On-line monitoring device and method for dissolved gas in transformer oil based on infrared spectrum

Abstract

The application discloses an on-line monitoring device and method for dissolved gas in transformer oil based on infrared spectrum, which comprises a transformer oil gas separation part, an infrared spectrum gas detection part and a data processing part. The method comprises the following steps: firstly, carrying out oil gas separation; then, sending the obtained dissolved gas in oil into a gas chamber and scanning the gas chamber by using an infrared spectrometer to obtain a spectrum diagram; and finally, obtaining the content of various target gases in oil by using a gas quantitative analysis model. The application uses infrared spectrum method as the gas detection part, has the advantages of fast spectrum data acquisition, no consumption of gas samples, wide detection range, high stability, maintenance-free, long-term operation monitoring cost, etc. compared with the most widely used gas chromatography method for on-line monitoring of gases in transformer oil. The device adopts a structure based on single-chamber vacuum degassing, solves the problem of spectrum baseline offset distortion, and does not need to consume nitrogen as background, thereby improving the automation degree of the device.

Description

Technical Field

[0001] This invention belongs to the field of power equipment monitoring technology, specifically relating to an online monitoring device and method for dissolved gases in transformer oil based on infrared spectroscopy. Background Technology

[0002] Transformers are a crucial component of the power grid system. A fault in a transformer can lead to a major disaster for the entire power system. Regular monitoring of transformer operation is essential for early warning of potential faults and timely detection of potential problems. This is vital for the safe and reliable operation of the entire power system, reducing repair time, and minimizing economic losses. The higher the voltage level of the transformer, the greater the likelihood of an accident, the wider the affected power supply area, and the more severe the economic losses. Currently, most high-voltage transformers both domestically and internationally are oil-immersed transformers. Transformer oil is a mixture extracted from petroleum containing long-chain or cyclic hydrocarbon chains. When certain faults occur in the transformer, characteristic gases are generated in the transformer oil. For example, electrical and thermal faults produce gases such as hydrogen, ethylene, and acetylene. Oxidation of the transformer oil can cause the accumulation of carbon monoxide and carbon dioxide. The analysis of gas composition and concentration in transformer oil is one of the bases for judging transformer faults. The "Preventive Testing Procedures for Power Equipment" lists dissolved gas analysis in oil as the primary method, and the national standard GB / T 7252-2001 uses CH4, C2H6, C2H4, C2H2, CO, CO2, and H2 gases as important criteria for judging internal faults in power transformers. Currently, the main methods for analyzing dissolved gases in transformer oil include gas chromatography, photoacoustic spectroscopy, Raman spectroscopy, and infrared spectroscopy. Gas chromatography is the most widely used method. However, gas chromatography columns are easily contaminated and require regular calibration or even replacement. Furthermore, gas chromatography requires a carrier gas, usually nitrogen, and the analytical results are affected by nitrogen pressure. Therefore, nitrogen cylinders need to be replaced regularly. This method is not only costly to operate but also relatively troublesome to maintain, and inadequate maintenance can easily lead to false alarms or missed alarms. Summary of the Invention

[0003] This invention provides an online monitoring device and method for dissolved gases in transformer oil based on infrared spectroscopy, which is used for transformer fault early warning and diagnostic analysis. It can effectively solve the problems of high operating costs and inconvenient maintenance caused by the need for chromatographic columns and regular calibration, as well as the need for regular replacement of nitrogen cylinders in current main transformer oil dissolved gas analysis.

[0004] To achieve the above objectives, the present invention provides an online monitoring device for dissolved gases in transformer oil based on infrared spectroscopy, comprising a transformer oil-gas separation section, an infrared spectroscopy gas detection section, and a data processing section. The transformer oil-gas separation section includes a degassing device and a transformer oil cylinder. The degassing device is used to extract oil from the transformer oil cylinder and separate dissolved gases from the transformer oil. The degassing device, the transformer oil cylinder, and the detection chamber are connected via pipelines. The infrared spectroscopy gas detection section includes an infrared spectrometer and a detection chamber. The detection chamber is connected to the degassing device and is used to store the separated dissolved gases from the transformer oil. The infrared spectrometer is used to scan the infrared spectrum of the gas in the detection chamber. The data processing section includes a main controller connected to the infrared spectrometer. The main controller is used to analyze the types and contents of dissolved gases in the transformer oil based on the infrared spectrum.

[0005] Furthermore, the degassing device includes a degassing tank, which is connected to the oil inlet of the transformer cylinder via an oil inlet pipe, and to the oil outlet of the transformer cylinder via a return oil pipe; both the oil inlet pipe and the return oil pipe are equipped with valves; a first pressure sensor is installed in the degassing tank; the degassing tank is connected to a detection chamber via a third pipe, on which valves and a vacuum pump are installed; a temperature sensor and a second pressure sensor are installed at the interface of the detection chamber.

[0006] Furthermore, a gas collection cylinder is connected to the third pipe via a fifth pipe.

[0007] Furthermore, an air pump is installed on the degassing tank, the air pump is connected to the first end of an air pipe, and the second end of the air pipe extends into the bottom of the degassing tank.

[0008] Furthermore, a hydrogen sensor is installed on the degassing tank.

[0009] Furthermore, the valve is a solenoid valve, and the degassing tank is equipped with a liquid level sensor.

[0010] The monitoring method based on the above-mentioned online monitoring device for dissolved gases in transformer oil using infrared spectroscopy includes the following steps:

[0011] Step 1: Clean the detection chamber with a degassing device;

[0012] Step 2: Use a degassing device to separate the dissolved gases from the transformer oil and send them to the detection gas chamber;

[0013] Step 3: Use a Fourier transform infrared spectrometer to scan the detection chamber containing dissolved gases in the oil to obtain a single-wavelength image of the sample. Subtract the background single-wavelength image from the sample single-wavelength image to obtain the infrared spectrum of the gas.

[0014] Step 4: Analyze the above infrared spectrum using a gas analysis model to obtain the dissolved gas content in the oil.

[0015] Furthermore, step 2 is as follows:

[0016] Step 2.1: Use an infrared spectrometer to scan the detection cell to obtain a background single-wave pattern;

[0017] Step 2.2, Oil-gas separation: Extract transformer oil from the transformer cylinder into the degassing tank; begin oil-gas separation, dissolved gas in the transformer oil is separated to the space above the degassing tank, turn on the air pump to send the gas above to the bottom of the degassing tank and it will emerge from the bottom of the degassing tank, causing dissolved gas in the oil to precipitate out, accelerating the separation of gas in the oil. After the gas separation is completed, turn off the air pump; open solenoid valves V5 and V7, close the other solenoid valves, and the piston of the gas collection cylinder moves, collecting the separated gas into the gas collection cylinder;

[0018] The piston of the driving gas collection cylinder moves to send gas into the detection chamber. At this time, the detection chamber is filled with separated gas. The temperature sensor and the second pressure sensor detect the temperature and pressure in the detection chamber at this time; and calculate the pressure compensation coefficient.

[0019] Furthermore, step 4 involves inputting the infrared spectrum into the gas quantitative analysis model, which outputs the gas concentration of each component. Temperature and pressure compensation is then performed using the pressure compensation coefficient to obtain the compensated gas concentration.

[0020] The gas quantitative analysis model is based on infrared spectral scanning experiments of a large number of standard gases. The quantitative relationship between the absorption intensity of each target gas infrared spectral absorption peak or absorption band and its concentration is obtained based on the experimental results.

[0021] Furthermore, after step 4 is completed, the three-ratio method is used to determine whether the transformer may be faulty based on the compensated gas concentration results. If so, the device will issue an alarm.

[0022] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0023] The device of this invention employs infrared spectroscopy as the gas detection component. Compared to the widely used gas chromatography method, this device offers advantages such as higher sensitivity, no gas sample consumption, a wider detection range, low cost for expanding to new components, high stability, maintenance-free operation, and low long-term monitoring costs. In this invention, the gas chamber simultaneously serves as both the background gas and the gas chamber for the sample being tested. Compared to existing infrared spectroscopy methods that require dual-chamber switching, this avoids stability issues that may arise from mechanical failures during chamber switching.

[0024] Furthermore, the measurement frequency can be set by the host computer to achieve continuous measurement, and the results of each measurement are automatically saved in the computer, thus improving the level of automation.

[0025] Furthermore, considering that hydrogen has no infrared activity, a hydrogen sensor based on electrochemical principles is used for measurement.

[0026] This invention is based on a gas spectral analysis method using a single-chamber vacuum degassing system. A vacuum pump is used to evacuate the gas chamber to 0.02 MPa, which is then used as a background for scanning the background spectrum. The degassed gas is then sent to the vacuum chamber to scan the sample spectrum. A gas quantitative analysis model is used to analyze the spectral data to obtain the gas concentration. This method solves the problems of spectral baseline shift and distortion, as well as the increased cost caused by using nitrogen to clean the gas chamber and scan the background each time.

[0027] Furthermore, the method described in this invention involves cleaning the sample gas chamber before each measurement. The inlet and outlet of the degassing device are connected to the sample gas chamber, and the vacuum pump inside the degassing device is connected to the inlet. The vacuum pump operates to expel the gas obtained from the previous separation from the gas chamber, avoiding interference with the current detection. When degassing is complete, the gas is discharged from the inlet to the sample gas chamber. After the gas is discharged, the inlet is immediately closed, and the outlet is also in a closed state, ensuring good sealing of the gas path. Compared with the existing Fourier transform infrared spectroscopy method for detecting dissolved gases in transformer oil, this device ensures that the gas in the gas chamber for the current measurement is completely separated from the oil and gas in the current transformer oil and is not affected by the gas detected in the previous measurement.

[0028] Furthermore, existing Fourier transform infrared spectroscopy methods for detecting dissolved gases in transformer oil require periodic nitrogen purging of the gas chamber and rescanning of the background spectrum to cope with environmental changes, wasting manpower and resources. Some infrared spectroscopy methods have improved this by employing a dual-chamber switching device, which includes a sealed gas chamber filled with nitrogen. Before each detection, the system switches to the nitrogen chamber to scan the background. In contrast, this device first uses a vacuum method to purge the gas chamber until the internal pressure is reduced to 0.02 MPa, and then scans the chamber at this point as the background. After the gases in the oil are completely separated, the separated gases... The gas is then expelled back into the original gas chamber. Since the gas chamber is sealed, the gas inside is the sum of a small portion of the gas at 0.02 MPa pressure and the gas separated in this process. The spectrum obtained by subtracting the background single-wavelength image from the single-wavelength image obtained by scanning the gas chamber in this state is the spectrum obtained by separating the gas. Compared with the two original infrared spectroscopy methods mentioned above, this device only requires one sample gas chamber and does not consume nitrogen or a nitrogen gas chamber. It avoids the errors caused by scanning the nitrogen gas chamber in the traditional method and avoids the errors caused by different windows in the dual-gas chamber switching device. This not only reduces costs but also simplifies operation.

[0029] To address the issue of spectral baseline drift and distortion in quantitative gas analysis using infrared spectroscopy, this device rescans the background for each detection, ensuring that the background and sample are scanned under identical conditions, without consuming nitrogen. Compared to gas chromatography, this method offers faster spectral analysis, shorter detection times, degassing time within 15 minutes, and results obtained from spectrometer scanning and gas quantitative analysis model analysis within 5 minutes, resulting in a total run time of less than 20 minutes.

[0030] Furthermore, this invention conducted numerous infrared spectral scanning experiments on standard gases to study the quantitative relationship between the absorption intensity of infrared spectral absorption peaks or bands of various target gases and their concentrations. It also investigated the basic methods of quantitative infrared spectral analysis. To reduce the dimensionality of the raw data, the TR regularization method was used to extract the characteristic variables of the spectra. A quantitative analysis model based on partial least squares was established using the experimentally obtained data. During formal measurements, the concentrations of various gases in transformer oil can be quickly obtained using this quantitative analysis model, improving monitoring efficiency. Attached Figure Description

[0031] Figure 1 This is a structural diagram of the online monitoring device for dissolved gases in transformer oil based on infrared spectroscopy according to the present invention;

[0032] Figure 2 This is a diagram of the internal structure of the degassing device.

[0033] In the attached diagram: 1. Transformer oil cylinder; 2. Degassing device; 3. Data processing section; 4. Infrared spectroscopy gas detection section; 11. First gas pipe; 12. Air pump; 13. Second gas pipe; 14. Third pipe; 15. Fourth pipe; 16. Vacuum pump; 17. Fifth pipe; 18. Second pressure sensor; 19. Gas collection cylinder; 20. Air pump; 21. Oil inlet pipe; 22. Degassing tank; 23. Liquid level sensor; 24. First pressure sensor; 25. Hydrogen sensor; 26. Oil pump; 27. Oil return pipe; 28. First pipe; 29. ​​Second pipe; 30. Nozzle; 31. Display; 32. Main controller; 41. Monitoring gas chamber; 42. Infrared spectrometer.

[0034] For solenoid valve Detailed Implementation

[0035] To make the objectives and technical solutions of this invention clearer and easier to understand, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0036] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more. In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0037] Reference Figure 1 An online monitoring device for dissolved gases in transformer oil based on infrared spectroscopy includes a transformer oil-gas separation section, an infrared spectroscopy gas detection section 4, and a data processing section 3.

[0038] The transformer oil-gas separation section consists of a degassing device 2 and a transformer cylinder 1. The degassing device 2 contains a degassing tank 22, a vacuum pump 16, an oil pump 26, a gas pump 12, a hydrogen sensor 25, a solenoid valve, a liquid level sensor 23, a pressure sensor, and a temperature sensor. The infrared spectroscopy gas detection section 4 consists of a thermal infrared spectrometer and a detection chamber. The data processing section 3 consists of a display 31 and a main controller 32, which incorporates a data communication unit, an equipment control unit, and a gas quantitative analysis unit. The degassing device 2, the thermal infrared spectrometer, and the display 31 are all connected to the main controller 32.

[0039] The communication unit is used for data communication with the degassing device 2 and to read data sent by each sensor.

[0040] The equipment control unit is used to control the start and stop of the degassing device 2 and the infrared spectrometer.

[0041] The gas quantitative analysis unit is used to analyze the infrared spectrum of a gas to obtain the content of each component in the gas to be detected.

[0042] The infrared spectrometer used is the Spectrum Two portable infrared spectrometer from PerkinElmer, USA, which employs AVC technology to automatically remove interference from background air components such as CO2 and water. The main controller 32 is an ARK-1220L industrial computer from Advantech, which has the advantages of small size and high performance.

[0043] The structural connections of each part of the online monitoring device for dissolved gases in transformer oil of the present invention are as follows:

[0044] The oil inlet of transformer cylinder 1 is connected to the first end of oil inlet pipe 21. The second end of oil inlet pipe 21 is connected to the right end interface of oil pump 26 and one end of first pipe 28. The other end of first pipe 28 extends into the bottom of degassing tank 22. Solenoid valve V1 is installed on oil inlet pipe 21, and solenoid valve V3 is installed on first pipe 28. The oil outlet of transformer cylinder 1 is connected to the first interface on the left end of oil pump 26 through return oil pipe 27. Solenoid valve V2 is installed on the return oil pipe. The second interface on the left end of oil pump 26 is connected to nozzle 3 above degassing tank 22 through second pipe 29. Connected to the first pipe, a solenoid valve V4 is installed on the second pipe 29. A level sensor 23 is installed inside the degassing tank 22. A hydrogen sensor 25 and a first pressure sensor 24 are connected above the degassing tank 22. A first gas pipe 11 above the degassing tank 22 connects to a gas pump 12. A second gas pipe 13 connected to the other end of the gas pump 12 extends to the bottom of the degassing tank 22. A third pipe 14 above the degassing tank 22 connects to one interface of the detection chamber. Solenoid valves V5, V7, and V8 are installed sequentially on the third pipe 14 from the degassing cylinder to the detection chamber. The other interface of the detection chamber is connected to a solenoid valve V9, a temperature sensor, and a second pressure sensor 18. The chamber is placed on the infrared spectrometer detection platform. The infrared spectrometer, degassing device 2, and controller are connected. Solenoid valve V9 is a reserved interface; this valve can be opened if it is necessary to remove gas from the chamber.

[0045] The third pipe 14 is connected to the fourth pipe 15 and the fifth pipe 17. The connection point between the third pipe 14 and the fourth pipe 15 is between solenoid valves V5 and V7; the connection point between the third pipe 14 and the fifth pipe 17 is between solenoid valves V7 and V8. A solenoid valve V6 is installed on the fourth pipe 15, and the other end of solenoid valve V6 is connected to a vacuum pump 16. A gas collection cylinder 19 and an air pump 20 are installed on the fifth pipe 17, and a solenoid valve V10 is installed between the gas collection cylinder 19 and the air pump 20. The air pump 20 provides power to the gas in the gas collection cylinder 19, causing it to move to the target position.

[0046] Based on the above scheme, the detection chamber 41 of this device adopts a double-cone chamber with a length of 110mm, an inner diameter of 10mm in the middle, and an inner diameter of 26mm at both ends of the double-cone space. Both ends have high-transmittance calcium fluoride windows with a specification of φ32mm×4mm, which can withstand 200N pressure. There are air inlets and outlets on the side.

[0047] Based on the above scheme, the target gas types, their minimum detection limit, detection range, and measurement error for this device are as follows:

[0048]

[0049] The online monitoring method for dissolved gases in transformer oil based on the above-mentioned device includes the following steps:

[0050] Step 1: Clean the sample gas chamber using the degassing device.

[0051] Power is supplied to all parts of the device. The detection chamber contains the detection gas that was previously separated from the oil. After startup, the detection chamber is first cleaned with the degassing device 2 to avoid interference from the previous gas in this detection. Solenoid valves V5, V6, V7, and V8 are opened, while the other solenoid valves are closed. At this time, the detection chamber and the degassing tank 22 are connected. Vacuum pump 16 is turned on and starts working to extract the gas from the chamber and the degassing tank 22. At this time, the chamber and the degassing tank 22 are connected. When the reading of the second pressure sensor 18 reaches 0.02 MPa, all solenoid valves are closed and vacuum pump 16 is turned off. At this time, the pressure in both the degassing tank 22 and the chamber is 0.02 MPa, and the chamber cleaning is complete.

[0052] Step 2: Use degassing device 2 to separate the dissolved gases from the transformer oil and send them to the sample gas chamber.

[0053] After vacuum pump 16 is turned off, wait 10 seconds and scan the gas chamber as a background using an infrared spectrometer.

[0054] Next, oil-gas separation is performed: After the background scan is completed, solenoid valves V1 and V4 and oil pump 26 are opened, and the remaining solenoid valves are closed. Oil pump 26 starts working and draws transformer oil from transformer cylinder 1. The transformer oil is sprayed into degassing tank 22 from nozzle 30 above degassing tank 22. When the liquid level sensor above degassing tank 22 detects the liquid level, oil pump 26 stops working. The oil inlet is 500m. Oil-gas separation begins. Due to the low pressure inside degassing tank 22, dissolved gas in transformer oil is separated to the space above degassing tank 22. Air pump 12 works to send the gas above to the bottom of degassing tank 22 and it comes out from the bottom of degassing tank 22, causing dissolved gas in the oil to precipitate out and accelerating the separation of gas in the oil. After 300 seconds, the gas separation is completed. Air pump 12 is closed, solenoid valves V5 and V7 are opened, and the remaining solenoid valves are closed. The piston of gas collection cylinder 19 moves and collects the separated gas into gas collection cylinder 19.

[0055] Open solenoid valve V8, close the other solenoid valves, and the piston of gas collection cylinder 19 moves to send gas into the detection chamber. After 2 seconds, close solenoid valve V8. At this time, the detection chamber is filled with separated gas (i.e., dissolved gas in oil used for this test). The temperature sensor and the second pressure sensor detect the temperature and pressure in the detection chamber at this time; calculate the pressure compensation coefficient.

[0056] After gas separation is complete, the transformer oil in the degassing tank 22 needs to be returned to the transformer oil cylinder. Open solenoid valves V2 and V3, close the other solenoid valves, and start the oil pump 26 to send the oil in the degassing tank 22 back to the transformer oil cylinder 1. When the liquid level sensor 23 below the degassing tank 22 can no longer detect the liquid level, it means that the oil in the degassing tank 22 has been drained. Stop the oil pump 26 and the entire degassing process is over.

[0057] In this step, the infrared spectrometer scanned the gas chamber after the pressure inside was pumped down to 0.02 MPa, producing a background single-wavelength image. At the end of this step, the separated gas had been returned to the gas chamber, and the infrared spectrometer scanned the gas chamber again to obtain a sample single-wavelength image.

[0058] Step 3: Use a Fourier transform infrared spectrometer to scan the detection chamber containing dissolved gases in the oil to obtain a single-wavelength image of the sample. Subtract the background from the single-wavelength image to obtain the infrared spectrum of the gas.

[0059] Step 4: Analyze the above infrared spectrum using a gas analysis model to obtain the dissolved gas content in the oil. The infrared spectrum obtained by scanning the gas chamber with an infrared spectrometer serves as the input to the quantitative gas analysis model, and the output is the gas concentration of each component. Temperature and pressure compensation is performed using a temperature and pressure compensation formula to obtain the compensated gas concentration. Based on the compensated gas concentration result, the three-ratio method is used to determine whether the transformer may have a fault. If a fault is found, the device will issue an alarm to alert the personnel.

[0060] The infrared spectrometer scans the gas chamber to obtain the spectrum, and the obtained spectral data is sent to the main controller 32. After receiving the spectral data, the main controller 32 processes the spectral data using a gas quantitative analysis model to obtain the content of each target gas, which is then displayed on the display 31. The gas quantitative analysis model is based on infrared spectral scanning experiments with a large number of standard gases to study the quantitative relationship between the absorption intensity of each target gas's infrared spectral absorption peak or absorption band and its concentration. The basic methods of infrared spectral quantitative analysis technology are studied. To reduce the dimensionality of the raw data, the TR regularization method is used to extract the characteristic variables of the spectrum. A quantitative analysis model based on partial least squares is established using the experimentally obtained data. The entire detection process can be completed within 20 minutes. In continuous measurement mode, the data obtained from each detection is automatically saved to a designated folder.

[0061] The method for establishing the above-mentioned quantitative gas analysis model is as follows:

[0062] Extensive standard gas experiments were conducted to obtain infrared spectral data of single and mixed components of various target gases at different concentrations. An absorbance matrix X was established as the input matrix, with its elements representing spectral characteristic variables. A concentration matrix Y was established as the output matrix. Both the absorbance and concentration matrices were standardized. Then, the first pair of components was extracted from the input and output matrices respectively. The components were then transformed into two principal components to maximize their correlation. The extracted principal components were compared with the original data matrix (the matrix used for component extraction; if it's the first pair of principal components, then the matrix is ​​the absorbance matrix). The gas quantitative analysis model is obtained by performing regression processing on the principal components and the concentration matrix (if it is not the first pair of principal components, then the matrix is ​​the residual data matrix mentioned below). If the regression accuracy has reached the required accuracy, the process is terminated, resulting in the gas quantitative analysis model. The gas quantitative analysis model is the relationship between the principal components and the input absorbance matrix and the concentration matrix. Otherwise, the residual data matrix after extracting the principal components is used for the next round of component extraction. The principal components extracted from the output matrix and the input matrix are then subjected to regression processing and expressed in the form of regression against the original variables. The concentration matrix can be calculated using the obtained principal components and absorbance matrix, thus establishing the gas quantitative analysis model. The specific steps are as follows:

[0063] a. Standardize X and Y respectively;

[0064] b. For each principal component, let h = 1 and u = y. i Where h is the number of iterations and u is an intermediate variable;

[0065] c. Perform the following processing on the X matrix: Let Normalization processing Finally, a principal component of X to be tested is obtained. Among them, wT As an intermediate variable;

[0066] d. Perform similar processing on Y: Let Normalization processing Finally, a principal component of Y to be tested is obtained. q T q is an intermediate variable;

[0067] e. Compare t in step c with the previous one. If the deviation between the two is outside the expected error allowable range, repeat steps c and d; otherwise, proceed to the next step.

[0068] f. to order For p T Normalization is performed. Obtain principal components p T As an intermediate variable, For p T The result after normalization;

[0069] g. Calculate the regression coefficients of the principal components, the X matrix, and the Y matrix.

[0070] h. The residual of the X matrix after the current principal component extraction The residual of the Y matrix E h E represents the residual of the updated X matrix. h-1 F is the residual of the X matrix obtained in the previous loop. h F represents the residual of the updated Y matrix. h-1 The residual of the Y matrix obtained in the previous loop; if h = 1, E h- 1 is a matrix X, F h-1 For Y;

[0071] i. Replace X and Y with the residual matrix of the extracted principal components, that is, let X = E h Y = F h h = h + 1, and then continue the above operation from step b for the next principal component.

[0072] After steps a to i, the model parameters t and b based on partial least squares were established using standardized sample data. Given a set of sample data X with unknown concentrations, the predicted value Y of the model can be calculated using the following steps:

[0073] a. Standardize the input data X, let h = 0, Y = 0; h is the number of loops;

[0074] bh = h + l;

[0075] c. Continue until h is greater than the number of principal components; otherwise, proceed to step b.

[0076] d. Finally, the standardized Y is obtained. The mean and standard deviation during model training are used to deduce the specific unstandardized output matrix Y, i.e., the gas concentration data.

[0077] The temperature and pressure compensation calculation formula in step 2 is as follows:

[0078]

[0079] In the formula, C 补偿前 This represents the uncompensated gas concentration.

[0080] C 补偿后 The gas concentration obtained after compensation;

[0081] T represents the temperature sensor reading, in Kelvin.

[0082] p is the reading of the second pressure sensor, in MPa.

[0083] The above content provides a further detailed description of the present invention patent in conjunction with specific implementation schemes. It should be noted that the specific implementation schemes of the present invention patent are not limited thereto. For those skilled in the art, various deductions and extensions can be made without departing from the concept of the present invention, but all such extensions should be considered within the scope of patent protection defined by the claims submitted herein.

Claims

1. An on-line monitoring device for dissolved gases in transformer oil based on infrared spectroscopy, characterized in that, It includes a transformer oil-gas separation section, an infrared spectroscopy gas detection section (4), and a data processing section (3); The transformer oil-gas separation section includes a degassing device (2) and a transformer cylinder (1). The degassing device (2) is used to take oil from the transformer cylinder (1) and separate the dissolved gas in the transformer oil. The degassing device (2), the transformer cylinder (1), and the gas detection chamber are connected by a pipeline. The infrared spectroscopy gas detection section (4) includes an infrared spectrometer and a detection gas chamber. The detection gas chamber is connected to the degassing device (2) and is used to store the dissolved gas separated from the transformer oil. The infrared spectrometer is used to scan the infrared spectrum of the gas in the detection gas chamber. The data processing section (3) includes a main controller (32) connected to an infrared spectrometer. The main controller (32) is used to analyze the type and content of dissolved gases in transformer oil based on the external spectral diagram. The degassing device (2) includes a degassing tank (22), which is connected to the oil inlet of the transformer cylinder (1) via an oil inlet pipe (21), and the degassing tank (22) is connected to the oil outlet of the transformer cylinder (1) via a return oil pipe (27); both the oil inlet pipe (21) and the return oil pipe (27) are equipped with valves; The degassing tank (22) is equipped with a first pressure sensor (24); the degassing tank (22) is connected to the detection chamber through a third pipe (14), and a valve and a vacuum pump (16) are installed on the third pipe (14); a temperature sensor and a second pressure sensor (18) are installed at the interface of the detection chamber. A hydrogen sensor (25) is installed on the degassing tank (22).

2. The device for on-line monitoring of dissolved gases in transformer oil based on infrared spectroscopy according to claim 1, characterized in that, A gas collection cylinder (19) is connected to the third pipe (14) via a fifth pipe (17).

3. The device for on-line monitoring of dissolved gases in transformer oil based on infrared spectroscopy according to claim 1, characterized in that, An air pump (12) is installed on the degassing tank (22). The air pump (12) is connected to the first end of the air pipe, and the second end of the air pipe extends into the bottom of the degassing tank (22).

4. The device for on-line monitoring of dissolved gases in transformer oil based on infrared spectroscopy according to claim 1, characterized in that, The valve is a solenoid valve, and the degassing tank (22) is equipped with a liquid level sensor (23).

5. The monitoring method of the on-line monitoring device for dissolved gases in transformer oil based on infrared spectrum according to claim 1, characterized in that, Includes the following steps: Step 1: Clean the detection chamber with the degassing device (2); Step 2: Use the degassing device (2) to separate the dissolved gas from the transformer oil and send it to the detection gas chamber; Step 3: Use a Fourier transform infrared spectrometer to scan the detection chamber containing dissolved gases in the oil to obtain a single-wavelength image of the sample. Subtract the background single-wavelength image from the sample single-wavelength image to obtain the infrared spectrum of the gas. Step 4: Analyze the above infrared spectrum using a gas quantitative analysis model to obtain the dissolved gas content in the oil; Step 2.1: Use an infrared spectrometer to scan the detection cell to obtain a background single-wave pattern; Step 2.2, Perform oil-gas separation: Draw transformer oil from transformer cylinder (1) into degassing tank (22); start oil-gas separation, the dissolved gas in the transformer oil is separated into the space above the degassing tank (22), turn on the air pump (12) to send the gas above to the bottom of the degassing tank (22) and it will emerge from the bottom of the degassing tank (22), causing the dissolved gas in the oil to precipitate out, accelerating the separation of gas in the oil. After the gas separation is completed, turn off the air pump (12); open solenoid valve V5 and solenoid valve V7, close the other solenoid valves, and the piston of the gas collection cylinder (19) moves to collect the separated gas into the gas collection cylinder (19); Drive the piston of the gas collection cylinder (19) to move and send the gas into the detection chamber. At this time, the detection chamber is filled with separated gas. The temperature sensor and the second pressure sensor detect the temperature and pressure in the detection chamber at this time; calculate the pressure compensation coefficient. The method for establishing the quantitative gas analysis model is as follows: A standard gas experiment is conducted to obtain infrared spectral data of single and mixed components of each target gas at different concentrations. An absorbance matrix X is established as the input matrix, with the elements of the absorbance matrix representing the characteristic variables of the spectrum. A concentration matrix is ​​established as the output matrix Y. The absorbance and concentration matrices are standardized. Then, the first pair of components is extracted from the input and output matrices respectively. Two principal components are extracted from the components through matrix transformation, maximizing their correlation. Regression processing is performed on the extracted principal components and the original data matrix. If the regression accuracy meets the requirements, the process is terminated, resulting in the quantitative gas analysis model. The quantitative gas analysis model represents the relationship between the principal components, the input absorbance matrix, and the concentration matrix. Otherwise, the residual data matrix after principal component extraction is used for the next round of component extraction. Regression processing is performed on the principal components extracted from the output and input matrices, and the result is expressed as a regression against the original variables.

6. The monitoring method according to claim 5, characterized in that, The process of step 4 is as follows: input the infrared spectrum into the gas quantitative analysis model, the gas quantitative analysis model outputs the gas concentration of each component, and temperature and pressure compensation is performed through the pressure compensation coefficient to obtain the compensated gas concentration. The gas quantitative analysis model is based on infrared spectral scanning experiments of a large number of standard gases. The quantitative relationship between the absorption intensity of each target gas infrared spectral absorption peak or absorption band and its concentration is obtained based on the experimental results.

7. The monitoring method of claim 5, wherein, After step 4 is completed, the three-ratio method is used to determine whether the transformer may be faulty based on the compensated gas concentration results. If it is, the device will issue an alarm.

Images