Online anomaly detection method, system and equipment for analyzing dissolved gas in transformer oil and medium

Abstract

The invention provides an online anomaly detection method, system and equipment for analyzing dissolved gas in transformer oil and a medium. The online anomaly detection method comprises the following steps: step 1, acquiring characteristic data required for analyzing the dissolved gas in the transformer oil; step 2, processing the obtained feature data by using stream-oriented computation to obtain a data set containing a plurality of data samples; 3, calculating an abnormal score value corresponding to each data sample in the data set by using an RRCF algorithm; 4, judging abnormal data according to the obtained abnormal score value; according to the method, the accuracy and robustness of anomaly detection are remarkably improved, and the possibility of false alarm is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of transformer status monitoring, and in particular relates to an online abnormality detection method, system, equipment and medium for analyzing dissolved gas in transformer oil. Background Art

[0002] Power transformers are an important part of the production and operation of hydropower stations. The analysis of dissolved gases in transformer oil is an important means of transformer status monitoring. By analyzing the composition and content of dissolved gases in oil, the operating status and potential faults of the transformer can be determined. Traditional dissolved gas analysis methods are usually performed offline using laboratory equipment such as gas chromatographs. Although these methods are highly accurate, they cannot achieve real-time monitoring, and are complex and time-consuming to operate.

[0003] At present, most of the anomaly detection algorithms for dissolved gas data in transformer oil are trained based on offline data, usually relying on historical data sets, and using machine learning or deep learning methods to train models to achieve anomaly detection and fault diagnosis of dissolved gas in transformer oil. These algorithms generally have shortcomings such as reliance on historical data, lack of real-time performance, and sample imbalance. Summary of the invention

[0004] The purpose of the present invention is to provide an online anomaly detection method, system, equipment and medium for analysis of dissolved gas in transformer oil, which solves the defects of the existing anomaly detection method of dissolved gas data in transformer oil, such as dependence on historical data, lack of real-time performance and unbalanced samples.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: The present invention provides an online abnormality detection method for analysis of dissolved gas in transformer oil, comprising the following steps: Step 1, obtaining characteristic data required for analyzing dissolved gas in transformer oil; Step 2: Use streaming computing to process the obtained feature data to obtain a data set containing multiple data samples; Step 3, use the RRCF algorithm to calculate the anomaly score corresponding to each data sample in the data set; Step 4: Identify abnormal data based on the obtained abnormal score value.

[0006] Preferably, in step 1, the required characteristic data include a hydrogen sub-table, a carbon monoxide sub-table, a carbon dioxide sub-table, a methane sub-table, an acetylene sub-table, an ethylene sub-table, an ethane sub-table, a total hydrocarbon sub-table, a trace water sub-table and an oil temperature sub-table, and the table structure of each sub-table includes a timestamp and a numerical value.

[0007] Preferably, the required feature data is post-processed using the built-in streaming computing engine of the TDengine time series database to obtain a data set containing multiple data samples.

[0008] Preferably, in step 3, the RRCF algorithm is used to calculate the abnormal score value corresponding to each data sample in the data set. The specific method is: The obtained data set is combined with the RRCF algorithm to construct a set of random cut forests consisting of random cut trees, where an attribute is randomly selected from each data sample in the data set, a random cut tree is constructed for each attribute, and the random cut trees corresponding to all attributes constitute a random cut forest; The complexity of the random cut forest is defined by the depth of the random cut tree, and the calculation formula of the anomaly score value is defined by the complexity of the random cut forest; The abnormal score calculation formula is used to calculate the abnormal score value corresponding to each data sample.

[0009] Preferably, the abnormal score value calculation formula is as follows:

[0010] in, is the abnormal score value.

[0011] Preferably, in step 4, the abnormal data is identified according to the obtained abnormal score value, and the specific method is: Set up a sliding window mechanism; If the anomaly score value of a data sample is the highest anomaly score value in multiple consecutive windows, then the data sample is high-risk anomaly data.

[0012] An online abnormality detection system for analysis of dissolved gas in transformer oil, comprising: A characteristic data acquisition unit, used to acquire characteristic data required for analyzing dissolved gas in transformer oil; A data processing unit, used to process the obtained feature data by using streaming computing to obtain a data set containing multiple data samples; An abnormal score value calculation unit, used to calculate the abnormal score value corresponding to each data sample in the data set using the RRCF algorithm; The abnormal data identification unit is used to identify abnormal data according to the obtained abnormal score value.

[0013] A computer device comprising: a processor suitable for executing a computer program; A computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the method described is performed.

[0014] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0015] A computer program product comprises a computer program, wherein the computer program implements the method when executed by a processor.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides an online anomaly detection method for analysis of dissolved gas in transformer oil. The streaming computing technology processes the real-time generated monitoring data of dissolved gas in transformer oil. The streaming computing technology can process high-speed data streams with its efficient real-time data processing capability, provide real-time analysis and prediction, and can respond to events faster than the traditional batch processing mode, avoid data accumulation and delay, and provide strong data support for the stable operation of the transformer. The RRCF algorithm is a real-time anomaly detection algorithm suitable for data streams, which can train and detect anomalies in real time in data streams. The streaming computing provides real-time processing capability of data streams. Combined with the RRCF algorithm, it can effectively monitor and detect anomalies of dissolved gas data in transformer oil in real time, thereby improving the accuracy and response speed of fault detection. Finally, the window mechanism is used to group the data stream, and the time window is used to aggregate and analyze the data, providing efficient and real-time data support for the anomaly detection algorithm. The method not only focuses on the anomaly score in a single window, but observes the performance of data points in a series of continuous windows. If a data point continues to obtain high anomaly scores in multiple continuous windows, then the data point will be finally determined as high-risk anomaly data. This comprehensive judgment mechanism significantly improves the accuracy and robustness of anomaly detection and reduces the possibility of false alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of the method described in the embodiment of the present invention. DETAILED DESCRIPTION

[0018] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0019] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0020] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0021] Example 1 This embodiment provides an online abnormality detection method for dissolved gas analysis in transformer oil, which combines stream computing and RRCF algorithm to solve the problem of online abnormality detection of dissolved gas analysis in transformer oil, wherein: 1. Stream computing can process the real-time monitoring data of dissolved gas in transformer oil, group the data stream using the window mechanism, aggregate and analyze the data in the time window, and provide efficient and real-time data support for the anomaly detection algorithm; 2. The RRCF algorithm is a real-time anomaly detection algorithm suitable for data streams. It can train and detect anomalies in real time in data streams. Stream computing provides real-time processing capabilities for data streams. Combined with the RRCF algorithm, it can effectively monitor and detect anomalies in real time for dissolved gas data in transformer oil, thereby improving the accuracy and response speed of fault detection.

[0022] Specifically, Figure 1 As shown, the online anomaly detection method for dissolved gas analysis in transformer oil provided in this embodiment is further carried out based on the TDengine time series data warehouse that has been built, including the following steps: Step 1: Based on the existing time series data warehouse, obtain data related to dissolved gas in transformer oil.

[0023] In the built TDengine time series data warehouse, the storage status of the data on dissolved gases in transformer oil is that each gas data is stored as an independent subtable, and these subtables have the same table structure, that is, they contain two dynamic fields: timestamp and value. The reason for adopting this design is that it follows TDengine's efficient processing mechanism for time series data and takes advantage of its powerful data compression and fast query capabilities.

[0024] From the established TDengine time series data warehouse, the characteristic data required for online monitoring of transformer oil chromatography are extracted. The required characteristic data include ten sub-tables: hydrogen, carbon monoxide, carbon dioxide, methane, acetylene, ethylene, ethane, total hydrocarbons, trace water, and oil temperature, which will be used for subsequent work.

[0025] Step 2: Use streaming computing to process the data into the data structure required by the RRCF algorithm.

[0026] The obtained feature data is processed using streaming computing to obtain a data set.

[0027] Stream processing, also known as real-time stream computing, is a computing mode that can process and analyze continuously generated data streams in real time. Unlike the traditional batch processing mode, stream computing does not need to wait for all data to arrive before starting processing. Instead, it is driven by data generation and processes data while receiving it. This mode is often used in scenarios such as the Internet of Things and finance that require processing of large amounts of continuous data.

[0028] In this embodiment, the built-in stream computing engine of the TDengine time series database product is used, which provides the ability to process written data streams in real time. SQL can be used to define real-time stream transformations. When data is written to the source table of the stream, the data will be automatically processed in the defined manner and the results will be pushed to the destination table according to the defined trigger mode. It provides a lightweight solution to replace complex stream processing systems and can provide millisecond-level calculation result delays in high-throughput data writing.

[0029] Step 3: Construct and maintain random cut forest RCF according to the RRCF algorithm.

[0030] This embodiment uses the RRCF algorithm to detect anomalies in the data of dissolved gas in transformer oil. The RRCF algorithm (robust random cut forest) is designed for processing dynamic data streams, so it has significant advantages in detecting anomalies in time series data. It constructs a set of random cut trees to detect anomalies or outliers in the data stream, which is very suitable for application scenarios such as transformer online anomaly detection that require immediate response. The following is the detailed process of the algorithm to build and maintain the random cut forest: Step 3.1: Create a random cut tree RRCT Take each record in the data set as a data sample, randomly select an attribute A from the data sample, randomly select a value j between the maximum and minimum values ​​of this attribute A, and use j as the split value to divide the value of this attribute A into two parts. Values ​​less than this split value are divided into the left node, and values ​​greater than the split value are divided into the right node. Repeat the above steps and recursively construct the left and right child nodes of each node in the cutting tree until the termination condition is met, that is, the left and right nodes contain only one sample or multiple identical samples, or the depth of the tree reaches the specified height.

[0031] Step 3.2: Update the cutting tree RRCT in real time When new data samples are added or old data samples are removed, the structure of the cut tree will be updated and adjusted accordingly. This feature is also the theoretical basis for the RRCF algorithm to be good at processing streaming data.

[0032] ① Search for the node Q corresponding to the data sample p to be deleted in the original RRCT and the sibling node R of the Q node; ② Delete the parent node of Q and replace it with R; ③Get the modified RRCT'.

[0033] ① Calculate the maximum and minimum values ​​of each attribute in the non-empty sample set Z; ② Add sample p to sample set Z to form a new sample set Z', and recalculate the maximum and minimum values ​​of each attribute; ③ According to the process in step 3.1, randomly select a split value between the new maximum and minimum values, compare the size of sample p with the split value, and insert sample p into the corresponding node of the segmentation tree.

[0034] Step 4: Calculate the anomaly score for each data sample in each dataset In the RRCF algorithm, a random cut tree RRCT that can be updated in real time is first constructed, and the depth of the leaf nodes of each tree is used to measure the complexity of the current random cut forest. That is, if the addition of the data sample will significantly increase the complexity of RRCT, then this data sample is considered an outlier, and it is further scored for anomaly. Finally, the outliers are screened out according to the scoring threshold.

[0035] The complexity of random cut forest is defined as:

[0036] Where y is a data point belonging to the non-empty sample set Z; is the depth of the tree.

[0037] Therefore, the calculation formula of the abnormal score value can be expressed as:

[0038] in, is the abnormal score value; T represents the random cutting tree T, Represents a data set that does not contain sample x; The anomaly score is reflected in the comparison of the changes in the depth values ​​of all leaf nodes in the RRCT tree when the data sample x is deleted. The larger the score value, the more obvious the anomaly of the data sample x in the data set.

[0039] Step 5: Output the identified abnormal data samples.

[0040] In order to improve the accuracy of abnormality monitoring, this embodiment uses the sliding window mechanism of stream computing to judge the data stream of dissolved gas in transformer oil in multiple continuous time windows. If a data sample is rated as a high abnormality score in multiple consecutive windows, then the data point will be marked as high-risk abnormal data. Specifically: According to the step 4 Sort the data samples from 1 to n in descending order of value and output the top 2% sample point set D 1 , this is the first round of outlier judgment; Output the top 2% sample point set D after the anomaly scores are sorted by the N sliding windows in sequence 2 ~D N ; Compare these N abnormal sample sets, and output a set G of high-risk abnormal points that appear N / 3 times from the N abnormal sample sets; Finally, the last abnormal point set D is output N And high-risk gathering point G.

[0041] The working principle of this embodiment: Most of the current abnormality detection algorithms for dissolved gas data in transformer oil are based on offline data training, and generally have problems such as dependence on historical data, lack of real-time performance, and sample imbalance. This embodiment proposes an online abnormality detection method for dissolved gas analysis in transformer oil that integrates streaming computing and RRCF algorithm. The advantages of this method are mainly reflected in the following points: Stream computing technology, with its efficient real-time data processing capabilities, can process high-speed data streams and provide real-time analysis and prediction. Compared with traditional batch processing, stream computing can respond to events faster and avoid data accumulation and delays. In the monitoring of dissolved gas in transformer oil, stream computing can realize real-time analysis of continuous data streams, help anomaly detection algorithms detect anomalies, and provide strong data support for the stable operation of transformers.

[0042] The RRCF algorithm is a classic streaming anomaly detection algorithm. Its advantage is that it can update the model in real time and has good processing capabilities for high-dimensional data streams. It randomly selects features and cut points, builds multiple random trees, and evaluates the degree of abnormality of data points through the statistical information of the common cut points.

[0043] 3. The comprehensive sliding window judgment mechanism increases the robustness of the detection results and reduces the possibility of false alarms.

[0044] This method obtains data streams through the sliding window method and inputs them into the RRCF algorithm for processing. By integrating the analysis results of multiple windows, high-risk abnormal data is marked, thereby enhancing the robustness of the detection results. This method not only focuses on the abnormality score in a single window, but also observes the performance of data points in a series of continuous windows. If a data point continues to obtain high abnormality scores in multiple continuous windows, then this data point will be ultimately judged as high-risk abnormal data. This comprehensive judgment mechanism significantly improves the accuracy and robustness of anomaly detection and reduces the possibility of false alarms.

[0045] Example 2 This embodiment provides an online abnormality detection system for analyzing dissolved gas in transformer oil, comprising: A characteristic data acquisition unit, used to acquire characteristic data required for analyzing dissolved gas in transformer oil; A data processing unit, used to process the obtained feature data by using streaming computing to obtain a data set containing multiple data samples; An abnormal score value calculation unit, used to calculate the abnormal score value corresponding to each data sample in the data set using the RRCF algorithm; An abnormal data identification unit is used to identify abnormal data based on the obtained abnormal score value Example 3 This embodiment 3 provides a computer device, including: a memory for storing a computer program; and a processor for implementing the steps of a computer method when executing the computer program.

[0046] When the processor executes the computer program, the steps of the above-mentioned computer method are implemented, for example: An online abnormality detection method for analysis of dissolved gas in transformer oil, the method comprising the following steps: Step 1, obtaining characteristic data required for analyzing dissolved gas in transformer oil; Step 2: Use streaming computing to process the obtained feature data to obtain a data set containing multiple data samples; Step 3, use the RRCF algorithm to calculate the anomaly score corresponding to each data sample in the data set; Step 4: Identify abnormal data based on the obtained abnormal score value.

[0047] Alternatively, when the processor executes the computer program, the functions of each module in the above system are realized, for example: A characteristic data acquisition unit, used to acquire characteristic data required for analyzing dissolved gas in transformer oil; A data processing unit, used to process the obtained feature data by using streaming computing to obtain a data set containing multiple data samples; An abnormal score value calculation unit, used to calculate the abnormal score value corresponding to each data sample in the data set using the RRCF algorithm; The abnormal data identification unit is used to identify abnormal data according to the obtained abnormal score value.

[0048] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing preset functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0049] For example, the computer program can be divided into: a feature data acquisition unit, a data processing unit, an abnormal score value calculation unit and an abnormal data identification unit, and the specific functions of each module are as follows: A characteristic data acquisition unit, used to acquire characteristic data required for analyzing dissolved gas in transformer oil; A data processing unit, used to process the obtained feature data by using streaming computing to obtain a data set containing multiple data samples; An abnormal score value calculation unit, used to calculate the abnormal score value corresponding to each data sample in the data set using the RRCF algorithm; The abnormal data identification unit is used to identify abnormal data according to the obtained abnormal score value.

[0050] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above are examples of computer devices and do not constitute a limitation on computer devices. The computer device may include more components than the above, or a combination of certain components, or different components. For example, the computer device may also include input and output devices, network access devices, buses, etc.

[0051] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the computer device, and uses various interfaces and lines to connect various parts of the entire computer device.

[0052] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.

[0053] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0054] Example 4 This embodiment 4 also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, an online abnormality detection method for dissolved gas analysis in transformer oil is implemented, the method comprising the following steps: Step 1, obtaining characteristic data required for analyzing dissolved gas in transformer oil; Step 2: Use streaming computing to process the obtained feature data to obtain a data set containing multiple data samples; Step 3, use the RRCF algorithm to calculate the anomaly score corresponding to each data sample in the data set; Step 4: Identify abnormal data based on the obtained abnormal score value.

[0055] If the module / unit integrated in the computer system is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0056] Based on such understanding, the present invention implements all or part of the processes in the above method, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above computer method can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or preset intermediate form, etc.

[0057] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0058] It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0059] Example 5 This embodiment 5 provides a computer product, the computer program product includes a computer program, the computer program is stored in a computer-readable storage medium; a processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device can perform an online abnormality detection method for dissolved gas analysis in transformer oil, the method comprising the following steps: Step 1, obtaining characteristic data required for analyzing dissolved gas in transformer oil; Step 2: Use streaming computing to process the obtained feature data to obtain a data set containing multiple data samples; Step 3, use the RRCF algorithm to calculate the anomaly score corresponding to each data sample in the data set; Step 4: Identify abnormal data based on the obtained abnormal score value.

[0060] It should be noted that a person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0061] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An online abnormality detection method for analysis of dissolved gas in transformer oil, characterized in that: The following steps are involved: Step 1, obtaining characteristic data required for analyzing dissolved gas in transformer oil; Step 2: Use streaming computing to process the obtained feature data to obtain a data set containing multiple data samples; Step 3, use the RRCF algorithm to calculate the anomaly score corresponding to each data sample in the data set; Step 4: Identify abnormal data based on the obtained abnormal score value.

2. The online abnormality detection method for dissolved gas analysis in transformer oil according to claim 1, characterized in that: In step 1, the required characteristic data include hydrogen sub-table, carbon monoxide sub-table, carbon dioxide sub-table, methane sub-table, acetylene sub-table, ethylene sub-table, ethane sub-table, total hydrocarbon sub-table, trace water sub-table and oil temperature sub-table, and the table structure of each sub-table includes a timestamp and a numerical value.

3. The online abnormality detection method for dissolved gas analysis in transformer oil according to claim 1, characterized in that: The required feature data is post-processed using the built-in streaming computing engine of the TDengine time series database to obtain a data set containing multiple data samples.

4. The online abnormality detection method for dissolved gas analysis in transformer oil according to claim 1, characterized in that: In step 3, the RRCF algorithm is used to calculate the anomaly score corresponding to each data sample in the data set. The specific method is: The obtained data set is combined with the RRCF algorithm to construct a set of random cut forests consisting of random cut trees, where an attribute is randomly selected from each data sample in the data set, a random cut tree is constructed for each attribute, and the random cut trees corresponding to all attributes constitute a random cut forest; The complexity of the random cut forest is defined by the depth of the random cut tree, and the calculation formula of the anomaly score value is defined by the complexity of the random cut forest; The abnormal score calculation formula is used to calculate the abnormal score value corresponding to each data sample.

5. The online abnormality detection method for dissolved gas analysis in transformer oil according to claim 4, characterized in that: The calculation formula of abnormal score value is as follows: in, is the abnormal score value.

6. The online abnormality detection method for dissolved gas analysis in transformer oil according to claim 1, characterized in that: In step 4, the abnormal data is identified based on the obtained abnormal score value. The specific method is: Set up a sliding window mechanism; If the anomaly score value of a data sample is the highest anomaly score value in multiple consecutive windows, then the data sample is high-risk anomaly data.

7. An online abnormality detection system for analysis of dissolved gas in transformer oil, characterized in that: include: A characteristic data acquisition unit, used to acquire characteristic data required for analyzing dissolved gas in transformer oil; A data processing unit, used to process the obtained feature data by using streaming computing to obtain a data set containing multiple data samples; An abnormal score value calculation unit, used to calculate the abnormal score value corresponding to each data sample in the data set using the RRCF algorithm; The abnormal data identification unit is used to identify abnormal data according to the obtained abnormal score value.

8. A computer device, characterized in that: include: a processor suitable for executing a computer program; A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the method according to any one of claims 1 to 6 is performed.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images