Fault early warning method for photoacoustic spectrometry oil gas detection

Through photoacoustic spectroscopy technology in oil and gas detection, through degassing and photoacoustic cell detection, combined with sound wave signal analysis and characteristic evaluation index, the problem of inaccurate judgment of fault types in oil and gas detection is solved, and high-accurate fault diagnosis and Early warning improves maintenance efficiency and reduces maintenance costs.

CN119064282APending Publication Date: 2024-12-03WUHAN HAOMAI OPTOELECTRONICS TECH CO LTD
View PDF 0 Cites 14 Cited by

Patent Information

Application Number
CN202411180938.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

In complex industrial environments, photoacoustic spectroscopy technology is easily interfered by multiple factors in oil and gas detection, resulting in deviations in the detection results, resulting in false alarms or missed alarms, and inaccurate judgment of fault types.

Method used

By extracting oil samples from the transformer, performing degassing treatment, and then using a laser to irradiate the gas in the photoacoustic cell, the gas molecules absorb laser energy and produce thermal expansion, capture the sound wave signals caused by gas pressure fluctuations, perform digital pre-processing and analysis, and extract oil and gas Detected characteristic data, build an oil and gas identification model, calculate the gas characteristic evaluation index and fault warning coefficient, divide the fault warning level and issue early warning signals.

Benefits of technology

Improve the accuracy of fault type prediction during oil and gas detection, reduce false alarms and missed reports, provide high-accurate fault diagnosis and early warning, enhance the fault prediction ability in complex working conditions, improve maintenance efficiency and Reduced maintenance costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064282A_ABST
    Figure CN119064282A_ABST
Patent Text Reader

Abstract

The invention discloses a fault early warning method for photoacoustic spectrometry oil gas detection, and relates to the technical field of oil gas detection, and the method comprises the following steps: extracting a transformer oil sample from a transformer, sending the oil sample into a degassing device, carrying out degassing treatment on the extracted transformer oil sample, and releasing gas dissolved in the oil. The high-accuracy fault diagnosis method is provided by analyzing the type and concentration of the gas dissolved in the transformer oil, the heat effect is generated by selective absorption of gas molecules to the laser with the specific wavelength, then the pressure fluctuation is generated, the gas components are identified by accurately measuring the pressure fluctuation, and the fault diagnosis accuracy is improved. According to the method, the fault types are rapidly and accurately identified by constructing a database containing different gas characteristics and combining signal processing and identification, the oil gas fault early warning coefficient is calculated based on the obtained data, and the potential fault risk is identified in advance, so that the reliability of fault early warning is improved, and the safety risk caused by the fault is effectively avoided or reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of oil and gas detection, and in particular to a fault early warning method for photoacoustic spectroscopy oil and gas detection. Background Art

[0002] Photoacoustic spectroscopy is a technology that generates acoustic signals for detection based on the interaction between light and matter. In the field of oil and gas detection, the acoustic signals generated by oil and gas samples under specific light irradiation are measured to analyze the oil and gas components and their contents. With the continuous development of photoacoustic spectroscopy, its detection accuracy, stability and reliability are constantly improving, providing a solid technical foundation for oil and gas detection. Therefore, photoacoustic spectroscopy is increasingly widely used in the online monitoring of dissolved gas in oil due to its high precision, wide detection range, and no consumption of the measured gas and carrier gas.

[0003] In the existing technology, although photoacoustic spectroscopy technology can detect a variety of gas components, it is still interfered by various factors in a complex industrial environment. In addition, the frequencies of different component gases are close, which is prone to cross-interference, resulting in deviations in the detection results, and then false alarms or missed alarms, and inaccurate fault type judgment. Therefore, how to improve the accuracy of fault type prediction in the oil and gas detection process is the problem we need to solve. To this end, a fault warning method for photoacoustic spectroscopy oil and gas detection is proposed. Summary of the invention

[0004] The present invention aims to provide a fault warning method for photoacoustic spectroscopy oil and gas detection to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A fault warning method for photoacoustic spectroscopy oil and gas detection comprises the following steps:

[0007] Step 1, extracting a transformer oil sample from a transformer, and sending the oil sample into a degassing device, degassing the extracted transformer oil sample to release the gas dissolved in the oil, releasing the gas dissolved in the oil, the degassing process needs to be efficient and the oil sample does not contact with the air to avoid cross contamination, collecting the released gas and sending it into a photoacoustic cell, performing subsequent photoacoustic spectrum detection in the photoacoustic cell, and using a laser to irradiate the gas in the photoacoustic cell, so that the gas molecules absorb the laser energy, generate a thermal effect and cause thermal expansion, so that the thermal expansion generates gas pressure fluctuations;

[0008] Step 2, using a photoacoustic detector to capture the acoustic wave signal data caused by the gas pressure fluctuation, and digitally preprocessing and analyzing the collected acoustic wave data;

[0009] Step 3: extract characteristic data of oil and gas detection based on the preprocessed acoustic wave data, and associate gas types, concentrations, components and gas ratios to obtain an oil and gas classification sequence;

[0010] Step 4: Combine the characteristic data of oil and gas detection and the oil and gas classification sequence to build an oil and gas identification model, analyze the gas type and composition, obtain the gas characteristic evaluation index, analyze the gas ratio and gas type and composition, obtain the gas ratio evaluation index, analyze the gas concentration changes of different types of gas, obtain the gas change rate evaluation index, and evaluate the gas state and characteristics of the oil sample;

[0011] Step 5: Based on the associated data of the oil and gas classification sequence, the gas characteristic evaluation index, the gas ratio evaluation index and the gas change rate evaluation index are combined to obtain the oil and gas fault warning coefficient, and the fault warning level is divided according to the fault type, location and severity, and the corresponding fault threshold is matched to the fault warning level;

[0012] Step 6: Input the detected data, analyze and obtain the results of the oil and gas fault warning coefficient, and combine it with the preset fault threshold to determine whether there is a fault risk, and send a warning signal of the corresponding level to the operation and maintenance personnel. According to the warning information, take on-site corrective measures for the corresponding equipment.

[0013] A further improvement of the technical solution of the present invention is that in step 1, the process of generating gas pressure fluctuation is:

[0014] Step 101, ensure that the operator wears protective equipment, use the transformer oil sampling tool to extract oil samples from the transformer oil tank, clean the sampling port before sampling to prevent contamination by impurities, and store the extracted oil samples in a clean and dry glass or metal container to ensure the representativeness of the oil sample during the sampling process and avoid the mixing of external pollutants;

[0015] Step 102, transferring the oil sample to a degassing device, ensuring that the degassing process is carried out in a sealed or inert gas environment to reduce the chance of the oil sample coming into contact with air, applying vacuum or nitrogen purge to release the gas dissolved in the oil so as to reduce the gas concentration in the oil sample to a minimum level, collecting the released gas and sending it to a photoacoustic cell, which is a container that can hold a gas sample and allow laser penetration, and the photoacoustic cell needs to be pre-cleaned and filled with helium to reduce background noise;

[0016] Step 103, in the photoacoustic pool, a tunable laser is used to emit a laser beam to irradiate the gas, so that the gas molecules absorb the laser energy through irradiation, a thermal effect occurs and the local temperature rises, thereby generating pressure fluctuations in the photoacoustic pool. The thermally expanded gas molecules cause the surrounding gas to form instantaneous high-pressure areas and low-pressure areas, thereby generating gas pressure fluctuations.

[0017] A further improvement of the technical solution of the present invention is that in step 2, the process of digital preprocessing and analysis of acoustic wave data is as follows:

[0018] Step 201, using a photoacoustic detector, such as a piezoelectric sensor, pre-placed in the photoacoustic cell to capture the acoustic wave signal generated by the gas pressure fluctuation in an analog form;

[0019] Step 202, digital preprocessing the sound wave data, wherein the digital preprocessing process includes sampling, quantization, filtering, denoising and signal correction;

[0020] Step 203, converting the analog signal of the digitally preprocessed sound wave data into a digital signal, and storing the digitized signal data in a data warehouse for further analysis.

[0021] A further improvement of the technical solution of the present invention is that in step 3, the process of obtaining the oil and gas classification sequence is as follows:

[0022] Step 301, using fast Fourier transform to convert the time domain signal to the frequency domain, so that the acoustic wave signal is converted into a spectrum, and the spectrum features are extracted, the characteristic frequency components generated by different gases in the signal are identified, and the energy distribution of the signal at different frequencies is displayed;

[0023] Step 302, identifying peaks in the frequency domain signal, corresponding to specific acoustic frequencies generated by gas molecules after absorbing laser energy, and extracting features related to gas type and concentration, including peak frequency, peak height, and peak area;

[0024] Step 303, matching the extracted features with the spectral database of known gases to identify the gas type, and using the Beer-Lambert law, based on the molar absorptivity, absorbance and optical path length of the gas, to determine the gas concentration, and correlating the calculated gas concentration with the peak characteristic parameters to establish a quantitative relationship between the gas type and concentration;

[0025] Furthermore, the calculation formula of gas concentration is:

[0026] Where c is the gas concentration, A is the absorbance, estimated by the peak area or peak height, ∈ is the molar absorptivity, an intrinsic property of the gas, and l is the light path length, which represents the path length of the light in the detection medium;

[0027] Step 304, based on the spectrum and energy distribution analysis, identify the characteristics of different gases in the acoustic signal, and establish a mapping relationship between gas concentration and acoustic signal characteristics, and then calculate the ratio of different gas characteristic peaks to distinguish indications of different types of faults;

[0028] Step 305: classify the oil sample gas according to the result of feature recognition analysis, and generate an oil and gas classification sequence in combination with the timestamp.

[0029] A further improvement of the technical solution of the present invention is that in step 4, the construction process of the oil and gas identification model is:

[0030] Step 401, traverse the original data of the acoustic wave signal of oil and gas detection, ensure that the data contains the characteristics generated by gases of different types and components, and pre-process the original data of the acoustic wave signal, including filtering, denoising, and baseline correction operations. The baseline correction is used to adjust the signal baseline to a stable level to facilitate subsequent analysis, so as to improve data quality and accuracy;

[0031] Step 402, extracting the features of peak frequency, peak height, and peak area from the preprocessed raw data of the acoustic wave signal, and creating a database containing different gas features, matching the detected features, each gas sample containing its corresponding feature vector and classification label, and then dividing the associated data in the database to obtain a training set and a test set;

[0032] Step 403, using a neural network model combined with the associated data of the oil and gas classification sequence, taking the extracted features as the input of the model, and using a training set with classification labels to train the model, optimizing the classification performance by adjusting the model parameters, and using an independent test data set for verification to obtain an oil and gas identification model;

[0033] Step 404, using the trained model to classify new oil and gas samples, identify the gas type, and analyze the proportion and content of each component in the gas based on the characteristic peaks in the spectral data;

[0034] Step 405, based on the extracted characteristic data of each gas, calculate the gas characteristic evaluation index to reflect the intensity and distribution of the gas characteristics; calculate the ratio between different gas characteristics to obtain the gas ratio evaluation index to assist in gas type differentiation and fault diagnosis; analyze the gas concentration data at different time points, calculate the rate of change of the gas concentration, and calculate the gas change rate evaluation index based on the rate of change and trend of the gas concentration to evaluate the trend of gas concentration change over time.

[0035] A further improvement of the technical solution of the present invention is that the calculation formula of the gas characteristic evaluation index is:

[0036]

[0037] Among them, A i represents the gas characteristic evaluation index, x i Indicates the characteristic value of the i-th gas, i represents the gas type index, from 1 to n, n represents the total number of gas types, Min(xi ) represents the minimum value of all gas characteristic values, Max(x i ) represents the maximum value among all gas characteristic values, WM(x i ) represents the average value of gas characteristic value, S(x i ) represents the standard deviation of all gas characteristic values, which measures the degree of data dispersion. α represents the adjustment factor, which is used to control the influence of the standard deviation on the evaluation index. represents the sum of all gas eigenvalues;

[0038] The calculation formula of the gas ratio evaluation index is:

[0039]

[0040] Among them, C ij represents the gas ratio evaluation index, x i and x j denote the characteristic values ​​of the i-th and j-th gases, respectively, where i≠j, represents the ratio of the eigenvalues ​​between the i-th gas and the j-th gas, represents the average value of the ratio between all different gas pairs, represents the standard deviation of the ratios between all different gas pairs, and Respectively represent the minimum and maximum values ​​of the ratios between all different gas pairs;

[0041] The calculation formula of the gas change rate evaluation index is:

[0042]

[0043] Among them, G t represents the gas change rate evaluation index, x t represents the gas characteristic value at the tth time point, T represents the total number of time points, represents the rate of change from time point t to time point t+1, represents the absolute value of the rate of change, and Respectively represent the minimum and maximum values ​​of all change rates, Represents extreme value ratio.

[0044] A further improvement of the technical solution of the present invention is that in step 5, the process of obtaining the oil and gas fault warning coefficient is:

[0045] Step 501, summarizing the associated data of the oil and gas classification sequence, and calculating the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index of each gas according to the acquired data;

[0046] Step 502, analyzing the relative position of each evaluation index, normalizing it, determining the importance of each evaluation index in fault warning, assigning a weight to each index, multiplying each normalized evaluation index by its corresponding weight, summing them, analyzing the change rate factor of the evaluation index, multiplying the weighted sum result by the change rate factor, and obtaining the oil and gas fault warning coefficient;

[0047] Step 503, collect historical fault data and analyze the severity of the fault in combination with the size of the oil and gas fault warning coefficient, and associate it with the fault type and location, and divide the fault warning levels into four different levels based on the fault type, location, and severity, namely, the first fault warning level, the second fault warning level, the third fault warning level, and the fourth fault warning level, and the fault warning level increases from the first level to the fourth level in severity;

[0048] Step 504: For each fault warning level, a corresponding fault threshold is matched, and corresponding warning measures are set according to the fault warning level.

[0049] A further improvement of the technical solution of the present invention is that the calculation formula of the oil and gas fault warning coefficient is:

[0050]

[0051] Among them, OG m Indicates the oil and gas fault warning coefficient, I m represents any one of the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index, m = 1, 2, 3 respectively corresponds to, w m Represents the weight of each index, satisfying Min(I) and Max(I) represent the minimum and maximum values ​​of the three evaluation indices, respectively. ∈ represents a very small positive number to prevent the denominator from being zero, and is set to 10. -6 , |ΔI Max | represents the absolute value of the maximum change in the three evaluation indices, |ΔI avg | represents the absolute value of the average change of the three evaluation indexes.

[0052] A further improvement of the technical solution of the present invention is that: the multiple fault warning levels correspond to multiple fault thresholds, wherein the fault thresholds include an upper threshold and a lower threshold;

[0053] The multiple fault warning levels and the multiple fault thresholds satisfy the following relationship:

[0054] The first level fault warning level is the normal level: 0 <OG m ≤OG ej ;

[0055] The second level fault warning level is the attention level: OG ej <OG m ≤OG sj ;

[0056] The third level fault warning level is the warning level: OG sj <OG m ≤OG fj ;

[0057] The fourth level of fault warning is the emergency level: OG m >OG fj ;

[0058] Among them, OG m is the oil and gas fault warning coefficient, OG ej is the lower threshold corresponding to the second-level fault warning level and the upper threshold corresponding to the first-level fault warning level, OG sj is the lower threshold corresponding to the third level fault warning level and the upper threshold corresponding to the second level fault warning level, OG fj is the lower threshold corresponding to the fourth level fault warning level and the upper threshold corresponding to the third level fault warning level, OG ej =0.2, OG sj =0.5, OG fj =0.8.

[0059] A further improvement of the technical solution of the present invention is that in step 6, the process of issuing the warning signal is:

[0060] Step 601, input the detected oil and gas acoustic wave signal data, perform feature extraction, and calculate the gas feature evaluation index, gas ratio evaluation index and gas change rate evaluation index based on the extracted feature data, and comprehensively analyze and obtain the oil and gas fault warning coefficient;

[0061] Step 602, comparing the oil and gas fault warning coefficient obtained by analysis with the preset fault threshold, determining the current fault warning level, and generating warning information including fault type, location, severity and recommended measures according to the fault warning level;

[0062] Step 603, the warning information is sent to the operation and maintenance personnel through SMS, email, and system notification. The operation and maintenance personnel receive the warning information and understand the current fault warning level and specific situation, decide whether to prepare corresponding equipment for on-site correction, and then take corresponding corrective measures according to the fault situation.

[0063] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:

[0064] 1. The present invention provides a fault warning method for photoacoustic spectroscopy oil and gas detection. By analyzing the type and concentration of dissolved gas in transformer oil, a high-accuracy fault diagnosis method is provided. The selective absorption of a specific wavelength of laser by gas molecules is used to produce a thermal effect, thereby generating pressure fluctuations. The gas composition is identified by accurately measuring the pressure fluctuations. A database containing different gas characteristics is constructed. Signal processing and recognition are combined to achieve rapid and accurate identification of fault types. The oil and gas fault warning coefficient is calculated based on the acquired data, and potential fault risks are identified in advance, thereby improving the reliability of fault warnings and effectively avoiding or reducing safety risks caused by faults.

[0065] 2. The present invention provides a fault warning method for photoacoustic spectroscopy oil and gas detection. By detecting the dissolved gas in the transformer oil, potential fault signs can be discovered in time. By fast Fourier transform and spectrum analysis of the acoustic signal, changes in gas composition can be quickly identified. By calculating the gas characteristic evaluation index, the gas ratio evaluation index and the gas change rate evaluation index, real-time tracking of the fault development trend can be achieved. Combined with the neural network model and historical fault data, the fault development trend can be predicted, and a warning can be issued before the fault occurs. The source of the fault can be accurately located, which can buy precious time for operation and maintenance personnel to handle the fault, provide clear maintenance directions, and quickly locate and handle the fault, greatly improving maintenance efficiency, thereby reducing losses caused by the fault.

[0066] 3. The present invention provides a fault warning method for photoacoustic spectroscopy oil and gas detection. Through the real-time monitoring and early warning functions of photoacoustic spectroscopy oil and gas detection technology, unnecessary disassembly and repair of the transformer can be reduced, thereby reducing maintenance costs and avoiding excessive maintenance due to misjudgment. At the same time, it also reduces emergency repairs caused by failure to discover faults in time. In addition, through multi-factor fault warning, the overall characteristics of the acoustic wave signal are comprehensively considered to more comprehensively evaluate the operating status of the equipment, which not only enhances the fault prediction capability under complex working conditions, but also improves the recognition rate of different types of faults, providing more reliable warning information for operation and maintenance personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0068] Figure 1 is a flow chart of the method of the present invention;

[0069] Figure 2 The flowchart of obtaining the oil and gas classification sequence of the present invention is as follows;

[0070] Figure 3 The flowchart of the oil and gas identification model of the present invention is as follows;

[0071] Figure 4 This is a flow chart for obtaining the oil and gas fault warning coefficient of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution 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 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 are within the scope of protection of the present invention.

[0073] Examples, such as Figure 1-4 As shown, the present invention provides a fault warning method for photoacoustic spectroscopy oil and gas detection, comprising the following steps:

[0074] Step 1: Extract transformer oil samples from the transformer and send the oil samples to the degassing device to degas the extracted transformer oil samples to release the gas dissolved in the oil. The degassing process must be efficient and the oil samples must not come into contact with the air to avoid cross contamination. The released gas is collected and sent to the photoacoustic cell for subsequent photoacoustic spectroscopy detection. In the photoacoustic cell, a laser is used to irradiate the gas so that the gas molecules absorb the laser energy, generate a thermal effect and cause thermal expansion, causing the thermal expansion to produce gas pressure fluctuations. Ensure that the operator wears protective equipment and use the transformer oil sampling tool to extract oil samples from the transformer oil tank. Before sampling, the sampling port must be cleaned to prevent contamination by impurities, and the extracted oil samples must be stored in a clean and dry glass or metal container to ensure the representativeness of the oil sample during the sampling process. and avoid the mixing of external pollutants, transfer the oil sample to the degassing device, ensure that the degassing process is carried out in a sealed or inert gas environment to reduce the chance of the oil sample contacting the air, apply vacuum or nitrogen purge to release the gas dissolved in the oil, so as to reduce the gas concentration in the oil sample to the lowest level, collect the released gas and send it to the photoacoustic cell, which is a container that can hold gas samples and allow laser penetration. The photoacoustic cell needs to be cleaned in advance and filled with helium to reduce background noise. In the photoacoustic cell, a tunable laser is used to emit a laser beam to irradiate the gas. The irradiation causes the gas molecules to absorb the laser energy, generate a thermal effect and increase the local temperature, thereby generating pressure fluctuations in the photoacoustic cell. It should be noted that the thermally expanded gas molecules will cause the surrounding gas to form instantaneous high-pressure and low-pressure areas, thereby generating gas pressure fluctuations;

[0075] Step 2, using a photoacoustic detector to capture the acoustic wave signal data caused by gas pressure fluctuations, and digitally preprocessing and analyzing the collected acoustic wave data, using a photoacoustic detector such as a piezoelectric sensor placed in the photoacoustic pool in advance, to capture the acoustic wave signal generated by the gas pressure fluctuation in an analog form for subsequent processing, and digitally preprocessing the acoustic wave data, wherein the digital preprocessing process includes sampling, quantization, filtering, denoising and signal correction, the specific sampling process is to discretize the continuously changing acoustic wave signal in time, that is, to collect a sample value on the acoustic wave signal at equal time intervals, and the sampling frequency satisfies the Nyquist theorem, that is, the sampling frequency should be greater than twice the highest frequency of the signal to ensure that the original signal can be accurately reconstructed; the specific quantization process is to convert the amplitude value of the sampled acoustic wave signal into a discrete Digital value is achieved by dividing the signal amplitude into several quantization levels and assigning each sample value to the nearest quantization level. The purpose of quantization is to enable analog signals to be represented and processed in digital form in computers; the specific process of filtering is to use digital filters to filter the collected sound wave signals to remove high-frequency noise and interference signals; the specific process of denoising is to use denoising techniques such as wavelet transform and empirical mode decomposition to further improve the signal-to-noise ratio and clarity of the signal; the specific process of signal correction is to perform baseline correction and amplitude correction on the sound wave signals to eliminate the influence of instrument errors and background noise, which helps to more accurately reflect the true situation of gas pressure fluctuations, convert the analog signals of the digitally preprocessed sound wave data into digital signals, and store the digitized signal data in the data warehouse for further analysis;

[0076] Step 3: Based on the pre-processed acoustic wave data, extract the characteristic data of oil and gas detection, and associate the gas type, concentration, composition and gas ratio to obtain the oil and gas classification sequence. Use fast Fourier transform to convert the time domain signal to the frequency domain so that the acoustic wave signal is converted into a spectrum diagram, and extract the spectrum characteristics, identify the characteristic frequency components generated by different gases in the signal, and display the energy distribution of the signal at different frequencies. Identify the peak in the frequency domain signal, which corresponds to the specific acoustic wave frequency generated by the gas molecules after absorbing the laser energy, and extract the characteristics related to the gas type and concentration, including peak frequency, peak height, peak area, and compare the extracted characteristics with the known gas The spectral database is matched to identify the gas type, and the Beer-Lambert law is used to determine the gas concentration based on the gas's molar absorption coefficient, absorbance and light path length. The calculated gas concentration is associated with the peak characteristic parameters to establish a quantitative relationship between the gas type and concentration. Based on the spectrum and energy distribution analysis, the characteristics generated by different gases in the acoustic signal are identified, and a mapping relationship between gas concentration and acoustic signal characteristics is established. The ratio of different gas characteristic peaks is calculated to distinguish the indications of different types of faults. According to the results of feature recognition analysis, the oil sample gas is classified, and combined with the timestamp, an oil and gas classification sequence is generated;

[0077] Furthermore, the calculation formula of gas concentration is:

[0078] Where c is the gas concentration, A is the absorbance, estimated by the peak area or peak height, ∈ is the molar absorptivity, an intrinsic property of the gas, and l is the light path length, which represents the path length of the light in the detection medium;

[0079] Step 4, combining the characteristic data of oil and gas detection and the oil and gas classification sequence, constructing an oil and gas identification model, analyzing the gas type and composition, obtaining the gas characteristic evaluation index, analyzing the gas ratio and the gas type and composition, obtaining the gas ratio evaluation index, analyzing the gas concentration changes of different types of gases, obtaining the gas change rate evaluation index, evaluating the gas state and characteristics of the oil sample, traversing the original data of the acoustic wave signal of oil and gas detection, ensuring that the data contains the characteristics generated by gases of different types and components, and preprocessing the original data of the acoustic wave signal, including filtering, denoising, and baseline correction operations. The baseline correction is used to adjust the signal baseline to a stable level to facilitate subsequent analysis to improve data quality and accuracy. The peak frequency, peak height, and peak area features are extracted from the preprocessed original data of the acoustic wave signal, and a database containing different gas features is created to match the detected features. Each gas sample contains its corresponding feature vector and classification label, and then classified. The associated data in the database are divided into training sets and test sets. The neural network model is combined with the associated data of the oil and gas classification sequence, the extracted features are used as the input of the model, and the model is trained using the training set with classification labels. The classification performance is optimized by adjusting the model parameters, and an independent test data set is used for verification to obtain the oil and gas identification model. The trained model is used to classify new oil and gas samples, identify the gas type, and analyze the proportion and content of each component in the gas based on the characteristic peaks in the spectral data. According to the extracted characteristic data of each gas, the gas characteristic evaluation index is calculated to reflect the intensity and distribution of the gas characteristics; the ratio between different gas characteristics is calculated to obtain the gas ratio evaluation index to assist in the distinction of gas types and fault diagnosis; the gas concentration data at different time points are analyzed, the rate of change of gas concentration is calculated, and the gas change rate evaluation index is calculated based on the rate of change and trend of gas concentration to evaluate the trend of gas concentration change over time;

[0080] Furthermore, the calculation formula of the gas characteristic evaluation index is:

[0081]

[0082] Among them, A i represents the gas characteristic evaluation index, x i Indicates the characteristic value of the i-th gas, i represents the gas type index, from 1 to n, n represents the total number of gas types, Min(x i ) represents the minimum value of all gas characteristic values, Max(x i ) represents the maximum value among all gas characteristic values, WM(x i ) represents the average value of gas characteristic value, S(x i ) represents the standard deviation of all gas characteristic values, which measures the degree of data dispersion. α represents the adjustment factor, which is used to control the influence of the standard deviation on the evaluation index. represents the sum of all gas eigenvalues;

[0083] The calculation formula of the gas ratio evaluation index is:

[0084]

[0085] Among them, C ij represents the gas ratio evaluation index, x i and x j denote the characteristic values ​​of the i-th and j-th gases, respectively, where i≠j, represents the ratio of the eigenvalues ​​between the i-th gas and the j-th gas, represents the average value of the ratio between all different gas pairs, represents the standard deviation of the ratios between all different gas pairs, and Respectively represent the minimum and maximum values ​​of the ratios between all different gas pairs;

[0086] The calculation formula of the gas change rate assessment index is:

[0087]

[0088] Among them, G t represents the gas change rate evaluation index, x t represents the gas characteristic value at the tth time point, T represents the total number of time points, represents the rate of change from time point t to time point t+1, represents the absolute value of the rate of change, and Respectively represent the minimum and maximum values ​​of all change rates, It is necessary to explain that when the gas characteristic value changes greatly over time and the changes are significantly different in different time periods, G t The value of G is relatively high. If the gas characteristic value hardly changes with time, or changes consistently, that is, the rate of change in all time periods is very close, then G t The value of will be lower;

[0089] Step 5, based on the associated data of the oil and gas classification sequence, combined with the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index, the oil and gas fault warning coefficient is obtained, and the fault warning level is divided based on the fault type, location and severity, and the corresponding fault threshold is matched to the fault warning level. The associated data of the oil and gas classification sequence is summarized, and the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index of each gas are calculated according to the acquired data, and the relative position of each evaluation index is analyzed and normalized to determine the importance of each evaluation index in fault warning, and a weight is assigned to each index. Each normalized evaluation index is multiplied by its corresponding The weights of the index are summed up, the change rate factor of the evaluation index is analyzed, and the weighted summation result is multiplied by the change rate factor to obtain the oil and gas fault warning coefficient. The historical fault data is collected and combined with the size of the oil and gas fault warning coefficient, the severity of the fault is analyzed, and it is associated with the fault type and location. Four different levels of fault warning levels are divided based on the fault type, location, and severity, namely, the first level fault warning level, the second level fault warning level, the third level fault warning level, and the fourth level fault warning level. The fault warning level increases step by step from the first level to the fourth level. For each fault warning level, the corresponding fault threshold is matched, and the corresponding warning measures are set according to the fault warning level;

[0090] Furthermore, the calculation formula of the oil and gas fault warning coefficient is:

[0091]

[0092] Among them, OG m Indicates the oil and gas fault warning coefficient, I m represents any one of the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index, m = 1, 2, 3 respectively corresponds to, w m Represents the weight of each index, satisfying Min(I) and Max(I) represent the minimum and maximum values ​​of the three evaluation indices, respectively. ∈ represents a very small positive number to prevent the denominator from being zero, and is set to 10. -6 , |ΔI Max | represents the absolute value of the maximum change in the three evaluation indices, |ΔI avg | represents the absolute value of the average change of the three evaluation indexes. It should be noted that when the value of the evaluation index is close to its historical maximum value, the normalized value will be close to 1, thereby increasing OG m When the rate of change of the evaluation index is high, that is, |ΔI Max |relative to |ΔI avg |Very large, the rate of change factor will increase significantly, further pushing up OG m Therefore, OG mThe value of increases as the evaluation index approaches its limit or the rate of change increases, reflecting the increased risk of oil and gas failure;

[0093] Furthermore, the multiple fault warning levels correspond to multiple fault thresholds, wherein the fault threshold includes an upper threshold and a lower threshold;

[0094] Multiple fault warning levels and multiple fault thresholds satisfy the following relationship:

[0095] The first level fault warning level is the normal level: 0 <OG m ≤OG ej , the equipment is operating normally, no special attention is needed, no special measures are needed, just maintain normal monitoring;

[0096] The second level fault warning level is the attention level: OG ej <OG m ≤OG sj , the equipment has a minor abnormality that requires attention but has not yet reached an emergency level. Increase the monitoring frequency and perform preventive maintenance. Check whether the relevant equipment has any abnormal signs and prepare an emergency plan, but do not start it for the time being;

[0097] The third level fault warning level is warning level: OG sj <OG m ≤OG fj , the equipment has obvious abnormalities and may be about to fail. Immediate action is required. The emergency plan should be activated immediately and relevant personnel should be notified to go to the site for inspection and prepare for repairs;

[0098] The fourth level of fault warning is the emergency level: OG m >OG fj , it is necessary to immediately shut down for inspection or emergency repair to prevent serious equipment failure or safety accidents, notify the professional maintenance team to come to the scene immediately and report to the superior management department;

[0099] Among them, OG m is the oil and gas fault warning coefficient, OG ej is the lower threshold corresponding to the second-level fault warning level and the upper threshold corresponding to the first-level fault warning level, OG sj is the lower threshold corresponding to the third level fault warning level and the upper threshold corresponding to the second level fault warning level, OG fj is the lower threshold corresponding to the fourth level fault warning level and the upper threshold corresponding to the third level fault warning level, OG ej =0.2, OG sj =0.5, OG fj =0.8;

[0100] Step 6: Input the detected data, analyze and obtain the result of the oil and gas fault warning coefficient, and combine it with the preset fault threshold to determine whether there is a fault risk, and send a warning signal of the corresponding level to the operation and maintenance personnel. According to the warning information, take on-site corrective measures for the corresponding equipment, input the detected oil and gas acoustic signal data, extract features, and calculate the gas feature evaluation index, gas ratio evaluation index and gas change rate evaluation index based on the extracted feature data, and comprehensively analyze to obtain the oil and gas fault warning coefficient. The oil and gas fault warning coefficient obtained by the analysis is compared with the preset fault threshold to determine the current fault warning level, and generate warning information including the fault type, location, severity and recommended measures according to the fault warning level. The warning information is sent to the operation and maintenance personnel through SMS, email, and system notification. The operation and maintenance personnel receive the warning information, understand the current fault warning level and specific situation, decide whether to prepare the corresponding equipment for on-site correction, and then take corresponding corrective measures according to the fault situation.

[0101] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A fault warning method for photoacoustic spectroscopy oil and gas detection, characterized in that: The following steps are involved: Step 1, extracting a transformer oil sample from a transformer, and sending the oil sample into a degassing device, degassing the extracted transformer oil sample to release the gas dissolved in the oil, collecting the released gas and sending it into a photoacoustic cell, using a laser to irradiate the gas, so that its thermal expansion generates gas pressure fluctuations; Step 2, using a photoacoustic detector to capture the acoustic wave signal data caused by the gas pressure fluctuation, and digitally preprocessing and analyzing the collected acoustic wave data; Step 3: extract characteristic data of oil and gas detection based on the preprocessed acoustic wave data, and associate gas types, concentrations, components and gas ratios to obtain an oil and gas classification sequence; Step 4: Combine the characteristic data of oil and gas detection and the oil and gas classification sequence to build an oil and gas identification model, analyze the gas type and composition, obtain the gas characteristic evaluation index, analyze the gas ratio and gas type and composition, obtain the gas ratio evaluation index, analyze the gas concentration changes of different types of gas, obtain the gas change rate evaluation index, and evaluate the gas state and characteristics of the oil sample; Step 5: Based on the associated data of the oil and gas classification sequence, the gas characteristic evaluation index, the gas ratio evaluation index and the gas change rate evaluation index are combined to obtain the oil and gas fault warning coefficient, and the fault warning level is divided according to the fault type, location and severity, and the corresponding fault threshold is matched to the fault warning level; Step 6: Input the detected data, analyze and obtain the results of the oil and gas fault warning coefficient, and combine it with the preset fault threshold to determine whether there is a fault risk, and send a warning signal of the corresponding level to the operation and maintenance personnel. According to the warning information, take on-site corrective measures for the corresponding equipment.

2. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 1 is characterized in that: In step 1, the process of generating gas pressure fluctuation is as follows: Step 101, using a transformer oil sampling tool to extract an oil sample from a transformer oil tank, and storing the extracted oil sample in a clean, dry glass or metal container; Step 102, transferring the oil sample to a degassing device to release the gas dissolved in the oil, collecting the released gas and sending it to a photoacoustic cell; Step 103, in the photoacoustic cell, a tunable laser is used to emit a laser beam to irradiate the gas, so that the gas molecules absorb the laser energy through irradiation, a thermal effect occurs and the local temperature rises, thereby generating pressure fluctuations in the photoacoustic cell.

3. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 2 is characterized in that: In step 2, the process of digital preprocessing and analysis of acoustic wave data is as follows: Step 201, using a photoacoustic detector pre-placed in the photoacoustic cell, to capture the acoustic wave signal generated by the gas pressure fluctuation in an analog form; Step 202, digital preprocessing the sound wave data, wherein the digital preprocessing process includes sampling, quantization, filtering, denoising and signal correction; Step 203, converting the analog signal of the digitally preprocessed sound wave data into a digital signal, and storing the digitized signal data in a data warehouse.

4. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 3 is characterized by: In step 3, the process of obtaining the oil and gas classification sequence is as follows: Step 301, using fast Fourier transform to convert the time domain signal to the frequency domain, so that the acoustic wave signal is converted into a spectrum, and the spectrum features are extracted, the characteristic frequency components generated by different gases in the signal are identified, and the energy distribution of the signal at different frequencies is displayed; Step 302, identifying peaks in the frequency domain signal, corresponding to specific acoustic frequencies generated by gas molecules after absorbing laser energy, and extracting features related to gas type and concentration, including peak frequency, peak height, and peak area; Step 303, matching the extracted features with the spectral database of known gases to identify the gas type, and using the Beer-Lambert law, based on the molar absorptivity, absorbance and optical path length of the gas, to determine the gas concentration, and correlating the calculated gas concentration with the peak characteristic parameters to establish a quantitative relationship between the gas type and concentration; Step 304, based on the spectrum and energy distribution analysis, identify the characteristics of different gases in the acoustic signal, and establish a mapping relationship between gas concentration and acoustic signal characteristics, and then calculate the ratio of different gas characteristic peaks to distinguish indications of different types of faults; Step 305: classify the oil sample gas according to the result of feature recognition analysis, and generate an oil and gas classification sequence in combination with the timestamp.

5. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 4 is characterized in that: In step 4, the construction process of the oil and gas identification model is as follows: Step 401, traverse the original data of the acoustic wave signal of oil and gas detection, and pre-process the original data of the acoustic wave signal, including filtering, denoising, and baseline correction operations; Step 402, extracting the features of peak frequency, peak height, and peak area from the preprocessed raw data of the acoustic wave signal, and creating a database containing different gas features, matching the detected features, each gas sample containing its corresponding feature vector and classification label, and then dividing the associated data in the database to obtain a training set and a test set; Step 403, using a neural network model combined with the associated data of the oil and gas classification sequence, taking the extracted features as the input of the model, and using a training set with classification labels to train the model, optimizing the classification performance by adjusting the model parameters, and using an independent test data set for verification to obtain an oil and gas identification model; Step 404, using the trained model to classify new oil and gas samples, identify the gas type, and analyze the proportion and content of each component in the gas based on the characteristic peaks in the spectral data; Step 405, based on the extracted characteristic data of each gas, calculate the gas characteristic evaluation index to reflect the intensity and distribution of the gas characteristics; calculate the ratio between different gas characteristics to obtain the gas ratio evaluation index to assist in gas type differentiation and fault diagnosis; analyze the gas concentration data at different time points, calculate the rate of change of the gas concentration, and calculate the gas change rate evaluation index based on the rate of change and trend of the gas concentration to evaluate the trend of gas concentration change over time.

6. A fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 5, characterized in that: The calculation formula of the gas characteristic evaluation index is: Among them, A i represents the gas characteristic evaluation index, x i Indicates the characteristic value of the i-th gas, i represents the gas type index, from 1 to n, n represents the total number of gas types, Min(x i ) represents the minimum value of all gas characteristic values, Max(x i ) represents the maximum value among all gas characteristic values, WM(x i ) represents the average value of gas characteristic value, S(x i ) represents the standard deviation of all gas characteristic values, which measures the degree of data dispersion. α represents the adjustment factor. represents the sum of all gas eigenvalues; The calculation formula of the gas ratio evaluation index is: Among them, C ij represents the gas ratio evaluation index, x i and x j denote the characteristic values ​​of the i-th and j-th gases, respectively, where i≠j, represents the ratio of the eigenvalues ​​between the i-th gas and the j-th gas, represents the average value of the ratio between all different gas pairs, represents the standard deviation of the ratios between all different gas pairs, and Respectively represent the minimum and maximum values ​​of the ratios between all different gas pairs; The calculation formula of the gas change rate evaluation index is: Among them, G t represents the gas change rate evaluation index, x t represents the gas characteristic value at the tth time point, T represents the total number of time points, represents the rate of change from time point t to time point t+1, represents the absolute value of the rate of change, and Respectively represent the minimum and maximum values ​​of all change rates, Represents extreme value ratio.

7. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 6 is characterized by: In step 5, the process of obtaining the oil and gas fault warning coefficient is as follows: Step 501, summarizing the associated data of the oil and gas classification sequence, and calculating the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index of each gas according to the acquired data; Step 502, analyzing the relative position of each evaluation index, normalizing it, determining the importance of each evaluation index in fault warning, assigning a weight to each index, multiplying each normalized evaluation index by its corresponding weight, summing them, analyzing the change rate factor of the evaluation index, multiplying the weighted sum result by the change rate factor, and obtaining the oil and gas fault warning coefficient; Step 503, collect historical fault data and analyze the severity of the fault in combination with the size of the oil and gas fault warning coefficient, and associate it with the fault type and location, and divide the fault warning levels into four different levels based on the fault type, location, and severity, namely, the first fault warning level, the second fault warning level, the third fault warning level, and the fourth fault warning level, and the fault warning level increases from the first level to the fourth level in severity; Step 504: For each fault warning level, a corresponding fault threshold is matched, and corresponding warning measures are set according to the fault warning level.

8. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 7 is characterized by: The calculation formula of the oil and gas fault warning coefficient is: Among them, OG m Indicates the oil and gas fault warning coefficient, I m represents any one of the gas characteristic evaluation index, gas ratio evaluation index and gas change rate evaluation index, m = 1, 2, 3 respectively corresponds to, w m Represents the weight of each index, satisfying Min(I) and Max(I) represent the minimum and maximum values ​​of the three evaluation indices, respectively. ∈ represents a very small positive number to prevent the denominator from being zero, and is set to 10. -6 , |ΔI Max | represents the absolute value of the maximum change in the three evaluation indices, |ΔI avg | represents the absolute value of the average change of the three evaluation indexes.

9. A fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 8, characterized in that: A plurality of the fault warning levels correspond to a plurality of the fault thresholds, wherein the fault thresholds include an upper threshold and a lower threshold; The multiple fault warning levels and the multiple fault thresholds satisfy the following relationship: Level 1 fault warning level 0 <OG m ≤OG ej ; Second level fault warning level OG ej <OG m ≤OG sj ; Level 3 fault warning level OG sj <OG m ≤OG fj ; Four-level fault warning level OG m >OG fj ; Among them, OG m is the oil and gas fault warning coefficient, OG ej is the lower threshold corresponding to the second-level fault warning level and the upper threshold corresponding to the first-level fault warning level, OG sj is the lower threshold corresponding to the third level fault warning level and the upper threshold corresponding to the second level fault warning level, OG fj is the lower threshold corresponding to the fourth level fault warning level and the upper threshold corresponding to the third level fault warning level, OG ej =0.2, OG sj =0.5, OG fj =0.

8.

10. The fault warning method for photoacoustic spectroscopy oil and gas detection according to claim 1, characterized in that: In step 6, the process of issuing the warning signal is: Step 601, input the detected oil and gas acoustic wave signal data, perform feature extraction, and calculate the gas feature evaluation index, gas ratio evaluation index and gas change rate evaluation index based on the extracted feature data, and comprehensively analyze and obtain the oil and gas fault warning coefficient; Step 602, comparing the oil and gas fault warning coefficient obtained by analysis with the preset fault threshold, determining the current fault warning level, and generating warning information including fault type, location, severity and recommended measures according to the fault warning level; Step 603, the warning information is sent to the operation and maintenance personnel through SMS, email, and system notification. The operation and maintenance personnel receive the warning information and understand the current fault warning level and specific situation, decide whether to prepare corresponding equipment for on-site correction, and then take corresponding corrective measures according to the fault situation.

Citation Information

Cited By

  • Method for detecting gas in transformer oil based on laser spectrum

    CN119757250A

  • A method for detecting gas in transformer oil based on laser spectroscopy

    CN119757250B

  • Transformer oil detection method and device, computer equipment, storage medium and computer program product

    CN119804340A

  • Transformer oil detection method and device, computer device, storage medium and computer program product

    CN119804340B

  • Method for measuring gas concentration in transformer oil

    CN119881276A