Abnormality detection method, device and equipment of isolation switch cabinet and storage medium

Abstract

The embodiment of the invention discloses an anomaly detection method, device and equipment for an isolation switch cabinet and a storage medium, and the method comprises the steps: sequentially constructing triple data according to the operation data of the isolation switch cabinet, and enabling the triple data to comprise initial attribute data, target attribute data and a distance value, the target attribute data is preset attribute data directly related to the initial attribute data, the distance value is a distance value between the target attribute data and the first initial attribute data, and the first initial attribute data is current attribute data; the target attribute data of the previously constructed triple data is used as the initial attribute data of the adjacent subsequent triple data; performing format conversion on the triple data to obtain target triple data, wherein the target triple data comprises a correlation value; performing anomaly detection processing on the target triple data through a preset anomaly detection algorithm to obtain a corresponding anomaly detection result; the problem of low working efficiency and accuracy of anomaly detection is solved, and the working efficiency and accuracy of anomaly detection are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence technology, and in particular to an abnormality detection method, device, equipment and storage medium for an isolating switch cabinet. Background Art

[0002] The isolating switch cabinet is an important electrical equipment in the rail transit system. It has the functions of isolating power supply, switching lines and switching operations to ensure the normal operation and safe maintenance of the line. It is a switchgear without arc extinguishing function. The isolating switch in the isolating switch cabinet is closed and opened by a motor-driven knife switch. Due to the jitter or instability of electrical quantities such as motor current and voltage, the isolating switch may have poor contact or be out of place when opening and closing, making the current-carrying cross-sectional area small when the isolating switch is closed, thereby causing safety hazards of overcurrent and overheating. In severe cases, the isolating switch will be burned, causing train operation accidents. Therefore, it is particularly important to determine whether the isolating switch cabinet is operating abnormally.

[0003] Existing abnormality detection for isolating switch cabinets is usually carried out regularly through electrical testing. This method can quantitatively evaluate the electrical performance of the isolating switch and discover potential electrical faults. However, the detection process of this method is relatively complicated and time-consuming, and the detection usually depends on the work experience of the inspectors. Therefore, the detection work efficiency and detection accuracy corresponding to this electrical testing method are relatively low. Summary of the invention

[0004] The embodiments of the present application provide an abnormality detection method, device, equipment and storage medium for an isolating switch cabinet, which can solve the technical problems of low abnormality detection efficiency and accuracy of the isolating switch cabinet, and improve the abnormality detection efficiency and accuracy of the isolating switch cabinet.

[0005] In a first aspect, an embodiment of the present application provides an abnormality detection method for an isolation switch cabinet, comprising:

[0006] Obtaining operation data of the isolation switch cabinet, the operation data including current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, the switch attribute data including opening response time, closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal;

[0007] Constructing triplet data in sequence according to the operation data, the triplet data includes starting attribute data, target attribute data and distance value, the target attribute data is preset attribute data directly related to the starting attribute data, the distance value is the distance value between the target attribute data and the first starting attribute data, the first starting attribute data is the current attribute data, and the target attribute data of the triplet data constructed previously is used as the starting attribute data of the adjacent subsequent triplet data;

[0008] Performing format conversion on the triple data to obtain target triple data, the target triple data includes starting attribute data, target attribute data and correlation value, and the correlation value is calculated according to the distance value of the corresponding triple data;

[0009] The target triplet data is processed for anomaly detection using a preset anomaly detection algorithm to obtain the corresponding anomaly detection result.

[0010] In one embodiment, triple data is sequentially constructed according to the operation data, including:

[0011] Constructing a first triplet of data according to the operation data, the first triplet of data including first starting attribute data, first target attribute data and a first distance value, the first target attribute data being preset attribute data directly related to the current attribute data, and the first distance value being one;

[0012] Constructing a second triplet of data according to the operation data, the second triplet of data including second starting attribute data, second target attribute data, and a second distance value, wherein the second starting attribute data is one of the first target attribute data, the second target attribute data is preset attribute data directly related to the second starting attribute data, and the second distance value is the first distance value plus one;

[0013] The remaining triple data are sequentially constructed according to the operation data, the remaining triple data include starting attribute data, target attribute data and distance value, the starting attribute data is one of the target attribute data of the adjacent previously constructed triple data, the target attribute data is the preset attribute data directly related to the starting attribute data in the corresponding triple data, the distance value is the distance value of the adjacent previously constructed triple data with different starting attribute data plus one, and the distance values ​​of the triple data with the same starting attribute data are the same;

[0014] The triple data is formatted to obtain target triple data, which includes starting attribute data, target attribute data and correlation values, including:

[0015] The first triplet data, the second triplet data and the remaining triplet data are format-converted based on corresponding distance values ​​to obtain target triplet data, which includes starting attribute data, target attribute data and correlation values.

[0016] In one embodiment, obtaining operation data of the isolation switch cabinet includes:

[0017] Acquire current signals and voltage signals, transmit the current signals and voltage signals to a wave recording device for digital processing, and obtain current attribute data, voltage attribute data, resistance attribute data, and power attribute data;

[0018] The switch quantity signal is obtained, and the switch quantity signal is transmitted to the recording device for digital processing to obtain the switch quantity attribute data.

[0019] In one embodiment, format conversion is performed on triple data to obtain target triple data, including:

[0020] According to the distance values ​​in the triple data, determine the longest distance value in all triple data;

[0021] Obtain a first quotient value according to the quotient of the distance value of the current triple data and the longest distance value;

[0022] Obtaining a correlation value corresponding to the current triple data according to a difference between the first and the first quotient values;

[0023] The correlation value replaces the distance value in the current triple data to obtain the corresponding target triple data.

[0024] In one embodiment, the target triplet data is subjected to anomaly detection processing by a preset anomaly detection algorithm to obtain a corresponding anomaly detection result, including:

[0025] Sort the target triplet data from high to low according to the correlation value, and divide the target triplet data with the same correlation value into a group;

[0026] For each group of target triplet data, the target attribute data in the corresponding target triplet data is extracted, and the kurtosis coefficient of the corresponding group is calculated according to the target attribute data;

[0027] The group with the same kurtosis coefficient as the preset threshold is determined as the target group;

[0028] The target triplet data of the target group is processed for anomaly detection using a preset anomaly detection algorithm to obtain the corresponding anomaly detection result.

[0029] In one embodiment, an abnormality detection process is performed on the target triplet data of the target group by a preset abnormality detection algorithm to obtain a corresponding abnormality detection result, including:

[0030] For the target triplet data in the target group, extract the starting attribute data in the corresponding target triplet data;

[0031] Divide the initial attribute data of the same attribute into a sub-data set to obtain a plurality of sub-data sets;

[0032] Select one of the sub-datasets and apply spatial mapping to cluster similar data items to obtain a clustered data set;

[0033] Perform clustering processing on each clustered data set to obtain the corresponding center point and the distance value between each data point in the clustered data set and the center point;

[0034] Delete the data points whose distance value is less than or equal to the maximum distance value and replace them with the center point;

[0035] The central point is processed for anomaly detection using a preset anomaly detection algorithm to obtain the corresponding anomaly detection result.

[0036] In one embodiment, the anomaly detection algorithm is an isolation forest algorithm;

[0037] The central point is processed for anomaly detection through the preset anomaly detection algorithm to obtain the corresponding anomaly detection results, including:

[0038] Create a root node based on the aggregated data set;

[0039] A split value is randomly determined based on the maximum and minimum values ​​of the center points of all clustered data sets;

[0040] Place the center points of the corresponding aggregated data set that are less than the split value in the left subtree of the corresponding root node, and place the center points of the corresponding aggregated data set that are greater than or equal to the split value in the right subtree of the corresponding root node, until the leaf node contains only one node;

[0041] Build an isolated tree based on the corresponding root node and all leaf nodes;

[0042] Re-randomly determine a split value and construct a corresponding isolated tree until the number of corresponding isolated trees reaches a preset threshold;

[0043] Integrate all isolated trees to build an isolation forest;

[0044] The isolation forest is processed for anomaly detection using the preset isolation forest algorithm to obtain the corresponding anomaly judgment result.

[0045] In a second aspect, an embodiment of the present application provides an abnormality detection device for an isolation switch cabinet, comprising:

[0046] An operation data acquisition module is used to acquire the operation data of the isolation switch cabinet, the operation data includes current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, the switch attribute data includes opening response time, closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal;

[0047] A first data construction module is used to sequentially construct triplet data according to the operation data, wherein the triplet data includes starting attribute data, target attribute data and a distance value, wherein the target attribute data is preset attribute data directly related to the starting attribute data, and the distance value is a distance value between the target attribute data and the first starting attribute data, wherein the first starting attribute data is the current attribute data, and the target attribute data of the triplet data constructed previously is used as the starting attribute data of the adjacent subsequent triplet data;

[0048] A format conversion module is used to convert the format of the triple data to obtain target triple data, where the target triple data includes starting attribute data, target attribute data and a correlation value, where the correlation value is calculated based on the distance value of the corresponding triple data;

[0049] The anomaly detection module is used to perform anomaly detection processing on the target triplet data through a preset anomaly detection algorithm to obtain the corresponding anomaly detection result.

[0050] In a third aspect, an embodiment of the present application provides an abnormality detection device for an isolation switch cabinet, including:

[0051] memory and one or more processors;

[0052] The memory is used to store one or more programs;

[0053] When the one or more programs are executed by the one or more processors, the one or more processors implement the abnormality detection method for the isolation switch cabinet as described in the first aspect.

[0054] In a fourth aspect, an embodiment of the present application provides a storage medium storing computer executable instructions, which, when executed by a computer processor, are used to execute the abnormality detection method for the isolation switch cabinet as described in the first aspect.

[0055] The embodiment of the present application obtains the operation data of the isolation switch cabinet when performing abnormality detection on the operation of the isolation switch cabinet, sequentially constructs triple data according to the operation data, and converts the triple data into target triple data related to the correlation value, and performs abnormality detection processing on the target triple data through a preset abnormality detection algorithm to obtain a corresponding abnormality detection result. The above-mentioned technical means can be used to automatically detect faults on the operation data of the isolation switch cabinet, thereby avoiding the technical problems of low work efficiency and accuracy caused by manual detection. The target triple data corresponding to the operation data is processed through a preset abnormality detection algorithm to obtain an abnormality detection result, thereby realizing automated abnormality detection and improving the work efficiency of abnormality detection of the isolation switch cabinet; and abnormality detection is performed through a preset abnormality detection algorithm, without relying on the work experience of personnel, and abnormality detection is performed through a machine learning detection algorithm to improve the accuracy of abnormality detection of the isolation switch cabinet.

[0056] The beneficial effects of the abnormality detection device of the isolation switch cabinet, the abnormality detection equipment of the isolation switch cabinet, and the storage medium provided above can refer to the beneficial effects of the abnormality detection method of the isolation switch cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a flow chart of an abnormality detection method for an isolation switch cabinet provided in an embodiment of the present application;

[0058] Figure 2 It is a schematic diagram of the electrical principle of an isolating switch cabinet provided in an embodiment of the present application;

[0059] Figure 3 It is a structural schematic diagram of an abnormality detection device for an isolation switch cabinet provided in an embodiment of the present application;

[0060] Figure 4 It is a structural schematic diagram of an abnormality detection device for an isolation switch cabinet provided in an embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical scheme and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for the convenience of description, only the part related to the present application but not all the contents are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow chart describes each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to a method, a function, a procedure, a subroutine, a subprogram, etc.

[0062] Existing abnormality detection for isolating switch cabinets is usually carried out regularly through electrical testing. This method can quantitatively evaluate the electrical performance of the isolating switch and discover potential electrical faults. However, the detection process of this method is relatively complicated and time-consuming, and the detection usually depends on the work experience of the inspectors. Therefore, the detection work efficiency and detection accuracy corresponding to this electrical testing method are relatively low.

[0063] Based on this, an abnormality detection method, device, equipment and storage medium of an isolating switch cabinet of an embodiment of the present application are provided, which is intended to obtain the operation data of the isolating switch cabinet, sequentially construct triple data according to the operation data, and convert the triple data into target triple data related to the correlation value when performing abnormality detection on the operation of the isolating switch cabinet, and perform abnormality detection processing on the target triple data through a preset abnormality detection algorithm to obtain a corresponding abnormality detection result. By adopting the above technical means, automatic fault detection can be performed on the operation data of the isolating switch cabinet, thereby avoiding the technical problems of low work efficiency and accuracy caused by manual detection. The target triple data corresponding to the operation data is subjected to abnormality detection processing through a preset abnormality detection algorithm to obtain an abnormality detection result. Compared with the existing manual detection method, this embodiment can realize automatic abnormality detection and improve the work efficiency of abnormality detection of the isolating switch cabinet; and abnormality detection is performed through a preset abnormality detection algorithm. Compared with the existing manual detection method, this embodiment does not need to rely on the work experience of personnel, and performs abnormality detection through a machine learning detection algorithm to improve the accuracy of abnormality detection of the isolating switch cabinet.

[0064] Figure 1A flowchart of an abnormality detection method for an isolating switch cabinet provided in an embodiment of the present application is given. The abnormality detection method for an isolating switch cabinet provided in this embodiment can be executed by an abnormality detection device of the isolating switch cabinet, and the abnormality detection device of the isolating switch cabinet can be implemented by software and / or hardware. The abnormality detection device of the isolating switch cabinet can be composed of two or more physical entities, or can be composed of one physical entity. Generally speaking, the abnormality detection device of the isolating switch cabinet can be a computer device or a cloud server.

[0065] The following description is made by taking a computer device as an example to implement the abnormality detection method of the isolation switch cabinet. Figure 1 , the abnormality detection method of the isolation switch cabinet specifically includes:

[0066] S101. Obtaining the operating data of the isolating switch cabinet, the operating data including current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, the switch attribute data including opening response time, closing response time, the time point of the relay switch signal, the time point of the closing switch signal and the time point of the opening switch signal.

[0067] Real-time operation data, such as current signal, voltage signal, temperature, switch status and other operation data, are obtained through sensors and / or detection equipment in the isolation switch cabinet. Subsequent abnormality detection and analysis are performed based on the real-time operation data. Current signals and voltage signals can be obtained through corresponding sensors in the isolation switch cabinet, and the current signals and current signals are transmitted to the recording device for digital processing to obtain current attribute data, voltage attribute data, resistance attribute data and power attribute data. Among them, the resistance attribute data can be obtained by the ratio of the current attribute data and the corresponding voltage attribute data, and the power attribute data can be obtained by the product of the current attribute data and the corresponding voltage attribute data. The switch signal and the switch quantity signal are obtained, and the switch signal and the switch quantity signal are transmitted to the recording device for digital processing to obtain the switch quantity attribute data, wherein the switch quantity attribute data includes the opening response time, the closing response time, the time point of the relay switch quantity signal, the time point of the closing switch quantity signal and the time point of the opening switch quantity signal, etc.

[0068] Figure 2 This is a schematic diagram of the electrical principle of an isolating switch cabinet provided in an embodiment of the present application, referring to Figure 2 The isolating switch cabinet includes a closing control contactor K1E, an opening control contactor K1A, a motor M, a closing travel switch S1E, an opening travel switch S1A and an intermediate relay J1. The closing control contactor K1E is used to control the closing main circuit. Figure 2When all the closing control contactors K1E are closed, the closing main circuit is in the on state. The opening control contactor K1A is used to control the opening main circuit. The motor M is used to drive the closing. The closing travel switch S1E is used to disconnect the 1-2 contacts after the closing is in place, so that the closing circuit is de-energized. The opening travel switch S1A is used to change the 1-2 contacts from closed to open after the opening work is completed, so that the opening circuit is de-energized.

[0069] The current signal is obtained through the sensor, and the current signal includes the motor current signal of the isolation switch cabinet (i.e., the current signal at K1E-6) and the control loop current signal (i.e., the current signal at K1E-A2). The current signal at K1E-6 is measured by a sensor (such as a Hall sensor), that is, the motor current is monitored. The current signal at K1E-A2 is measured by a sensor (such as a Hall sensor), that is, the control loop current is monitored.

[0070] The sensor for detecting the current signal may be a Hall sensor. The Hall sensor with a small wire diameter and high precision can be used to monitor the current of the closing coil when the motor in the (electric) disconnecting switch cabinet is started.

[0071] The voltage signal is detected by the sensor, wherein the voltage signal includes the voltage signal of the motor in the isolation switch cabinet (ie, the voltage signal at X1-1) and the voltage signal of the control power supply circuit. The voltage signal of the motor M is monitored by measuring the voltage signal at X1-1.

[0072] Among them, the voltage signal of the control power supply circuit includes the external locking voltage signal (i.e., the voltage signal at X1-15), the SCADA signal, the closing control contactor input voltage signal, and the opening control contactor input voltage signal. The voltage signal at the closing and opening locking point is obtained by measuring the voltage signal at X1-15. The SCADA signal includes the voltage signal of remote closing and the voltage signal of remote opening. The voltage signal of remote closing can be obtained by measuring the voltage at X1-8, and the voltage signal of remote opening can be obtained by measuring the voltage at X1-6. The input voltage signal of the closing controller contactor is obtained by measuring the voltage signal at K1E-A1, and the input voltage signal of the opening control contactor is obtained by measuring the voltage signal at K1A-A1.

[0073] When acquiring switch signal data, the disconnector opening and closing auxiliary contacts can be connected at the disconnector of the disconnector cabinet, and the switch signal data can be acquired through the disconnector opening and closing auxiliary contacts. The switch signal data includes the switch opening response time and the switch closing response time.

[0074] The switch quantity signal includes the first switch quantity signal (i.e. switch quantity I), the second switch quantity signal (i.e. switch quantity II) and the third switch quantity signal (i.e. switch quantity III). The voltage signal can be obtained through the auxiliary contact on the intermediate relay J1 to obtain the corresponding first switch quantity signal. The second switch quantity signal can be obtained through the voltage signal obtained through the auxiliary contact on the closing travel switch and the closing interlocking condition switch (i.e. K1A and S1E). The third switch quantity signal can be obtained through the voltage signal obtained on the auxiliary contact on the opening travel switch and the closing interlocking condition switch (i.e. K1E and S1A).

[0075] The detected current signal and voltage signal are transmitted to the analog input point of the recording device, so as to digitally process the current signal and voltage signal in the recording device to obtain the corresponding current attribute data, voltage attribute data, resistance attribute data and power attribute data. M =I K1E-6 *U X1-1 Calculate the corresponding power attribute data, where P M Represents power attribute data, I K1E-6 Represents current attribute data, U X1-1 Represents voltage attribute data.

[0076] The detected switch signal and switch quantity signal are transmitted to the switch quantity input point of the wave recording device, so as to digitally process the switch quantity signal in the wave recording device to obtain the corresponding switch quantity attribute data, wherein the switch quantity attribute data includes the opening response time, the closing response time, the time point of the relay switch quantity signal, the time point of the closing switch quantity signal and the time point of the opening switch quantity signal. Among them, the opening response time is the time required from the actual remote or local opening command to the actual wave recording device receiving the signal of the switch opening and closing auxiliary contact feedback back that the switch is in place; the closing response time is the time required from the actual remote or local closing command to the actual wave recording device receiving the signal of the switch opening and closing auxiliary contact feedback back that the switch is in place.

[0077] As mentioned above, after obtaining the operating data of the isolation switch cabinet, the operating data can be transmitted to the cloud server for subsequent abnormality detection and processing on the cloud server, or it can be performed on the local corresponding computer device (such as a host computer, etc.) for subsequent abnormality detection and processing.

[0078] S102. Construct triplet data in sequence according to the operating data, the triplet data including starting attribute data, target attribute data and distance value, the target attribute data is preset attribute data directly related to the starting attribute data, the distance value is the distance value between the target attribute data and the first starting attribute data, the first starting attribute data is the current attribute data, and the target attribute data of the triplet data constructed previously is used as the starting attribute data of the adjacent subsequent triplet data.

[0079] Exemplarily, a first triplet of data may be constructed based on the operation data, the first triplet of data including first starting attribute data, first target attribute data and first distance value, the first starting attribute data is current attribute data, the first target attribute data is preset attribute data directly related to the current attribute data, and the first distance value is 1. Afterwards, a second triplet of data may be constructed based on the operation data, the second triplet of data including second starting attribute data, second target attribute data and second distance value, the second starting attribute data is one of the attribute data in the first target attribute data, the second target attribute data is preset attribute data directly related to the second starting attribute data, and the second distance value is the first distance value plus 1. Afterwards, the remaining triple data are constructed in sequence according to the running data. The remaining triple data include starting attribute data, target attribute data and distance values. The starting attribute data is a type of attribute data in the target attribute data of the adjacent previously constructed triple data. The target attribute data is preset attribute data that is directly related to the starting attribute data in the corresponding triple data. The distance value is the distance value of the adjacent previously constructed triple data with different starting attribute data plus one. The distance values ​​of the triple data of the same starting attribute data are the same.

[0080] Exemplarily, from all the attributes of all the running data, current attribute data is extracted as the first starting attribute data. The preset attribute data directly related to the current attribute data is determined as the first target attribute data. For example, voltage attribute data, resistance attribute data, power attribute data, etc. are determined as attribute data directly related to the current attribute data, that is, the first target attribute data. The first distance value between the first target attribute data directly related to the current attribute data and the current attribute data is defined as one, that is, the distance value between the voltage attribute data, the resistance attribute data, the power attribute data and the current attribute data is one. Exemplarily, the format of the first triplet data can be set to: <first starting attribute data, first target attribute data, first distance value>. Therefore, the first triplet data that can be constructed can be <current attribute data, voltage attribute data, 1>, <current attribute data, resistance attribute data, 1> and <current attribute data, power attribute data, 1>. Afterwards, from the first target attribute data, a target attribute data is selected as the second starting attribute data. For example, the voltage attribute data is first selected as the second starting attribute data, and the preset attribute data directly related to the voltage attribute data is used as the second target attribute data. For example, the (second) target attribute data corresponding to the voltage attribute data is resistance attribute data, power attribute data, and switch attribute data, and the corresponding second distance value is the first distance value plus one; the corresponding second triplet data is: <voltage attribute data, resistance attribute data, 2> and <voltage attribute data, power attribute data, 2> and <voltage attribute data, switch attribute data, 2>.

[0081] Similarly, when the resistance attribute data is used as the starting attribute data, the corresponding second triplet data is: <resistance attribute data, voltage attribute data, 2>, <resistance attribute data, power attribute data, 2>, and <resistance attribute data, switch attribute data, 2>. This can be deduced by analogy until all the first target attribute data are used as the second starting attribute data to construct the corresponding second triplet data.

[0082] Afterwards, one is selected from the second target attribute data as the third starting attribute data, and the preset attribute data directly related to the third starting attribute data is used as the third target attribute data, and the third distance value is the second distance value plus one, that is, the third distance value is 3. Similarly, the remaining triple data are sequentially constructed according to the operation data, and the remaining triple data include starting attribute data, target attribute data and distance values, the starting attribute data is one attribute data in the target attribute data of the adjacent triple data constructed previously, the target attribute data is the preset attribute data directly related to the starting attribute data in the corresponding triple data, and the distance value is the distance value of the adjacent triple data of different starting attribute data constructed previously plus one.

[0083] It should be noted that the distance values ​​of the triplet data of the same starting attribute data are the same. For example, the distance values ​​of the first triplet data: <current attribute data, voltage attribute data, 1>, <current attribute data, resistance attribute data, 1> and <current attribute data, power attribute data, 1> are all 1.

[0084] As mentioned above, by processing the operating data of the isolation cabinet switch, triplet data is constructed, and the operating data is stored through structured triplet data for easy management; and based on the triplet data in a unified format, the subsequent data format conversion or data processing can improve the efficiency of data processing, thereby improving the work efficiency of anomaly detection.

[0085] S103, converting the format of the triple data to obtain target triple data, where the target triple data includes starting attribute data, target attribute data and a correlation value, and the correlation value is calculated based on the distance value of the corresponding triple data.

[0086] To facilitate the subsequent data processing of the anomaly detection algorithm, the triple data obtained above can be format converted to convert the distance value in the triple data into a correlation value. The longest distance value in all triple data can be determined based on the distance value in the triple data. Usually, the distance value of the triple data created last should be the longest distance value. According to the quotient of the distance value of the current triple data and the longest distance value, a first quotient value is obtained. According to the difference between the first quotient value and the first quotient value, the correlation value corresponding to the current triple data is obtained. The correlation value replaces the distance value in the current triple data to obtain the corresponding target triple data. Exemplarily, the first triple data, the second triple data and the remaining triple data can be format converted based on the corresponding distance values ​​in turn to obtain the target triple data, and the target triple data includes starting attribute data, target attribute data and correlation value.

[0087] For example, according to The correlation value is calculated, where d max =max(d(p 0 , p t1 ),d(p 0 , p t2 )...d(p 0 , p tn )), γ t represents the correlation value between the starting attribute data and the target attribute data, p 0 Represents the starting attribute data, p t represents the target attribute data, d represents the distance value from the starting attribute data to the target attribute data, and p ti represents the i-th target attribute data, d maxRepresents the maximum distance value among all triple data.

[0088] For example, assuming that the longest distance is 10, for the triple data: <current attribute data, voltage attribute data, 1>, the first quotient value can be 1 / 10, and the corresponding correlation value can be 1-(1 / 10)=90%, so the corresponding target triple data can be: <current attribute data, voltage attribute data, 90%>. By analogy, through the aforementioned format conversion, the target triple data corresponding to all triple data can be obtained.

[0089] S104: Perform anomaly detection processing on the target triplet data using a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

[0090] The preset anomaly detection algorithm is used to perform feature analysis on all target triplet data to obtain corresponding feature values, and anomaly detection processing is performed based on the feature values ​​to obtain the anomaly detection results corresponding to each running data.

[0091] Sort the target triple data from high to low according to the correlation value, and divide the target triple data with the same correlation value into a group. For each group of target triple data, extract the target attribute data in the corresponding target triple data, and calculate the kurtosis coefficient of the corresponding group based on the target attribute data. The group with the same kurtosis coefficient as the preset threshold is determined as the target group. Perform anomaly detection processing on the target triple data of the target group through the preset anomaly detection algorithm to obtain the corresponding anomaly detection result. It should be noted that the preset threshold is a range threshold.

[0092] The kurtosis coefficient is a statistic that describes the distribution of data. It reflects the sharpness of the data distribution, especially the sharpness of the peak of the data. The kurtosis coefficient can help identify whether the data is sharper or flatter than the normal distribution. In the data preprocessing or feature selection stage, the kurtosis coefficient is used to decide whether to use a certain attribute (feature) as the input of the algorithm. For each attribute, its kurtosis coefficient is calculated; based on business needs, the expected range (threshold) of the kurtosis coefficient is defined. For example, you may want to select attributes with a kurtosis coefficient within a certain range, because too high or too low kurtosis may indicate that the data is abnormal or unsuitable for subsequent algorithms.

[0093] For example, a feature extraction method is performed on each target triplet data to select features that conform to the Gaussian distribution. When selecting features, the random selection is abandoned and attribute screening is performed based on a certain method. For example, firstly, the target triplet data R{r 1 ,r 2 ,r 3 ...r n}, sorted from high to low according to the correlation value. Each time, the target triple data with the highest correlation is selected, its target attribute data is extracted, the kurtosis coefficient of the target triple data under the target attribute data is calculated, and the preset threshold is used to determine whether to use the target triple data as the input of the anomaly detection algorithm. If it meets expectations (that is, the kurtosis coefficient is the same as the preset threshold), the target attribute data is added to the selected attribute data; if it does not meet expectations (that is, the kurtosis coefficient is different from the preset threshold), the target attribute data is skipped. Until the target triple data with the same kurtosis coefficient as the preset threshold is screened out, the corresponding group of target triple data is determined as the target group.

[0094] The kurtosis coefficient is based on the normal distribution and represents the characteristic coefficient of the peak value of the probability density distribution curve at the mean. The calculation formula of the kurtosis coefficient is as follows:

[0095]

[0096] Among them, Kurt (R) represents the kurtosis coefficient. Assume that there is a set of samples R{r 1 ,r 2 ,r 3 ...r n}, each element in the sample is a sample value, and r in the above formula i is the i-th sample value, represents the mean value of the sample, and n is the number of samples.

[0097] After the target group is determined according to the kurtosis coefficient, the starting attribute data in the corresponding target triple data is extracted for the target triple data in the target group. The starting attribute data of the same attribute are divided into a sub-data set to obtain a plurality of sub-data sets. One of the sub-data sets is selected, and the spatial mapping is applied to cluster similar data items to obtain a clustered data set. For example, when the spatial mapping is applied to cluster similar data items, if d(x 1 ,x 2 )≤d 1 , then h(x 1 )=h(x 2 ) has a probability of at least p 1 ; if d(x 1 ,x 2 )≥d 2 , then h(x 1 )=h(x 2 ) has a probability of at most p 2 Among them, x 1 , x 2 are two points in the original data space, d(x 1 ,x 2 ) is the distance or similarity between two points, d1 d 2 、p 1 and p 2 It is a preset value and can be set according to actual conditions. 1 )=h(x 2 ) means when x 1 , x 2 When they are "close" in some sense (i.e., their distance is less than or equal to d 1 ), the output value of the function h(x) for these two inputs is also the same. Therefore, the corresponding function (d 1 d 2 、p 1 、p 2 0. Combine multiple such functions together and map the corresponding raw data into a list to obtain an aggregated data set.

[0098] Perform clustering on each clustered data set to obtain the corresponding center point and the distance value between each data point in the clustered data set and the center point. The data in the clustered data set can be clustered using the K-means algorithm according to k=2 to determine the center point of each cluster, and the center of each cluster is determined as the center point of the corresponding clustered data set. Select the point closest to the center point of the cluster, compare the distance between the point and the center point of the cluster, select the maximum distance value as the threshold, delete the data points in the corresponding clustered data set whose distance value is less than or equal to the maximum distance value, and replace them with the corresponding center point. Afterwards, perform anomaly detection on all center points using the preset anomaly detection algorithm to obtain the corresponding anomaly detection results.

[0099] Exemplarily, the preset anomaly detection algorithms include Isolation Forest test algorithm, SVM (support vector machine) test algorithm, Grubbs test algorithm, Z-score test algorithm, Tukey (post hoc multiple comparisons) test algorithm, Cook's distance test, Mahalanobis distance test algorithm and / or LOF (local outlier factor) test algorithm, etc.

[0100] This embodiment is described by taking the preset anomaly detection algorithm as the isolation forest algorithm as an example. After the center point of the cluster corresponding to each clustered data set is obtained based on clustering, a root node is created according to the clustered data set. A split value is randomly determined based on the maximum and minimum values ​​of the center points of all clustered data sets. The center point of the corresponding clustered data set whose center point is less than the split value is placed in the left subtree of the corresponding root node, and the center point of the corresponding clustered data set whose center point is greater than or equal to the split value is placed in the right subtree of the corresponding root node until the leaf node contains only one node. An isolated tree is constructed based on the corresponding root node and all leaf nodes. A split value is randomly determined based on the maximum and minimum values ​​of the center points of all clustered data sets and a new isolated tree is constructed until the number of corresponding isolated trees reaches a preset threshold. All isolated trees are integrated to construct an isolated forest. The isolated forest is subjected to anomaly detection processing by the preset isolation forest algorithm to obtain the corresponding anomaly judgment result.

[0101] According to the abnormal judgment results, data analysis is performed to obtain the operating health status of the isolation switch cabinet, and the abnormal judgment results and the operating health status of the isolation switch cabinet are visualized. Among them, the health status includes: power detection results, isolation knife closing indication, key state quantity detection results in the cabinet, signal quality analysis of external control and locking signals, and detection results of the characteristics of the opening and closing relays. The above health status is obtained through real-time analysis of the operating data of the isolation switch cabinet, and the health status monitoring of the isolation switch cabinet is realized.

[0102] When abnormal results are detected, an alarm is issued and / or relevant measures are executed remotely, and an abnormal log is generated. The abnormal log includes an abnormal briefing and an abnormal recording file. In addition, the monitoring data and the corresponding health status are recorded for easy tracing.

[0103] As mentioned above, by acquiring the operating data such as the current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data of the isolating switch cabinet, real-time monitoring and intelligent analysis are carried out to realize the monitoring of motor power, the analysis and judgment of whether the isolating knife switch is in place, the monitoring of key state quantities in the cabinet, the analysis of the signal quality of the externally introduced control quantity and the locking signal, the detection of the characteristics of the opening and closing relays and the analysis of the power quality, etc., thereby realizing the monitoring of the health status of the isolating switch cabinet, which is of great significance to the normal operation of the system using the isolating switch cabinet (such as the subway power supply system).

[0104] In the above, when performing abnormality detection on the operation of the isolation switch cabinet, the operation data of the isolation switch cabinet is obtained, triple data is constructed in sequence according to the operation data, and the triple data is converted into target triple data related to the correlation value, and the target triple data is processed for abnormality detection through a preset abnormality detection algorithm to obtain the corresponding abnormality detection result. The above technical means can be used to automatically detect faults on the operation data of the isolation switch cabinet, thereby avoiding the technical problems of low work efficiency and accuracy caused by manual detection. The target triple data corresponding to the operation data is processed for abnormality detection through a preset abnormality detection algorithm to obtain the abnormality detection result, realize automatic abnormality detection, and improve the work efficiency of abnormality detection of the isolation switch cabinet; and abnormality detection through a preset abnormality detection algorithm does not need to rely on the work experience of personnel, and abnormality detection through a machine learning detection algorithm improves the accuracy of abnormality detection of the isolation switch cabinet.

[0105] Based on the above embodiments, Figure 3 This is a schematic diagram of the structure of an abnormality detection device for an isolation switch cabinet provided in an embodiment of the present application. Figure 3 The abnormality detection device for the isolation switch cabinet provided in this embodiment specifically includes: an operation data acquisition module 21, a first data construction module 22, a format conversion module 23 and an abnormality detection module 24.

[0106] Among them, the operation data acquisition module 21 is used to obtain the operation data of the isolation switch cabinet, the operation data includes current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, and the switch attribute data includes the opening response time, the closing response time, the time point of the relay switch signal, the time point of the closing switch signal and the time point of the opening switch signal;

[0107] A first data construction module 22 is used to sequentially construct triplet data according to the operation data, wherein the triplet data includes starting attribute data, target attribute data and a distance value, wherein the target attribute data is preset attribute data directly related to the starting attribute data, and the distance value is a distance value between the target attribute data and the first starting attribute data, wherein the first starting attribute data is current attribute data, and the target attribute data of the triplet data constructed previously is used as the starting attribute data of the adjacent subsequent triplet data;

[0108] The format conversion module 23 is used to convert the format of the triple data to obtain the target triple data, the target triple data includes the starting attribute data, the target attribute data and the correlation value, and the correlation value is calculated according to the distance value of the corresponding triple data;

[0109] The anomaly detection module 24 is used to perform anomaly detection processing on the target triplet data through a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

[0110] In one embodiment, the first data construction module 22 includes: a first construction submodule, a second construction submodule and a third construction submodule;

[0111] A first construction submodule is used to construct a first triplet of data according to the operation data, the first triplet of data including first starting attribute data, first target attribute data and a first distance value, the first target attribute data is preset attribute data directly related to the current attribute data, and the first distance value is one;

[0112] A second construction submodule is used to construct a second triplet of data according to the operation data, the second triplet of data including second starting attribute data, second target attribute data, and a second distance value, the second starting attribute data is a type of attribute data in the first target attribute data, the second target attribute data is preset attribute data directly related to the second starting attribute data, and the second distance value is the first distance value plus one;

[0113] The third construction submodule is used to sequentially construct the remaining triple data according to the operation data, the remaining triple data including starting attribute data, target attribute data and distance value, the starting attribute data is one of the target attribute data of the adjacent previously constructed triple data, the target attribute data is the preset attribute data directly related to the starting attribute data in the corresponding triple data, the distance value is the distance value of the adjacent previously constructed triple data with different starting attribute data plus one, and the distance values ​​of the triple data with the same starting attribute data are the same;

[0114] Based on the foregoing embodiment, the format conversion module 23 is further used to perform format conversion on the first triplet data, the second triplet data and the remaining triplet data based on corresponding distance values ​​to obtain target triplet data, wherein the target triplet data includes starting attribute data, target attribute data and correlation value.

[0115] In one embodiment, the operation data acquisition module 21 includes: a first data acquisition submodule and a second data acquisition submodule;

[0116] The first data acquisition submodule is used to acquire current signals and voltage signals, transmit the current signals and voltage signals to a wave recording device for digital processing, and obtain current attribute data, voltage attribute data, resistance attribute data, and power attribute data;

[0117] The second data acquisition submodule is used to acquire the switch quantity signal, transmit the switch quantity signal to the wave recording device for digital processing, and obtain the switch quantity attribute data.

[0118] In one embodiment, the format conversion module 23 includes: a longest distance determination submodule, a quotient calculation submodule, a correlation calculation submodule, and a value replacement submodule;

[0119] The longest distance determination submodule is used to determine the longest distance value among all triple data according to the distance value in the triple data;

[0120] A quotient value calculation submodule, used to obtain a first quotient value according to the quotient of the distance value of the current triple data and the longest distance value;

[0121] A correlation calculation submodule, used for obtaining a correlation value corresponding to the current triple data according to a difference between the first and second quotient values;

[0122] The value replacement submodule is used to replace the distance value in the current triple data with the correlation value to obtain the corresponding target triple data.

[0123] In one embodiment, the anomaly detection module 24 includes: a sorting and grouping submodule, a kurtosis coefficient calculation submodule, a target group determination submodule and an anomaly detection submodule;

[0124] The sorting and grouping submodule is used to sort the target triplet data from high to low according to the correlation value, and divide the target triplet data with the same correlation value into a group;

[0125] The kurtosis coefficient calculation submodule is used to extract the target attribute data in the corresponding target triple data for each group of target triple data, and calculate the kurtosis coefficient of the corresponding group according to the target attribute data;

[0126] A target group determination submodule, used to determine a group with the same kurtosis coefficient as a preset threshold as a target group;

[0127] The anomaly detection submodule is used to perform anomaly detection processing on the target triplet data of the target group through a preset anomaly detection algorithm to obtain the corresponding anomaly detection result.

[0128] In one embodiment, the anomaly detection submodule includes: an attribute extraction unit, a data subset partitioning unit, a data aggregation unit, a center calculation unit, a center replacement unit, and a detection unit;

[0129] An attribute extraction unit, used for extracting starting attribute data in the corresponding target triple data in the target group;

[0130] A data subset division unit is used to divide the initial attribute data of the same attribute into a sub-data set to obtain a plurality of sub-data sets;

[0131] A data aggregation unit, used for selecting one of the sub-data sets, applying spatial mapping to aggregate similar data items, and obtaining an aggregated data set;

[0132] A center calculation unit is used to perform clustering processing on each clustered data set to obtain a corresponding center point and a distance value between each data point in the clustered data set and the center point;

[0133] A center replacement unit is used to delete data points whose distance values ​​are less than or equal to the maximum distance value and replace them with the center point;

[0134] The detection unit is used to perform anomaly detection processing on the central point through a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

[0135] In one embodiment, the anomaly detection algorithm is an isolation forest algorithm;

[0136] The detection unit includes: a root node creation subunit, a first segmentation value determination subunit, a node allocation subunit, an isolated tree creation subunit, a second segmentation value determination subunit, an isolated forest creation subunit and a detection subunit;

[0137] A root node creation subunit, used to create a root node according to an aggregated data set;

[0138] A first segmentation value determination subunit is used to randomly determine a segmentation value according to the maximum value and the minimum value of the center points of all the clustered data sets;

[0139] A node allocation subunit, for placing the center point of the corresponding aggregated data set that is less than the center point of the split value in the left subtree of the corresponding root node, and placing the center point of the corresponding aggregated data set that is greater than or equal to the split value in the right subtree of the corresponding root node, until the leaf node contains only one node;

[0140] An isolated tree creation subunit is used to construct an isolated tree according to the corresponding root node and all leaf nodes;

[0141] The second segmentation value determination subunit is used to randomly determine a segmentation value again and construct a corresponding isolated tree until the number of the corresponding isolated trees reaches a preset threshold;

[0142] Isolation forest creates subunits to integrate all isolated trees and build an isolation forest;

[0143] The detection subunit is used to perform anomaly detection processing on the isolation forest through a preset isolation forest algorithm to obtain a corresponding anomaly judgment result.

[0144] The abnormality detection device for the isolating switch cabinet provided in the embodiment of the present application can be used to execute the abnormality detection method for the isolating switch cabinet provided in the above embodiment, and has corresponding functions and beneficial effects.

[0145] The present application embodiment provides an abnormality detection device for an isolation switch cabinet, referring to Figure 4 The abnormality detection device of the isolation switch cabinet includes: a processor 31, a memory 32, a communication module 33, an input device 34 and an output device 35. The number of processors in the abnormality detection device of the isolation switch cabinet can be one or more, and the number of memories in the abnormality detection device of the isolation switch cabinet can be one or more. The processor, memory, communication module, input device and output device of the abnormality detection device of the isolation switch cabinet can be connected through a bus or other methods.

[0146] The memory 32, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the abnormality detection method of the isolation switch cabinet in any embodiment of the present application (for example, the operation data acquisition module, the first data construction module, the format conversion module and the abnormality detection module in the abnormality detection device of the isolation switch cabinet). The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function; the storage data area may store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device or other non-volatile solid-state storage device. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0147] The communication module 33 is used for data transmission.

[0148] The processor 31 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory, that is, realizes the above-mentioned abnormality detection method of the isolation switch cabinet.

[0149] The input device 34 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 35 may include a display device such as a display screen.

[0150] The abnormality detection device for the isolating switch cabinet provided above can be used to execute the abnormality detection method for the isolating switch cabinet provided in the above embodiment, and has corresponding functions and beneficial effects.

[0151] The embodiment of the present application also provides a storage medium storing computer executable instructions, which are used to execute a method for detecting an abnormality of an isolating switch cabinet when executed by a computer processor. The method for detecting an abnormality of an isolating switch cabinet comprises: obtaining operation data of the isolating switch cabinet, the operation data comprising current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, the switch attribute data comprising opening response time, closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal; sequentially constructing triplet data according to the operation data, the triplet data comprising starting, closing and closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal; sequentially constructing triplet data according to the operation data, the triplet data comprising starting, closing and closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal Initial attribute data, target attribute data and distance value, the target attribute data is preset attribute data directly related to the initial attribute data, the distance value is the distance value between the target attribute data and the first initial attribute data, the first initial attribute data is the current attribute data, and the target attribute data of the previously constructed triple data is used as the initial attribute data of the adjacent subsequent triple data; the triple data is formatted to obtain target triple data, the target triple data includes the initial attribute data, the target attribute data and the correlation value, and the correlation value is calculated according to the distance value of the corresponding triple data; the target triple data is processed for anomaly detection to obtain the corresponding anomaly detection result.

[0152] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (for example, in different computer systems connected by a network). The storage medium may store program instructions (for example, embodied as a computer program) that can be executed by one or more processors.

[0153] Of course, the storage medium storing computer executable instructions provided in an embodiment of the present application is not limited to the abnormality detection method for the isolation switch cabinet as described above, and can also execute related operations in the abnormality detection method for the isolation switch cabinet provided in any embodiment of the present application.

[0154] The abnormality detection device, storage medium and abnormality detection equipment of the isolating switch cabinet provided in the above embodiments can execute the abnormality detection method of the isolating switch cabinet provided in any embodiment of the present application. For technical details not described in detail in the above embodiments, please refer to the abnormality detection method of the isolating switch cabinet provided in any embodiment of the present application.

[0155] The above are only preferred embodiments of the present application and the technical principles used. The present application is not limited to the specific embodiments described herein, and various obvious changes, readjustments and substitutions that can be made by those skilled in the art will not deviate from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the claims.

Claims

1. A method for detecting abnormality of an isolation switch cabinet, characterized in that: include: Acquire operation data of the isolation switch cabinet, wherein the operation data includes current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, wherein the switch attribute data includes opening response time, closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal; Constructing triplet data in sequence according to the operation data, the triplet data including starting attribute data, target attribute data and a distance value, the target attribute data being preset attribute data directly related to the starting attribute data, the distance value being a distance value between the target attribute data and the first starting attribute data, the first starting attribute data being current attribute data, and the target attribute data of the triplet data constructed previously being used as the starting attribute data of the adjacent subsequent triplet data; Performing format conversion on the triple data to obtain target triple data, wherein the target triple data includes starting attribute data, target attribute data and a correlation value, wherein the correlation value is calculated according to the distance value of the corresponding triple data; Anomaly detection processing is performed on the target triplet data through a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

2. The method according to claim 1, characterized in that The step of sequentially constructing triplet data according to the operation data includes: Constructing a first triplet of data according to the operation data, the first triplet of data including first starting attribute data, first target attribute data and a first distance value, the first target attribute data being preset attribute data directly related to the current attribute data, and the first distance value being one; Constructing a second triplet of data according to the operation data, the second triplet of data including second starting attribute data, second target attribute data, and a second distance value, the second starting attribute data being one of the first target attribute data, the second target attribute data being preset attribute data directly related to the second starting attribute data, and the second distance value being the first distance value plus one; The remaining triple data are sequentially constructed according to the operation data, wherein the remaining triple data include starting attribute data, target attribute data and a distance value, wherein the starting attribute data is a type of attribute data in the target attribute data of the adjacent previously constructed triple data, the target attribute data is preset attribute data directly related to the starting attribute data in the corresponding triple data, and the distance value is the distance value of the adjacent previously constructed triple data with different starting attribute data plus one, and the distance values ​​of the triple data of the same starting attribute data are the same; The format conversion of the triple data is performed to obtain target triple data, wherein the target triple data includes starting attribute data, target attribute data and a correlation value, including: The first triplet data, the second triplet data and the remaining triplet data are format-converted based on corresponding distance values ​​to obtain target triplet data, wherein the target triplet data includes starting attribute data, target attribute data and a correlation value.

3. The method according to claim 1, characterized in that: The obtaining of the operating data of the isolation switch cabinet includes: Acquire a current signal and a voltage signal, transmit the current signal and the voltage signal to a wave recording device for digital processing, and obtain the current attribute data, the voltage attribute data, the resistance attribute data, and the power attribute data; The switch quantity signal is acquired, and the switch quantity signal is transmitted to a wave recording device for digital processing to obtain the switch quantity attribute data.

4. The method according to claim 1, characterized in that The step of converting the format of the triple data to obtain target triple data includes: Determine the longest distance value among all triple data according to the distance values ​​in the triple data; Obtaining a first quotient value according to the quotient of the distance value of the current triple data and the longest distance value; Obtaining a correlation value corresponding to the current triple data according to a difference between the first quotient value and the first quotient value; The correlation value is used to replace the distance value in the current triple data to obtain the corresponding target triple data.

5. The method according to claim 1, characterized in that The method of performing anomaly detection processing on the target triplet data by using a preset anomaly detection algorithm to obtain a corresponding anomaly detection result includes: Sort the target triplet data from high to low according to the correlation value, and divide the target triplet data with the same correlation value into a group; For each group of target triplet data, extract the target attribute data in the corresponding target triplet data, and calculate the kurtosis coefficient of the corresponding group according to the target attribute data; The group with the same kurtosis coefficient as the preset threshold is determined as the target group; Anomaly detection processing is performed on the target triplet data of the target group using a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

6. The method according to claim 5, characterized in that The method of performing anomaly detection processing on the target triplet data of the target group by using a preset anomaly detection algorithm to obtain a corresponding anomaly detection result includes: For the target triplet data in the target group, extracting the starting attribute data in the corresponding target triplet data; Divide the initial attribute data of the same attribute into a sub-data set to obtain a plurality of sub-data sets; Select one of the sub-datasets and apply spatial mapping to cluster similar data items to obtain a clustered data set; Perform clustering processing on each clustered data set to obtain a corresponding center point and a distance value between each data point in the clustered data set and the center point; The data points whose distance values ​​are less than or equal to the maximum distance value are deleted and replaced with the center point; Anomaly detection processing is performed on the central point using a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

7. The method according to claim 6, characterized in that The anomaly detection algorithm is an isolation forest algorithm; The abnormality detection processing is performed on the central point by using a preset abnormality detection algorithm to obtain a corresponding abnormality detection result, including: Creating a root node based on the aggregated data set; A split value is randomly determined based on the maximum and minimum values ​​of the center points of all clustered data sets; Place the center point of the corresponding aggregated data set that is less than the center point of the split value in the left subtree of the corresponding root node, and place the center point of the corresponding aggregated data set that is greater than or equal to the split value in the right subtree of the corresponding root node, until the leaf node contains only one node; Build an isolated tree based on the corresponding root node and all leaf nodes; Re-randomly determine a split value and construct a corresponding isolated tree until the number of corresponding isolated trees reaches a preset threshold; Integrate all isolated trees to build an isolation forest; The isolation forest is subjected to anomaly detection processing by using a preset isolation forest algorithm to obtain a corresponding anomaly judgment result.

8. An abnormality detection device for an isolation switch cabinet, characterized in that: include: An operation data acquisition module is used to acquire the operation data of the isolation switch cabinet, wherein the operation data includes current attribute data, voltage attribute data, resistance attribute data, power attribute data and switch attribute data, wherein the switch attribute data includes opening response time, closing response time, time point of relay switch signal, time point of closing switch signal and time point of opening switch signal; A first data construction module is used to sequentially construct triplet data according to the operation data, wherein the triplet data includes starting attribute data, target attribute data and a distance value, wherein the target attribute data is preset attribute data directly related to the starting attribute data, and the distance value is a distance value between the target attribute data and the first starting attribute data, wherein the first starting attribute data is current attribute data, and the target attribute data of the triplet data constructed previously is used as the starting attribute data of the adjacent subsequent triplet data; A format conversion module, used for performing format conversion on the triple data to obtain target triple data, wherein the target triple data includes starting attribute data, target attribute data and a correlation value, wherein the correlation value is calculated based on the distance value of the corresponding triple data; The anomaly detection module is used to perform anomaly detection processing on the target triplet data through a preset anomaly detection algorithm to obtain a corresponding anomaly detection result.

9. An abnormality detection device for an isolation switch cabinet, characterized in that: include: memory and one or more processors; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A storage medium storing computer executable instructions, characterized in that: The computer executable instructions are used to perform the method according to any one of claims 1 to 7 when executed by a processor.

Images