A method for denoising laser absorption spectroscopy signals based on improved local mean decomposition

Abstract

The present invention discloses a laser absorption spectrum signal denoising method based on improved local mean decomposition, comprising: obtaining a standard amplitude and a standard second harmonic signal, and obtaining the amplitude and the second harmonic signal of the laser absorption spectrum signal of the trace gas to be measured x ( k ) ; respectively for the second harmonic signal x ( k ) add two groups of white noise with equal size and opposite sign, perform local mean decomposition on the two groups of signals after adding white noise, and calculate the second harmonic signal according to the corresponding main factor component set x ( k ) corresponding to the main factor component set, from which useful main factor components are selected for reconstruction to obtain the second harmonic signal after noise reduction y ( k ), calculate the second harmonic signal y ( k ), and then correct the gas concentration corresponding to the trace gas to be measured according to the standard amplitude and the amplitude of the trace gas to be measured to obtain the corrected gas concentration of the trace gas to be measured. The present invention can improve the detection accuracy and the lower limit of measurement, and has a small amount of calculation, which is convenient for engineering application.

Description

Technical Field

[0001] The invention relates to the technical field of laser gas detection, and in particular to a laser absorption spectrum signal denoising method based on improved local mean decomposition. Background Art

[0002] Tunable diode laser absorption spectroscopy (TDLAS) technology has the advantages of non-contact measurement and anti-electromagnetic interference, and is widely used in environmental testing, industrial process control, national defense security, pathological diagnosis, etc. In recent years, it has also been gradually used in online monitoring of power equipment such as transformers and reactors. Gas analysis in insulating oil of transformers, reactors, etc. is an effective method to evaluate their operating status and identify early faults. For transformers, the main fault characteristic value gases in their insulating oil include acetylene, methane, ethylene, ethane, carbon monoxide and carbon dioxide. In view of the problems of high maintenance cost and susceptibility to vibration or noise in existing methods for monitoring dissolved gas in transformer insulating oil such as gas chromatography and photoacoustic spectroscopy, the detection of dissolved gas in transformer insulating oil based on TDLAS technology has the advantages of maintenance-free and good stability, especially in the near-infrared band. It also has the advantages of low cost and simple application development, and has a good development prospect.

[0003] The TDLAS measurement process is disturbed by the noise from the background environment, laser, detector, and "etalon effect". Especially in the measurement of trace gases, the signal-to-noise ratio is reduced due to the small effective component, making it difficult for the system to achieve high-precision detection of trace gases, and reducing the accuracy of early fault warning of transformers. In order to extract a stronger effective signal and improve the detection accuracy, the laser absorption spectrum signal needs to be denoised. In recent years, some studies have proposed methods such as principal component analysis, Gabor transform, and singular value decomposition matrix analysis to improve the signal-to-noise ratio of TDLAS signals, but the calculation is large. Other studies have proposed a spectral denoising method based on empirical mode decomposition (EMD), but the EMD method has problems such as endpoint effect, mode aliasing, over-envelope and under-envelope. Local mean decomposition (LMD), as a new adaptive signal decomposition method that has emerged in recent years, can effectively improve the above problems of EMD, but there are still problems of mode aliasing and endpoint effect.

[0004] When applying TDLAS technology, many factors such as long-term frequent use, ambient temperature fluctuations, changes in photoelectric conversion efficiency, changes in internal fiber loss such as microbending and aging can cause changes in laser optical power, resulting in inaccurate gas concentration detection. The existing method uses a phase-locked amplifier to demodulate the photoelectric conversion signal to obtain the first harmonic and second harmonic, and uses the second harmonic / first harmonic to correct the impact caused by the change in optical power. The use of second harmonic / first harmonic requires the first harmonic and second harmonic to be demodulated by a phase-locked amplifier. This method has a large amount of calculation and is difficult to implement in embedded applications, making it difficult to achieve engineering applications. Summary of the invention

[0005] The technical problem to be solved by the present invention is as follows: In view of the above problems in the prior art, a spectral denoising method based on improved local mean decomposition is provided, which can improve detection accuracy and measurement lower limit, and has small calculation amount and is convenient for engineering application.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A laser absorption spectrum signal denoising method based on improved local mean decomposition comprises the following steps:

[0008] Get the standard amplitude A s And the standard second harmonic signal s ( k ), and obtain the amplitude of the laser absorption spectrum signal of the trace gas to be measured A m And the second harmonic signal x ( k );

[0009] The second harmonic signal x ( k ) to add two groups of white noise with equal magnitude and opposite signs, and obtain two groups of signals after adding white noise x 1( k ), x 2( k );

[0010] For the signal x 1( k ), x 2( k ) perform local mean decomposition respectively and obtain the signal x 1( k ), x 2( k ) corresponds to the set of main factor components, according to the signal x 1( k ), x 2( k ) The corresponding main factor component set is calculated to obtain the second harmonic signal x ( k ) corresponds to the set of principal factor components;

[0011] From the second harmonic signal x ( k ) Select useful principal factor components from the corresponding principal factor component set for reconstruction to obtain the second harmonic signal after noise reduction y ( k), the useful main factor component specifically refers to the standard second harmonic signal s ( k )’s correlation meets the required main factor components;

[0012] Calculate the second harmonic signal y ( k ) corresponds to the gas concentration C m , and then according to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m Correction is performed to obtain the corrected gas concentration of the trace gas to be measured C s .

[0013] Furthermore, the standard amplitude A s And the standard second harmonic signal s ( k ) specifically refers to the amplitude of the laser absorption spectrum signal and the second harmonic signal of the high-concentration gas with a concentration greater than the preset threshold, and the standard amplitude is obtained A s And the standard second harmonic signal s ( k ) and obtain the amplitude of the laser absorption spectrum detection result of the trace gas to be tested A m And the second harmonic signal x ( k ), including:

[0014] The laser absorption spectroscopy technique is used to detect the gas to obtain a corresponding photoelectric conversion signal, the photoelectric conversion signal is subjected to signal amplitude calculation to obtain a corresponding amplitude, and the photoelectric conversion signal is input into a phase-locked amplifier for signal analysis and processing to obtain a corresponding second harmonic signal.

[0015] Further, when the photoelectric conversion signal is subjected to signal amplitude calculation to obtain a corresponding amplitude, it specifically includes:

[0016] Select a specified number of measurement cycles without absorption grooves in the photoelectric conversion signal, calculate the difference between the maximum and minimum values ​​of the photoelectric conversion signal in each selected measurement cycle as the amplitude of each selected measurement cycle, and calculate the average value of the amplitudes of all selected measurement cycles as the amplitude corresponding to the photoelectric conversion signal.

[0017] Furthermore, the measurement frequency expression is as follows:

[0018] T = f s / f 0

[0019] in, T To measure the frequency, f s is the sampling frequency of the photoelectric conversion signal, f 0 is the frequency of the photoelectric conversion signal.

[0020] Furthermore, according to the signal x 1( k ), x 2( k ) The corresponding main factor component set is calculated to obtain the second harmonic signal x ( k ) is the corresponding principal factor component set, the expression is as follows:

[0021] PF m =[ + ] / 2

[0022] Among them, PF m is the second harmonic signal x ( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 1( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 2( k ) in the set of principal factor components corresponding to m main factor components.

[0023] Furthermore, from the second harmonic signal x ( k ) When selecting useful principal factor components from the corresponding principal factor component set for reconstruction, it includes:

[0024] Calculate the second harmonic signal separately x ( k ) Each principal factor component of the corresponding principal factor component set is consistent with the standard second harmonic signal s ( k )’s correlation coefficient;

[0025] The optimal correlation coefficient threshold is selected from the specified correlation coefficient threshold range, and the optimal correlation coefficient threshold is selected according to the correlation coefficient between the main factor component and the standard second harmonic signal. s ( k) and the optimal correlation coefficient threshold, and set the weight value of each main factor component;

[0026] Calculate the second harmonic signal x ( k ) The weighted sum of each principal factor component of the corresponding principal factor component set is used to obtain the denoised second harmonic signal y ( k ).

[0027] Furthermore, when selecting the optimal correlation coefficient threshold from the specified correlation coefficient threshold range, it includes:

[0028] From each main factor component and the standard second harmonic signal s ( k ) are selected from the correlation coefficients greater than 0 to form a correlation coefficient set;

[0029] Traversing the specified correlation coefficient threshold range, adding correlation coefficients in the correlation coefficient set that are greater than or equal to the current correlation coefficient threshold to the first correlation coefficient set, adding correlation coefficients in the correlation coefficient set that are less than the current correlation coefficient threshold to the second correlation coefficient set, and calculating the variance of the first correlation coefficient set and the second correlation coefficient set corresponding to the current correlation coefficient threshold;

[0030] The correlation coefficient threshold with the largest variance in the specified correlation coefficient threshold range is selected as the optimal correlation coefficient threshold.

[0031] Furthermore, when calculating the variance of the first correlation coefficient set and the second correlation coefficient set corresponding to the current correlation coefficient threshold, the expression is as follows:

[0032]

[0033] in, and are the weights of the first correlation coefficient set and the second correlation coefficient set, respectively. The result of dividing the number of elements in the first correlation coefficient set by the number of elements in the correlation coefficient set. is the result of dividing the number of elements in the second correlation coefficient set by the number of elements in the correlation coefficient set, is the average value of the elements in the first correlation coefficient set, is the average value of the elements in the second correlation coefficient set, i It is the sequence number of the current correlation coefficient threshold in the correlation coefficient threshold range.

[0034] Further, according to the main factor component and the standard second harmonic signal s ( k) and the optimal correlation coefficient threshold, when setting the weight value of each main factor component, including: if the main factor component is s ( k ) is greater than or equal to the optimal correlation coefficient threshold, the weight value of the main factor component is set to 1, if the main factor component is s ( k ) is less than the optimal correlation coefficient threshold, the weight value of the main factor component is set to 0.

[0035] Furthermore, according to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m When making corrections, specifically calculate the gas concentration C m With standard amplitude A s Multiply the product and divide by the amplitude of the trace gas to be measured A m , and obtain the corrected gas concentration of the trace gas to be measured C s .

[0036] Compared with the prior art, the advantages of the present invention are:

[0037] The present invention makes the data extreme point distribution more uniform by adding white noise to the original signal before performing local mean decomposition on the second harmonic signal of the laser absorption spectrum signal of the gas to be measured, thereby improving the problems of modal aliasing, and by adding a pair of mutually opposite positive and negative white noises as auxiliary noises to the original signal, the auxiliary white noise remaining in the reconstructed signal caused by adding white noise can be eliminated, and has the advantages of small reconstruction error and small calculation amount. In addition, by selecting the effective components in the local mean decomposition result to complete the signal reconstruction, the signal noise can be effectively reduced.

[0038] After calculating the concentration of the gas to be measured, the present invention corrects the concentration of the gas to be measured by the photoelectric conversion signal amplitude of the gas to be measured and the standard amplitude. The influence caused by the optical power change is corrected based on the photoelectric conversion signal amplitude, the calculation amount is small, and it is convenient for engineering application. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the protection scope of the present invention is not limited thereby.

[0041] This embodiment proposes a laser absorption spectrum signal denoising method based on improved local mean decomposition, which realizes the denoising analysis of the laser spectrum signal by improving the local mean decomposition method, and corrects the influence of the optical power change based on the original signal amplitude, such as Figure 1 As shown, the method comprises the following steps:

[0042] S1) Get the standard amplitude A s And the standard second harmonic signal s ( k ), and obtain the amplitude of the laser absorption spectrum signal of the trace gas to be measured A m And the second harmonic signal x ( k );

[0043] S2) for the second harmonic signal x ( k ) to add two groups of white noise with equal magnitude and opposite signs, and obtain two groups of signals after adding white noise x 1( k ), x 2( k );

[0044] S3) for the signal x 1( k ), x 2( k ) perform local mean decomposition respectively and obtain the signal x 1( k ), x 2( k ) corresponds to the set of main factor components, according to the signal x 1( k ), x 2( k ) The corresponding main factor component set is calculated to obtain the second harmonic signal x ( k ) corresponds to the set of principal factor components;

[0045] S4) from the second harmonic signal x ( k ) Select useful principal factor components from the corresponding principal factor component set for reconstruction to obtain the second harmonic signal after noise reduction y ( k ), the useful main factor component specifically refers to the standard second harmonic signal s ( k)’s correlation meets the required main factor components;

[0046] S5) Calculate the second harmonic signal y ( k ) corresponds to the gas concentration C m , and then according to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m Correction is performed to obtain the corrected gas concentration of the trace gas to be measured C s .

[0047] Each step is described in detail below.

[0048] In step S1 of this embodiment, the standard amplitude A s And the standard second harmonic signal s ( k ) can be obtained through experiments. In this embodiment, the amplitude of the laser absorption spectrum signal of the high-concentration gas and the second harmonic signal are obtained as the standard amplitude. A s And the standard second harmonic signal s ( k ), the high concentration gas can refer to the upper limit of the gas detection range of each gas in the standard "QGDW 10536-2021-Technical Specifications for Online Monitoring Devices for Dissolved Gases in Transformer Oil", and step S1 specifically includes:

[0049] S11) Using laser absorption spectroscopy technology to detect high-concentration gas, obtain a corresponding photoelectric conversion signal, and calculate the signal amplitude of the photoelectric conversion signal through a signal amplitude calculation module to obtain a corresponding standard amplitude A s At the same time, the photoelectric conversion signal is processed by the signal analysis module to obtain the corresponding standard second harmonic signal s ( k ), k =1,2,…, N , N is the signal length;

[0050] S12) Using laser absorption spectroscopy technology to detect trace amounts of the gas to be tested, obtaining a corresponding photoelectric conversion signal, and performing signal amplitude calculation on the photoelectric conversion signal through a signal amplitude calculation module to obtain a corresponding standard amplitude A mAt the same time, the photoelectric conversion signal is processed by the signal analysis module to obtain the corresponding standard second harmonic signal x ( k ), k =1,2,…, N , N is the signal length.

[0051] In step S11 and step S12 of this embodiment, the signal analysis and processing module used specifically obtains the second harmonic signal of the laser absorption spectrum signal through a phase-locked amplifier, and its working principle is well known to those skilled in the art and will not be repeated in this embodiment.

[0052] In step S11 and step S12 of this embodiment, the process of the signal amplitude calculation module performing signal amplitude calculation specifically includes:

[0053] S101) Selecting a specified number of measurement cycles without absorption concave lines in the photoelectric conversion signal, specifically by selecting the measurement cycles without absorption concave lines in the photoelectric conversion signal. N 1 measurement cycle T , where the measurement frequency T = f s / f 0, f s is the sampling frequency of the photoelectric conversion signal, f 0 is the frequency of the photoelectric conversion signal, N 1 is a positive integer;

[0054] S102) Calculate the difference between the maximum and minimum values ​​of the photoelectric conversion signal in each selected measurement cycle as the amplitude of each selected measurement cycle. i The amplitude of a cycle is A i , i =1, 2, …, N 1;

[0055] S103) Calculate the average value of the amplitudes of all selected measurement cycles as the amplitude corresponding to the photoelectric conversion signal, and the expression is as follows:

[0056]

[0057] in, A is the amplitude corresponding to the photoelectric conversion signal, that is, the laser absorption spectrum signal.

[0058] In step S2 of this embodiment, the two groups of signals after adding white noise x 1( k ), x 2(k ) is expressed as follows:

[0059]

[0060]

[0061] in, σ ( k )and- σ ( k ) are two groups of white noise with equal magnitude and opposite signs.

[0062] In step S3 of this embodiment, the signal x 1( k ), x 2( k ) respectively perform local mean decomposition to obtain the primary factor PF (Primary Factor) component set and residual corresponding to the two groups of signals, where:

[0063] Signal x 1( k ) The corresponding principal factor component set can be expressed as { , m =1,2,…, M}, the residual is r1( k ),Signal x 2( k ) The corresponding principal factor component set can be expressed as { , m =1,2,…, M}, the residual is r2( k ), M Represents the number of principal factor components in the principal factor component set.

[0064] In this embodiment, the second harmonic signal x ( k ) The corresponding principal factor component set and residual are denoted as {PF m , m =1,2,…, M} and r( k ), according to the signal x 1( k ), x 2( k ) The corresponding main factor component set is calculated to obtain the second harmonic signal x ( k ) corresponding to the main factor component set, the expression is as follows:

[0065] PF m =[ + ] / 2

[0066] Among them, PF m is the second harmonic signal x ( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 1( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 2( k ) in the set of principal factor components corresponding to m main factor components.

[0067] In addition, the second harmonic signal x ( k ) corresponding to the residual r( k ) According to the signal x 1( k ), x 2( k ) The corresponding residual is calculated and the expression is as follows:

[0068] r( k )=[r1( k )+r2( k )] / 2

[0069] Among them, r1( k )、r2( k ) are signals x 1( k ), x 2( k ) corresponding to the residual.

[0070] In step S4 of this embodiment, from the second harmonic signal x ( k ) When selecting useful principal factor components from the corresponding principal factor component set for reconstruction, it includes:

[0071] S41) calculate the second harmonic signal separately x ( k ) Each principal factor component of the corresponding principal factor component set is consistent with the standard second harmonic signal s ( k )’s correlation coefficient;

[0072] S42) Filtering an optimal correlation coefficient threshold from a specified correlation coefficient threshold range, including:

[0073] S421) From each main factor component and the standard second harmonic signal s (k ) are selected from the correlation coefficients greater than 0 to form a correlation coefficient set, denoted as { c j , j =1,2,…, J};

[0074] S422) traverse the specified correlation coefficient threshold range, add the correlation coefficients in the correlation coefficient set that are greater than or equal to the current correlation coefficient threshold to the first correlation coefficient set, add the correlation coefficients in the correlation coefficient set that are less than the current correlation coefficient threshold to the second correlation coefficient set, and calculate the variance of the first correlation coefficient set and the second correlation coefficient set corresponding to the current correlation coefficient threshold. The expression is as follows:

[0075]

[0076] in, and are the weights of the first correlation coefficient set and the second correlation coefficient set, respectively. The result of dividing the number of elements in the first correlation coefficient set by the number of elements in the correlation coefficient set. is the result of dividing the number of elements in the second correlation coefficient set by the number of elements in the correlation coefficient set, is the average value of the elements in the first correlation coefficient set, is the average value of the elements in the second correlation coefficient set, i is the sequence number of the current correlation coefficient threshold in the correlation coefficient threshold range;

[0077] S423) Selecting the correlation coefficient threshold with the largest variance in the specified correlation coefficient threshold range as the optimal correlation coefficient threshold T h .

[0078] In this embodiment, considering that when the correlation coefficient is less than 0.3, the correlation between the two is very weak, and when the correlation coefficient is greater than 0.8, the correlation between the two is very strong, and in order to simplify the calculation, the value range of the correlation coefficient threshold is set between 0.4 and 0.8. T ={0.4, 0.5, 0.6, 0.7, 0.8}.

[0079] Correspondingly, the first correlation coefficient set ={ c j ≥ T i ( i =1,2,3,4,5)}, the second correlation coefficient set ={ cj < T i The relative coefficient set of}, weight = Number of elements in J , = Number of elements in J , = The average value of the elements in = The average value of the elements in the first correlation coefficient set is used to calculate the relative coefficient and in the second set of correlation coefficients The population variance in g i ,in i =1,2,3,4,5, select variance g i The maximum correlation coefficient threshold T i As the optimal correlation coefficient threshold T h ;

[0080] S43) According to the main factor component and the standard second harmonic signal s ( k ) and the optimal correlation coefficient threshold, setting the weight value of each main factor component, including: if the main factor component is s ( k ) is greater than or equal to the optimal correlation coefficient threshold T h , set the weight value of the main factor component to 1, if the main factor component is consistent with the standard second harmonic signal s ( k ) is less than the optimal correlation coefficient threshold T h , set the weight value of the main factor component to 0, so the weight of the main factor component w m It can be expressed as:

[0081]

[0082] in, c j Represents the second harmonic signal x ( k ) corresponds to the set of principal factor components j The main factor components and the standard second harmonic signal s ( k )’s correlation coefficient;

[0083] S44) Calculate the second harmonic signal x ( k ) The weighted sum of each principal factor component of the corresponding principal factor component set is used to obtain the denoised second harmonic signal y ( k ), the expression is as follows:

[0084]

[0085] Among them, PF m is the second harmonic signal x ( k ) in the set of principal factor components corresponding to m The main factor components, w m is the second harmonic signal x ( k ) in the set of principal factor components corresponding to m The weights of the main factor components.

[0086] In step S5 of this embodiment, the second harmonic signal is calculated. y ( k ) corresponds to the gas concentration C m When calculating the characteristic value of gas concentration using the absorption peak, that is, after gas absorption, an absorption peak will appear on the second harmonic, and the difference between the peak and the trough of the absorption peak is used as the characteristic value of gas concentration. Then, based on the calibration, the gas concentration measurement is completed. Specifically, several gases with different concentrations are used to calculate the corresponding characteristic values, and the relationship between the gas concentration and the characteristic value is established (generally a linear relationship). This process is called calibration. After the calibration is completed, it can be used to measure the gas concentration. It should be noted that the specific implementation process of calculating the gas concentration based on the second harmonic signal is well known to those skilled in the art, and the specific implementation process of the above content will not be described in detail in this embodiment.

[0087] In step S5 of this embodiment, according to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m When making corrections, specifically calculate the gas concentration C m With standard amplitude A s Multiply the product and divide by the amplitude of the trace gas to be measured A m , and obtain the corrected gas concentration of the trace gas to be measured C s, the formula is as follows:

[0088]

[0089] in, A s is the standard amplitude, A m is the amplitude of the trace gas to be measured, C m is the gas concentration.

[0090] In summary, the present invention discloses a laser absorption spectrum signal denoising method based on improved local mean decomposition. Aiming at the problems of modal aliasing and endpoint effect existing in local mean decomposition (LMD) during signal denoising, the present invention realizes denoising analysis of laser spectrum signal by improving the local mean decomposition method. Specifically, white noise is added to the original signal to make the distribution of data extreme points more uniform, thereby improving the problems of modal aliasing, and by adding a pair of mutually opposite positive and negative white noises as auxiliary noises to the original signal, the auxiliary white noise remaining in the reconstructed signal caused by adding only white noise can be eliminated, which has the advantage of small reconstruction error. In addition, by selecting the effective components in LMD to complete signal reconstruction, the signal noise can be effectively reduced.

[0091] In order to solve the problem of inaccurate gas concentration detection caused by laser optical power variation, the present invention corrects the influence of optical power variation based on the amplitude of photoelectric conversion signal, with small calculation amount and convenient for engineering application.

[0092] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A laser absorption spectroscopy signal denoising method based on improved local mean decomposition, characterized in that: The following steps are involved: Get the standard amplitude A s And the standard second harmonic signal s ( k ), the standard amplitude A s And the standard second harmonic signal s ( k ) specifically refers to the amplitude of the laser absorption spectrum signal and the second harmonic signal of the high-concentration gas whose concentration is greater than the preset threshold, and obtains the amplitude of the laser absorption spectrum signal of the trace gas to be measured A m And the second harmonic signal x ( k ), get the standard amplitude A s And the standard second harmonic signal s ( k ) and obtain the amplitude of the laser absorption spectrum detection result of the trace gas to be tested A m And the second harmonic signal x ( k ), including: The gas is detected using laser absorption spectroscopy technology to obtain a corresponding photoelectric conversion signal, the photoelectric conversion signal is subjected to signal amplitude calculation to obtain a corresponding amplitude, and the photoelectric conversion signal is input into a lock-in amplifier for signal analysis and processing to obtain a corresponding second harmonic signal; The second harmonic signal x ( k ) to add two groups of white noise with equal magnitude and opposite signs, and obtain two groups of signals after adding white noise x 1( k ), x 2( k ); For the signal x 1( k ), x 2( k ) perform local mean decomposition respectively and obtain the signal x 1( k ), x 2( k ) corresponds to the set of main factor components, according to the signal x 1( k ), x 2( k ) The corresponding main factor component set is calculated to obtain the second harmonic signal x ( k ) The corresponding main factor component set is expressed as follows: PF m =[ + ] / 2 Among them, PF m is the second harmonic signal x ( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 1( k ) in the set of principal factor components corresponding to m The main factor components, For signal x 2( k ) in the set of principal factor components corresponding to m principal factor components; From the second harmonic signal x ( k ) Select useful principal factor components from the corresponding principal factor component set for reconstruction to obtain the second harmonic signal after noise reduction y ( k ), the useful main factor component specifically refers to the standard second harmonic signal s ( k ) meets the required main factor components, from the second harmonic signal x ( k ) When selecting useful principal factor components from the corresponding principal factor component set for reconstruction, it includes: Calculate the second harmonic signal separately x ( k ) Each principal factor component of the corresponding principal factor component set is consistent with the standard second harmonic signal s ( k )’s correlation coefficient; The optimal correlation coefficient threshold is selected from the specified correlation coefficient threshold range, and the optimal correlation coefficient threshold is selected according to the correlation coefficient between the main factor component and the standard second harmonic signal. s ( k ) and the optimal correlation coefficient threshold, and set the weight value of each main factor component; Calculate the second harmonic signal x ( k ) The weighted sum of each principal factor component of the corresponding principal factor component set is used to obtain the denoised second harmonic signal y ( k ); Calculate the second harmonic signal y ( k ) corresponds to the gas concentration C m , and then according to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m Correction is performed to obtain the corrected gas concentration of the trace gas to be measured C s .

2. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 1, characterized in that: When the photoelectric conversion signal is subjected to signal amplitude calculation to obtain a corresponding amplitude, the method specifically includes: Select a specified number of measurement cycles without absorption grooves in the photoelectric conversion signal, calculate the difference between the maximum and minimum values ​​of the photoelectric conversion signal in each selected measurement cycle as the amplitude of each selected measurement cycle, and calculate the average value of the amplitudes of all selected measurement cycles as the amplitude corresponding to the photoelectric conversion signal.

3. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 2, characterized in that: The measurement frequency expression is as follows: T = f s / f 0 in, T To measure the frequency, f s is the sampling frequency of the photoelectric conversion signal, f 0 is the frequency of the photoelectric conversion signal.

4. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 1, characterized in that: When filtering the optimal correlation coefficient threshold from the specified correlation coefficient threshold range, include: From each main factor component and the standard second harmonic signal s ( k )'s correlation coefficients, select the correlation coefficients greater than 0 to form a correlation coefficient set; Traversing the specified correlation coefficient threshold range, adding correlation coefficients in the correlation coefficient set that are greater than or equal to the current correlation coefficient threshold to the first correlation coefficient set, adding correlation coefficients in the correlation coefficient set that are less than the current correlation coefficient threshold to the second correlation coefficient set, and calculating the variance of the first correlation coefficient set and the second correlation coefficient set corresponding to the current correlation coefficient threshold; The correlation coefficient threshold with the largest variance in the specified correlation coefficient threshold range is selected as the optimal correlation coefficient threshold.

5. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 4, characterized in that: When calculating the variance of the first correlation coefficient set and the second correlation coefficient set corresponding to the current correlation coefficient threshold, the expression is as follows: in, and are the weights of the first correlation coefficient set and the second correlation coefficient set, respectively. The result of dividing the number of elements in the first correlation coefficient set by the number of elements in the correlation coefficient set. is the result of dividing the number of elements in the second correlation coefficient set by the number of elements in the correlation coefficient set, is the average value of the elements in the first correlation coefficient set, is the average value of the elements in the second correlation coefficient set, i It is the sequence number of the current correlation coefficient threshold in the correlation coefficient threshold range.

6. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 1, characterized in that: According to the main factor component and the standard second harmonic signal s ( k ) and the optimal correlation coefficient threshold, when setting the weight value of each main factor component, including: if the main factor component is s ( k ) is greater than or equal to the optimal correlation coefficient threshold, the weight value of the main factor component is set to 1, if the main factor component is s ( k ) is less than the optimal correlation coefficient threshold, the weight value of the main factor component is set to 0.

7. The laser absorption spectrum signal denoising method based on improved local mean decomposition according to claim 1, characterized in that: According to the standard amplitude A s , the amplitude of the trace gas to be measured A m Gas concentration C m When making corrections, specifically calculate the gas concentration C m With standard amplitude A s Multiply the product and divide by the amplitude of the trace gas to be measured A m , and obtain the corrected gas concentration of the trace gas to be measured C s .

Images