Transformer on-line monitoring method and system

Abstract

The invention discloses a transformer on-line monitoring method and system, and relates to the technical field of power system on-line monitoring, and the method comprises the steps: collecting transformer data, and carrying out the data preprocessing; performing deep analysis on the preprocessed data based on a machine learning algorithm; and carrying out equipment operation state identification, and predicting a fault trend. The method provided by the invention realizes cleaning and optimization of original monitoring data, improves the accuracy and reliability of the data, provides a high-quality data basis for subsequent analysis, improves the consistency of data analysis, effectively removes interference components in signals, improves the accuracy of fault detection, and improves the reliability of fault detection. The device is ensured to operate in a safe range, the safety of the system is improved, the preventative maintenance is realized, the loss caused by sudden faults is reduced, and the stability and the economical efficiency of the power system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring of power systems, and in particular to a method and system for online monitoring of transformers. Background Art

[0002] With the development trend of automation and intelligence in power systems, transformer online monitoring technology has received widespread attention and application. By real-time acquisition of key parameters such as temperature, current, voltage, and gas in oil of transformers, combined with communication and computer technologies, remote monitoring of transformer operating status is achieved, greatly improving the operation and maintenance efficiency and safety level of power systems. However, with the expansion of power grid scale and the complexity of transformer structure, existing monitoring technologies are facing new challenges.

[0003] However, existing transformer online monitoring technologies still face a series of challenges in specific applications. First, in the data preprocessing stage, existing technologies generally lack targeted and effective data cleaning and feature selection methods. For example, for the massive amount of data collected by sensors, existing technologies are often unable to effectively eliminate outliers caused by environmental interference or equipment failures, resulting in noise interference in the subsequent analysis process. In terms of in-depth analysis, existing technologies often rely on simple threshold judgments or linear models. These methods have low state recognition accuracy when faced with the complex nonlinear operating characteristics of transformers and cannot meet the needs of high-precision monitoring. In particular, when the transformer is in a sub-healthy state, existing technologies often cannot accurately capture the precursor signals of faults. In terms of fault trend prediction, existing technologies lack prediction models based on historical data and machine learning algorithms, which cannot achieve early warning and trend prediction of potential transformer faults. For the changing trend of gas concentration in transformer oil, existing technologies have difficulty accurately predicting the occurrence time and severity of faults through time series analysis. The present invention improves the efficiency and accuracy of transformer monitoring by adopting efficient data preprocessing algorithms, advanced machine learning models and reliable fault prediction technologies. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing transformer online monitoring technology has the following problems: single data processing method, low accuracy in identifying equipment operating status, insufficient fault prediction model, and how to conduct in-depth data analysis and accurately predict fault trends.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a transformer online monitoring method, comprising collecting transformer data and performing data preprocessing; performing in-depth analysis of the preprocessed data based on a machine learning algorithm; identifying the equipment operating status and predicting fault trends.

[0007] As a preferred solution of the transformer online monitoring method described in the present invention, the transformer data includes temperature signals, electrical parameters, mechanical vibrations, gas in oil, and partial discharge signals.

[0008] Temperature signals include winding temperature and oil temperature.

[0009] Electrical parameters include voltage, current, and power.

[0010] Mechanical vibration data includes vibration spectrum and shock wave.

[0011] Gas in oil includes dissolved gas analysis (DGA).

[0012] Partial discharge signals include ultrasonic and ultra-high frequency.

[0013] As a preferred solution of the transformer online monitoring method described in the present invention, the data preprocessing includes, for transformer data, eliminating outliers outside the range of 1.5 times the interquartile range, eliminating data points outside the mean ±3σ, using K-nearest neighbor filling to clean up the multidimensional data of oil gas concentration, and using linear interpolation to process data with stable trends.

[0014] As a preferred solution of the transformer online monitoring method described in the present invention, the in-depth analysis includes normalizing and standardizing the data.

[0015] Normalized processing, expressed as:

[0016]

[0017] Among them, X ′ represents the standardized variable, X represents the original variable, and X min represents the minimum value of the original variable, X max Indicates the maximum value of the original variable.

[0018] Normalization processing is expressed as:

[0019]

[0020] Among them, μ represents the mean of the original variable, and σ represents the standard deviation of the original variable.

[0021] Filtering and denoising includes using low-pass filtering to remove low-frequency noise in temperature and current.

[0022] Use bandpass filtering to remove noise within a specific frequency range.

[0023] Wavelet transform is used to decompose the signal into components of different frequencies and remove the noise components in the partial discharge signal.

[0024] As a preferred solution of the transformer online monitoring method described in the present invention, the in-depth analysis also includes feature extraction.

[0025] Extract time domain features and calculate the mean μ, which is expressed as:

[0026]

[0027] Where N is the number of samples, x i is the data value of the ith sample, and n is the total number of data points in the dataset.

[0028] Calculate the variance σ 2 , expressed as:

[0029]

[0030] Calculate the skewness S, which measures the asymmetry of the data distribution, and is expressed as:

[0031]

[0032] Among them, σ is the standard deviation, which indicates the degree of dispersion of the data and is used to standardize (x i -μ) difference value.

[0033] Calculate the kurtosis K to evaluate the steepness of the data distribution, expressed as:

[0034]

[0035] Extracting frequency domain features involves performing Fourier transform on vibration and partial discharge signals to extract spectrum features X(f), which can be expressed as:

[0036]

[0037] Where x(t) is the signal function in the time domain, j represents the imaginary unit, f represents the frequency variable, and t is the time variable.

[0038] Extract the gas characteristics in the oil and calculate the ratio R1 of acetylene C2H2 to ethylene C2H4, which is expressed as:

[0039]

[0040] Calculate the ratio R2 of methane CH4 and hydrogen H2, expressed as:

[0041]

[0042] The gas ratio is calculated, and if the ratio exceeds a preset threshold, a discharge or overheating fault exists.

[0043] As a preferred solution of the transformer online monitoring method described in the present invention, the equipment operation status identification can identify the normal operation status, overheating status, insulation aging status, arc discharge status, and partial discharge status of the equipment in real time.

[0044] When all monitored parameters remain within the safe range, the equipment is in normal operation, characterized by low temperature, low harmonics and no partial discharge.

[0045] When the winding temperature is greater than 90°C or the oil temperature is greater than 80°C during equipment operation, the equipment is in an overheating state, characterized by increased equipment temperature and increased carbon dioxide concentration.

[0046] When CH4 / H2>0.1 or C2H4 / C2H6>1 is detected, a potential fault warning is issued, and the equipment is in an insulation aging state, characterized by increased methane and ethylene concentrations.

[0047] When C2H2 / C2H4>0.5, the equipment is in an arc discharge state, characterized by an abnormal increase in the concentration of acetylene.

[0048] When the detected discharge signal amplitude exceeds 500mV, the device is in a partial discharge state, characterized by a UHF signal reaching a peak.

[0049] As a preferred embodiment of the transformer online monitoring method of the present invention, the prediction of fault trends includes automatically triggering a fault warning based on the equipment data analysis results when equipment parameters exceed the normal range of the health index or show abnormal trends, and calculating the health index HI, which is expressed as:

[0050] HI=ω1T+ω2V+ω3G+ω4D+ω5P

[0051] Among them, T is the temperature feature, V is the vibration feature, G is the gas feature, D is the partial discharge feature, P is the electrical feature, and ω1, ω2, ω3, ω4, and ω5 represent the corresponding feature weight coefficients respectively.

[0052] Another object of the present invention is to provide a transformer online monitoring system that can identify the operating status of the equipment and predict fault trends, thereby solving the problems of current transformer online monitoring technology, such as imperfect fault warning and prediction mechanisms and the inability to effectively implement preventive maintenance.

[0053] As a preferred solution of the transformer online monitoring system described in the present invention, it includes: a data preprocessing module, a depth analysis module, and a fault prediction module.

[0054] The data preprocessing module is used to collect transformer data and perform data preprocessing; the in-depth analysis module is used to perform in-depth analysis on the preprocessed data based on a machine learning algorithm; and the fault prediction module is used to identify the equipment operating status and predict fault trends.

[0055] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a transformer online monitoring method.

[0056] A computer-readable storage medium stores a computer program, which implements the steps of a transformer online monitoring method when executed by a processor.

[0057] Beneficial effects of the present invention: The transformer online monitoring method provided by the present invention collects transformer data and performs data preprocessing, realizes the cleaning and optimization of the original monitoring data, improves the accuracy and reliability of the data, and provides a high-quality data basis for subsequent analysis. The preprocessed data is deeply analyzed based on the machine learning algorithm, which improves the consistency of data analysis, effectively removes the interference components in the signal, improves the accuracy of fault detection, identifies the equipment operating status, and predicts the fault trend, ensuring that the equipment operates within a safe range, improving the safety of the system, helping to achieve preventive maintenance, reducing the losses caused by sudden faults, and thus improving the stability and economy of the power system. The present invention achieves better results in improving data quality, fault identification accuracy and fault prediction efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0059] Figure 1 This is an overall flow chart of a transformer online monitoring method provided by the first embodiment of the present invention.

[0060] Figure 2 This is an overall flow chart of a transformer online monitoring system provided by the third embodiment of the present invention. DETAILED DESCRIPTION

[0061] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0062] Example 1, with reference to Figure 1 , as an embodiment of the present invention, provides a transformer online monitoring method, comprising:

[0063] S1: Collect transformer data and perform data preprocessing.

[0064] Furthermore, transformer data includes temperature signals, electrical parameters, mechanical vibrations, gas in oil, and partial discharge signals.

[0065] Temperature signals include winding temperature and oil temperature.

[0066] Electrical parameters include voltage, current, and power.

[0067] Mechanical vibration data includes vibration spectrum and shock wave.

[0068] Gas in oil includes dissolved gas analysis (DGA).

[0069] Partial discharge signals include ultrasonic and ultra-high frequency.

[0070] It should be noted that data preprocessing includes removing outliers outside the range of 1.5 times the interquartile range for transformer data, removing data points beyond the mean ±3σ, cleaning the multidimensional data of gas concentration in oil using K-nearest neighbor filling, and using linear interpolation to process data with stable trends.

[0071] It should also be noted that through precise data acquisition and efficient data preprocessing processes, accurate and reliable data input is provided for the transformer online monitoring system. The data collected includes multi-dimensional data such as temperature signals (winding temperature, oil temperature), electrical parameters (voltage, current, power), mechanical vibration data (vibration spectrum, shock wave), gas in oil (dissolved gas analysis DGA) and partial discharge signals (ultrasonic wave, ultra-high frequency). These data comprehensively cover the possible fault types of the transformer. The data preprocessing link includes data cleaning, outlier processing, data normalization and standardization steps. These processing methods improve the consistency and availability of the data, ensure the quality of the data, reduce errors in subsequent analysis, and improve the overall performance of the monitoring system; through efficient data preprocessing, the speed of data analysis is improved, making real-time monitoring possible; it provides a stable data foundation for subsequent deep learning and state recognition, thereby improving the accuracy and efficiency of fault diagnosis.

[0072] S2: Perform in-depth analysis on the preprocessed data based on machine learning algorithms.

[0073] Furthermore, in-depth analysis includes normalizing and standardizing the data.

[0074] Normalized processing, expressed as:

[0075]

[0076] Among them, X ′ represents the standardized variable, X represents the original variable, and X min represents the minimum value of the original variable, X max Indicates the maximum value of the original variable.

[0077] Normalization processing is expressed as:

[0078]

[0079] Among them, μ represents the mean of the original variable, and σ represents the standard deviation of the original variable.

[0080] Filtering and denoising includes using low-pass filtering to remove low-frequency noise in temperature and current.

[0081] Use bandpass filtering to remove noise within a specific frequency range.

[0082] Wavelet transform is used to decompose the signal into components of different frequencies and remove the noise components in the partial discharge signal.

[0083] It should be noted that in-depth analysis also includes feature extraction.

[0084] Extract time domain features and calculate the mean μ, which is expressed as:

[0085]

[0086] Where N is the number of samples, x i is the data value of the ith sample, and n is the total number of data points in the dataset.

[0087] Calculate the variance σ 2 , expressed as:

[0088]

[0089] Calculate the skewness S, which measures the asymmetry of the data distribution, and is expressed as:

[0090]

[0091] Among them, σ is the standard deviation, which indicates the degree of dispersion of the data and is used to standardize (x i -μ) difference value.

[0092] Calculate the kurtosis K to evaluate the steepness of the data distribution, expressed as:

[0093]

[0094] Extracting frequency domain features involves performing Fourier transform on vibration and partial discharge signals to extract spectrum features X(f), which can be expressed as:

[0095]

[0096] Where x(t) is the signal function in the time domain, j represents the imaginary unit, f represents the frequency variable, and t is the time variable.

[0097] Extract the gas characteristics in the oil and calculate the ratio R1 of acetylene C2H2 to ethylene C2H4, which is expressed as:

[0098]

[0099] Calculate the ratio R2 of methane CH4 and hydrogen H2, expressed as:

[0100]

[0101] The gas ratio is calculated, and if the ratio exceeds a preset threshold, a discharge or overheating fault exists.

[0102] It should also be noted that advanced machine learning algorithms are used to conduct in-depth analysis of pre-processed data. Through intelligent processing of the algorithm, valuable information is extracted from complex data to provide a basis for state identification and fault prediction. The use of various machine learning methods including time domain analysis, frequency domain analysis, feature extraction and pattern recognition can reveal the operating characteristics of the transformer from different angles. For example, by calculating the characteristics of gas in oil, such as the ratio of acetylene to ethylene, and the ratio of methane to hydrogen, potential faults such as discharge or overheating can be discovered in a timely manner; through in-depth analysis, the understanding of the complex operating state of the transformer is improved, providing a more accurate basis for fault diagnosis; the application of machine learning algorithms improves the level of automation of data analysis, reduces manual intervention, and improves the intelligence of the monitoring system; through continuous learning and optimization of the algorithm, the monitoring system has better adaptability and scalability, and can meet the monitoring needs under different working conditions.

[0103] S3: Identify the equipment operating status and predict failure trends.

[0104] Furthermore, the equipment operation status identification is performed to identify the normal operation status, overheating status, insulation aging status, arc discharge status, and partial discharge status of the equipment in real time.

[0105] When all monitored parameters remain within the safe range, the equipment is in normal operation, characterized by low temperature, low harmonics and no partial discharge.

[0106] When the winding temperature is greater than 90°C or the oil temperature is greater than 80°C during equipment operation, the equipment is in an overheating state, characterized by increased equipment temperature and increased carbon dioxide concentration.

[0107] When CH4 / H2>0.1 or C2H4 / C2H6>1 is detected, a potential fault warning is issued, and the equipment is in an insulation aging state, characterized by increased methane and ethylene concentrations.

[0108] When C2H2 / C2H4>0.5, the equipment is in an arc discharge state, characterized by an abnormal increase in the concentration of acetylene.

[0109] When the detected discharge signal amplitude exceeds 500mV, the device is in a partial discharge state, characterized by a UHF signal reaching a peak.

[0110] It should be noted that the prediction of fault trends includes automatically triggering fault warnings based on the results of equipment data analysis when equipment parameters exceed the normal range of the health index or show abnormal trends, and calculating the health index HI, which is expressed as:

[0111] HI=ω1T+ω2V+ω3G+ω4D+ω5P

[0112] Among them, T is the temperature feature, V is the vibration feature, G is the gas feature, D is the partial discharge feature, P is the electrical feature, and ω1, ω2, ω3, ω4, and ω5 represent the corresponding feature weight coefficients respectively.

[0113] It should also be noted that through the identification of equipment operating status and prediction of fault trends, real-time monitoring of the health status of the transformer and prediction of its future status are achieved; by real-time identification of the equipment's normal operating status, overheating status, insulation aging status, arc discharge status, and partial discharge status, immediate equipment status information is provided to operation and maintenance personnel. This status identification can not only detect problems in a timely manner, but also take corresponding maintenance measures based on the characteristics of different states; predicting fault trends predicts the future operating status of the equipment by calculating the health index. When the equipment parameters exceed the normal range or there is an abnormal trend, a fault warning is automatically triggered, which not only improves the equipment's operation and maintenance efficiency, but also reduces potential downtime risks and maintenance costs through preventive maintenance, thereby extending the service life of the equipment, ensuring the stable operation of the power system, and improving the reliability and economy of the power system.

[0114] Example 2 is an embodiment of the present invention, which provides a transformer online monitoring method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0115] First, the experiment selected six transformers with different operating conditions as experimental objects. These transformers were marked as A, B, C, D, E and F, where A and F were normal operating conditions, B was overheating condition, C was insulation aging condition, D was arc discharge condition, and E was partial discharge condition. Before the experiment, comprehensive data collection was carried out on the temperature signal, electrical parameters, mechanical vibration, gas in oil and partial discharge signal of each transformer. In the data preprocessing stage, outliers outside the range of 1.5 times the interquartile range were first removed, and then data points outside the mean ±3σ were removed. For the multidimensional data of gas concentration in oil, the K nearest neighbor filling method was used for data cleaning. For data with stable trend, linear interpolation was used to ensure the continuity and integrity of the data. In the in-depth analysis stage, the preprocessed data were normalized and standardized. The system is processed to eliminate the influence of different dimensions; filtering and denoising techniques, including low-pass filtering, band-pass filtering and wavelet transform, are used to remove noise in the signal; in the feature extraction process, the mean, variance, skewness and kurtosis of the time domain features, as well as the Fourier transform results of the frequency domain features, are calculated; for gas in oil, the ratio of acetylene to ethylene and the ratio of methane to hydrogen are calculated, and the presence of discharge or overheating faults is judged based on preset thresholds; equipment operating status identification and fault trend prediction are achieved through real-time monitoring and health index calculation; the health index is a comprehensive indicator that comprehensively considers temperature, vibration, gas, partial discharge and electrical characteristics, and adjusts the importance of each feature through weight coefficients; the present invention improves the accuracy of fault identification and the reliability of prediction, provides a valuable time window for preventive maintenance, and is creative and novel.

[0116] Example 3, reference Figure 2 , as an embodiment of the present invention, provides a transformer online monitoring system, including a data preprocessing module, a depth analysis module, and a fault prediction module.

[0117] The data preprocessing module is used to collect transformer data and perform data preprocessing; the deep analysis module is used to perform deep analysis on the preprocessed data based on machine learning algorithms; and the fault prediction module is used to identify the equipment operating status and predict fault trends.

[0118] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0119] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0120] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0121] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc. It should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention and are not intended to be limiting. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced with equivalents without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications should be encompassed by the claims of the present invention.

[0122] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A transformer online monitoring method, characterized in that: include: Collect transformer data and perform data preprocessing; Perform in-depth analysis of pre-processed data based on machine learning algorithms; Identify equipment operating status and predict failure trends.

2. The transformer online monitoring method according to claim 1, wherein: The transformer data includes temperature signals, electrical parameters, mechanical vibrations, gas in oil, and partial discharge signals; Temperature signals include winding temperature and oil temperature; Electrical parameters include voltage, current, and power; Mechanical vibration data includes vibration spectrum and shock wave; Gas in oil including dissolved gas analysis DGA; Partial discharge signals include ultrasonic and ultra-high frequency.

3. The method for computing platform load balancing based on particle swarm genetic algorithm according to claim 2, characterized in that: The data preprocessing includes removing outliers outside the range of 1.5 times the interquartile range for transformer data, removing data points outside the mean ±3σ, cleaning the multidimensional data of gas concentration in oil using K-nearest neighbor filling, and processing data with stable trends using linear interpolation.

4. The transformer online monitoring method according to claim 3, wherein: The in-depth analysis includes normalizing and standardizing the data; Normalized processing, expressed as: Among them, X ′ represents the standardized variable, X represents the original variable, and X min represents the minimum value of the original variable, X max Indicates the maximum value of the original variable; Normalization processing is expressed as: Among them, μ represents the mean of the original variable, σ represents the standard deviation of the original variable; Filtering and denoising include using low-pass filtering to remove low-frequency noise in temperature and current; Use bandpass filtering to remove noise within a specific frequency range; Wavelet transform is used to decompose the signal into components of different frequencies and remove the noise components in the partial discharge signal.

5. The transformer online monitoring method according to claim 4, characterized in that: The in-depth analysis also includes performing feature extraction; Extract time domain features and calculate the mean μ, which is expressed as: Where N is the number of samples, x i is the data value of the i-th sample, and n is the total number of data points in the data set; Calculate the variance σ 2 , expressed as: Calculate the skewness S, which measures the asymmetry of the data distribution, and is expressed as: Among them, σ is the standard deviation, which indicates the degree of dispersion of the data and is used to standardize (x i -μ) difference value; Calculate the kurtosis K to evaluate the steepness of the data distribution, expressed as: Extracting frequency domain features involves performing Fourier transform on vibration and partial discharge signals to extract spectrum features X(f), which can be expressed as: Where x(t) is the signal function in the time domain, j represents the imaginary unit, f represents the frequency variable, and t is the time variable; Extract the gas characteristics in the oil and calculate the ratio R1 of acetylene C2H2 to ethylene C2H4, which is expressed as: Calculate the ratio R2 of methane CH4 and hydrogen H2, expressed as: The gas ratio is calculated, and if the ratio exceeds a preset threshold, a discharge or overheating fault exists.

6. The transformer online monitoring method according to claim 5, characterized in that: The device operation status identification is performed to identify the normal operation status, overheating status, insulation aging status, arc discharge status, and partial discharge status of the device in real time; When all monitored parameters remain within the safe range, the equipment is in normal operation, characterized by low temperature, low harmonics and no partial discharge; When the winding temperature of the equipment is greater than 90°C or the oil temperature is greater than 80°C during operation, the equipment is in an overheating state, characterized by increased equipment temperature and increased carbon dioxide concentration; When CH4 / H2>0.1 or C2H4 / C2H6>1 is detected, a potential fault warning is issued, and the equipment is in an insulation aging state, characterized by increased methane and ethylene concentrations; When C2H2 / C2H4>0.5, the equipment is in an arc discharge state, characterized by an abnormal increase in the concentration of acetylene; When the detected discharge signal amplitude exceeds 500mV, the device is in a partial discharge state, characterized by a UHF signal reaching a peak.

7. The transformer online monitoring method according to claim 6, characterized in that: The prediction of fault trends includes automatically triggering a fault warning based on the equipment data analysis results when the equipment parameters exceed the normal range of the health index or have an abnormal trend, and calculating the health index HI, which is expressed as: HI=ω1T+ω2V+ω3G+ω4D+ω5P Among them, T is the temperature feature, V is the vibration feature, G is the gas feature, D is the partial discharge feature, P is the electrical feature, and ω1, ω2, ω3, ω4, and ω5 represent the corresponding feature weight coefficients respectively.

8. A system using the transformer online monitoring method according to any one of claims 1 to 7, characterized in that: Including data preprocessing module, in-depth analysis module, and fault prediction module; The data preprocessing module is used to collect transformer data and perform data preprocessing; The in-depth analysis module is used to perform in-depth analysis on the pre-processed data based on a machine learning algorithm; The fault prediction module is used to identify the equipment operating status and predict fault trends.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the transformer online monitoring method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the transformer online monitoring method according to any one of claims 1 to 7 are implemented.

Images