A transformer fault alarm method, a terminal device and a computer readable storage medium

Abstract

The application is suitable for the technical field of transformer fault monitoring, and discloses a transformer fault alarm method, a terminal device and a computer readable storage medium. The transformer fault alarm method comprises the following steps: obtaining online chromatographic data uploaded by a transformer oil dissolved gas online monitoring device; judging whether the online chromatographic data is valid chromatographic data; if yes, inputting the online chromatographic data into a trained transformer fault alarm model, wherein the trained transformer fault alarm model comprises an alarm module and a fault identification module; judging whether the input valid chromatographic data is fault chromatographic data through the alarm module; if yes, inputting the online chromatographic data into the fault identification module, identifying the fault type of the transformer through the fault identification module, and outputting the fault type of the transformer and alarming. The efficiency and accuracy of transformer fault alarm are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of transformer fault monitoring, and particularly relates to a transformer fault alarm method, a terminal device and a computer readable storage medium. BACKGROUND

[0002] As the core hinge equipment of the power system, the running state of the transformer is directly related to the safety and stability level of the entire power system. Oil dissolved gas composition content analysis is widely used as one of the most effective methods for sensing transformer latent faults. With the development of sensor technology, online monitoring technology of transformer oil dissolved gas has become one of the main ways to monitor the running state of the transformer in real time.

[0003] However, due to the influence of the surrounding environment and the limitation of the prior art, the accuracy and stability of the online chromatographic data are still some distance from the laboratory offline chromatographic data, and it is difficult to determine a strong universal alarm value. In addition, some practical applications find that the threshold margin of the dissolved gas composition content of the transformer in the national standard is high, the composition is slightly single, the classification is not fine, and there are problems such as not flexible enough and occasional missed reports when applied to online monitoring. Furthermore, when the chromatographic data exceeds the alarm value, there are problems such as incomplete coding when analyzing the fault by using the traditional three-ratio method. Therefore, the prior art has problems such as single threshold setting, high margin and incomplete three-ratio coding when judging the fault of the dissolved gas composition content in the transformer oil. SUMMARY

[0004] In order to overcome the problems in the related art, the embodiments of the present application provide a transformer fault alarm method, which realizes rapid and accurate identification of the fault type of the transformer.

[0005] The present application is realized by the following technical solutions:

[0006] In a first aspect, the embodiments of the present application provide a transformer fault alarm method, which comprises: obtaining online chromatographic data uploaded by an online monitoring device of transformer oil dissolved gas; judging whether the online chromatographic data is valid chromatographic data, if yes, inputting the online chromatographic data into a trained transformer fault alarm model, wherein the trained transformer fault alarm model comprises an alarm module and a fault identification module; judging whether the input valid chromatographic data is fault chromatographic data by the alarm module, if yes, inputting the online chromatographic data into the fault identification module, identifying the fault type of the transformer by the fault identification module, outputting the fault type of the transformer and alarming.

[0007] In some embodiments based on the first aspect, determining whether the online chromatographic data is valid chromatographic data comprises: determining whether the gas component value and the total hydrocarbon value in the online chromatographic data satisfy an abnormal data determination condition, and if the abnormal data determination condition is not satisfied, determining that the online chromatographic data is valid chromatographic data.

[0008] In some embodiments based on the first aspect, if the online chromatographic data is determined to be valid chromatographic data, the online chromatographic data is input into a trained transformer fault alarm model, wherein the trained transformer fault alarm model comprises an alarm module and a fault identification module; whether the input valid chromatographic data is fault chromatographic data is determined by the alarm module, if the input valid chromatographic data is fault chromatographic data, the online chromatographic data is input into the fault identification module, the fault type of the transformer is identified by the fault identification module, and before the fault type of the transformer is output and an alarm is given, the transformer fault alarm method further comprises: obtaining historical online chromatographic data and offline chromatographic data, establishing an alarm threshold database; training the alarm module based on the alarm threshold database to obtain a trained alarm module; the trained alarm module comprises the fault alarm threshold; obtaining fault chromatographic data of which the fault type of the transformer has been determined, and establishing a fault case library, wherein the fault chromatographic data comprises the dissolved gas component content in transformer oil, voltage grade and fault type characteristic parameters; determining a plurality of target algorithm models based on the fault case library, and constructing the fault identification module, wherein the final identification result of the fault identification module is determined by the preliminary identification results of the plurality of target algorithm models.

[0009] In the embodiments of the present application, the online chromatographic data obtained by the dissolved gas online monitoring device in transformer oil is utilized by setting the alarm threshold database, the chromatographic data fault alarm threshold with stronger practicability and applicability is obtained, and the accuracy of transformer fault alarm determination is improved.

[0010] In some embodiments based on the first aspect, the fault case library comprises a training set and a test set, and determining a plurality of target algorithm models based on the fault case library comprises: training a plurality of candidate algorithm models based on the training set, verifying the fault classification error rate of each candidate algorithm model by the test set, and determining a preset number of target algorithm models and the priority of each target algorithm model in the order from low to high of the fault classification error rate.

[0011] In some embodiments based on the first aspect, the fault identification module comprises a plurality of target algorithm models; the final identification result of the fault identification module is the identification result with the highest proportion among the preliminary identification results of all target algorithm models; when there are a plurality of preliminary identification results with the same proportion, the final identification result is determined according to the priority of the target algorithm model from which the plurality of preliminary identification results with the same proportion are obtained.

[0012] Based on the first aspect, in some embodiments, the alarm module comprises a quartile algorithm model and a quantile point algorithm model; based on the alarm threshold database, the alarm module is trained to obtain a trained alarm module, comprising: inputting the data in the alarm threshold database into the quartile algorithm model to obtain a first fault alarm threshold; inputting the data in the alarm threshold database into the quantile point algorithm model to obtain a second fault alarm threshold; taking the average of the first fault alarm threshold and the second fault alarm threshold as the fault alarm threshold to obtain the trained alarm module.

[0013] Based on the first aspect, in some embodiments, the transformer fault alarm method further comprises: periodically increasing the historical online chromatographic data and the offline chromatographic data, updating the alarm threshold database; based on the updated alarm threshold database, the alarm module of the transformer fault alarm model is retrained.

[0014] Based on the first aspect, in some embodiments, the transformer fault alarm method further comprises: after outputting the fault type of the transformer and alarming, the fault transformer corresponding to the fault chromatographic data is repaired to obtain the actual fault type of the fault transformer, and the actual fault type and the fault chromatographic data of the transformer are associated and added to the fault case library; based on the updated fault case library, the transformer fault identification module is retrained.

[0015] In the embodiment of the present application, the transformer fault alarm model comprising the alarm module and the fault identification module is trained by using the alarm threshold database and the fault case library, and more accurate alarm threshold and fault identification result are determined through the screening and combination of algorithms. Only the online chromatographic data uploaded by the transformer oil dissolved gas online monitoring device needs to be obtained, and the transformer fault rapid alarm can be realized, thereby effectively improving the efficiency and accuracy of the transformer fault alarm.

[0016] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present specification. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 is a flowchart of the transformer fault alarm method provided by an embodiment of the present application;

[0019] Figure 2is a whole process diagram of a transformer fault alarm method provided by an embodiment of the present application;

[0020] Figure 3 is a training flow diagram of a transformer fault alarm model provided by an embodiment of the present application;

[0021] Figure 4 is a quartile algorithm flow diagram provided by an embodiment of the present application;

[0022] Figure 5 is a quantile point algorithm flow diagram provided by an embodiment of the present application;

[0023] Figure 6 is an alarm module calculating an alarm threshold flow diagram provided by an embodiment of the present application;

[0024] Figure 7 is a fault recognition module outputting a fault recognition result flow diagram provided by an embodiment of the present application;

[0025] Figure 8 is a terminal device schematic diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0026] In the following description, for the purposes of explanation and not limitation, specific details are set forth, such as particular sequences of acts, techniques, etc. in order to provide a thorough understanding of the embodiments of the application. However, it will be apparent to those skilled in the art that the application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the application with unnecessary detail.

[0027] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or "comprising", when used in this specification and in the following claims, indicates the presence of the stated features, integers, steps, operations, elements, and / or components but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0028] It is also to be understood that the terminology "and / or" when used in this specification and in the following claims, refers to at least one of the items, or any combination of the items, listed after the term in the various aspects.

[0029] As used in the description of the application and the appended claims, the term “if’ can be interpreted to mean “when” or “upon” or “in response to determining” or “in response to ascertaining,” depending on the context. Similarly, the phrase “if it is determined” or “if [a described condition or event] is detected” can be interpreted to mean “upon determining” or “in response to determining” or “upon [the described condition or event] being detected” or “in response to [the described condition or event] being detected,” depending on the context.

[0030] In addition, the terms “first,” “second,” “third,” etc. as used in the description of the application and the appended claims are merely to distinguish descriptions and are not to be construed to imply or suggest relative importance.

[0031] Reference in the specification to “one embodiment” or “some embodiments” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrase “in one embodiment” or “in some embodiments” in various places in the specification are not necessarily all referring to the same embodiment, although it can. The terms “including,” “containing,” “having,” and variations thereof are meant to encompass the terms “including but not limited to.”

[0032] To solve the above problems, embodiments of the application provide a transformer fault alarm method, as shown in Figure 1 The above method includes steps 101 to 103.

[0033] Step 101: Obtain online chromatogram data uploaded by a transformer oil dissolved gas online monitoring device.

[0034] A transformer is an important hub device of a power system, and its operation reliability directly affects the safe and stable operation level of the entire power system. Once a fault occurs, the loss is huge. A transformer oil dissolved gas online monitoring device is widely used because it can sense the running state of a transformer in real time and discover and monitor latent faults. Online chromatogram data uploaded by a transformer online chromatogram device includes gas component content and total hydrocarbon values.

[0035] Step 102: Determine whether the online chromatogram data is valid chromatogram data.

[0036] Determine whether the gas component value and the total hydrocarbon value in the online chromatogram data satisfy an abnormal data determination condition. If the abnormal data determination condition is not satisfied, the online chromatogram data is determined to be valid chromatogram data.

[0037] In some embodiments, the abnormal data determination condition includes all gas component data values being 0, total hydrocarbon value data being 0, all component data and total hydrocarbon value data being 99999 or -99999, or being completely repeated with the previous data, and other obviously abnormal data, which is usually caused by the failure of the oil dissolved gas online monitoring device.

[0038] Step 103: If the online chromatographic data is valid chromatographic data, the online chromatographic data is input into the trained transformer fault alarm model to determine whether the transformer has failed. If the transformer has failed, the type of failure is determined, and the type of failure of the transformer is output and an alarm is given.

[0039] The online chromatographic data is input into the trained transformer fault alarm model to determine whether the online chromatographic data is fault chromatographic data. If the online chromatographic data is fault chromatographic data, it indicates that the transformer has failed. The type of failure of the transformer corresponding to the fault chromatographic data is determined based on the fault chromatographic data. The alarm module is used to determine whether the valid chromatographic data is fault chromatographic data based on the trained fault alarm threshold. Fault chromatographic data is data generated when the transformer fails. If the online chromatographic data is fault chromatographic data, the online chromatographic data is input into the fault identification module. The fault identification module is used to identify the type of failure of the transformer based on the online chromatographic data and give an alarm.

[0040] Specifically, the valid chromatographic data is classified according to voltage levels, and each voltage level corresponds to a set of fault alarm thresholds. If a gas component data value is higher than the fault alarm threshold corresponding to the gas under the voltage level, the valid chromatographic data is determined to be fault chromatographic data. The fault chromatographic data is input into the fault identification module, and the type of failure of the transformer is identified and an alarm is given through the transformer fault alarm model. Maintenance personnel receive the alarm information and perform transformer maintenance according to the type of failure.

[0041] As shown in Figure 2 After obtaining the online chromatographic data, it is determined whether the online chromatographic data is valid chromatographic data. If the online chromatographic data is valid chromatographic data, the valid chromatographic data is input into the transformer fault alarm model. If the online chromatographic data is not valid chromatographic data, it indicates that the transformer oil dissolved gas online monitoring device has failed, and the data is not accepted. The online chromatographic data input into the transformer fault alarm model is determined by the alarm module to be fault chromatographic data. If the online chromatographic data is determined to be fault chromatographic data, the fault chromatographic data is input into the fault identification module for fault category determination. If the online chromatographic data is not fault chromatographic data, it indicates that the transformer is operating normally.

[0042] Before the online chromatographic data is input into the trained transformer fault alarm model, the type of failure of the transformer is output and an alarm is given, the above transformer fault alarm method further includes Figure 3The steps 201 to 203 shown.

[0043] Step 201: Obtain historical online chromatographic data and offline chromatographic data, and establish an alarm threshold database.

[0044] The historical online chromatographic data is historical data uploaded by an online monitoring device for dissolved gas in transformer oil, and the offline chromatographic data is chromatographic data tested in a laboratory after on-site sampling by an experimental personnel. The historical online chromatographic data is analyzed, and abnormal data caused by device failure and the like is removed, and only valid chromatographic data is retained. The historical online chromatographic data after removal of abnormal data is combined with the offline chromatographic data, classified according to transformer voltage grades, and an alarm threshold database is established. A set of chromatographic data in the alarm threshold database includes gas component data, total hydrocarbon value, and voltage grade of the transformer corresponding to the set of data, wherein the gas components include H2, CO, CO2, CH4, C2H4, C2H6, C2H2, and the like.

[0045] Step 202: Based on the alarm threshold database, an alarm module is trained to obtain a trained alarm module; the trained alarm module contains a fault alarm threshold.

[0046] When a transformer fails, the insulating oil of the transformer will change in composition, for example, when a thermal fault occurs, the insulating oil decomposes, and different temperatures will also cause different amounts of H2, CO, CO2, CH4, C2H4, C2H6, C2H2, and the like to dissolve in the oil. The transformer fault can be alarmed by analyzing the chromatographic data at this time.

[0047] However, due to the influence of the surrounding environment and the limitations of the prior art, the accuracy and stability of the online chromatographic data are still some distance from the laboratory offline chromatographic data, and it is difficult to determine a strong universal alarm value for the same. In addition, some practical applications find that the threshold margin of the content of some transformer oil dissolved gas components in the national standard is too high, the components are slightly single, the classification is not fine, and there are problems such as not flexible enough, occasional false negatives, and the like when applied to online monitoring.

[0048] To this end, the present application establishes an alarm threshold database based on online chromatographic data and offline chromatographic data to determine a suitable alarm threshold, such as Figure 4As shown, in some embodiments, alarm threshold database data is input into the quartile algorithm model to calculate the first fault alarm threshold. Specifically, the data in the alarm threshold database is classified by voltage level. For data at the same voltage level, it is further divided into eight parameter data tables based on gas component categories: H2, CO, CO2, CH4, C2H4, C2H6, C2H2, and total hydrocarbon value. Z-score transformation and standardization are performed. Data greater than 3 and less than -3 after standardization are deleted as invalid data. The dataset after deleting invalid data is restored to obtain the data before Z-score standardization, which is then used as the data to be analyzed. The first fault alarm threshold for each gas component is calculated using the quartile method. Arrange the data in each parameter data table in ascending order. Let the total number of data in a parameter data table be n, let i represent the integer part of the first quantile 0.25(n+1), let j represent its fractional part, and let x[i] be an array, i = 1, 2, 3, ..., n. Then the first quartile Q1 = (1-j)x[i] + jx[i+1]. Find the third quartile Q3 corresponding to the third quantile 0.75(n+1), and calculate the interquartile range I. QR =Q3-Q1, calculate the first fault alarm threshold a = Q3+mI using the interquartile range and the third interquartile range. QR 0.3 < m < 0.5, where the value of m is determined based on the specific database to ensure that the fault alarm threshold does not exceed the highest value in the data table.

[0049] In some embodiments, alarm threshold database data is input into the quantile algorithm model to calculate a second fault alarm threshold. For example... Figure 5 As shown, the data in the alarm threshold database is also classified by voltage level. For data at the same voltage level, it is further divided into eight parameter data tables based on gas component categories: H2, CO, CO2, CH4, C2H4, C2H6, C2H2, and total hydrocarbon value. Each data point in the parameter data table is checked for a value of 0. If it is 0, it is stored in the database, and the number of data points with a value of 0 is recorded. If it is not 0, the Z-score method is used for transformation and standardization. Data points greater than 3 and less than -3 after standardization are deleted as invalid data. The dataset after deleting invalid data is restored to its original state, and the data before Z-score standardization is used as the data to be analyzed. Non-parametric tests are used to fit the distribution of each gas parameter to obtain quantiles. Based on the proportion of 0-value data, the quantiles of all data, including 0 and non-zero values, are obtained. The value of this quantile is the second alarm threshold b.

[0050] like Figure 6As shown, after the first fault alarm threshold and the second fault alarm threshold are calculated by the quartile algorithm and the quantile algorithm, the average of the first fault alarm threshold and the second fault alarm threshold is taken as the fault alarm threshold, and the fault alarm threshold

[0051] Step 203: Determine a plurality of target algorithm models based on the fault case library, and construct a fault identification module, wherein the final identification result of the fault identification module is determined by the preliminary identification results of the plurality of target algorithm models.

[0052] In some embodiments, fault chromatogram data of which the transformer fault type has been determined is acquired, and a fault case library is established, wherein the fault chromatogram data includes transformer oil dissolved gas component content, voltage grade and fault type characteristic parameters. The fault chromatogram data in the fault case library can be acquired by literature research or field operation data accumulation, etc. The fault case library is divided into a training set and a test set, and a plurality of candidate algorithm models are trained based on the training set. As shown, Figure 7 As shown, the candidate algorithm models include Fisher discriminant method, K-nearest neighbor algorithm, logistic regression algorithm, decision tree, support vector machine, naive Bayes classification, random forest, three-ratio method, etc. The trained candidate algorithm models are verified by the test set for fault classification error rate, and a plurality of target algorithm models of a preset number are determined in order from low to high fault classification error rate. The preset number can be any integer less than or equal to the total number of the above algorithms, such as 1, 2, 3, etc. At the same time, the priority of each target algorithm model is determined according to the fault classification error rate, and the target algorithm model with the lower fault classification error rate has the higher priority. For example, the first 5 candidate algorithm models with the lowest fault classification error rate are taken as the target algorithm models, then the target algorithm model with the lowest fault classification error rate has the priority 1, and the remaining target algorithm models are ranked from high to low fault classification error rate, and the priorities are 2, 3, 4 and 5 in turn.

[0053] The preliminary identification results of the plurality of target algorithm models are calculated, and the identification result with the highest proportion in all preliminary identification results is determined as the final identification result of the transformer fault alarm model training method. When there are multiple preliminary identification results with the same proportion, the final identification result is determined according to the priority of the target algorithm model from which the multiple preliminary identification results with the same proportion are obtained.

[0054] Specifically, the sum of the priority values of the target recognition algorithms corresponding to the recognition results with the same proportion is calculated, and the recognition result with the lowest sum of the priority values is the final recognition result. For example, there are 5 target algorithm models, 2 target algorithm models have a fault recognition result m1, 2 target algorithm models have a fault recognition result m2, and 1 target algorithm model has a fault recognition result m3. The fault recognition result m1 and the fault recognition result m2 have the highest proportion in all preliminary recognition results. At this time, the sum of the priority values of the target algorithm model with the fault recognition result m1 is calculated, one target algorithm priority value is 1, and the other target algorithm priority value is 2. Therefore, the sum of the priority values of the target algorithm model with the fault recognition result m1 is 3. At the same time, the sum of the priority values of the target algorithm model with the fault recognition result m2 is calculated, one target algorithm priority value is 3, and the other target algorithm priority value is 4. Therefore, the sum of the priority values of the target algorithm model with the fault recognition result m2 is 7. It is determined that the recognition result m1 with the lowest sum of the priority values is the final recognition result.

[0055] The transformer fault alarm method further includes periodically increasing historical online chromatographic data and offline chromatographic data, and updating the alarm threshold database. Based on the updated alarm threshold database, the alarm module of the transformer fault alarm model is updated.

[0056] In some embodiments, the updated fault alarm threshold of the alarm module is calculated based on the updated alarm threshold database. The online chromatographic data collected by the chromatographic online monitoring device of the dissolved gas in the field transformer oil or the offline chromatographic data tested in the laboratory after the field sampling by the experimental personnel is stored in a specific location for preservation, in order to be extracted by the transformer fault alarm model for data periodically. Among them, the offline chromatographic data is directly stored in the alarm threshold database, and the online chromatographic data is recorded in the alarm threshold database after completing the abnormal value cleaning. The periodic update time can be real-time, one month, one year, etc. Self-set time.

[0057] The transformer fault alarm method further includes outputting the fault type of the transformer and alarming, and then repairing the fault transformer corresponding to the fault chromatographic data, obtaining the actual fault type of the fault transformer and the fault chromatographic data, and associating the actual fault type with the fault chromatographic data of the transformer and adding it to the fault case library.

[0058] It should be understood that the size of the serial number of each step in the above-mentioned embodiments does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0059] The embodiment of the present application also provides a terminal device, which is used for implementing the above-mentioned method and comprises a processor and a memory. Figure 8The terminal device 800 can include at least one processor 810, a memory 820, and a computer program stored in the memory 820 and executable on the at least one processor 810, wherein the processor 810 implements the steps in any of the above method embodiments when executing the computer program, for example Figure 1 steps 101 to 103 in the illustrated embodiment, or Figure 3 steps 201 to 203 in the illustrated embodiment.

[0060] By way of example, the computer program can be segmented into one or more modules / units, which are stored in the memory 820 and executed by the processor 810 to complete the present application. The one or more modules / units can be a series of computer program segments capable of completing a specific function, which are used to describe the execution process of the computer program in the terminal device 800.

[0061] Those skilled in the art can understand that Figure 8 The terminal device is merely an example and does not constitute a limitation on the terminal device, and can include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0062] The processor 810 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0063] The memory 820 can be an internal storage unit of the terminal device, or an external storage device of the terminal device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. The memory 820 is used to store the computer program and other programs and data required by the terminal device. The memory 820 can also be used to temporarily store data that has been output or will be output.

[0064] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, the bus in the drawings of the present application does not limit to only one bus or one type of bus.

[0065] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps in each embodiment of the transformer fault alarm method.

[0066] The embodiment of the present application provides a computer program product, when the computer program product is run on a mobile terminal, so that the mobile terminal executes to realize the steps in each embodiment of the transformer fault alarm method.

[0067] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the present application realizes all or part of the processes in the above-mentioned embodiments, which can be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can realize the steps in each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer readable medium at least includes any entity or device capable of carrying the computer program code to a photographing device / terminal equipment, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium cannot be an electrical carrier signal and a telecommunication signal.

[0068] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0069] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0070] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / network device and method can be implemented in other manners. For example, the described apparatus / network device embodiments are merely schematic. For example, the division of the modules or units is merely logical function division. There can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between the units can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0071] The units described as separated components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e. can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments of the present application.

[0072] The above-described embodiments are merely used to illustrate the technical solutions of the present application, rather than limit them; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A transformer fault alerting method, characterized by, The method comprises the following steps: obtaining online chromatographic data uploaded by an online transformer oil dissolved gas monitoring device; determining whether the online chromatographic data is valid chromatographic data, and if so, inputting the online chromatographic data into a trained transformer fault alarm model, wherein the trained transformer fault alarm model comprises an alarm module and a fault identification module; determining whether the input valid chromatographic data is fault chromatographic data through the alarm module, and if so, inputting the online chromatographic data into the fault identification module, identifying the fault type of the transformer through the fault identification module, outputting the fault type of the transformer and alarming; wherein the fault identification module comprises a plurality of target algorithm models; the final identification result of the fault identification module is the identification result with the highest proportion among the preliminary identification results of all the target algorithm models; when there are multiple preliminary identification results with the same proportion, the final identification result is determined according to the priority of the target algorithm models that obtain the multiple preliminary identification results with the same proportion; the alarm module comprises a quartile algorithm model and a quantile point algorithm model; wherein, before determining whether the online chromatographic data is valid chromatographic data, the method further comprises: obtaining historical online chromatographic data and offline chromatographic data, and establishing an alarm threshold database; training the alarm module based on the alarm threshold database to obtain a trained alarm module; the trained alarm module contains the fault alarm threshold; obtaining fault chromatographic data with a determined transformer fault type, and establishing a fault case library, wherein the fault chromatographic data comprises transformer oil dissolved gas component content, voltage grade and fault type characteristic parameters; determining a plurality of target algorithm models based on the fault case library to construct the fault identification module, wherein the final identification result of the fault identification module is determined by the preliminary identification results of the plurality of target algorithm models; wherein the candidate algorithm models include Fisher discriminant method, K-nearest neighbor algorithm, logistic regression algorithm, decision tree, support vector machine, naive Bayes classification, random forest, and three-ratio method; the trained candidate algorithm models are verified by a test set to determine a plurality of target algorithm models with a preset number in the order of fault classification error rate from low to high; wherein, the training of the alarm module based on the alarm threshold database to obtain a trained alarm module comprises: inputting the data in the alarm threshold database into the quartile algorithm model to obtain a first fault alarm threshold; inputting the data in the alarm threshold database into the quantile point algorithm model to obtain a second fault alarm threshold; taking the average of the first fault alarm threshold and the second fault alarm threshold as the fault alarm threshold to obtain the trained alarm module; wherein, inputting the data in the alarm threshold database into the quartile algorithm model to obtain a first fault alarm threshold; inputting the data in the alarm threshold database into the quantile point algorithm model to obtain a second fault alarm threshold, comprises: The data in the alarm threshold database is classified according to voltage levels, and the data of the same voltage level is classified into eight parameter data tables of H2, CO, CO2, CH4, C2H4, C2H6, C2H2 and total hydrocarbon values according to gas groups; whether each data in the parameter data table is 0 is judged in turn, if it is 0, it is stored in the database, and the number of all data values of 0 is recorded; if it is not 0, the Z-score method is used for transformation and standardization processing, and the data greater than 3 and less than-3 after standardization is deleted as invalid data, and the data set after deleting invalid data is restored to obtain the data before Z-score standardization as the data to be analyzed; The first fault alarm threshold of each gas component is obtained by quartile method; the data in each parameter data table is arranged in ascending order, the total number of data in a parameter data table is n, i represents the integer part of the first quartile point 0.25(n+1), j represents the decimal part, x[i] is an array, i=1, 2, 3, …, n, then the first quartile Q1=(1-j)x[i]+jx[i+1], the third quartile Q3 corresponding to the third quartile point 0.75(n+1) is obtained, and the interquartile range I QR =Q3-Q1 is calculated, and the first fault alarm threshold a=Q3+mI QR is calculated by the interquartile range and the third quartile point, 0.3 The distribution form of each gas parameter is fitted by using non-parametric test to obtain quantile points, and the quantile points of all data including 0 value and non-0 value are obtained according to the proportion of 0 value data, and the value of the quantile point is the second alarm threshold b.

2. The transformer fault warning method of claim 1, wherein, The judgment of whether the online chromatographic data is valid chromatographic data includes: Judging whether the gas component value and the total hydrocarbon value in the online chromatographic data meet the abnormal data determination condition, if not, determining that the online chromatographic data is valid chromatographic data.

3. The transformer fault warning method of claim 1, wherein, The fault case library includes a training set and a test set, and the determination of a plurality of target algorithm models based on the fault case library includes: Training a plurality of candidate algorithm models based on the training set, verifying the fault classification error rate of each candidate algorithm model through the test set, and determining a preset number of target algorithm models and the priority of each target algorithm model in order from low to high according to the fault classification error rate.

4. The transformer fault warning method of claim 1, wherein, The transformer fault alarm method further includes: Periodically increasing the historical online chromatographic data and the offline chromatographic data, updating the alarm threshold database, and retraining the alarm module of the transformer fault alarm model based on the updated alarm threshold database. The transformer fault alarm method further includes:

5. The transformer fault warning method of claim 4, wherein, After outputting the fault type of the transformer and alarming, the fault transformer corresponding to the fault chromatographic data is repaired to obtain the actual fault type of the fault transformer, and the actual fault type and the fault chromatographic data of the transformer are associated and added to the fault case library; Based on the updated fault case library, the transformer fault identification module is retrained. The processor executes the computer program to realize the steps of the transformer fault alarm method according to any one of claims 1 to 5.

6. A terminal device comprising a memory and a processor and a computer program stored in the memory and executable on the processor, characterized in that, The computer program is executed by the processor to realize the steps of the transformer fault alarm method according to any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: receiving a request for a resource from a client; determining whether the client is authorized to access the resource; and if the client is authorized to access the resource, providing the resource to the client. ​

Images