Transformer fault identification method and device and electronic equipment

Abstract

The application discloses a transformer fault identification method and device and electronic equipment. The method comprises the following steps: obtaining current online characteristic gas data and current offline characteristic gas data of a transformer in a current fault period; determining current index values of characteristic indexes corresponding to a plurality of characteristic gases generated in the current fault period based on the current online characteristic gas data and the current offline characteristic gas data; correcting the current online characteristic gas data by using a pre-determined strong correlation rule according to the current index values of the characteristic indexes corresponding to the plurality of characteristic gases, to obtain corrected online characteristic gas data; and obtaining a fault identification result of the transformer in the current fault period by using a pre-trained fault identification model based on the corrected online characteristic gas data. The application solves the technical problem of low fault identification accuracy in directly identifying transformer faults by using an oil chromatography online detection device.

Description

Technical Field

[0001] This invention relates to the field of transformer testing, and more specifically, to a method, apparatus, and electronic device for identifying transformer faults. Background Technology

[0002] Among related technologies, online oil chromatography detection systems are the most widely used transformer fault early warning systems. They judge faults by comparing the content of characteristic gases dissolved in the oil with specified values. These characteristic gases mainly include hydrogen (H2), methane (CH4), ethylene (C2H2), acetylene (C2H4), ethane (C2H6), carbon monoxide (CO), carbon dioxide (CO2), and total hydrocarbons. Since different transformer faults produce different gases, a preliminary judgment can be made on the presence and nature of the fault based on the characteristic gases generated by different fault types. However, in practical applications, online oil chromatography detection devices experience numerous and recurring faults, resulting in extensive daily maintenance. Nearly half of these faults are false alarms caused by online detection device malfunctions (including host device failures and data upload mode failures). Therefore, directly using online oil chromatography detection devices for transformer fault identification has a low accuracy rate.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides a transformer fault identification method, apparatus, and electronic device to at least solve the technical problem of low fault identification accuracy when directly identifying transformer faults using an online oil chromatography detection device.

[0005] According to one aspect of the present invention, a transformer fault identification method is provided, comprising: acquiring current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, the current online characteristic gas data being collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data being obtained based on an offline device; determining the current index values ​​of characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the current online and offline characteristic gas data; correcting the current online characteristic gas data using a pre-determined strong association rule based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, to obtain corrected online characteristic gas data, wherein the strong association rule is used to indicate the association relationship between the various characteristic gases; and obtaining the fault identification result of the transformer during the current fault period using a pre-trained fault identification model based on the corrected online characteristic gas data of the transformer during multiple historical fault periods and the fault types corresponding to the multiple historical fault periods.

[0006] According to another aspect of the present invention, a transformer fault identification device is also provided, comprising: an acquisition module, configured to acquire current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, the current online characteristic gas data is collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on detection by an offline device; and a determination module, configured to determine, based on the current online characteristic gas data and the current offline characteristic gas data, the current characteristic indicators corresponding to the various characteristic gases generated during the current fault period. The system includes: an index value correction module, which corrects the current online characteristic gas data based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, using pre-determined strong correlation rules to obtain corrected online characteristic gas data, wherein the strong correlation rules are used to indicate the correlation between the various characteristic gases; and a detection module, which uses a pre-trained fault identification model based on the corrected online characteristic gas data to obtain the fault identification result of the transformer during the current fault period, wherein the fault identification model is trained based on historical online characteristic gas data corresponding to the transformer during multiple historical fault periods and the fault types corresponding to the multiple historical fault periods.

[0007] According to another aspect of the present invention, an electronic device is also provided, including one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement any of the above-described transformer fault identification methods.

[0008] In this embodiment of the invention, by acquiring current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, the current online characteristic gas data is collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on an offline device detection; based on the current online and offline characteristic gas data, the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period are determined; according to the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, the current online characteristic gas data is corrected using a pre-determined strong correlation rule to obtain... The corrected online characteristic gas data, wherein the aforementioned strong correlation rules are used to indicate the correlation between the aforementioned multiple characteristic gases; based on the aforementioned corrected online characteristic gas data, a pre-trained fault identification model is used to obtain the fault identification result of the aforementioned transformer during the aforementioned current fault period, wherein the aforementioned fault identification model is trained based on the historical online characteristic gas data corresponding to the aforementioned transformer during multiple historical fault periods, and the fault types corresponding to the aforementioned multiple historical fault periods, thereby achieving the purpose of comprehensively identifying transformer faults based on the strong correlation rules between various characteristic gases and the fault identification model, thus realizing the technical effect of improving the accuracy of transformer fault identification, and thus solving the technical problem of low fault identification accuracy when directly using an online oil chromatography detection device for transformer fault identification. Attached Figure Description

[0009] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0010] Figure 1 This is a schematic diagram of a transformer fault identification method according to an embodiment of the present invention;

[0011] Figure 2 This is a schematic diagram of an optional transformer fault identification method according to an embodiment of the present invention;

[0012] Figure 3 This is a schematic diagram of a transformer fault identification device according to an embodiment of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0015] As the core equipment in the power grid system, the transformer's operating performance directly affects the reliability and quality of power supply. With the rapid expansion of urban power grids, the number of operating transformers has increased dramatically, highlighting the growing pressure on equipment maintenance. Therefore, it is necessary to monitor the internal main components of transformers during operation to provide early warning of potential faults. Experience shows that over 70% of common internal transformer faults can be effectively detected and warned of through analysis of dissolved gases in the oil. This technology is currently the most common method for assessing the operating status of transformers. Currently, the main method for internal detection and early warning of transformer tank faults is the online dissolved gas detection system based on gas chromatography. As one of the important means of transformer insulation testing in the power system, it can effectively detect latent faults and their development trends within the transformer. It is generally believed that the content, composition, and ratio of the measured characteristic gases can be used to determine the type of fault that may exist inside the transformer. This method has advantages such as accurate results, fast detection speed, and small sample size.

[0016] Among related technologies, online oil chromatography detection systems are the most widely used transformer fault early warning systems. They judge faults by comparing the content of characteristic dissolved gases in the oil with specified values. These characteristic gases mainly include H2, CH4, C2H2, C2H4, C2H6, CO, CO2, and total hydrocarbons. Since different transformer faults produce different gases, a preliminary judgment can be made regarding the presence and nature of the fault based on the characteristic gases generated by different fault types. However, in practical applications, online oil chromatography detection devices experience numerous and recurring faults, resulting in extensive daily maintenance. Nearly half of these faults are false alarms caused by malfunctions in the online detection device (including host failures and data upload mode failures). Therefore, directly using online oil chromatography detection devices for transformer fault identification has a low accuracy rate.

[0017] Based on the problem, this invention provides a method embodiment for transformer fault identification. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0018] Figure 1 This is a flowchart of a transformer fault identification method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0019] Step S102: Obtain the current online characteristic gas data and the current offline characteristic gas data of the transformer during the current fault period. The current online characteristic gas data includes the current online characteristic data corresponding to various characteristic gases generated during the current fault period. The current online characteristic gas data is collected based on the online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on the offline device detection.

[0020] Optionally, the various characteristic gases may include, but are not limited to, gases such as H2, CH4, C2H2, C2H4, C2H6, CO, and CO2. Currently, the online characteristic gas data is obtained through online oil chromatography detection of the substation transformer; currently, the offline characteristic gas data is obtained through offline detection in the laboratory. The current online characteristic data may include, but is not limited to, concentration data of the characteristic gases.

[0021] Step S104: Based on the current online characteristic gas data and the current offline characteristic gas data, determine the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period.

[0022] In one optional embodiment, based on the current online characteristic gas data and the current offline characteristic gas data, the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period are determined, including: denoising the current online characteristic gas data and the current offline characteristic gas data to obtain denoised current online characteristic gas data and denoised current offline characteristic gas data; and determining the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the denoised current online characteristic gas data and the denoised current offline characteristic gas data.

[0023] Optionally, spectral analysis can be used to denoise the current online and offline characteristic gas data. It can be understood that useful signals are typically low-frequency or relatively stable, while noise signals are typically high-frequency. After decomposing the noisy original signal using spectral analysis, the noisy portion is mainly concentrated in the high-frequency signal. First, Fourier processing is performed on the time series of the online and offline characteristic gas data to convert the discrete signal into a continuous signal. Power spectrum analysis is then performed on the continuous signal, its frequency domain waveform is plotted, and the spectral waveform characteristics of the continuous signal are analyzed. The waveform is then fitted to obtain the power spectrum distribution. After changing the parameters, an inverse transform is performed to obtain the filtered result, i.e., the denoised current online and offline characteristic gas data.

[0024] In an optional embodiment, when there are multiple feature indicators, and these multiple feature indicators include at least: relative length, relative height, and relative error, based on the denoised current online feature gas data and the denoised current offline feature gas data, the current indicator values ​​of the feature indicators corresponding to the various feature gases generated during the current fault period are determined, including:

[0025] The current relative length corresponding to any characteristic gas generated during the current fault period is determined as follows: determine the duration of any characteristic gas during the current fault period, and the total duration of the current fault period; take the proportion of the duration to the total duration as the current relative length corresponding to any characteristic gas generated during the current fault period.

[0026] The current relative altitude corresponding to any one of the characteristic gases generated during the current fault period is determined as follows: the first difference between the maximum and minimum values ​​of the current online characteristic data corresponding to any one of the characteristic gases during the current fault period, and the second difference between the maximum and minimum values ​​of the current online characteristic data corresponding to each of the multiple characteristic gases are determined; the proportion of the first difference to the second difference is taken as the current relative altitude corresponding to any one of the characteristic gases generated during the current fault period.

[0027] The current relative error corresponding to any characteristic gas generated during the current fault period is determined as follows: the third difference between the current online characteristic data and the corresponding current offline characteristic data for any characteristic gas is determined, and the third difference is used as the current relative error corresponding to any characteristic gas generated during the current fault period.

[0028] Optionally, there may be multiple characteristic indicators, including at least: relative length, relative height, and relative error. Relative length refers to the ratio of the time length corresponding to any one characteristic gas in the entire fault to the total sequence length corresponding to the entire fault. Relative height refers to the ratio of the absolute value of the difference between the maximum and minimum values ​​of any one characteristic gas within the fault band to the absolute value of the difference between the maximum and minimum values ​​in the entire sequence. Relative error refers to the relative error between the characteristic gas data from the online oil chromatography detection device and the offline characteristic gas data from the laboratory within the change band of any one characteristic gas.

[0029] Step S106: Based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, the current online characteristic gas data is corrected using a pre-determined strong correlation rule to obtain the corrected online characteristic gas data. The strong correlation rule is used to indicate the correlation between the various characteristic gases.

[0030] In one optional embodiment, based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, a pre-determined strong correlation rule is used to correct the current online characteristic gas data to obtain corrected online characteristic gas data. This includes: based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, using a strong correlation rule, determining similar characteristic gas data corresponding to the current online characteristic gas data from historical online characteristic gas data corresponding to multiple historical fault periods; and correcting the current online characteristic gas data based on the similar characteristic gas data to obtain corrected online characteristic gas data.

[0031] Optionally, the characteristic data (such as gas concentration) corresponding to the characteristic gas of each type of fault usually have certain regularities. These regularities can be queried through strong association rules, thereby finding the similar characteristic gas data with the most similar structure to the online characteristic gas data of the current fault period. Based on the similar characteristic gas data, the online characteristic gas data of the current fault period can be corrected.

[0032] In an optional embodiment, the historical online characteristic gas data includes historical online characteristic data corresponding to various characteristic gases generated during the corresponding fault period. Before determining similar characteristic gas data corresponding to the current online characteristic gas data from the historical online characteristic gas data corresponding to the various characteristic gases based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases and using strong association rules, the method further includes: determining the historical index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the various historical fault periods based on the historical online characteristic gas data corresponding to the various historical fault periods and the historical offline characteristic gas data corresponding to the various historical fault periods; obtaining frequent itemsets with Boolean association rules corresponding to the characteristic indicators based on the historical index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the various historical fault periods; and determining strong association rules based on the frequent itemsets.

[0033] Optionally, the historical online characteristic gas data corresponding to multiple historical fault periods are obtained from the online oil chromatography detection device of the substation transformer; the historical offline characteristic gas data corresponding to multiple historical fault periods are obtained from the offline detection device in the laboratory. Spectral analysis is used to denoise the historical online and offline characteristic gas data corresponding to multiple historical fault periods. Based on the processed historical online and offline characteristic gas data corresponding to multiple historical fault periods, historical values ​​of three indicators—relative length, relative height, and relative error—are calculated. Based on these historical indicator values, a frequent itemset of Boolean association rules is generated using an association rule mining algorithm (i.e., the Apriori algorithm). Strong association rules satisfying the minimum confidence level are constructed using these frequent itemsets.

[0034] Optionally, strong association rules can be generated from frequent itemsets. Strong association rules must simultaneously satisfy a minimum support and a minimum confidence. Specifically, all frequent itemsets are iterated over, and then 1, 2, ..., k elements are sequentially selected from each frequent itemset as consequents, with the remaining elements in the itemset as antecedents. The confidence of the rule is then calculated for selection. It is possible, but not limited to, selecting a minimum support (min_sup) of 5% and a minimum confidence (min_conf) of 60% as the criteria for establishing strong association rules. The definitions of support and confidence are as follows:

[0035] Regarding support and association rules Support refers to the percentage of transactions in a frequent itemset D that include A∪B (i.e., transactions that simultaneously contain the predecessor A and the successor B).

[0036]

[0037] For confidence level, the confidence level of an association rule refers to the percentage of transactions containing both A and B to the total number of transactions containing B. The specific formula is as follows.

[0038]

[0039] in, The confidence level of the association rule is represented by Support(A∪B), which represents the number of transactions between A and B, and Support(A) represents the number of transactions between B.

[0040] In one optional embodiment, the online characteristic gas data is corrected based on similar characteristic gas data to obtain corrected online characteristic gas data, including: determining a distance index value between similar characteristic gas data and the current online characteristic gas data, wherein the distance index value is used to indicate the degree of difference between similar characteristic gas data and the current online characteristic gas data; and correcting the online characteristic gas data based on the distance index value to obtain corrected online characteristic gas data.

[0041] Optionally, similar characteristic gas data is first determined based on pre-mined strong association rules. This involves selecting the most similar characteristic gas data from historical online characteristic gas data and using a distance index to characterize the similarity between the two. A smaller distance index indicates greater similarity. The similarity can then be used to correct the online characteristic gas data. For example, if the distance index is small and less than a preset distance threshold, the similar characteristic gas data is used as the corrected online characteristic gas data. If the distance index is large and greater than the preset distance threshold, other methods are used to correct the online characteristic gas data; for instance, the similar characteristic gas data can be multiplied by a corresponding correction coefficient to obtain the corrected online characteristic gas data.

[0042] Optionally, historical online characteristic gas data has high accuracy. By using the above method, similar characteristic gas data can be selected from the historical online characteristic gas data to correct the current online characteristic gas data, so as to eliminate possible error interference in the current online characteristic gas data.

[0043] Optionally, the distance index value between similar characteristic gas data and currently online characteristic gas data is determined, including: determining the distance index value between similar characteristic gas data and currently online characteristic gas data in the following manner:

[0044]

[0045] Among them, Ft A represents the current online characteristic gas data at any sampling time t within the current fault period; t' For the historical fault periods corresponding to the current fault period, ||F| represents the similar characteristic gas data at historical sampling time t′ corresponding to any sampling time t. t A t' || represents the distance index between similar characteristic gas data and currently online characteristic gas data, N v w is the number of physical quantities that affect the distance index value, which is set in advance. i Let be the weight of any physical quantity that affects the distance index value. It is half the duration of the fault period; F represents the standard deviation between the sampled values ​​of any physical quantity during the current fault period. i,t+j Let A be the index value of the i-th characteristic gas among multiple characteristic gases in the current online characteristic gas data at time t+j. i,t'+j Let t'+i be the index value of the i-th characteristic gas among multiple characteristic gases in the similar characteristic gas data.

[0046] Optionally, the historical fault periods corresponding to the current fault period are those with similar characteristics to the current fault period. It should be noted that since a fault is a process, when determining the distance index between similar characteristic gas data at a certain moment (e.g., time t) and the currently online characteristic gas data, similarity is sought based on the first half of time t (i.e., the predetermined duration before time t, which is half the duration corresponding to the fault period) and the second half of the time period, rather than just time t itself. The distance index values ​​obtained through this method are more realistic and have higher accuracy.

[0047] In one optional embodiment, when there are multiple sets of similar characteristic gas data, the current online characteristic gas data is corrected based on the similar characteristic gas data to obtain corrected online characteristic gas data, including: determining the weight values ​​corresponding to the multiple sets of similar characteristic gas data respectively; and obtaining the corrected online characteristic gas data based on the multiple sets of similar characteristic gas data and the weight values ​​corresponding to the multiple sets of similar characteristic gas data respectively.

[0048] Optionally, when there are multiple sets of similar characteristic gas data, the corrected online characteristic gas data can be determined by comprehensively considering all sets of similar characteristic gas data. Specifically, based on multiple sets of similar characteristic gas data and their respective weight values, the corrected online characteristic gas data is obtained through a weighted calculation method, expressed by the following formula:

[0049]

[0050] Among them, AN t This represents any set of similar characteristic gas data from multiple sets of similar characteristic gas data, with units of μL / L; N a This represents the total number of sets of similar characteristic gas data; This represents multiple sets of similar characteristic gas data; t i This represents the historical fault periods corresponding to multiple sets of similar characteristic gas data. The weight values ​​for each set of similar characteristic gas data are obtained using the following formula.

[0051]

[0052] Step S108: Based on the corrected online characteristic gas data, a pre-trained fault identification model is used to obtain the fault identification result of the transformer during the current fault period. The fault identification model is trained based on the historical online characteristic gas data corresponding to the transformer during multiple historical fault periods and the fault types corresponding to the multiple historical fault periods.

[0053] Using the methods described above, fault identification is performed based on the corrected online characteristic gas data through model prediction. The corrected online characteristic gas data has higher reliability, thus making the final fault identification results more accurate and reliable.

[0054] Through steps S102 to S108, the goal of transformer fault identification can be achieved by combining strong correlation rules between various characteristic gases and fault identification models, thereby improving the technical effect of transformer fault identification accuracy and solving the technical problem of low fault identification accuracy when directly identifying transformer faults using an online oil chromatography detection device.

[0055] Based on the embodiments and optional embodiments, the present invention proposes an optional implementation method. Figure 2 This is a flowchart of an optional transformer fault identification method according to an embodiment of the present invention, such as... Figure 2 As shown, the method includes a strong association rule and fault identification model establishment phase and a testing phase. Specifically, the strong association rule and fault identification model establishment phase includes:

[0056] Step S11: Based on the online oil chromatography detection device of the substation transformer, acquire historical online characteristic gas data corresponding to multiple historical fault periods, and obtain historical offline characteristic gas data corresponding to multiple historical fault periods based on laboratory offline measurements; use spectrum analysis to denoise the historical online characteristic gas data and the historical offline characteristic gas data corresponding to multiple historical fault periods (specifically including wavelet transform multi-scale decomposition, wavelet coefficient denoising at each scale, inverse wavelet transform signal reconstruction, and wavelet denoising) to obtain the processed historical online characteristic gas data and the processed historical offline characteristic gas data corresponding to multiple historical fault periods in the form of a smooth sequence. It can be understood that the useful signal is usually a low-frequency signal or relatively stable.

[0057] Step S12 involves performing frequent itemset analysis and association rule mining based on the processed historical online and offline characteristic gas data corresponding to multiple historical fault periods. Specifically, this includes: calculating historical values ​​for relative length, relative height, and relative error based on the processed historical online and offline characteristic gas data corresponding to multiple historical fault periods; generating frequent itemsets of Boolean association rules using the Apriori algorithm based on these historical values; constructing strong association rules that satisfy the minimum confidence level using these frequent itemsets; and determining strong association rules that satisfy a minimum support and minimum confidence level based on the frequent itemsets.

[0058] Step S13: Based on the historical online characteristic gas data corresponding to the transformer in multiple historical fault periods, and the fault types corresponding to the multiple historical fault periods, a fault identification model is trained.

[0059] The testing phase specifically includes:

[0060] Step S21: Using the same processing method as step S1, the online characteristic gas data of the transformer during the current fault period is obtained using an online oil chromatography monitoring device, and the offline characteristic gas data of the transformer during the current fault period is obtained using a laboratory offline device. The online and offline characteristic gas data are then denoised using a spectrum analysis method (specifically including wavelet transform multi-scale decomposition, wavelet coefficient denoising at each scale, inverse wavelet transform signal reconstruction, and wavelet denoising) to obtain the processed online and offline characteristic gas data.

[0061] Step S22: Based on the processed current online characteristic gas data and the processed current offline characteristic gas data, calculate the current index values ​​of the three indicators of relative length, relative height, and relative error corresponding to the various characteristic gases generated during the current fault period.

[0062] Step S23: Based on the current index values ​​of the relative length, relative height, and relative error corresponding to the various characteristic gases generated during the current fault period, a strong association rule is used to correct the error of the processed current online characteristic gas data. Specifically, this includes: determining similar characteristic gas data corresponding to the current online characteristic gas data from the historical online characteristic gas data corresponding to multiple historical fault periods based on the strong association rule; determining the distance index value between the similar characteristic gas data and the current online characteristic gas data, wherein the distance index value is used to indicate the degree of difference between the similar characteristic gas data and the current online characteristic gas data; and correcting the online characteristic gas data based on the distance index value to obtain the corrected online characteristic gas data.

[0063] Step S24: Based on the corrected online characteristic gas data, a pre-trained fault identification model is used to obtain the fault identification result of the transformer during the current fault period.

[0064] It should be noted that existing online transformer oil chromatography monitoring devices frequently experience false alarms. Data interruptions and abnormal malfunctions account for the highest proportion of these failures. Communication or network issues can cause data interruptions or anomalies in certain gas components, or even data stagnation, where data remains unchanged for extended periods. These are all abnormal data points, and these false alarms often lead to substation outages for maintenance, requiring significant manpower and resources. Since there are discrepancies between online and offline data, it is necessary to correct characteristic gas values. Based on the characteristic gas values ​​from past faults, the fault type can be directly determined, guiding maintenance personnel to the next steps. This invention fully considers the changes in transformer gases caused by internal transformer issues. It utilizes the Apriori algorithm to mine the multidimensional correlations between measurement errors of various characteristic gases. Finally, based on process similarity, it searches the historical sample set for the characteristic gas content of similar faults to correct similarity errors in the substation's characteristic gas data for each time period, improving the accuracy of fault identification.

[0065] The embodiments of this invention can achieve at least the following technical effects: 1) The characteristic gases of the online oil chromatography detection device play a crucial role in transformer fault identification. These gases exhibit no obvious regularity, are difficult to distinguish intuitively, and severely impact subsequent fault identification results. Therefore, it is essential to identify, eliminate, and reconstruct abnormal data caused by equipment malfunctions. 2) Based on historical characteristic gas data from the online oil chromatography detection device, the Apriori algorithm is used to classify and mine not only the characteristic gas sequence but also the relationships between the relative errors of each characteristic gas. 3) Data mining methods are used not only to correct historical characteristic gas data from the online oil chromatography detection device but also to identify faults in subsequent online oil chromatography detection device characteristic gases.

[0066] This embodiment also provides a transformer fault identification device, which is used to implement the embodiments and preferred embodiments; details already described will not be repeated. As used below, the terms "module" and "device" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0067] According to embodiments of the present invention, an embodiment of an apparatus for implementing a transformer fault identification method is also provided. Figure 3 This is a schematic diagram of the structure of a transformer fault identification device according to an embodiment of the present invention, as shown below. Figure 3 As shown, the transformer fault identification device includes: an acquisition module 300, a determination module 302, a correction module 304, and a detection module 306, wherein:

[0068] The acquisition module 300 is used to acquire the current online characteristic gas data and the current offline characteristic gas data of the transformer during the current fault period. The current online characteristic gas data includes the current online characteristic data corresponding to various characteristic gases generated during the current fault period. The current online characteristic gas data is collected based on the online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on the detection of the offline device.

[0069] The determination module 302, connected to the acquisition module 300, is used to determine the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the current online characteristic gas data and the current offline characteristic gas data.

[0070] The correction module 304, connected to the determination module 302, is used to correct the current online characteristic gas data based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, using a pre-determined strong correlation rule, to obtain the corrected online characteristic gas data. The strong correlation rule is used to indicate the correlation relationship between the various characteristic gases.

[0071] The detection module 306, connected to the correction module 304, is used to obtain the fault identification result of the transformer during the current fault period based on the corrected online characteristic gas data and a pre-trained fault identification model. The fault identification model is trained based on the historical online characteristic gas data corresponding to the transformer during multiple historical fault periods and the fault types corresponding to the multiple historical fault periods.

[0072] In this embodiment of the invention, an acquisition module 300 is configured to acquire current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period. The current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period. The current online characteristic gas data is collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on an offline device. A determination module 302, connected to the acquisition module 300, is configured to determine the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the current online and offline characteristic gas data. A correction module 304, connected to the determination module 302, is configured to, based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, use a pre-determined strong correlation rule to correct the current online characteristic gas data. The characteristic gas data is corrected to obtain corrected online characteristic gas data. Strong correlation rules are used to indicate the correlation between various characteristic gases. The detection module 306, connected to the correction module 304, is used to obtain the fault identification result of the transformer in the current fault period based on the corrected online characteristic gas data and a pre-trained fault identification model. The fault identification model is trained based on the historical online characteristic gas data corresponding to the transformer in multiple historical fault periods and the fault types corresponding to the multiple historical fault periods. This achieves the goal of comprehensively identifying transformer faults based on strong correlation rules between various characteristic gases and the fault identification model, thereby improving the technical effect of transformer fault identification accuracy and solving the technical problem of low fault identification accuracy when directly using an online oil chromatography detection device for transformer fault identification.

[0073] It should be noted that the modules can be implemented by software or hardware. For example, the latter can be implemented in the following ways: the modules can be located in the same processor; or the modules can be located in different processors in any combination.

[0074] It should be noted that the acquisition module 300, determination module 302, correction module 304, and detection module 306 correspond to steps S102 to S108 in the embodiments. The modules and corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the embodiments. It should also be noted that the modules, as part of the device, can run on a computer terminal.

[0075] It should be noted that the optional or preferred implementation methods of this embodiment can be found in the relevant descriptions in the embodiments, and will not be repeated here.

[0076] The transformer fault identification device may also include a processor and a memory. The acquisition module 300, determination module 302, correction module 304, detection module 306, etc. are all stored in the memory as program modules. The processor executes the program modules stored in the memory to realize the corresponding functions.

[0077] The processor contains a core that retrieves the corresponding program modules from memory. One or more cores can be configured. Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory includes at least one memory chip.

[0078] According to an embodiment of this application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein, during program execution, the device where the non-volatile storage medium is located executes any transformer fault identification method.

[0079] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals, and the non-volatile storage medium includes stored programs.

[0080] Optionally, during program execution, the device containing the non-volatile storage medium performs the following functions: acquiring current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on detection by an offline device; determining the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the current online and offline characteristic gas data; correcting the current online characteristic gas data using pre-determined strong association rules based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, obtaining corrected online characteristic gas data, wherein the strong association rules are used to indicate the correlation between the various characteristic gases; and obtaining the fault identification result of the transformer during the current fault period based on the corrected online characteristic gas data, using a pre-trained fault identification model, wherein the fault identification model is trained based on historical online characteristic gas data corresponding to multiple historical fault periods of the transformer, and fault types corresponding to multiple historical fault periods.

[0081] According to an embodiment of this application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein the program executes any transformer fault identification method during runtime.

[0082] According to an embodiment of this application, an embodiment of a computer program product is also provided, which, when executed on a data processing device, is adapted to execute a program that initializes any one of the steps of a transformer fault identification method.

[0083] Optionally, when executed on a data processing device, the computer program product is suitable for executing an initialization program with the following method steps: acquiring current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, the current online characteristic gas data is collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data is obtained based on detection by an offline device; based on the current online and offline characteristic gas data, determining the current index values ​​of the characteristic indicators corresponding to the various characteristic gases generated during the current fault period; according to the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, correcting the current online characteristic gas data using a pre-determined strong association rule to obtain corrected online characteristic gas data, wherein the strong association rule is used to indicate the association relationship between the various characteristic gases; based on the corrected online characteristic gas data, using a pre-trained fault identification model to obtain the fault identification result of the transformer during the current fault period, wherein the fault identification model is trained based on historical online characteristic gas data corresponding to multiple historical fault periods of the transformer and fault types corresponding to multiple historical fault periods.

[0084] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps: acquiring current online characteristic gas data and current offline characteristic gas data of the transformer during the current fault period, wherein the current online characteristic gas data includes current online characteristic data corresponding to various characteristic gases generated during the current fault period, the current online characteristic gas data being collected based on an online oil chromatography monitoring device, and the current offline characteristic gas data being obtained based on detection by an offline device; determining the current index values ​​of characteristic indicators corresponding to the various characteristic gases generated during the current fault period based on the current online and offline characteristic gas data; correcting the current online characteristic gas data using a pre-determined strong association rule based on the current index values ​​of the characteristic indicators corresponding to the various characteristic gases, to obtain corrected online characteristic gas data, wherein the strong association rule is used to indicate the correlation between the various characteristic gases; and obtaining the fault identification result of the transformer during the current fault period based on the corrected online characteristic gas data, using a pre-trained fault identification model, wherein the fault identification model is trained based on historical online characteristic gas data corresponding to multiple historical fault periods of the transformer, and fault types corresponding to multiple historical fault periods.

[0085] The order of the embodiments in this invention is merely for illustrative purposes and does not represent the superiority or inferiority of the embodiments.

[0086] In the embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0087] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of modules can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interface; the indirect coupling or communication connection between modules may be electrical or other forms.

[0088] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0089] Furthermore, the functional modules in the various embodiments of this invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or as a software functional module.

[0090] If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned non-volatile storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0091] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A transformer fault identification method, characterized by, The method comprises: obtaining current online characteristic gas data of the transformer in the current fault period, and current offline characteristic gas data, wherein the current online characteristic gas data comprises current online characteristic data corresponding to a plurality of characteristic gases generated in the current fault period, the current online characteristic gas data is collected based on an oil chromatographic online monitoring device, and the current offline characteristic gas data is detected based on an offline device; based on the current online characteristic gas data and the current offline characteristic gas data, determining current index values of characteristic indexes corresponding to the plurality of characteristic gases generated in the current fault period; According to the current index value of the feature index corresponding to each of the plurality of characteristic gases, a predetermined strong association rule is used to correct the current online characteristic gas data to obtain corrected online characteristic gas data, including: according to the current index value of the feature index corresponding to each of the plurality of characteristic gases, using the strong association rule to determine the similar characteristic gas data corresponding to the current online characteristic gas data from the historical online characteristic gas data of a plurality of historical failure periods, wherein the strong association rule is used to indicate the association relationship between the plurality of characteristic gases; the distance index value between the similar characteristic gas data and the current online characteristic gas data is determined by the following way: , wherein, is the current online characteristic gas data of any one sampling time t in the current failure period; is the similar characteristic gas data corresponding to the historical sampling time in the historical failure period corresponding to the current failure period, is the distance index value between the similar characteristic gas data and the current online characteristic gas data, is the number of physical quantities that affect the distance index value, is the corresponding weight of any one of the physical quantities that affect the distance index value, is half of the length of the failure period; is the standard deviation between the sampling values of the arbitrary physical quantity in the current failure period; is the index value of the i-th characteristic gas in the plurality of characteristic gases corresponding to the current online characteristic gas data at time t+j, is is the index value of the i-th characteristic gas in the plurality of characteristic gases corresponding to the similar characteristic gas data at time t-j; based on the distance index value, the online characteristic gas data is corrected to obtain the corrected online characteristic gas data; based on the corrected online characteristic gas data, using a pre-trained fault recognition model to obtain a fault recognition result of the transformer in the current fault period, wherein the fault recognition model is trained based on historical online characteristic gas data corresponding to the transformer in a plurality of historical fault periods, and fault types corresponding to the plurality of historical fault periods.

2. The method of claim 1, wherein, The historical online characteristic gas data comprises historical online characteristic data corresponding to the plurality of characteristic gases generated in the corresponding fault period, and before the step of determining similar characteristic gas data corresponding to the current online characteristic gas data from the plurality of historical online characteristic gas data corresponding to the plurality of historical fault periods based on the current index values of the characteristic indexes corresponding to the plurality of characteristic gases, the method further comprises: determining historical index values of the characteristic indexes corresponding to the plurality of characteristic gases generated in the plurality of historical fault periods based on the historical online characteristic gas data corresponding to the plurality of historical fault periods and historical offline characteristic gas data corresponding to the plurality of historical fault periods; based on the historical index values of the characteristic indexes corresponding to the plurality of characteristic gases generated in the plurality of historical fault periods, obtaining a frequent item set with a Boolean association rule corresponding to the characteristic indexes; determining the strong association rule based on the frequent item set.

3. The method of claim 2, wherein, In the case that the characteristic indexes are a plurality of indexes, and the plurality of indexes at least comprise a relative length, a relative height, and a relative error, the step of determining the historical index values of the characteristic indexes corresponding to the plurality of characteristic gases generated in the plurality of historical fault periods based on the historical online characteristic gas data corresponding to the plurality of historical fault periods and the historical offline characteristic gas data corresponding to the plurality of historical fault periods comprises: determining a historical relative length of any one of the plurality of characteristic gases generated in any one of the historical fault periods by the following method: determining a duration of the any one of the characteristic gases in the any one of the historical fault periods, and a total duration of the any one of the historical fault periods; and taking a proportion of the duration in the total duration as the historical relative length of the any one of the characteristic gases generated in the any one of the historical fault periods. The historical relative height corresponding to the arbitrary characteristic gas generated in the arbitrary historical failure period is determined in the following manner: A first difference between a maximum value and a minimum value in the historical online characteristic data corresponding to the arbitrary characteristic gas in the arbitrary historical failure period and a second difference between a maximum value and a minimum value in the historical online characteristic data corresponding to the multiple characteristic gases are determined, and a proportion of the first difference to the second difference is taken as the historical relative height corresponding to the arbitrary characteristic gas generated in the arbitrary historical failure period; The historical relative error corresponding to the arbitrary characteristic gas generated in the arbitrary historical failure period is determined in the following manner: A third difference between the historical online characteristic data corresponding to the arbitrary characteristic gas and the historical offline characteristic data corresponding to the arbitrary characteristic gas in the arbitrary historical failure period is determined, and the third difference is taken as the historical relative error corresponding to the arbitrary characteristic gas generated in the arbitrary historical failure period.

4. The method of claim 1, wherein, In the case that the similar characteristic gas data is multiple groups, the correcting the current online characteristic gas data according to the similar characteristic gas data to obtain the corrected online characteristic gas data comprises: Determining weight values corresponding to the multiple groups of similar characteristic gas data respectively; Obtaining the corrected online characteristic gas data based on the multiple groups of similar characteristic gas data and the weight values corresponding to the multiple groups of similar characteristic gas data respectively.

5. The method according to any one of claims 1 to 4, characterized in that, The determining the current index value of the characteristic index corresponding to the multiple characteristic gases generated in the current failure period based on the current online characteristic gas data and the current offline characteristic gas data comprises: Performing denoising processing on the current online characteristic gas data and the current offline characteristic gas data to obtain denoised current online characteristic gas data and denoised current offline characteristic gas data; Determining the current index value of the characteristic index corresponding to the multiple characteristic gases generated in the current failure period based on the denoised current online characteristic gas data and the denoised current offline characteristic gas data.

6. A transformer fault recognition apparatus characterized by comprising: The device is used to implement the transformer failure identification method in any one of claims 1 to 5, and the device comprises: An acquisition module is configured to acquire current online characteristic gas data and current offline characteristic gas data of a transformer in a current failure period, wherein the current online characteristic gas data comprises current online characteristic data corresponding to multiple characteristic gases generated in the current failure period, the current online characteristic gas data is acquired based on an oil chromatographic online monitoring device, and the current offline characteristic gas data is acquired based on an offline device; A determination module is configured to determine current index values of characteristic indexes corresponding to the multiple characteristic gases generated in the current failure period based on the current online characteristic gas data and the current offline characteristic gas data. A correction module is configured to correct the current online characteristic gas data according to the current index values of the characteristic indexes corresponding to the multiple characteristic gases respectively, and obtain corrected online characteristic gas data, wherein the strong correlation rule is used to indicate the correlation between the multiple characteristic gases. A detection module is configured to obtain a fault identification result of the transformer at the current fault period based on the corrected online characteristic gas data and a pre-trained fault identification model, wherein the fault identification model is trained based on historical online characteristic gas data corresponding to multiple historical fault periods of the transformer respectively and fault types corresponding to the multiple historical fault periods respectively.

7. An electronic device, comprising: One or more processors and a memory are included, and the memory is configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the transformer fault identification method in any one of claims 1 to 5.

Images