Transformer fault monitoring method and system

Abstract

The present invention relates to the technical field of transformer fault detection, and specifically discloses a transformer fault monitoring method and system. The method comprises: obtaining historical fault parameters of the transformer, drawing a transformer fault parameter change curve based on the historical fault parameters of the transformer; clustering the curve period based on the transformer fault parameter change curve, and determining a fluctuation tolerance threshold of the curve period based on the clustering result; determining a transformer volatility index based on the fluctuation tolerance threshold, setting an initial #imgabs0# value, and correcting the initial #imgabs1# value based on the transformer volatility index; extracting abnormal data points of the curve period based on the corrected #imgabs2# value using a LOF algorithm, and monitoring the current operating parameters of the transformer based on the abnormal data points of the fault period. By monitoring the operating parameters of the transformer in real time, abnormal data points can be quickly located, and operating faults can be discovered in a timely manner.

Description

Technical Field

[0001] The present application relates to the technical field of transformer fault detection, and more specifically, to a transformer fault monitoring method and system. Background Art

[0002] Power transformers are essential equipment in power plants and substations. Transformers perform a variety of functions: not only stepping up voltage to deliver electricity to users, but also stepping it down to various voltages to meet demand. In short, both stepping up and stepping down voltage are performed by transformers. Transformers are core components of power systems. The safe operation of transformers not only impacts the operating costs and safety of power systems but also, to a significant extent, the normal operation of society and the safety of people and property. Fault monitoring technology can promptly detect faults during transformer operation and enable timely action to mitigate accidents.

[0003] In the prior art, transformer monitoring is generally performed through manual detection and judgment, resulting in low fault monitoring efficiency, inability to detect transformer operating faults in a timely manner, and low reliability. Summary of the Invention

[0004] The present invention provides a transformer fault monitoring method and system to solve the problem of delayed transformer fault detection in the prior art, including:

[0005] Obtain historical fault parameters of the transformer, and draw a transformer fault parameter change curve based on the historical fault parameters of the transformer;

[0006] Clustering the curve period according to the transformer fault parameter change curve, and determining the fluctuation tolerance threshold of the curve period according to the clustering result;

[0007] Determine the transformer's fluctuation index based on the fluctuation tolerance threshold and set the initial Value, according to the fluctuation index of the transformer The value is corrected;

[0008] According to the revised The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle.

[0009] Furthermore, clustering the curve periods according to the transformer fault parameter change curve includes:

[0010] Obtaining a preset curve period, and dividing the transformer fault parameter change curve according to the preset curve period to obtain a plurality of first sub-fault parameter curves;

[0011] Obtaining a vibration signal of the transformer within a curve period corresponding to the first sub-fault parameter curve, and drawing a vibration signal change curve according to the vibration signal of the transformer within the curve period corresponding to the first sub-fault parameter curve;

[0012] The correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve in the corresponding curve period is calculated, and the curve period is clustered according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve.

[0013] Furthermore, clustering the curve periods according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve includes:

[0014] A sample data set is established based on the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve, and k initial cluster centers of the sample data set are randomly selected;

[0015] Calculate the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center, and divide the curve period corresponding to the correlation coefficient into the corresponding cluster cluster according to the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center;

[0016] Calculate the average value of the correlation coefficient within each cluster, and recalculate the cluster center based on the average value of the correlation coefficient within each cluster;

[0017] Repeat the above steps until the cluster center no longer changes or the number of iterations reaches the preset maximum number of iterations, and obtain the clustering result of the curve period.

[0018] Furthermore, determining the fluctuation tolerance threshold of the curve period according to the clustering result includes:

[0019] Determine the cluster center of each curve period based on the clustering results of the curve period, and calculate the average value of all cluster centers;

[0020] The absolute value of the difference between each cluster center and the average value is calculated based on the average value of all cluster centers, the absolute value of the difference between each cluster center and the average value is normalized, and the normalized absolute value of the difference is determined as the tolerance threshold coefficient;

[0021] A preset fluctuation tolerance threshold is obtained, and the preset fluctuation tolerance threshold is corrected according to the tolerance threshold coefficient to obtain the fluctuation tolerance threshold.

[0022] Furthermore, determining the transformer fluctuation index according to the fluctuation tolerance threshold includes:

[0023] Obtaining a preset sliding time window, and segmenting the transformer fault parameter change curve according to the preset sliding time window to obtain a plurality of second sub-fault parameter curves;

[0024] Calculating an average value of the second sub-fault parameter curve, and drawing an average value change curve according to the average value of the second sub-fault parameter curve;

[0025] Obtain the absolute value of the slope of two adjacent average values ​​in the average value change curve, and calculate the difference between the absolute value of the slope and the fluctuation tolerance threshold;

[0026] The average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is calculated, and the average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is determined as the volatility index.

[0027] Furthermore, the initial The values ​​are modified, including:

[0028] according to The value correction formula is for the initial The value is corrected, The specific value correction formula is:

[0029] ,

[0030] in, For the corrected value, For the initial value, is the preset indicator tolerance threshold, is a volatility indicator, is the preset range coefficient, is the natural exponential function.

[0031] Furthermore, the modified The value is based on the LOF algorithm to extract abnormal data points of the fault cycle, including:

[0032] The LOF algorithm is used for each curve period to obtain the local anomaly factor of each data point;

[0033] Extract outlier data points based on the local outlier factor of each data point.

[0034] Furthermore, extracting abnormal data points according to the local abnormal factor of each data point includes:

[0035] If the local anomaly factor is greater than 1, the corresponding local anomaly factor is extracted as the abnormal data point;

[0036] If the local anomaly factor is less than or equal to 1, the corresponding local anomaly factor is determined to be a normal data point.

[0037] Furthermore, the monitoring of the current operating parameters of the transformer according to the abnormal data points of the fault cycle includes:

[0038] Obtain abnormal data points and corresponding curve periods of historical fault parameters of the transformer, and establish a training sample set based on the abnormal data points and corresponding curve periods of historical fault parameters of the transformer;

[0039] Establishing an initial fault monitoring model based on the training sample set and training the initial fault monitoring model to obtain a trained fault monitoring model;

[0040] Obtain the current operating parameters of the transformer and the corresponding curve period, input the current operating parameters of the transformer and the corresponding curve period into the trained fault monitoring model to determine whether the transformer has a fault.

[0041] In order to achieve the above object, the present invention further provides a transformer fault monitoring system, comprising:

[0042] An acquisition module is used to acquire historical fault parameters of the transformer and draw a transformer fault parameter change curve based on the historical fault parameters of the transformer;

[0043] A setting module is used to cluster the curve period according to the transformer fault parameter change curve, and determine the fluctuation tolerance threshold of the curve period according to the clustering result;

[0044] Correction module, used to determine the fluctuation index of transformer according to the fluctuation tolerance threshold, set the initial Value, according to the fluctuation index of the transformer The value is corrected;

[0045] Monitoring module, used to The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle.

[0046] The beneficial effects of the present invention are:

[0047] By applying the above technical solution, the present invention corrects the LOF algorithm by the volatility index of the transformer. The abnormal data points in the historical fault parameters of the transformer are extracted based on the LOF algorithm, which effectively improves the accuracy of abnormal data point extraction. The operating parameters of the transformer are monitored by the extracted abnormal data points, which can quickly identify abnormal data points in the operating parameters and improve the fault detection efficiency of the transformer. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 A schematic diagram showing a flow chart of a transformer fault monitoring method proposed in an embodiment of the present invention is shown;

[0050] Figure 2 The figure shows the overall structure of a transformer fault monitoring system proposed in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0052] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. Throughout this application, unless otherwise specified, "plurality" means two or more.

[0053] In the description of this application, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0054] The present application embodiment provides a transformer fault monitoring method, such as Figure 1 Shown, including:

[0055] S101, obtaining historical fault parameters of the transformer, and drawing a transformer fault parameter change curve based on the historical fault parameters of the transformer;

[0056] S102, clustering the curve period according to the transformer fault parameter change curve, and determining the fluctuation tolerance threshold of the curve period according to the clustering result;

[0057] In some embodiments of the present application, the clustering of curve periods according to the transformer fault parameter change curve includes: obtaining a preset curve period, segmenting the transformer fault parameter change curve according to the preset curve period to obtain a plurality of first sub-fault parameter curves; obtaining the vibration signal of the transformer within the curve period corresponding to the first sub-fault parameter curve, and drawing a vibration signal change curve according to the vibration signal of the transformer within the curve period corresponding to the first sub-fault parameter curve; calculating the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve within the corresponding curve period, and clustering the curve periods according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve.

[0058] In this embodiment, when the transformer is operating normally, all internal components operate according to the set operating rules, so its operating parameters have a strong correlation with the vibration signal. When the transformer fails, the correlation between the operating parameters and the vibration signal will decrease.

[0059] In this embodiment, the historical fault parameters of the transformer are the voltage or current change parameters when the transformer has a historical fault. The multiple segmented curve periods are clustered by calculating the Pearson correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve in the same curve period, and the state of the transformer in each curve period is determined by the correlation coefficient.

[0060] In some embodiments of the present application, the clustering of curve periods according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve includes: establishing a sample data set according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve, and randomly selecting k initial clustering centers of the sample data set; calculating the Euclidean distance from the correlation coefficient in the sample data set to the initial clustering center, and dividing the curve periods corresponding to the correlation coefficient into corresponding cluster clusters according to the Euclidean distance from the correlation coefficient in the sample data set to the initial clustering center; calculating the average value of the correlation coefficient within each cluster, and recalculating the cluster center according to the average value of the correlation coefficient within each cluster; repeating the above steps until the cluster center no longer changes or the number of iterations reaches a preset maximum number of iterations, to obtain the clustering results of the curve periods.

[0061] In this embodiment, the curve periods are clustered based on the k-means clustering algorithm. The preset k value of the clustering algorithm can be set according to the number of curve periods. In this embodiment, the k value is set to 10.

[0062] In some embodiments of the present application, the determination of the fluctuation tolerance threshold of the curve period based on the clustering results includes: determining the cluster center of each curve period based on the clustering results of the curve period, and calculating the average value of all cluster centers; calculating the absolute value of the difference between each cluster center and the average value based on the average value of all cluster centers, normalizing the absolute value of the difference between each cluster center and the average value, and determining the normalized absolute value of the difference as the tolerance threshold coefficient; obtaining a preset fluctuation tolerance threshold, and correcting the preset fluctuation tolerance threshold according to the tolerance threshold coefficient to obtain the fluctuation tolerance threshold.

[0063] In this embodiment, the cluster center of each curve cycle represents the average correlation coefficient of the cluster within which the curve cycle resides. By calculating the absolute value of the difference between the cluster center of each curve cycle and the average value and normalizing the difference, the range of the absolute value of the difference is limited to 1-2, thereby obtaining a tolerance threshold coefficient. A higher absolute value of the curve cycle difference indicates that the correlation coefficient of that curve cycle is further away from the average value, and the probability of transformer failure within that curve cycle is greater, thus resulting in a larger calculated tolerance threshold coefficient. The preset fluctuation tolerance threshold is adjusted using the tolerance threshold coefficient. A larger tolerance threshold coefficient results in a smaller adjusted fluctuation tolerance threshold, thereby obtaining the final fluctuation tolerance threshold for each curve cycle.

[0064] S103, determine the transformer fluctuation index according to the fluctuation tolerance threshold, and set the initial Value, according to the fluctuation index of the transformer The value is corrected;

[0065] In some embodiments of the present application, the determination of the volatility index of the transformer based on the fluctuation tolerance threshold includes: obtaining a preset sliding time window, dividing the transformer fault parameter change curve according to the preset sliding time window to obtain a plurality of second sub-fault parameter curves; calculating the average value of the second sub-fault parameter curve, and drawing the average value change curve according to the average value of the second sub-fault parameter curve; obtaining the absolute value of the slope of two adjacent average values ​​in the average value change curve, and calculating the difference between the absolute value of the slope and the fluctuation tolerance threshold; calculating the average value of the difference between all the absolute values ​​of the slope and the fluctuation tolerance threshold, and determining the average value of the difference between all the absolute values ​​of the slope and the fluctuation tolerance threshold as the volatility index. The volatility index represents the degree of fluctuation of the transformer in the corresponding curve period. The larger the volatility index, the higher the corresponding degree of fluctuation.

[0066] In this embodiment, the fluctuation coefficient of the transformer is obtained by calculating the absolute value of the slope of two adjacent average values ​​in the average value change curve of the second sub-fault parameter curve, and the average value of the difference between the fluctuation coefficient and the fluctuation tolerance threshold is determined as the fluctuation index, thereby obtaining the fluctuation index of each curve period.

[0067] In some embodiments of the present application, the initial The value is modified, including: The value correction formula is for the initial The value is corrected, The specific value correction formula is:

[0068] ,

[0069] in, For the corrected value, For the initial value, is the preset indicator tolerance threshold, is a volatility indicator, is the preset range coefficient, is the natural exponential function.

[0070] In this embodiment, the volatility index is used to The value is corrected to obtain the period of each curve The larger the volatility index, the higher the corrected The smaller the value, the smaller the The value can more accurately identify abnormal data points, prevent the phenomenon of missing abnormal data points, and effectively improve the accuracy of the algorithm.

[0071] S104, according to the revised The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle.

[0072] In some embodiments of the present application, the modified The method extracts abnormal data points of the fault cycle based on the LOF algorithm, including: applying the LOF algorithm to each curve cycle to obtain a local abnormal factor of each data point; and extracting abnormal data points according to the local abnormal factor of each data point.

[0073] In some embodiments of the present application, the extracting of abnormal data points based on the local anomaly factor of each data point includes: if the local anomaly factor is greater than 1, extracting the corresponding local anomaly factor as an abnormal data point; if the local anomaly factor is less than or equal to 1, determining the corresponding local anomaly factor as a normal data point.

[0074] In this embodiment, the LOF algorithm is applied to each curve period to obtain the local anomaly factor of each data point in each curve period, and the abnormal data points are extracted.

[0075] In some embodiments of the present application, the monitoring of the current operating parameters of the transformer based on the abnormal data points of the fault cycle includes: obtaining the abnormal data points of the historical fault parameters of the transformer and the corresponding curve cycles, and establishing a training sample set based on the abnormal data points of the historical fault parameters of the transformer and the corresponding curve cycles; establishing an initial fault monitoring model based on the training sample set and training the initial fault monitoring model to obtain a trained fault monitoring model; obtaining the current operating parameters of the transformer and the corresponding curve cycles, inputting the current operating parameters of the transformer and the corresponding curve cycles into the trained fault monitoring model to determine whether a fault has occurred in the transformer.

[0076] In this embodiment, an initial fault monitoring model is established based on a deep learning neural network model. The initial fault monitoring model is trained using abnormal data points of the transformer's historical fault parameters and the corresponding curve cycles, so that the fault monitoring model can accurately identify abnormal data points in the curve cycle, thereby timely discovering fault information of the transformer's current operating parameters and improving the transformer's monitoring efficiency.

[0077] Based on the same technical concept, such as Figure 2 As shown, the present invention also provides a transformer fault monitoring system, including: an acquisition module for acquiring historical fault parameters of the transformer, and drawing a transformer fault parameter change curve based on the historical fault parameters of the transformer; a setting module for clustering the curve period according to the transformer fault parameter change curve, and determining the fluctuation tolerance threshold of the curve period according to the clustering result; a correction module for determining the fluctuation index of the transformer according to the fluctuation tolerance threshold, and setting the initial Value, according to the fluctuation index of the transformer The value is corrected; the monitoring module is used to The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle.

[0078] By applying the above technical solution, the present invention obtains the historical fault parameters of the transformer and draws the transformer fault parameter change curve according to the historical fault parameters of the transformer; clusters the curve period according to the transformer fault parameter change curve, and determines the fluctuation tolerance threshold of the curve period according to the clustering result; determines the volatility index of the transformer according to the fluctuation tolerance threshold, and sets the initial Value, according to the fluctuation index of the transformer The value is corrected; according to the corrected The value is based on the LOF algorithm to extract abnormal data points in the curve cycle and monitor the current operating parameters of the transformer based on the abnormal data points in the fault cycle. By monitoring the operating parameters of the transformer in real time, abnormal data points can be quickly located and operational faults can be discovered in a timely manner.

[0079] Through the above description of the embodiments, those skilled in the art will clearly understand that the present invention can be implemented via hardware or via software combined with a necessary general-purpose hardware platform. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product. This software product can be stored on a non-volatile storage medium (such as a CD-ROM, USB flash drive, or external hard drive) and includes instructions for enabling a computer device (such as a personal computer, server, or network device) to execute the methods described in various implementation scenarios of the present invention.

[0080] Finally, it should be noted that the above embodiments 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, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, 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.

Claims

1. A transformer fault monitoring method, characterized in that: The method comprises: Obtain historical fault parameters of the transformer, and draw a transformer fault parameter change curve based on the historical fault parameters of the transformer; Clustering the curve period according to the transformer fault parameter change curve, and determining the fluctuation tolerance threshold of the curve period according to the clustering result; Determine the transformer's fluctuation index based on the fluctuation tolerance threshold and set the initial Value, according to the fluctuation index of the transformer The value is corrected; According to the revised The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle; The clustering of curve periods according to the transformer fault parameter change curve includes: Obtaining a preset curve period, and dividing the transformer fault parameter change curve according to the preset curve period to obtain a plurality of first sub-fault parameter curves; Obtaining a vibration signal of the transformer within a curve period corresponding to the first sub-fault parameter curve, and drawing a vibration signal change curve according to the vibration signal of the transformer within the curve period corresponding to the first sub-fault parameter curve; Calculate the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve within the corresponding curve period, and cluster the curve period according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve; Clustering the curve periods according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve includes: A sample data set is established based on the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve, and k initial cluster centers of the sample data set are randomly selected; Calculate the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center, and divide the curve period corresponding to the correlation coefficient into the corresponding cluster cluster according to the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center; Calculate the average value of the correlation coefficient within each cluster, and recalculate the cluster center based on the average value of the correlation coefficient within each cluster; Repeat the above steps until the cluster center no longer changes or the number of iterations reaches the preset maximum number of iterations, and obtain the clustering result of the curve period; Determining the fluctuation tolerance threshold of the curve period according to the clustering result includes: Determine the cluster center of each curve period based on the clustering results of the curve period, and calculate the average value of all cluster centers; The absolute value of the difference between each cluster center and the average value is calculated based on the average value of all cluster centers, the absolute value of the difference between each cluster center and the average value is normalized, and the normalized absolute value of the difference is determined as the tolerance threshold coefficient; Obtaining a preset fluctuation tolerance threshold, and correcting the preset fluctuation tolerance threshold according to a tolerance threshold coefficient to obtain a fluctuation tolerance threshold; Determining the transformer's fluctuation index according to the fluctuation tolerance threshold includes: Obtaining a preset sliding time window, and segmenting the transformer fault parameter change curve according to the preset sliding time window to obtain a plurality of second sub-fault parameter curves; Calculating an average value of the second sub-fault parameter curve, and drawing an average value change curve according to the average value of the second sub-fault parameter curve; Obtain the absolute value of the slope of two adjacent average values ​​in the average value change curve, and calculate the difference between the absolute value of the slope and the fluctuation tolerance threshold; The average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is calculated, and the average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is determined as the volatility index.

2. The transformer fault monitoring method according to claim 1, characterized in that: The initial The values ​​are modified, including: according to The value correction formula is for the initial The value is corrected, The specific value correction formula is: ， in, For the corrected value, For the initial value, is the preset indicator tolerance threshold, is a volatility indicator, is the preset range coefficient, is the natural exponential function.

3. The transformer fault monitoring method according to claim 2, characterized in that: According to the revised The value is based on the LOF algorithm to extract abnormal data points of the fault cycle, including: The LOF algorithm is used for each curve period to obtain the local anomaly factor of each data point; Extract outlier data points based on the local outlier factor of each data point.

4. The transformer fault monitoring method according to claim 3, characterized in that: The extracting of abnormal data points according to the local abnormal factor of each data point includes: If the local anomaly factor is greater than 1, the corresponding local anomaly factor is extracted as the abnormal data point; If the local anomaly factor is less than or equal to 1, the corresponding local anomaly factor is determined to be a normal data point.

5. The transformer fault monitoring method according to claim 4, characterized in that: The monitoring of the current operating parameters of the transformer according to the abnormal data points of the fault cycle includes: Obtain abnormal data points and corresponding curve periods of historical fault parameters of the transformer, and establish a training sample set based on the abnormal data points and corresponding curve periods of historical fault parameters of the transformer; Establishing an initial fault monitoring model based on the training sample set and training the initial fault monitoring model to obtain a trained fault monitoring model; Obtain the current operating parameters of the transformer and the corresponding curve period, input the current operating parameters of the transformer and the corresponding curve period into the trained fault monitoring model to determine whether the transformer has a fault.

6. A transformer fault monitoring system, characterized in that: include: An acquisition module is used to acquire historical fault parameters of the transformer and draw a transformer fault parameter change curve based on the historical fault parameters of the transformer; A setting module is used to cluster the curve period according to the transformer fault parameter change curve, and determine the fluctuation tolerance threshold of the curve period according to the clustering result; Correction module, used to determine the fluctuation index of transformer according to the fluctuation tolerance threshold, set the initial Value, according to the fluctuation index of the transformer The value is corrected; Monitoring module, used to The abnormal data points of the curve cycle are extracted based on the LOF algorithm, and the current operating parameters of the transformer are monitored according to the abnormal data points of the fault cycle; The setting module clusters curve periods according to the transformer fault parameter change curve, including: Obtaining a preset curve period, and dividing the transformer fault parameter change curve according to the preset curve period to obtain a plurality of first sub-fault parameter curves; Obtaining a vibration signal of the transformer within a curve period corresponding to the first sub-fault parameter curve, and drawing a vibration signal change curve according to the vibration signal of the transformer within the curve period corresponding to the first sub-fault parameter curve; Calculate the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve within the corresponding curve period, and cluster the curve period according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve; Clustering the curve periods according to the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve includes: A sample data set is established based on the correlation coefficient between the first sub-fault parameter curve and the vibration signal change curve, and k initial cluster centers of the sample data set are randomly selected; Calculate the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center, and divide the curve period corresponding to the correlation coefficient into the corresponding cluster cluster according to the Euclidean distance from the correlation coefficient in the sample data set to the initial cluster center; Calculate the average value of the correlation coefficient within each cluster, and recalculate the cluster center based on the average value of the correlation coefficient within each cluster; Repeat the above steps until the cluster center no longer changes or the number of iterations reaches the preset maximum number of iterations, and obtain the clustering result of the curve period; The setting module determines the fluctuation tolerance threshold of the curve period according to the clustering result, including: Determine the cluster center of each curve period based on the clustering results of the curve period, and calculate the average value of all cluster centers; The absolute value of the difference between each cluster center and the average value is calculated based on the average value of all cluster centers, the absolute value of the difference between each cluster center and the average value is normalized, and the normalized absolute value of the difference is determined as the tolerance threshold coefficient; Obtaining a preset fluctuation tolerance threshold, and correcting the preset fluctuation tolerance threshold according to a tolerance threshold coefficient to obtain a fluctuation tolerance threshold; The correction module determines the transformer's fluctuation index according to the fluctuation tolerance threshold, including: Obtaining a preset sliding time window, and segmenting the transformer fault parameter change curve according to the preset sliding time window to obtain a plurality of second sub-fault parameter curves; Calculating an average value of the second sub-fault parameter curve, and drawing an average value change curve according to the average value of the second sub-fault parameter curve; Obtain the absolute value of the slope of two adjacent average values ​​in the average value change curve, and calculate the difference between the absolute value of the slope and the fluctuation tolerance threshold; The average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is calculated, and the average value of the difference between the absolute values ​​of all slopes and the volatility tolerance threshold is determined as the volatility index.

Images