Rectifier transformer fault diagnosis method and system

Abstract

The invention relates to the field of rectifier transformers, in particular to a rectifier transformer fault diagnosis method and system. The method comprises the following steps: constructing a temperature sequence of the rectifier transformer according to a preset sampling interval, setting an influence window of a target temperature, and calculating an abnormal degree of the target temperature; constructing an isolated tree, constructing a scatter diagram by taking the abnormal degree of the temperature data points in the temperature data set as an abscissa and taking the path length of the temperature data points in the isolated tree as an ordinate, and calculating the credibility of the target temperature according to the scatter diagram; and taking a result of credibility normalization as a weight, and taking a result of weighted summation of the path length as an average path length of the target temperature in the isolated forest, so as to calculate an abnormal score and complete fault diagnosis. According to the technical scheme, the accuracy of the temperature anomaly diagnosis result of the rectifier transformer can be improved.

Description

Technical Field

[0001] The present invention relates to the field of rectifier transformers, and more specifically, to a rectifier transformer fault diagnosis method and system. Background Art

[0002] A rectifier transformer is a special transformer whose main function is to convert alternating current into direct current. This type of transformer is widely used in applications where a DC power supply is required, such as electrochemical industry, electric traction, steel rolling, electroplating, electrolysis, and other fields. They play a vital role in converting electrical energy into a form suitable for specific load requirements. Failure of a rectifier transformer can not only cause power supply interruptions, affecting production and life, but can also cause safety accidents, resulting in casualties and property losses. Therefore, effective fault diagnosis and preventive maintenance of rectifier transformers are key to ensuring stable operation of the power system.

[0003] The existing Chinese patent application document with publication number CN113139610A discloses a method and device for detecting anomalies in transformer monitoring data. The method comprises: obtaining a sequence to be detected of transformer online monitoring data; constructing an abnormal data recognition model by using time series modeling and an isolation forest algorithm; constructing an abnormal type recognition pattern by using an improved multidimensional SAX vector representation method; identifying abnormal data of the sequence to be detected by using the abnormal data recognition model; determining the abnormal type of the abnormal data by using the abnormal type recognition pattern, wherein the abnormal type comprises an invalid abnormal pattern and a valid abnormal pattern; and when the abnormal type is the invalid abnormal pattern, performing a correlation check on the abnormal type by using a sequence association analysis.

[0004] However, the anomaly score of the monitoring data in the isolation forest is related to the path length of the data. Since there is randomness in the extraction process of the monitoring data, the monitoring data extraction results are inaccurate, which will lead to inaccurate anomaly detection results for the transformer. Summary of the invention

[0005] In order to solve the problem of inaccurate abnormality detection results for transformers, the present invention provides a rectifier transformer fault diagnosis method and system.

[0006] In a first aspect, the present invention discloses a rectifier transformer fault diagnosis method, comprising: constructing a temperature sequence of the rectifier transformer according to a preset sampling interval, taking any temperature data point in the temperature sequence as a target temperature, setting an influence window of the target temperature, and calculating the abnormality of the target temperature; randomly obtaining a preset number of temperature data sets with the same length from the temperature sequence, constructing an isolated tree, wherein one temperature data set corresponds to one isolated tree, taking the abnormality of the temperature data point in the temperature data set as the abscissa and the path length of the temperature data point in the isolated tree as the ordinate to construct a scatter plot, and calculating the credibility of the target temperature according to the scatter plot; taking the result of the credibility normalization as a weight, and taking the result of the weighted summation of the path length as the average path length of the target temperature in the isolated forest, so as to calculate the abnormality score and complete the fault diagnosis; wherein the average path length satisfies the relationship: , Represents a temperature data point The average path length, Indicates that it contains temperature data points The number of isolated trees, Represents a temperature data point In the The path length in an isolated tree, Represents a temperature data point In the The credibility of an isolated tree, Represents the normalization function.

[0007] By setting the impact window and calculating the abnormal degree of the target temperature, a basis is provided for real-time monitoring of the temperature data of the rectifier transformer. Secondly, by randomly obtaining multiple data sets of the same length from the temperature sequence and constructing isolated trees, these isolated trees are used to evaluate the credibility of the target temperature, thereby enhancing the ability to identify abnormal behavior of temperature data points. Then, the credibility is used as a weight to perform a weighted summation of the path lengths, and the average path length of the target temperature in the isolated forest is calculated as the anomaly score, which provides a quantitative indicator for anomaly detection of temperature data. It not only improves the accuracy of anomaly detection, but also can adapt to different data distributions and changes, which helps to timely discover and respond to potential temperature anomalies in rectifier transformers and ensure the safety of rectifier transformer operation.

[0008] Preferably, the impact window includes: constructing the impact window with the target temperature as the center and a preset length as the window size.

[0009] Preferably, calculating the degree of abnormality includes: taking the temperature value range of the temperature data points in the impact window as the first span; calculating the temperature mean of all temperature data points in the impact window, removing the temperature data points with the largest difference from the temperature mean in the impact window, and taking the range of the remaining temperature data points as the second span; calculating the degree of abnormality according to the first span and the second span, and the degree of abnormality satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

[0010] The influence of extreme values ​​on the abnormality evaluation is effectively reduced, and the accuracy of the evaluation is improved. The difference between the first span and the second span is used to calculate the abnormality through a specific exponential relationship, where the use of the exponential function enhances the sensitivity to small differences, making the calculation of the abnormality more accurate.

[0011] Preferably, calculating the abnormality degree also includes: taking the temperature value variance of the temperature data points in the impact window as the first span; removing the temperature data point with the largest temperature value and the temperature data point with the smallest temperature value in the impact window, and taking the variance of the remaining temperature data points as the second span; calculating the abnormality degree according to the first span and the second span, and the abnormality degree satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

[0012] By calculating the variance of the temperature data points as the first span and the variance after removing the extreme values ​​as the second span, the solution can accurately evaluate the fluctuation of the temperature data points and identify possible outliers. By removing the extreme values ​​with greater influence, the interference of extreme data on the overall abnormality assessment is reduced, and the credibility of the rectifier transformer temperature data is improved.

[0013] Preferably, calculating the credibility includes: using the Dijkstra algorithm to obtain the global shortest path of all temperature data points in the scatter plot, and taking the ratio of the total number of temperature data points to the global shortest path as the global density; calculating the Euclidean distance between each temperature data point in the scatter plot and the target temperature, taking a preset number of temperature data points closest to the target temperature as reference data points, using the Dijkstra algorithm to obtain the local shortest path between the reference data points and the target temperature, and taking the ratio of the total number of reference data points and the target temperature to the local shortest path as the local density; using the absolute value of the difference between the global density and the local density as the credibility using the result of negative correlation mapping.

[0014] It reflects the relationship between temperature data points and their relative position to the target temperature, so that data points that are abnormally close to or far from the target temperature can be identified. The calculation of global density takes into account all temperature data points, providing an overall view of the data distribution, while local density focuses on data points around the target temperature, capturing more subtle temperature changes. By calculating the local shortest path between the reference data point and the target temperature, the solution can more accurately locate temperature anomalies and improve the accuracy of fault diagnosis. It can help operation and maintenance personnel quickly locate and handle temperature-related faults of rectifier transformers, thereby ensuring the safe and stable operation of the power system and reducing equipment damage and accident risks caused by temperature anomalies.

[0015] Preferably, calculating the credibility also includes: calculating the local density of each temperature data point in the scatter plot using a K-nearest neighbor algorithm; calculating the local density mean of all temperature data points in the scatter plot; and taking the ratio of the local density of the target temperature to the local density mean as the credibility.

[0016] The K-nearest neighbor algorithm is used to calculate the local density of each temperature data point in the scatter plot, taking into account the sample distribution in the vicinity of the temperature data point, thereby more accurately reflecting the local characteristics of the temperature data point. The ratio of the local density of the target temperature to the local density mean is used as the credibility. This method not only considers the local density of the temperature data point itself, but also compares it with the average level of the overall data, thereby more comprehensively evaluating the credibility of the data point, effectively identifying outliers in the temperature data, and improving the accuracy and reliability of temperature monitoring.

[0017] Preferably, the calculation of the abnormality score satisfies the relationship: , Represents a temperature data point The anomaly score, Represents a temperature data point The average path length, Represents the average path length of trees in the isolation forest.

[0018] In a second aspect, the present invention discloses a rectifier transformer fault diagnosis system, comprising: a processor; and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the system executes the above-mentioned rectifier transformer fault diagnosis method.

[0019] Beneficial effects of the present invention: 1. The present invention realizes efficient detection of rectifier transformer faults by combining temperature series analysis and isolation forest algorithm. First, the temperature series is constructed by presetting the sampling interval, and then the impact window of the target temperature is set to calculate its abnormality. Not only the extreme value difference of the temperature data points is considered, but also the extreme values ​​are eliminated to reduce the influence of the abnormal values, thereby improving the accuracy of the abnormality calculation. The physical significance of this step is that by analyzing the statistical characteristics of the temperature data, the abnormal pattern of temperature change can be captured, which provides important data support for fault diagnosis.

[0020] 2. The present invention uses a combination of Dijkstra algorithm and K-nearest neighbor algorithm to calculate the global and local density of each temperature data point in the scatter plot, thereby obtaining a more accurate credibility assessment. This method can not only reflect the positional relationship of the data points in the overall data set, but also capture the density of the data points in the local area, providing a more comprehensive perspective for the identification of abnormal points. This method of comprehensive use of global and local density information improves the accuracy and reliability of fault diagnosis.

[0021] 3. The present invention can help operation and maintenance personnel quickly locate and handle temperature-related faults of rectifier transformers, thereby ensuring the safe and stable operation of the power system and reducing the risk of equipment damage and accidents caused by abnormal temperature. Through this method, early warning and accurate diagnosis of rectifier transformer faults can be achieved, the efficiency and effect of fault handling can be improved, maintenance costs can be reduced, and equipment life can be extended. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 The present invention is a flowchart of a rectifier transformer fault diagnosis method according to an embodiment of the present invention.

[0023] Figure 2 It is a scatter plot constructed in a rectifier transformer fault diagnosis method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0026] The present invention provides a rectifier transformer fault diagnosis method. Figure 1 As shown, a rectifier transformer fault diagnosis method includes steps S1 to S3, which are described in detail below.

[0027] S1, construct the temperature sequence of the rectifier transformer according to the preset sampling interval, set the impact window of the target temperature, and calculate the abnormal degree of the target temperature.

[0028] It should be noted that temperature is a direct manifestation of transformer failure. During operation, the transformer will generate heat due to internal losses (such as copper loss and iron loss). Under normal circumstances, this heat will be dissipated to the surrounding environment through the cooling system to keep the transformer oil temperature and winding temperature within a safe range. When a transformer fails, such as internal short circuit, poor contact, overload, etc., it will cause additional heat loss and increase the temperature. Therefore, an abnormal increase in temperature is often a direct manifestation of a fault.

[0029] Winding overheating is a common and frequent fault of transformers, which poses a serious threat to the safe operation of transformers, because the winding temperature directly determines the service life of the transformer. In rectifier transformers, internal faults such as inter-turn short circuit or inter-layer short circuit, coil discharge to the enclosure, heating of internal lead joints, multi-point grounding of the core to increase eddy current and overheating, etc., will cause abnormal transformer winding temperature. The insulation material of the transformer winding will accelerate aging at high temperature, resulting in a decrease in insulation performance, thus causing failures. Therefore, controlling the winding temperature within a safe range is the key to ensuring the safe operation of the transformer.

[0030] In one embodiment, a person skilled in the art places a temperature sensor at a key position of the rectifier transformer and constructs a temperature sequence according to a preset sampling interval. For example, the preset sampling interval of the key position is set to 5 seconds, which can be specifically set by a person skilled in the art.

[0031] It should be noted that during the operation of the rectifier transformer, its temperature may be abnormal due to cooling system failure or internal failure (winding short circuit, insulation damage, etc.). Such anomalies are drastic changes in temperature data in a short period of time and a large temperature range span. Therefore, an influence window of temperature data can be constructed, and the temperature span within the influence window can be used to calculate the degree of abnormality. The larger the temperature span, the greater the degree of abnormality. At the same time, because in the process of temperature sequence acquisition, environmental factors will cause noise to be mixed in the temperature data collected by the temperature sensor, and noise will also cause the temperature span to become larger, so it is necessary to consider the influence of noise to make corrections to the degree of abnormality.

[0032] Noise often appears as outliers and individual data points. Therefore, the degree of abnormality can be corrected by removing the change in temperature span before and after the temperature data point that deviates most from the temperature mean in the impact window. The greater the change in temperature span before and after removal, the greater the impact of noise in the current impact window, the greater the possibility of false abnormalities caused by noise, and the need to reduce the degree of abnormality.

[0033] Centered on the target temperature, the influence window is constructed with a preset length as the window size. Exemplarily, the preset length is set to 7.

[0034] When the target temperature is located at the two ends of the temperature sequence, that is, the number of temperature data points on the left side or the number of temperature data points on the right side of the target temperature does not meet the preset length, the remaining data points are used as the window size of the influencing window.

[0035] For example, the target temperature is the third temperature data point in the temperature sequence, and the preset length is 7, but the number of temperature data points to the left of the target temperature is only 2, then the influence window length of the target temperature is 6, that is, the number of temperature data points to the left of the target temperature is 2, and the number of temperature data points to the right of the target temperature is 3.

[0036] The temperature range of the temperature data points in the impact window is taken as the first span; the temperature mean of all temperature data points in the impact window is calculated, the temperature data point with the largest difference from the temperature mean in the impact window is removed, and the range of the remaining temperature data points is taken as the second span; the degree of abnormality is calculated based on the first span and the second span, and the degree of abnormality satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

[0037] In one embodiment, the abnormality degree further includes: taking the temperature value variance of the temperature data points in the impact window as the first span; removing the temperature data point with the largest temperature value and the temperature data point with the smallest temperature value in the impact window, and taking the variance of the remaining temperature data points as the second span; calculating the abnormality degree according to the first span and the second span, and the abnormality degree satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

[0038] S2, construct an isolation tree, use the abnormal degree of the temperature data point in the temperature data set as the horizontal axis and the path length of the temperature data point in the isolation tree as the vertical axis to construct a scatter plot, and calculate the credibility of the target temperature based on the scatter plot.

[0039] In one embodiment, for a temperature sequence, a preset number of temperature data sets with the same length are selected by simple random sampling. For example, 50 temperature data sets with a length of 100 are sampled. The specific parameters can be set by a person skilled in the art to ensure that each data point in the temperature sequence is detected as much as possible. An isolation tree is constructed for each temperature data set, and one temperature data set corresponds to one isolation tree.

[0040] like Figure 2 As shown, a scatter plot is constructed by taking the abnormal degree of the temperature data points in the temperature data set as the horizontal axis and the path length of the temperature data points in the isolated tree as the vertical axis.

[0041] The global shortest path of all temperature data points in the scatter plot is obtained by using the Dijkstra algorithm, and the ratio of the total number of temperature data points to the global shortest path is used as the global density; the Euclidean distance between each temperature data point in the scatter plot and the target temperature is calculated, and the preset number of temperature data points closest to the target temperature are used as reference data points, and the local shortest path between the reference data point and the target temperature is obtained by using the Dijkstra algorithm, and the ratio of the total number of reference data points and the target temperature to the local shortest path is used as the local density; the absolute value of the difference between the global density and the local density is used as the credibility using the result of negative correlation mapping. Exemplarily, the negative correlation mapping can use an exponential function.

[0042] When the absolute value of the difference is larger, it means that there are abnormal temperature data points in the current tree, and the path length of the target temperature in the current tree is less credible.

[0043] It should be noted that the path length of the temperature data point in the isolated tree is negatively correlated with the degree of abnormality. Therefore, for an isolated tree, a coordinate system is established with the degree of abnormality as the horizontal axis and the path length as the vertical axis, and then the corresponding temperature data points in the tree are plotted in the coordinate system. The relationship between the abnormality of the data point and the path length in an isolated tree is shown as follows: Figure 2 As shown, most of the temperature data points are concentrated together, while a few temperature data points deviate, indicating that these few temperature data points are biased in the isolated tree, that is, the reliability of the tree in calculating the abnormal score value is low.

[0044] In one embodiment, the credibility also includes: calculating the local density of each temperature data point in the scatter plot using a K-nearest neighbor algorithm; calculating the local density mean of all temperature data points in the scatter plot; and taking the ratio of the local density of the target temperature to the local density mean as the credibility.

[0045] S3, the result of credibility normalization is used as the weight, and the result of weighted summation of path length is used as the average path length of target temperature in the isolated forest to calculate the anomaly score and complete fault diagnosis.

[0046] In one embodiment, for the target temperature, all isolated trees containing the target temperature are obtained, and the credibility sequence and path length sequence of the target temperature in each isolated tree are recorded.

[0047] The result of weighted summation of path lengths is taken as the average path length of the target temperature in the isolated forest. The average path length satisfies the relationship: , Represents a temperature data point The average path length, Indicates that it contains temperature data points The number of isolated trees, Represents a temperature data point In the The path length in an isolated tree, Represents a temperature data point In the The credibility of an isolated tree, Represents the normalization function.

[0048] The anomaly score of the target temperature is calculated using the isolation forest, and the calculation of the anomaly score satisfies the relationship: , Represents a temperature data point The anomaly score, Represents a temperature data point The average path length, Represents the average path length of trees in the isolation forest.

[0049] The abnormal score of each temperature data point in the temperature sequence is obtained by traversing, and the temperature data point with an abnormal score greater than the preset abnormal threshold is marked as abnormal data. At this time, it is considered that the rectifier transformer is faulty, and an alarm signal is immediately generated and sent. At the same time, the power is immediately cut off to prevent greater danger, reminding the staff to conduct a comprehensive overhaul of the rectifier transformer in time and complete the fault diagnosis of the rectifier transformer.

[0050] An embodiment of the present invention further discloses a rectifier transformer fault diagnosis system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a rectifier transformer fault diagnosis method according to the present invention is implemented.

[0051] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0052] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0053] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0054] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A rectifier transformer fault diagnosis method, It is characterized in that include: Construct a temperature sequence of the rectifier transformer according to a preset sampling interval, take any temperature data point in the temperature sequence as the target temperature, set the impact window of the target temperature, and calculate the abnormality degree of the target temperature; Randomly obtain a preset number of temperature data sets with the same length from the temperature sequence, and construct an isolation tree, where one temperature data set corresponds to one isolation tree. A scatter plot is constructed using the abnormality of the temperature data points in the temperature data set as the horizontal axis and the path length of the temperature data points in the isolation tree as the vertical axis. The credibility of the target temperature is calculated based on the scatter plot. The result of credibility normalization is used as the weight, and the result of weighted summation of path length is used as the average path length of the target temperature in the isolated forest to calculate the anomaly score and complete the fault diagnosis; Among them, the average path length satisfies the relationship: , Represents a temperature data point The average path length, Indicates that it contains temperature data points The number of isolated trees, Represents a temperature data point In the The path length in an isolated tree, Represents a temperature data point In the The credibility of an isolated tree, Represents the normalization function.

2. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that The impact window includes: The influence window is constructed with the target temperature as the center and the preset length as the window size.

3. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that Calculating the abnormality degree includes: The temperature range of the temperature data points in the impact window is taken as the first span; Calculate the temperature mean of all temperature data points in the impact window, remove the temperature data point with the largest difference from the temperature mean in the impact window, and use the range of the remaining temperature data points as the second span; The degree of abnormality is calculated based on the first span and the second span, and the degree of abnormality satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

4. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that Calculating the abnormality level further includes: The variance of the temperature values ​​of the temperature data points in the impact window is taken as the first span; Eliminate the temperature data point with the largest temperature value and the temperature data point with the smallest temperature value in the impact window, and use the variance of the remaining temperature data points as the second span; The degree of abnormality is calculated based on the first span and the second span, and the degree of abnormality satisfies the relationship: , Represents a temperature data point The degree of abnormality, Represents a temperature data point The first span of the window of influence, Represents a temperature data point The second span of the impact window, Represents an exponential function.

5. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that Calculating the credibility includes: The Dijkstra algorithm is used to obtain the global shortest path of all temperature data points in the scatter plot, and the ratio of the total number of temperature data points to the global shortest path is taken as the global density; Calculate the Euclidean distance between each temperature data point and the target temperature in the scatter plot, take a preset number of temperature data points closest to the target temperature as reference data points, use the Dijkstra algorithm to obtain the local shortest path between the reference data points and the target temperature, and take the ratio of the total number of reference data points and the target temperature to the local shortest path as the local density; The absolute value of the difference between the global density and the local density is used as the credibility using the result of negative correlation mapping.

6. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that Calculating the credibility also includes: The local density of each temperature data point in the scatter plot is calculated using the K-nearest neighbor algorithm; Calculate the local density mean of all temperature data points in the scatter plot; The ratio of the local density of the target temperature to the local density mean is taken as the credibility.

7. A rectifier transformer fault diagnosis method according to claim 1, It is characterized in that The calculation of the abnormality score satisfies the relationship: , Represents a temperature data point The anomaly score, Represents a temperature data point The average path length, Represents the average path length of trees in the isolation forest.

8. A rectifier transformer fault diagnosis system, It is characterized in that include: Processor; and A memory storing computer instructions, wherein when the computer instructions are executed by a processor, the system executes a rectifier transformer fault diagnosis method according to any one of claims 1 to 7.

Images