A gas detection system and method for transformer oil with high anti-interference capability
By using the oil-gas separation module and photoacoustic module in the transformer oil combined with the data analysis module, the detection results are corrected by using the interference influence coefficient of C3 gas, which solves the problem of gas detection in the transformer oil being easily disturbed by C3 gas, and improves the accuracy of detection sex.
Patent Information
- Application Number
- CN202210356855.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-06
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2042-04-06
AI Technical Summary
In the prior art, the gas detection method in transformer oil is susceptible to interference from C3 gas, resulting in inaccurate analysis results, which in turn affects the accuracy of fault diagnosis.
The oil and gas separation module, photoacoustic module and data analysis module are used to separate and detect the gas in the transformer oil through dissolution equilibrium method and photoacoustic spectroscopy. The gas detection results are corrected by using the interference influence coefficient of C3 gas and the concentration of C3 gas in the mixed gas. , weaken the interference effect of C3 gas.
It improves the accuracy of gas detection in transformer oil, reduces the interference of C3 gas on the gas analysis of target components, and obtains higher reliability of analysis results.
Smart Images

Figure CN114705628B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of online detection of electrical equipment, and in particular to a gas detection system and method in transformer oil with high anti-interference capability. Background Art
[0002] Transformers are vital electrical equipment in power systems. Once a failure occurs, it will cause huge losses. Detecting the operating status of the transformer and promptly discovering potential faults and abnormal operating conditions of the transformer are very important to ensure the normal operation of the transformer and the power grid.
[0003] At present, the vacuum degassing method is commonly used in power systems to separate the dissolved gas in transformer oil from the insulating oil, and then gas chromatography is used to detect the dissolved gas in the transformer oil. This gas chromatography detection method is prone to contamination of the chromatographic column, which leads to inaccurate analysis results, and then leads to false alarms and missed alarms in transformer fault diagnosis. Compared with gas chromatography, photoacoustic spectroscopy does not have the problem of chromatographic column contamination, is not prone to false alarms and missed alarms, and is conducive to reliably discovering the fault type of the transformer.
[0004] According to the similarity and solubility characteristics, in addition to the target component gases methane, ethane, ethylene, acetylene, carbon monoxide and carbon dioxide, some C3 gases (propane, propylene and propyne) are also separated from the oil, which interferes with the analysis of the target component gases. C4 and gases containing more carbon atoms are difficult to separate from the oil because their molecules contain more C atoms and are closer to the composition structure of oil. Their content in the mixed gas separated from the degasser is very small, and they have little interference with the target components. Therefore, C3 gas is the main source of interference for the photoacoustic spectroscopy analysis of the target component gas. In order to obtain higher accuracy of gas analysis results in transformer oil and improve the accuracy of transformer operation status detection, it is necessary to reduce the interference of C3 gas on the analysis of the target component gas as much as possible. Summary of the invention
[0005] The purpose of the present invention is to overcome the above-mentioned shortcomings of the prior art and provide a transformer oil gas detection system and method with high anti-interference ability.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] A transformer oil gas detection system with high anti-interference capability, comprising an oil-gas separation module, a photoacoustic module, a collection module and a data analysis module;
[0008] When testing, the oil-gas mixture in the transformer enters the oil-gas separation module, where it is separated based on the dissolution equilibrium method. The separated mixed gas enters the photoacoustic cell of the photoacoustic module. The mixed gas in the photoacoustic cell absorbs the energy of the modulated light source and generates sound waves. After the microphone in the photoacoustic cell detects the sound waves, it outputs the corresponding photoacoustic intensity signal. The photoacoustic intensity signal enters the data analysis module through the acquisition module. The data analysis module is used to calibrate the concentration of each gas in the oil-gas mixture in the transformer. The correction process of the data analysis module is as follows:
[0009] Based on the photoacoustic intensities of the six characteristic gases at the characteristic spectral lines, the six characteristic gases in the transformer oil were solved by photoacoustic spectroscopy. The six gases were CO, CO2, CH4, C2H6, C2H4 and C2H2, and the detection concentrations were y1, y2, ···, y6 respectively;
[0010] Based on the photoacoustic intensity of C3 gas corresponding to the characteristic spectral line, the concentration of C3 gas in the mixed gas in transformer oil is solved by photoacoustic spectroscopy. The concentrations of C3 gas propane, propylene and propyne are a1, a2 and a3 respectively.
[0011] The interference influence coefficient k of C3 gas on six characteristic gases 1i , k 2i and k 3i , to eliminate the interference of C3 gas, the gas detection concentration in oil is c i =y i -a1×k 1i +a2×k 2i +a3×k 3i ;
[0012] Among them, i=1,2,···,6.
[0013] Furthermore, the process of obtaining the detection concentration by the data analysis module is as follows:
[0014] The coefficient matrix D and the random error e2 in formula (3) are calculated by using a mixed gas composed of 9 gases with different concentrations and the least square method, where i is a 9×1 vector, representing the photoacoustic intensity at the 9 characteristic wavelengths; D is a 9×9 matrix, representing the proportional coefficient corresponding to each gas; z is a 9×1 vector, representing the concentration of each gas, and e2 is a 9×1 vector, representing the random error;
[0015] i=Dz+e2 (3)
[0016] Obtain the photoacoustic intensity corresponding to the characteristic spectral line of the mixed gas to be measured, and substitute it into formula (3) to obtain the characteristic gas concentrations y1, y2, . . . , y6 and C3 gas concentrations a1, a2 and a3 respectively.
[0017] Furthermore, the interference influence coefficient k of C3 gas on the six characteristic gases is 1i , k 2i and k 3i The method to obtain is:
[0018] n standard mixed gas samples are introduced into the oil-gas separation module respectively, and then pass through the photoacoustic module and the acquisition module in turn to obtain the photoacoustic intensity. The data analysis module uses the photoacoustic spectroscopy to solve the gas detection result in the mixed gas in the transformer oil as y k ;
[0019] The standard mixed gas sample is composed of characteristic gas with the same concentration and C3 gas with different concentrations. The characteristic gas concentration is e i , C3 gas concentration is a j1 , a j2 , a j3 are all known, where j = 1, 2, ···, n;
[0020] The specific process of the data analysis module calculating the interference influence coefficients k1, k2, k3 of a group of C3 gases on the mixed gas in oil is as follows:
[0021] The interference influence coefficients to be determined are k1, k2, k3, and the gas detection results in transformer oil of n standard mixed gas samples are y k , where k = 1, 2, ···, n, then
[0022]
[0023] After sorting, we get
[0024] Will Denoted as A, Denoted as K, Denoted as B, the above matrix is formalized as
[0025] AK=B (6)
[0026] For the linear equation system Ax=b, the recursive least squares equation is:
[0027]
[0028] x (k+1) =x (k) +P k+1 a k+1 (b k+1- a k+1 T x (k) ) (8)
[0029] where ak+1 and b k+1 To update data, both are scalars;
[0030] First, the initial values P0 and K are calculated based on the gas detection results of four sets of standard mixed gas samples and the known gas concentrations. (0) , the calculation formula is:
[0031] P0=(A0 T ×A0) -1 (9)
[0032] K (0) =P0×A0 T ×B0 (10)
[0033] Next, update the first row y1=e1+k1×a in the matrix equation AK=B 11 +k2×a 12 +k3×a 13 The updated data is A (1) and B (1) , where A (1) is a 3×1 vector, B (1) is a 1×1 vector; substitute into the recursive least squares equation
[0034]
[0035] K (1) =K (0) +P1×A (1) ×(B (1) -A (1) T ×K (0) ) (12)
[0036] Similarly, the second, third, and fourth rows of data are updated respectively and substituted into the recursive least squares equation. After four recursions are completed, the new gas detection results and the known gas concentration are used to update from the first row of data again until n groups of data are used up, and the final interference influence coefficients are obtained as k1, k2, and k3.
[0037] Through the above steps, the interference influence coefficients k1, k2, k3 of C3 gas on a characteristic gas in oil are obtained;
[0038] The same steps are used to obtain the interference influence coefficient k of C3 gas on the characteristic gas in oil. 1i , k 2i , k 3i , where i=1,2,…,6.
[0039] Furthermore, the wavelength of the modulated light source is 4.660-14.925 μm.
[0040] Furthermore, the data analysis module is also used to obtain the fault type or abnormal operating state of the transformer according to the concentration content of each gas by using the improved three-ratio method, and determine whether it exceeds the set threshold. If it exceeds the threshold, an alarm signal is sent.
[0041] Furthermore, it also includes a computer communication module, which is used to transmit the gas concentration and fault type data to the computer in the station through the IEC61850 / MODBUS TCP communication protocol.
[0042] A method for detecting gas in transformer oil with high anti-interference capability comprises the following operations:
[0043] The interference influence coefficient k of C3 gas on six characteristic gases under the use environment is obtained by using the standard mixed gas 1i , k 2i and k 3i , where i = 1, 2, ···, 6;
[0044] Photoacoustic spectroscopy is used to detect six characteristic gases in transformer oil. The six gases are CO, CO2, CH4, C2H6, C2H4 and C2H2, and the detection results are y1, y2, ···, y6 respectively;
[0045] Photoacoustic spectroscopy was used to determine the concentration of C3 gas in the mixed gas in transformer oil. The concentrations of C3 gas propane, propylene and propyne were a1, a2 and a3 respectively.
[0046] The gas concentration in oil after eliminating the interference of C3 gas is c i =y i -a1×k 1i +a2×k 2i +a3×k 3i ;
[0047] Where i=1,2,···,6.
[0048] Furthermore, the solution method for C3 gas concentration is:
[0049] The analysis model is established based on the proportional linear relationship between gas concentration and photoacoustic signal intensity as follows:
[0050] y=Xb+e1 (1)
[0051] Where: X is an n×m dimensional matrix, containing the values of n calibration samples on m spectral lines, i.e., photoacoustic intensity; b is an m×1 dimensional regression vector, b=[b1,b2,…,b m ] T; y is an n×1 dimensional vector, representing the concentration of the analyzed gas; e1 is an n×1 dimensional vector, representing the random error;
[0052] The above formula is transformed into the following by TR algorithm:
[0053]
[0054] Where: ||·|| p represents the p-norm, when p=2, it is the Euclidean norm; a and b represent the norms of the regression deviation and regression vector respectively, 1≤a, b<∞; L represents a regularization operator; λ represents the weight of controlling the second term relative to the first term;
[0055] According to the number of spectral lines j, the length of the regression vector is determined to be m=j, X is taken as the candidate spectral line value, the parameters a and b in formula (2) are set to 1, the candidate spectral value is selected as the regularization operator L, and the weight λ is set to 2; then use formula (2) to screen m for each C3 gas i characteristic spectral lines, where ∑m i ≥9, i=1,2,3;
[0056] The coefficient matrix D and the random error e2 in formula (3) are calculated by using a mixed gas composed of 9 gases with different concentrations and the least square method, wherein i is a 9×1 vector, which represents the photoacoustic intensity at the determined 9 characteristic wavelengths; D is a 9×9 matrix, which represents the proportional coefficient corresponding to each gas, and the proportional coefficient is the product of system parameters, absorption coefficient, etc. The system parameters and absorption coefficient use the known concentration and photoacoustic intensity, and then the product of all these partial coefficients is directly calculated according to the least square method, that is, the total coefficient, which is defined as the proportional coefficient; z is a 9×1 vector, which represents the concentration of each gas, and e2 is a 9×1 vector, which represents the random error;
[0057] i=Dz+e2 (3)
[0058] The photoacoustic intensity of the mixed gas to be tested corresponding to the characteristic spectrum line is detected and substituted into formula (3) to obtain the concentrations a1, a2 and a3 of the three C3 gases respectively.
[0059] Furthermore, the recursive least square method is used to calculate the interference influence coefficient of C3 gas on the mixed gas in oil;
[0060] Based on n standard mixed gas samples, the n gases consist of characteristic gases with the same concentration and C3 gases with different concentrations, and the characteristic gas concentration is e i 、Three C3 gas concentrations a j1 , a j2 , a j3 It is known that j = 1, 2, ···, n;
[0061] Calculate the interference influence coefficient of a group of C3 gases on the mixed gas in oil. The interference influence coefficients to be determined are k1, k2, k3. The gas detection results in transformer oil of these n standard mixed gas samples are y. k , k=1,2,···,n; then
[0062]
[0063] After sorting, we get
[0064] Will Denoted as A, Denoted as K, Denoted as B, the above matrix is formalized as
[0065] AK=B (6)
[0066] For the linear equation system Ax=b, the recursive least squares equation is:
[0067]
[0068] x (k+1) =x (k) +P k+1 a k+1 (b k+1- a k+1 T x (k) ) (8)
[0069] where a k+1 and b k+1 To update data, both are scalars;
[0070] Calculate the initial values P0 and K based on the gas detection results of four sets of standard mixed gas samples and the known gas concentrations (0) , the calculation formula is:
[0071] P0=(A0 T ×A0) -1 (9)
[0072] K (0) =P0×A0 T ×B0 (10)
[0073] Next, update the first row y1=e1+k1×a in the matrix equation AK=B 11 +k2×a 12 +k3×a 13 The updated data is A (1) and B (1) , where A (1)is a 3×1 vector, B (1) is a 1×1 vector; substitute into the recursive least squares equation
[0074]
[0075] K (1) =K (0) +P1×A (1) ×(B (1) -A (1) T ×K (0) ) (12)
[0076] Similarly, the second, third, and fourth rows of data are updated respectively and substituted into the recursive least squares equation. After four recursions, the new gas detection results and the known gas concentration are used to update from the first row of data again until n groups of data are used up, and the final interference influence coefficients are obtained as k1, k2, and k3.
[0077] Through the above steps, we get the interference influence coefficients k1, k2, k3 of C3 gas on a characteristic gas in oil; using the same steps, we get the interference influence coefficient k of C3 gas on all characteristic gases in oil. 1i , k 2i , k 3i , i=1,2,……,6.
[0078] Compared with the prior art, the present invention has the following beneficial effects:
[0079] The present invention discloses a gas detection system and method in transformer oil with high anti-interference capability. The gas dissolved in transformer oil is separated by the oil-gas separation module, photoacoustic module, acquisition module and data analysis module of the system, and the content of various fault characteristic gases in transformer oil is detected by photoacoustic spectroscopy. In view of the interference of C3 gas on characteristic gas in oil, the interference influence coefficient of C3 gas on characteristic gas and the concentration of C3 gas in mixed gas are used to correct the gas detection result; the present invention greatly reduces the interference influence of C3 gas, obtains gas detection results with reliable accuracy, and the detection result has high anti-interference capability for C3 gas. BRIEF DESCRIPTION OF THE DRAWINGS
[0080] Figure 1 It is a system structure diagram of the present invention. DETAILED DESCRIPTION
[0081] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments 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 ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0082] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0083] The present invention is further described in detail below in conjunction with the accompanying drawings:
[0084] See also Figure 1 , Figure 1 The system structure diagram of the present invention is a transformer oil gas detection system with high anti-interference ability, which includes a transformer, an oil-gas separation module, a photoacoustic module, a collection module, a data analysis module, a computer communication module and a station computer. The computer communication module is connected to the data analysis module to transmit the detected oil gas data to the station computer. Among them, the photoacoustic module includes a modulated light source, a photoacoustic cell and a microphone; the collection module includes a preamplifier, a lock-in amplifier and an A / D conversion unit.
[0085] The modulated light source in the photoacoustic module adopts an internally modulated electric pulse small-volume infrared modulated light source, and the wavelength of the modulated light source covers the range of 4.660 to 14.925 μm. The photoacoustic cell adopts a non-resonant photoacoustic cell. When the non-resonant photoacoustic cell is working, the modulation frequency of the modulated light source is lower than the lowest resonant frequency of the photoacoustic cell, and it works in a non-resonant state. The microphone in the photoacoustic module is a capacitive microphone, which is set at the antinode point of the photoacoustic cell.
[0086] The photoacoustic module is based on the photoacoustic effect. After gas molecules absorb incident light of a specific wavelength, they transition to an excited state. The excited state molecules convert the absorbed light energy into translational kinetic energy through radiationless relaxation. Macroscopically, this is manifested as an increase in the temperature of gas molecules, while in a closed environment such as a photoacoustic cell, the gas pressure increases with the increase in temperature. The photoacoustic effect mainly depends on the absorption of infrared radiation by the gas, and the absorption is selective. Different gases will absorb the energy of modulated light sources of different wavelengths. By changing the frequency of the modulated light source, the corresponding gas temperature fluctuates periodically, and the resulting pressure change is the sound wave.
[0087] The workflow of the present invention is:
[0088] The oil-gas separation module uses the dissolution equilibrium method to degas the transformer oil. The gas in the oil separated by the oil-gas separation module enters the photoacoustic module. In the photoacoustic pool, the gas absorbs the energy of the modulated light source to generate sound waves. After the microphone detects the sound waves, the corresponding electroacoustic signal data is output. The photoacoustic module transmits the data detected by the microphone to the acquisition module. After passing through the preamplifier, the phase-locked amplifier, and the A / D conversion unit in sequence, the converted data is obtained. After that, the photoacoustic detection results (i.e., the photoacoustic intensity) enter the data analysis module. The data analysis module corrects the content of each gas in the oil output from the acquisition module according to the C3 gas concentration. Specifically, the gas data in the transformer oil is analyzed using the least squares method. The six characteristic gases measured by the photoacoustic spectroscopy method are corrected using N samples of each gas. According to the concentration content of each gas, the improved three-ratio method is used to obtain the fault type or abnormal operating state of the transformer, and it is judged whether it exceeds the set threshold and sends an alarm signal. The results of the data analysis module are transmitted to the station computer through the computer communication module, and the obtained gas concentration and fault type data are transmitted to the station computer through the IEC61850 / MODBUS TCP communication protocol.
[0089] The data analysis process of the data analysis module is as follows:
[0090] First, calculate the C3 gas concentration. The method for calculating the C3 gas concentration is as follows:
[0091] The analysis model is established based on the proportional linear relationship between concentration and photoacoustic signal intensity as follows:
[0092] y=Xb+e1 (1)
[0093] Where: X is an n×m dimensional matrix, containing the values of n calibration samples on m spectral lines, i.e., photoacoustic intensity; b is an m×1 dimensional regression vector, b=[b1,b2,…,b m ] T ; y is an n×1 dimensional vector, representing the concentration of the analyzed gas; e1 is an n×1 dimensional vector, representing the random error;
[0094] The TR algorithm transforms formula (1) into the following:
[0095]
[0096] Where: ||·|| p represents the p-norm, when p=2, it is the Euclidean norm; a and b represent the norms of the regression deviation and regression vector respectively, 1≤a, b<∞; L represents a regularization operator; λ represents the weight of controlling the second term relative to the first term;
[0097] According to the number of selected spectral lines j, the length of the regression vector is determined to be m=j, X is taken as the candidate spectral line value, the parameters a and b in (2) are set to 1, the candidate spectral value is selected as the regularization operator L, and the weight λ is set to 2; then use formula (2) to screen m for each C3 gas i characteristic spectral lines with high sensitivity and high signal-to-noise ratio, where ∑m i ≥9, i=1,2,3;
[0098] The coefficient matrix D and the random error e2 in formula (3) are calculated by using a mixed gas composed of 9 gases with different concentrations and the least square method, where i is a 9×1 vector, representing the photoacoustic intensity at the 9 characteristic wavelengths, D is a 9×9 matrix, representing the proportional coefficient corresponding to each gas, z is a 9×1 vector, representing the concentration of each gas, and e2 is a 9×1 vector, representing the random error;
[0099] i=Dz+e2 (3)
[0100] The photoacoustic intensity corresponding to the selected characteristic spectral line of the mixed gas to be tested is detected and substituted into formula (3) to obtain the concentrations a1, a2 and a3 of the three C3 gases respectively.
[0101] Next, the recursive least squares method is used to calculate the interference coefficient of C3 gas on the mixed gas in oil; the result y obtained by detecting the gas in oil using photoacoustic spectroscopy i =d i +Δ 1i +Δ 2i +Δ 3i ; where i = 1, 2, ···, 6; where d i is the true value of the gas mixture in oil, Δ 1i is the interference value of propane on the i-th characteristic gas, Δ 2i is the interference value of propylene to the i-th characteristic gas, Δ 3iis the interference value of propyne on the i-th characteristic gas; the interference value is the product of the interference influence coefficient of C3 gas on the mixed gas in oil and the concentration of C3 gas; the system itself reserves n standard mixed gas samples, which contain characteristic gases of the same concentration and C3 gases of different concentrations. The concentrations of these gases are all known, that is, the concentration of the characteristic gas is e i It is known that the concentration of C3 gas is a j1 , a j2 , a j3 It is known that j = 1, 2, ···, n; taking the calculation of the interference influence coefficient of a group of C3 gases on the mixed gas in oil as an example, the interference influence coefficients to be determined are k1, k2, k3, and the gas detection results in the transformer oil of these n standard mixed gas samples are y k , where k = 1, 2, ···, n; then
[0102]
[0103] After sorting, we get
[0104] Will Denoted as A, Denoted as K, Denoted as B, the above matrix is formalized as
[0105] AK=B (6)
[0106] For the linear equation system Ax=b, the recursive least squares equation is:
[0107]
[0108] x (k+1) =x (k) +P k+1 a k+1 (b k+1 -a k+1 T x (k) ) (8)
[0109] where a k+1 and b k+1 To update data, both are scalars;
[0110] First, the initial values P0 and K are calculated based on the gas detection results of four sets of standard mixed gas samples and the known gas concentrations. (0) , the calculation formula is:
[0111] P0=(A0 T ×A0) -1 (9)
[0112] K (0)=P0×A0 T ×B0 (10)
[0113] Next, update the first row y1=e1+k1×a in the matrix equation AK=B 11 +k2×a 12 +k3×a 13 The updated data is A (1) and B (1) , where A (1) is a 3×1 vector, B (1) is a 1×1 vector; substitute into the recursive least squares equation
[0114]
[0115] K (1) =K (0) +P1×A (1) ×(B (1) -A (1) T ×K (0) ) (12)
[0116] Similarly, the second, third, and fourth rows of data are updated respectively and substituted into the recursive least squares equation. After four recursions are completed, the new gas detection results and the known gas concentration are used to update from the first row of data again until n groups of data are used up, and the final interference influence coefficients are obtained as k1, k2, and k3.
[0117] Through the above steps, the interference influence coefficients k1, k2, k3 of C3 gas on a characteristic gas in oil are obtained; using the same steps, the interference influence coefficient k of C3 gas on the characteristic gas in oil can be obtained. 1i , k 2i , k 3i , where i=1,2,…,6.
[0118] The data analysis module is based on the concentrations of C3 gas a1, a2, a3, combined with the interference coefficient k of C3 gas on the mixed gas in oil. 1i , k 2i , k 3i , where i = 1, 2, ..., 6; finally, according to c i =y i -a1×k 1i +a2×k 2i +a3×k 3i Calculate the gas detection results in oil after eliminating the interference of C3 gas.
[0119] The above contents are only for explaining the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.
Claims
1. A transformer oil gas detection system with high anti-interference ability, characterized in that: It includes oil-gas separation module, photoacoustic module, acquisition module and data analysis module; When testing, the oil-gas mixture in the transformer enters the oil-gas separation module, where it is separated based on the dissolution equilibrium method. The separated mixed gas enters the photoacoustic cell of the photoacoustic module. The mixed gas in the photoacoustic cell absorbs the energy of the modulated light source and generates sound waves. After the microphone in the photoacoustic cell detects the sound waves, it outputs the corresponding photoacoustic intensity signal. The photoacoustic intensity signal enters the data analysis module through the acquisition module. The data analysis module is used to calibrate the concentration of each gas in the oil-gas mixture in the transformer. The correction process of the data analysis module is as follows: Based on the photoacoustic intensity corresponding to the characteristic spectral lines of the six characteristic gases, the six characteristic gases in the transformer oil are solved by photoacoustic spectroscopy. The six gases are CO, CO2, CH4, C2H6, C2H4 and C2H2, and the detection concentrations are y1, y2, ..., y6 respectively; Based on the photoacoustic intensity of C3 gas corresponding to the characteristic spectral line, the concentration of C3 gas in the mixed gas in transformer oil is solved by photoacoustic spectroscopy. The concentrations of C3 gas propane, propylene and propyne are a1, a2 and a3 respectively. The interference influence coefficient k of C3 gas on six characteristic gases 1i , k 2i and k 3i , to eliminate the interference of C3 gas, the gas detection concentration in oil is c i =y i -a1×k 1i +a2×k 2i +a3×k 3i ; Where i=1,2,…,6; The interference coefficient k of C3 gas on the six characteristic gases 1i , k 2i and k 3i The method to obtain is: n standard mixed gas samples are introduced into the oil-gas separation module respectively, and then pass through the photoacoustic module and the acquisition module in turn to obtain the photoacoustic intensity. The data analysis module uses the photoacoustic spectroscopy to solve the gas detection result in the mixed gas in the transformer oil as y k ; The standard mixed gas sample is composed of characteristic gas with the same concentration and C3 gas with different concentrations. The characteristic gas concentration is e i , C3 gas concentration is a j1 , a j2 , a j3 All are known, where j=1,2,…,n; The specific process of the data analysis module calculating the interference influence coefficients k1, k2, k3 of a group of C3 gases on the mixed gas in oil is as follows: The interference influence coefficients to be determined are k1, k2, k3, and the gas detection results in transformer oil of n standard mixed gas samples are y k , where k=1,2,…,n, then (4) After sorting, we get (5) Will Denoted as A, Denoted as K, Denoted as B, the above matrix is formalized as AK=B(6) For linear equations , the recursive least squares equation is: (7) (8) in and To update data, both are scalars; First, the initial values P0 and K are calculated based on the gas detection results of four sets of standard mixed gas samples and the known gas concentrations. (0) , the calculation formula is: (9) (10) Next, update the first row of the matrix equation AK=B The updated data is A (1) and B (1) , where A (1) For 3 1 vector, B (1) is 1 1 vector; substitute into the recursive least squares equation (11) (12) Similarly, the second, third, and fourth rows of data are updated respectively and substituted into the recursive least squares equation. After four recursions are completed, the new gas detection results and the known gas concentration are used to update from the first row of data again until n groups of data are used up, and the final interference influence coefficients are obtained as k1, k2, and k3. Through the above steps, the interference influence coefficients k1, k2, k3 of C3 gas on a characteristic gas in oil are obtained; The same steps are used to obtain the interference influence coefficient k of C3 gas on the characteristic gas in oil. 1i , k 2i , k 3i , where i=1,2,…,6.
2. The transformer oil gas detection system with high anti-interference ability according to claim 1 is characterized in that: The process of obtaining the detection concentration by the data analysis module is as follows: The coefficient matrix D and the random error e2 in formula (3) are calculated by using a mixed gas composed of 9 gases with different concentrations and the least square method, where: i is a 9×1 vector, representing the photoacoustic intensity at the 9 characteristic wavelengths; D is a 9×9 matrix, representing the proportional coefficient corresponding to each gas; z is a 9×1 vector, representing the concentration of each gas, and e2 is a 9×1 vector, representing the random error; <h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> =Dz+e2(3) Obtain the photoacoustic intensity corresponding to the characteristic spectral line of the mixed gas to be measured, and substitute it into formula (3) to obtain the characteristic gas concentrations y1, y2, ..., y6 and C3 gas concentrations a1, a2 and a3 respectively.
3. The transformer oil gas detection system with high anti-interference ability according to claim 1 is characterized in that: The wavelength of the modulated light source is 4.660-14.925 μm.
4. The transformer oil gas detection system with high anti-interference ability according to claim 1 is characterized in that: The data analysis module is also used to obtain the fault type or abnormal operating state of the transformer according to the concentration content of each gas using the improved three-ratio method, and determine whether it exceeds the set threshold. If it exceeds the threshold, an alarm signal is sent.
5. The transformer oil gas detection system with high anti-interference capability according to claim 4 is characterized in that: It also includes a computer communication module, which is used to transmit the gas concentration and fault type data to the computer in the station through the IEC61850 / MODBUS TCP communication protocol.
6. A method for detecting gas in transformer oil with high anti-interference ability, characterized in that: The following operations are included: The interference influence coefficient k of C3 gas on six characteristic gases under the use environment is obtained by using the standard mixed gas 1i , k 2i and k 3i , where i=1,2,…,6; Photoacoustic spectroscopy is used to detect six characteristic gases in transformer oil. The six gases are CO, CO2, CH4, C2H6, C2H4 and C2H2, and the detection results are y1, y2, ..., y6 respectively; Photoacoustic spectroscopy was used to determine the concentration of C3 gas in the mixed gas in transformer oil. The concentrations of C3 gas propane, propylene and propyne were a1, a2 and a3 respectively. The gas concentration in oil after eliminating the interference of C3 gas is c i =y i -a1×k 1i +a2×k 2i +a3×k 3i ; Where i=1,2,…,6; The solution for C3 gas concentration is: The analysis model is established based on the proportional linear relationship between gas concentration and photoacoustic signal intensity as follows: y = Xb + e 1(1) in: X It is an n×m dimensional matrix, containing the values of n calibration samples on m spectral lines, i.e., the photoacoustic intensity; b is an m×1 dimensional regression vector, b =[ b 1, b 2,…, b m ] T ; y is an n×1-dimensional vector, representing the concentration of the analyzed gas; e 1 is an n×1 dimensional vector, representing random error; The above formula is transformed into the following by TR algorithm: (2) Where: ||·|| p express p -norm, p =2 is the Euclidean norm; a and b Represent the norm of regression deviation and regression vector respectively, 1≤ a , b <∞; L represents a regularization operator; λ It means controlling the weight of item 2 relative to item 1; According to the number of spectral lines j, the length of the regression vector is determined to be m=j. X The candidate spectral line values are taken, and the parameters a and b in formula (2) are set to 1. The candidate spectral line values are selected as the regularization operator L, and the weights λ Set to 2; then use formula (2) to filter m for each C3 gas i characteristic spectral lines, where ∑m i ≥9, i= 1,2,3; The coefficient matrix D and the random error e2 in formula (3) are calculated by using a mixed gas composed of 9 gases with different concentrations and the least squares method, where i is a 9×1 vector, representing the photoacoustic intensity at the 9 characteristic wavelengths; D is a 9×9 matrix, representing the proportional coefficient corresponding to each gas; z is a 9×1 vector, representing the concentration of each gas, and e2 is a 9×1 vector, representing the random error; <h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> =Dz+e2(3) Detect the photoacoustic intensity of the mixed gas to be tested corresponding to the characteristic spectrum line, substitute it into formula (3), and obtain the concentrations a1, a2 and a3 of the three C3 gases respectively; The recursive least square method is used to calculate the interference coefficient of C3 gas on the mixed gas in oil; Based on n standard mixed gas samples, the n gases consist of characteristic gases with the same concentration and C3 gases with different concentrations, and the characteristic gas concentration is e i 、Three C3 gas concentrations a j1 , a j2 , a j3 It is known that j=1,2,…,n; Calculate the interference influence coefficient of a group of C3 gases on the mixed gas in oil. The interference influence coefficients to be determined are k1, k2, k3. The gas detection results in transformer oil of these n standard mixed gas samples are y. k , k=1,2,…,n; then (4) After sorting, we get (5) Will Denoted as A, Denoted as K, Denoted as B, the above matrix is formalized as AK=B(6) For linear equations , the recursive least squares equation is: (7) (8) in and To update data, both are scalars; Calculate the initial values P0 and K based on the gas detection results of four sets of standard mixed gas samples and the known gas concentrations (0) , the calculation formula is: (9) (10) Next, update the first row of the matrix equation AK=B The updated data is A (1) and B (1) , where A (1) For 3 1 vector, B (1) is 1 1 vector; substitute into the recursive least squares equation (11) (12) Similarly, the second, third, and fourth rows of data are updated respectively and substituted into the recursive least squares equation. After four recursions, the new gas detection results and the known gas concentration are used to update from the first row of data again until n groups of data are used up, and the final interference influence coefficients are obtained as k1, k2, and k3. Through the above steps, we get the interference influence coefficients k1, k2, k3 of C3 gas on a characteristic gas in oil; using the same steps, we get the interference influence coefficient k of C3 gas on all characteristic gases in oil. 1i , k 2i , k 3i , i=1,2,…,6.
Citation Information
Patent Citations
A packaging system
IE61850B1
Spectral method-based multi-component gas concentration analysis method
CN108287141A
Analyzing fault gas concentration in liquid
CN109690290A
Online oil gas monitoring system of laser photoacoustic spectroscopy
CN110702611A
Double-chamber photoacoustic spectrum monitoring system for dissolved gas in transformer oil
CN113959956A
