An envelope extraction method for optical fiber sensing damage monitoring
By eliminating outliers and noise, using extreme value method and wavelet decomposition technology, the envelope lines in the fiber grating sensing system are extracted, which solves the problem of low efficiency caused by large data volume and high sampling frequency, and achieves rapid extraction of envelope lines and Maintain the effect of changing trend of the original signal.
Patent Information
- Application Number
- CN202211055838.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-31
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2042-08-31
AI Technical Summary
In fiber grating sensing systems, the amount of data collected during long-term damage monitoring is large and the sampling frequency is high, which makes the existing methods unable to effectively extract the envelope, low efficiency and cannot maintain the change trend of the original signal.
Through preprocessing, the outliers and noise are eliminated, the peaks and troughs are extracted using the extreme value method, and the cubic spline interpolation is performed, and the low-frequency signal is extracted in combination with wavelet decomposition to achieve rapid extraction of envelopes and maintain the change trend of the original signal.
Fast envelope extraction of large amounts of data and high sampling frequency signals is realized, which reduces monitoring time, effectively eliminates the temperature influence, and maintains the change trend of the original signal.
Smart Images

Figure CN115629074B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to an envelope extraction method for optical fiber sensing damage monitoring, and belongs to the field of signal processing in optical fiber sensing. Background Art
[0002] Fiber Bragg grating sensor is a new type of sensor with advantages such as light weight, small size, and resistance to electromagnetic interference. It is widely used in aerospace structural health monitoring, large civil construction health monitoring and other fields. In the fiber Bragg grating sensing system, the light emitted by the broadband light source is incident on the grating of the fiber Bragg grating sensor through the fiber coupler, and the light that meets the conditions will be reflected back. When a crack appears on the object being measured, the stress and strain distribution on its surface will change, causing the central wavelength of the reflected light to change, which has a great impact on the change trend of the signal envelope. Therefore, crack monitoring can be achieved by analyzing the change trend of the signal envelope.
[0003] In the patent of the equipment fault feature extraction method based on wavelet packet and Hilbert envelope spectrum analysis, a method of extracting the signal envelope spectrum of power transmission and transformation equipment using wavelet packet transform and Hilbert is proposed. The fault characteristics of the power transmission and transformation equipment are obtained by analyzing the envelope spectrum. However, in the long-term fiber grating sensor damage monitoring, the amount of data collected by the fiber grating sensor is too large and the sampling frequency is high, so that the Hilbert transform or wavelet transform cannot extract a relatively ideal envelope line, and this method cannot be used. At the same time, since the signal does not change dramatically in the initial stage of crack initiation, the extracted envelope line needs to keep the change trend of the original signal as much as possible, and the envelope line needs to be extracted based on the peak and valley values of the signal. In a patent of a signal envelope extraction method and device based on a pole-constructed low-pass filter and a patent of a spectrum envelope extraction method based on interpolation fitting, although the envelope line of the signal is also obtained by the extreme value extraction method. However, these methods need to transform the signal before extracting the extreme value. When the amount of data is large, these transformations require a lot of time, which reduces the monitoring efficiency. Summary of the invention
[0004] In view of the problems of low efficiency and unsatisfactory envelope curves in the existing methods when extracting the envelope of a large amount of data, the main purpose of the present invention is to propose an envelope extraction method for optical fiber sensing damage monitoring, which can realize the rapid extraction of envelope curves for signals containing a large amount of data and a high sampling frequency, and effectively maintain the changing trend of the original signal.
[0005] The objective of the present invention is achieved through the following technical solutions:
[0006] The invention discloses an envelope extraction method for optical fiber sensing damage monitoring. Firstly, the damage signal collected by the optical fiber grating strain sensor is preprocessed to remove abnormal values and noise in the signal, and the signal part used for envelope extraction is intercepted; all the peaks and troughs in the signal are extracted by using the extreme value method to construct the upper and lower envelopes of the signal; the upper and lower envelopes are subtracted to eliminate the influence of temperature change on the optical fiber grating strain sensor, and the extreme envelope is obtained; cubic spline interpolation is performed on the obtained extreme envelope to restore it to the length of the original signal to obtain a relatively rough envelope; finally, the low-frequency signal in the rough envelope is extracted by wavelet decomposition to obtain the signal envelope, so as to achieve the purpose of quickly extracting the envelope and effectively maintaining the change trend of the original signal.
[0007] The present invention discloses an envelope extraction method for optical fiber sensing damage monitoring, comprising the following steps:
[0008] Step 1. Use the box plot method to eliminate abnormal values in the signal collected by the fiber Bragg grating sensor.
[0009] ① Sort the data in the signal from small to large to obtain a new data set Xi;
[0010] ② Calculate the quartile Q of the data set Xi 1 , Q 2 (Median), Q 3 ;
[0011] ③Calculate the interquartile range IQR = Q 3 -Q 1 ;
[0012] ④ Calculate the upper whisker of the box plot A = Q 3 +1.5*IQR, lower whisker B=Q 1 -1.5*IQR;
[0013] ⑤ Compare each data set Xi with the upper and lower whiskers of the box plot. When a data set is larger than the upper whisker or smaller than the lower whisker, it is determined as an outlier and removed to obtain the data set Xi′ after removing the outliers.
[0014] Step 2. Use the wavelet threshold denoising method to remove burrs from the data set Xi′.
[0015] ①Use the Sym wavelet function to perform wavelet transform on the data set Xi′ to obtain the wavelet coefficients;
[0016] ② Process the obtained wavelet coefficients to remove the wavelet coefficients related to the high-frequency noise signal to remove the noise in the signal;
[0017] ③Perform inverse wavelet transform on the processed wavelet coefficients to obtain the denoised data set Xi″.
[0018] Step 3. Truncate the data set Xi″, retain the signals that can characterize the occurrence and evolution of damage, remove the signals after structural destruction, and obtain the data set Xi″′.
[0019] Step 4. Use the extreme value method to extract all the peaks and troughs of the data set Xi″′.
[0020] ① Find the opposite number of the data set Xi″′ and get the data set -Xi″′;
[0021] ②Use the findpeaks function in Matlab to obtain the maximum value X of the data set Xi″′ a and the maximum value X of the data set -Xi″′ b , then X a is the peak value of all waves in the data set Xi″′, X b are all the trough values of the data set Xi″′.
[0022] Step 5. Construct the upper envelope of the data set Xi″′ with all the peaks of the data set Xi″′, and construct the lower envelope of the data set Xi″′ with all the troughs of the data set Xi″′.
[0023] ① Take all the peak values X of the data set Xi″′ a and the upper envelope a of the data set Xi″′ is constructed based on the location of the wave peak;
[0024] ② Take all the trough values X of the data set Xi″′ b and the location of the trough to construct the lower envelope b of the data set Xi″′;
[0025] Step 6. Subtract the data of the upper envelope a from the lower envelope b to eliminate the influence of temperature on the fiber Bragg grating sensor and obtain the extreme envelope c.
[0026] Step 7. Perform cubic spline interpolation on the extreme envelope c to obtain the rough envelope d of the data set Xi″′.
[0027] ① Calculate the horizontal coordinate x of each data point of the extreme envelope c i With the vertical coordinate y i ;
[0028] ②Use the spline function in the numerical processing software to calculate the horizontal coordinate x i With the vertical coordinate y i Interpolation is performed between them to make the number of data of the extreme envelope c consistent with the number of data sets Xi″′, and the rough envelope d is obtained.
[0029] Step 8. Perform wavelet decomposition on the rough envelope d, extract the low-frequency signal, and obtain the signal envelope e of the intercepted signal, so as to achieve the purpose of quickly extracting the envelope and effectively maintaining the change trend of the original signal.
[0030] ①Use db4 wavelet function to perform wavelet decomposition on the rough envelope d to obtain the high-frequency component and low-frequency component in the signal;
[0031] ② Remove the high-frequency components in the signal and retain only the low-frequency components to obtain the signal envelope e.
[0032] Beneficial Effects
[0033] 1. The present invention discloses an envelope extraction method for optical fiber sensing damage monitoring, which uses an extreme value method to extract all peaks and troughs in a signal to construct an upper and lower envelope of the signal; the upper and lower envelopes are subtracted to eliminate the influence of temperature changes on the fiber grating strain sensor to obtain an extreme envelope; cubic spline interpolation is performed on the obtained extreme envelope to restore it to the length of the original signal to obtain a relatively rough envelope; on the basis of extreme value interpolation to obtain the envelope, wavelet decomposition is introduced to extract the low-frequency signal in the rough envelope to obtain the signal envelope, thereby realizing rapid extraction of the envelope of a signal containing a large amount of data and a high sampling frequency.
[0034] 2. The present invention discloses an envelope extraction method for optical fiber sensing damage monitoring. Based on beneficial effect 1, it does not require Hilbert transform or wavelet transform of the signal, and can effectively maintain the changing trend of the original signal, so that subsequent damage monitoring has a basis. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:
[0036] Figure 1 4 is an original signal waveform diagram of an embodiment of the present invention.
[0037] Figure 2 It is the envelope extracted by using Hilbert transform in the embodiment of the present invention.
[0038] Figure 3 It is the envelope extracted by the extreme value interpolation method in the embodiment of the present invention.
[0039] Figure 4 It is a burr inside the signal of the embodiment of the present invention.
[0040] Figure 5 It is a waveform diagram after outlier removal and wavelet threshold denoising according to an embodiment of the present invention.
[0041] Figure 6 It is a signal waveform diagram after interception according to an embodiment of the present invention.
[0042] Figure 7 It is the rough envelope curve after cubic spline interpolation in the embodiment of the present invention.
[0043] Figure 8 It is the signal envelope obtained after wavelet decomposition in the embodiment of the present invention.
[0044] Fig. 9 3 is a comparison between the signal envelope of the embodiment of the present invention and the intercepted signal.
[0045] Fig.10 4 is a flow chart of envelope extraction according to an embodiment of the present invention. DETAILED DESCRIPTION
[0046] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. The technical problems solved by the technical solution of the present invention and the beneficial effects are also described. It should be noted that the described embodiments are only intended to facilitate the understanding of the present invention and do not have any limiting effect on it.
[0047] When monitoring fatigue cracks in aluminum alloy plates, the original signal waveform collected by the fiber grating sensor is shown in the figure below: Figure 1 As shown in Figure 1, due to the chirp effect of the fiber Bragg grating, the signal contains a large number of abnormal values 0. When the envelope is extracted using the Hilbert transform, the following is obtained: Figure 2 The envelope shown in . Due to the large amount of data and high sampling frequency, the envelope overlaps with the original signal and cannot be further analyzed. When using extreme value interpolation to obtain the envelope, the following is obtained: Figure 3 The envelope curve shown in Figure 1 is shown in Figure 2. Also, due to the large amount of signal data, the envelope curve obtained not only has many internal fluctuations, but also contains burrs. In addition, there will be a small number of abnormal waveforms during the debugging stage of the sensor. These abnormal values and abnormal waveforms need to be eliminated during the preprocessing process.
[0048] Step 1. A fiber Bragg grating strain sensor with a central wavelength of 1546 nm is used for fatigue crack monitoring of aluminum alloy plates. During the monitoring process, the chirp effect of the fiber Bragg grating causes an abnormal value of 0. The upper whisker of the signal is calculated to be 1563.5 and the lower whisker is 1544.5 using the box plot method. The abnormal values exceeding the upper and lower whiskers of the box plot are eliminated to avoid misjudgment in the subsequent envelope analysis.
[0049] Step 2. In addition to the outlier value 0, there are also Figure 4 The burrs shown in the figure are removed by using the wavelet threshold denoising method to remove the burrs of the signal from which the outliers have been removed. The larger burrs inside the waveform are removed to make the waveform smooth, avoiding the extraction of the extreme values of the burr part when extracting the extreme values. The waveform after outlier removal and wavelet threshold denoising is shown in the figure below. Figure 5 shown.
[0050] Step 3. After the outliers are removed and the noise is removed, the signal is intercepted, and the data before 27,000 is retained, and the signal after the structure is damaged is removed. This data contains information about the occurrence and evolution of damage. At the same time, the amount of data is reduced and the speed of envelope extraction is improved. The intercepted signal is as follows: Figure 6 shown.
[0051] Step 4. Use the peak extraction function in Matlab to extract all the peaks and troughs of the intercepted signal.
[0052] Step 5. Construct an upper envelope of the signal by intercepting the peak of the signal, and construct a lower envelope of the signal by intercepting the trough of the signal.
[0053] Step 6. Subtract the data of the upper envelope and the lower envelope to eliminate the influence of temperature on the fiber Bragg grating sensor and obtain the extreme envelope.
[0054] Step 7. Since the extreme envelope is the line connecting the difference between the maximum and minimum values of the intercepted waveform, its length is different from the length of the intercepted signal, so it is necessary to perform cubic spline interpolation on the extreme envelope to restore the length of the extreme envelope to the length of the intercepted signal.
[0055] The rough envelope obtained after processing in steps 4, 5, 6 and 7 is as follows: Figure 7 shown.
[0056] Step 8. The rough envelope obtained by interpolation is still a sine waveform. It is necessary to perform wavelet decomposition on the rough envelope to extract the low-frequency signal and obtain the change trend of the rough envelope, so as to obtain Figure 8 The signal envelope of the captured signal is shown.
[0057] The intercepted waveform, upper and lower envelopes, and envelopes are processed as follows Fig. 9 As shown in the comparison, the envelope is slightly smoother than the intercepted waveform due to the weakening of the influence of temperature on the sensor, and the change trend of the original signal is maintained, which can reflect the process of structural damage from appearance, evolution and intensification to structural destruction.
[0058] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. An envelope extraction method for optical fiber sensing damage monitoring, Features: The method comprises the following steps: Step 1. Use the box plot method to remove abnormal values in the signal collected by the fiber Bragg grating sensor to obtain a data set Xi' without abnormal values; Step 2. Use the wavelet threshold denoising method to remove burrs from the data set Xi' to obtain the denoised data set Xi"; Step 3. Intercept the data set Xi”, retain the signals that can characterize the occurrence and evolution of damage, remove the signals after structural damage, and obtain the data set Xi”’; Step 4. Use the extreme value method to extract all the peaks and troughs of the data set Xi”’; Step 5. Construct the upper envelope of the data set Xi"' with all the peaks of the data set Xi"', and construct the lower envelope of the data set Xi"' with all the troughs of the data set Xi"'; Step 6. Subtract the data of the upper envelope a from the lower envelope b to eliminate the influence of temperature on the fiber Bragg grating sensor and obtain the extreme envelope c; Step 7. Perform cubic spline interpolation on the extreme envelope c to obtain the rough envelope d of the data set Xi''; Step 8. Perform wavelet decomposition on the rough envelope d, extract the low-frequency signal, and obtain the signal envelope e of the intercepted signal, so as to achieve the purpose of quickly extracting the envelope and effectively maintaining the change trend of the original signal.
2. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 1 is ① Sort the data in the signal from small to large to obtain a new data set Xi; ② Calculate the quartile Q of the data set Xi 1 , Q 2 , Q 3 ; ③Calculate the interquartile range IQR = Q 3 -Q 1 ; ④ Calculate the upper whisker of the box plot A=Q 3 +1.5*IQR, lower whisker B=Q 1 -1.5*IQR; ⑤ Compare each data in the data set Xi with the upper and lower whiskers of the box plot. When a data is larger than the upper whisker or smaller than the lower whisker, it is determined as an outlier and removed to obtain the data set Xi' without the outliers.
3. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 2 is ①Use the Sym wavelet function to perform wavelet transform on the data set Xi' to obtain the wavelet coefficients; ②Process the obtained wavelet coefficients to remove the wavelet coefficients related to the high-frequency noise signal to remove the noise in the signal; ③Perform inverse wavelet transform on the processed wavelet coefficients to obtain the denoised data set Xi".
4. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 4 is ① Find the opposite number of the data set Xi”’ and get the data set -Xi”’; ②Use the findpeaks function in the numerical processing software to obtain the maximum value X of the data set Xi”’ a And the maximum value X of the data set -Xi"' b , then X a is the peak value of all waves in the data set Xi”’, X b are all the trough values of the data set Xi”’.
5. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 5 is ① Take all the peak values X of the data set Xi”’ a The upper envelope a of the data set Xi"' is constructed based on the location of the wave peak; ② Take all the trough values X of the data set Xi”’ b The lower envelope b of the data set Xi"' is constructed based on the location of the trough.
6. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 7 is ① Calculate the horizontal coordinate x of each data point of the extreme envelope c i With the vertical coordinate y i ; ②Use the spline function in Matlab to calculate the horizontal coordinate x i With the vertical coordinate y i Interpolate between them to make the number of data of the extreme envelope c consistent with the number of data sets Xi”’, and obtain the rough envelope d.
7. The envelope extraction method for optical fiber sensing damage monitoring according to claim 1, Features: The implementation method of step 8 is ①Use db4 wavelet function to perform wavelet decomposition on the rough envelope d to obtain the high-frequency component and low-frequency component in the signal; ② Remove the high-frequency components in the signal and retain only the low-frequency components to obtain the signal envelope e.
