3 As shown;

[0073] Step (S4) uses the hash mapping voting algorithm to extract the nominal base frequency of the cable. The specific process is as follows: Figure 4 As shown here Figure 2 The spectrum shown is used as an example to illustrate the detailed process. It is worth noting that the spectrum has identified the 6th order cable vibration frequency through step (S3), but lacks the 5th order mode.

[0074] (S401) Calculate the frequency difference of the peak sequence obtained in step (S3) and calculate the amplitude weight. As shown in Table 1, the first and second columns are Figure 1 The frequencies and their amplitudes in the , and the third and fourth columns are the differences and amplitude weights.

[0075] Table 1 Summary of data in the hash map voting algorithm

[0076]

[0077] (S402)-(S403) Traverse each element in the difference sequence and multiply by 200 to obtain the hash value in column 5. Columns 6 and 7 contain the results of rounding down and rounding up the hash value, respectively. Calculate the distance weights of two adjacent integers, as shown in columns 8 and 9. Finally, vote for the two adjacent integers, with the number of votes being the product of the magnitude weight and the distance weight.

[0078] (S404) The results of the hash voting are counted, as shown in Table 2. The results show that the integer with the most votes is 66.

[0079] Table 2 Summary of hash voting results

[0080] Ranking Candidate integer Number of votes 1 66 0.5742 2 67 0.3242 3 133 0.0576 4 134 0.0271

[0081] (S405)-(S407) Decode the results of the hash vote. Traverse the difference sequence again and select all the differences that voted for the integer with the most votes (66). Take the product of the amplitude weight and the distance weight as the total weight and perform weighted average. The final result is:

[0082]

[0083] This is the nominal fundamental frequency.

[0084] Step (S5) determines the modal order of the characteristic frequency identified in step (S3) by the modal frequency order determination algorithm. The specific process is as follows: Figure 5 The results are shown in Table 3, where the fifth-order mode is missing.

[0085] Table 3 Summary of modal frequency order determination algorithm results

[0086] characteristic frequency Modal order 0.328 1 0.660 2 0.992 3 1.323 4 1.990 6 2.321 7

[0087] Step 6: Linearly fit the determined eigenfrequency and its corresponding modal order. The result is as follows: Figure 6 The slope of the fitted straight line is 0.3308 (Hz), which is the fundamental frequency of the cable. In this embodiment, the cable length is relatively large, and the cable tension theory of cable vibration is used to calculate the cable force:

[0088]

[0089] in,

[0090]

[0091] L=314.988m

[0092] f=0.3308Hz.

[0093] On the other hand, the cable force measured by the traditional accelerometer method is 5468 kN, which is very close to the result obtained by the method of this embodiment, proving that the method described in this embodiment is feasible.

[0094] The above description of the embodiments is intended to facilitate understanding and use of the invention by those skilled in the art. It will be apparent that those skilled in the art can readily make various modifications to these embodiments and apply the general principles described herein to other embodiments without requiring inventive effort. Therefore, the present invention is not limited to the above-described embodiments. Improvements and modifications made by those skilled in the art based on the explanations of the present invention without departing from the scope of the present invention should be within the scope of protection of the present invention.

Claims

1. A fully automatic cable frequency extraction, order determination and cable force identification method, characterized in that: The following steps are involved: (S1) Collecting vibration time history data of the cable; (S2) performing spectrum analysis on the vibration time history data measured in step (S1) to obtain a displacement spectrum; (S3) identifying the characteristic frequency in the displacement spectrum obtained in step (S2) by a memory-type sieve peak algorithm to obtain the modal frequency of the cable; (S4) extracting the nominal fundamental frequency of the cable from the modal frequency of the cable in step (S3) by a voting statistical algorithm of a hash map; (S5) determining the orders of all the cable modal frequencies extracted in step (S3) according to the nominal fundamental frequency of the cable obtained in step (S4) by a modal frequency order determination algorithm; (S6) extracting the cable force based on the cable modal frequency obtained in step (S3) and the order of the cable modal frequency obtained in step (S5) in combination with the cable-beam vibration theory; In step (S4), the hash mapping voting algorithm specifically includes the following steps: (S401) Calculate the difference sequence of the characteristic frequency sequence in step (S3), that is, the frequency difference sequence; and simultaneously calculate the amplitude weight corresponding to each frequency difference, whose value is the smaller value of the spectrum amplitudes corresponding to two adjacent frequencies. The corresponding formula is: AW i = min ( A i , HE i-1 )4 in, i-index number of the characteristic frequency sequence; A-amplitude of characteristic frequency; AW-amplitude weight; (S402) Traverse each element in the frequency difference sequence and multiply the element by an integer factor to obtain a hash value; for two integers with adjacent hash values, calculate their distance weights respectively, using the formula: PW floor (val)=ceil(val)-val; PW ceil (val)=val-floor(val); in, PW-distance weight; floor - round down; ceil - round up; val - the value of an element in the sequence; (S403) voting for two integers with adjacent hash values, with the number of votes being the product of the amplitude weight and the distance weight; (S404) looping through steps (S402) and (S403) until the traversal is complete, counting the hash voting results, and finding the hash value with the most votes; (S405) traverse each element in the differential sequence again, multiply the element by the integer factor in step (S402) to obtain a hash value, and compare the hash value with the hash value with the most votes in step (S404); (S406) If the difference in step (S405) does not exceed 1, the corresponding difference can participate in the weighted average calculation of the fundamental frequency, and the weight is the product of the amplitude weight and the distance weight; (S407) Steps (S405) and (S406) are repeated until the traversal is completed. At the same time, a weighted average is performed on the characteristic frequency difference, and the result is the nominal fundamental frequency.

2. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 1 is characterized in that: In step (S1), the vibration time history data of the cable is one of cable acceleration data or cable vibration displacement data.

3. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 2 is characterized in that: The cable acceleration data is obtained by measuring with a wired or wireless acceleration sensor.

4. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 2 is characterized in that: The cable vibration displacement data is obtained through microwave radar or video technology observation.

5. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 1 is characterized in that: In step (S3), the memory-type peak screening algorithm specifically includes the following steps: (S301) Setting the size S of the screening window; (S302) defining a Boolean array B whose initial values ​​are all true; (S303) Traverse each peak in the displacement spectrum sequence. For the current index i, if B[i] is false, skip the current value and go to the next value in the sequence. (S304) For the current value, compare it with the next S / 2 points, and update the B array; (S305) For the current value, compare its size with the previous S / 2 points; (S306) After comparison, if the current value is greater than the surrounding S values, then the point value is the peak point; (S307) Repeat the process from step (S303) to step (S306) until the traversal is completed and all peak points are output.

6. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 1 is characterized in that: The screening window size S is initialized to 1.5 times the design value of the cable fundamental frequency.

7. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 1, characterized in that: The hash value is obtained by multiplying the difference by 200. The number of votes is the result of the amplitude weight and the distance weight, which is generally not an integer.

8. The method for fully automatic cable frequency extraction, order determination and cable force identification according to claim 1 is characterized in that: In step (S5), the modal frequency order determination algorithm specifically includes the following steps: (S501) using the forward pointer to traverse the characteristic frequency sequence obtained in step (S3); (S502) Divide the current frequency by the nominal fundamental frequency. If the result is approximately an integer, the first valid modal order is obtained; otherwise, return to step (S501) to continue traversal. (S503) traversing the remaining frequency sequence using the back pointer; (S504) Subtracting the frequencies corresponding to the two front and rear pointers. If the difference divided by the nominal fundamental frequency is approximately an integer, a valid modal order is obtained, and the rear pointer is used to update the front pointer. (S505) Steps (S503) and (S504) are repeated repeatedly until the traversal is completed, and all identified modal orders are output.

9. An application of the fully automatic cable frequency extraction, order determination and cable force identification method according to any one of claims 1 to 8 in cable force identification in load testing, periodic inspection and long-term monitoring of cable-stayed bridges, arch bridges and suspension bridges.

Images