Cable partial discharge detection method, device, system, equipment and storage medium

Abstract

The present application discloses a method, device, system, equipment and storage medium for detecting partial discharge of a cable, which obtains signal data of the cable, the signal data is collected by one or more sensors and contains sensor tag information, selects a target classifier according to the sensor tag information, inputs the signal data into the target classifier to output an initial detection result, and then sets a weight vector for the initial detection result and inputs it into a trained Bagging classifier to output a final detection result, to determine whether partial discharge occurs in the cable. The scheme uses signal data collected by multiple sensors to train the AdaBoosting classifier and the Bagging classifier, and finally obtains a Bagging classifier that can be used to accurately detect whether partial discharge occurs in the cable. This method can accurately determine whether partial discharge occurs in the cable and improve detection accuracy.

Description

Technical Field

[0001] The present application relates to the field of power detection, and specifically to a method, device, system, equipment and storage medium for detecting partial discharge of a cable. Background Art

[0002] With the continuous development of science and technology, the demand for daily and industrial electricity is increasing, and the proportion of power cables used in power grids is gradually increasing. Cables will generate non-abnormal signals or partial discharge signals during operation, but cables often have hidden dangers of electricity use. When hidden dangers of electricity use occur in cables, they are often accompanied by partial discharge. Therefore, by analyzing the partial discharge signal, the hidden dangers of electricity use of cables can be inferred.

[0003] At present, the commonly used method for detecting partial discharge of cables is to use a sensor to collect signal data of the cable and analyze the signal data using feature extraction technology. This method cannot accurately determine whether partial discharge occurs in the cable during actual detection, resulting in low detection accuracy. Summary of the invention

[0004] In view of this, the present application provides a cable partial discharge detection method, device, system, equipment and readable storage medium to solve the defects of the existing cable partial discharge detection method that cannot accurately determine whether partial discharge occurs in the cable and has low detection accuracy.

[0005] To achieve the above objectives, the proposed solution is as follows:

[0006] In a first aspect, an embodiment of the present application provides a method for detecting partial discharge of a cable, the method being applied to a terminal device, comprising:

[0007] Acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information;

[0008] Selecting at least one target classifier from a plurality of trained AdaBoosting strong classifiers according to the sensor tag information;

[0009] Inputting the signal data into each of the target classifiers and outputting each initial detection result;

[0010] A weight vector is set for each of the initial detection results, and each of the initial detection results after the weight vector is set is input into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

[0011] In a second aspect, an embodiment of the present application provides a device for detecting partial discharge of a cable, comprising:

[0012] A signal data acquisition module, used to acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information;

[0013] A target classifier determination module, used to select at least one target classifier from a plurality of trained AdaBoosting strong classifiers according to the sensor tag information;

[0014] An initial detection result obtaining module, used for inputting the signal data into each target classifier and outputting each initial detection result;

[0015] The final detection result obtaining module is used to set a weight vector for each of the initial detection results, input each initial detection result after setting the weight vector into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

[0016] In a third aspect, an embodiment of the present application provides a cable partial discharge detection system, comprising:

[0017] A sensor, used for collecting signal data or sample signals of the cable;

[0018] The server is used to execute the cable partial discharge detection method provided in the first aspect.

[0019] In a fourth aspect, an embodiment of the present application provides a cable partial discharge detection device, including a memory and a processor;

[0020] The memory is used to store programs;

[0021] The processor is used to execute the program to implement the various steps of the cable partial discharge detection method provided in the first aspect.

[0022] In a fifth aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the various steps of the cable partial discharge detection method provided in the first aspect above are implemented.

[0023] It can be seen from the above technical solution that the present application first obtains the signal data of the cable, wherein the signal data is collected by one or more sensors, and the signal data contains sensor tag information; then, at least one target classifier is selected from multiple trained AdaBoosting strong classifiers according to the sensor tag information, the signal data is input into each target classifier and each initial detection result is output; finally, each initial detection result is set with a weight vector and then input into the trained Bagging classifier and the final detection result is output, and the final detection result is used to determine whether partial discharge occurs in the cable. This method can use one or more sensors to collect the signal data of the cable during actual detection, and then use the trained AdaBoosting classifier and Bagging classifier to detect the signal data, so as to more accurately determine whether partial discharge occurs in the cable and improve the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of an application scenario of a method for detecting partial discharge of a cable provided in an embodiment of the present application;

[0025] Figure 2 An optional flow chart of a method for detecting partial discharge of a cable provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of the correspondence between the signal data and the sensor provided in the embodiment of the present application;

[0027] Figure 4 A schematic diagram of the correspondence between sensors and sample signals during training of the AdaBoosting strong classifier provided in an embodiment of the present application;

[0028] Figure 5 A schematic diagram of the corresponding relationship between the sensor and the sample signal during actual detection provided by the embodiment of the present application;

[0029] Figure 6 An optional training flow chart of multiple AdaBoosting strong classifiers provided in an embodiment of the present application;

[0030] Figure 7 A schematic diagram of a cable partial discharge detection device provided in an embodiment of the present application;

[0031] Figure 8 A schematic diagram of the structure of a cable partial discharge detection device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] Please refer to Figure 1 , Figure 1 A schematic diagram of an application scenario of the cable partial discharge detection method provided in an embodiment of the present application is shown. The application scenario includes the cable 10, sensor 20 and server 30 provided in an embodiment of the present application. The cable 10 generates signal data, the sensor 20 can collect the signal data, and the server 30 can obtain the signal data collected by the sensor 20.

[0034] It can be understood that the sensor 20 may also be other acquisition devices.

[0035] In some implementation scenarios, the sensor 20 may be placed around the cable 10 .

[0036] In this embodiment, the server may obtain signal data in the following ways:

[0037] The sensor 20 inputs the signal data into the sampling device of the server, and the sampling device converts the signal data into data that can be recognized by the server 30. Alternatively, the sensor 20 has a bus interface, and the server 30 directly reads the data signal collected by the sensor 20 through the corresponding bus.

[0038] The present application provides a method for detecting partial discharge of a cable. Figure 2 , Figure 2 An optional flow chart of a method for detecting partial discharge of a cable provided in an embodiment of the present application is shown, comprising the following steps:

[0039] S1: Acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information.

[0040] Specifically, when the cable is in operation, if the electric field strength in a local area of ​​its insulator reaches the breakdown field strength, local discharge will occur in the area, thereby generating a local discharge signal. When the cable is in normal working condition, it will generate non-abnormal signals, and the local discharge signal and the non-abnormal signal can be collected by sensors. Optionally, there are many types of sensors used, including ultrasonic sensors, geomagnetic wave sensors, infrared sensors, radio frequency sensors, pulse current sensors, and ultra-high frequency sensors, etc. These sensors can collect ultrasonic signals, current signals, electromagnetic wave signals, temperature signals, etc. In the embodiments provided in the present application, the signal data acquired by the server can be collected by one or more sensors, one or more of which can be sensors of the same type or sensors of different types, and the collected signal data contains sensor tag information, refer to Figure 3 , the sensor tag information can be used to determine which sensor the signal data is collected by. Among them, signal data 1 contains sensor A tag information, and signal data 2 contains sensor B tag information.

[0041] S2: Select at least one target classifier from multiple trained AdaBoosting strong classifiers according to the sensor tag information.

[0042] After the signal data with sensor tag information is obtained in step S1, several AdaBoosting strong classifiers are pre-trained using sample signals, and target sensors are selected from the signal data according to the sensor tag information contained in the signal data, wherein the signal samples can be collected by several sensors.

[0043] Please refer to Figure 4 and Figure 5 , Figure 4 and Figure 5 It shows how to determine the target classifier according to the sensor used in actual detection, that is, the sample signal used by the target classifier during training contains the same sensor label information as the sensor label information contained in the signal data obtained in S1.

[0044] Reference Figure 4 , Figure 4 The corresponding relationship between the sensor and the sample signal during the training of the AdaBoosting strong classifier is shown:

[0045] During training, the AdaBoosting strong classifier a is trained using the sample signal 1 collected by the sensor A, and the sample signal 1 contains the sensor A label information.

[0046] The AdaBoosting strong classifier b is trained using the sample signal 1 collected by sensor A and the sample signal 2 collected by sensor B. The sample signal 2 contains the label information of sensor B.

[0047] The AdaBoosting strong classifier c is trained using the sample signal 3 collected by the sensor C, and the sample signal 3 contains the sensor C label information.

[0048] Reference Figure 5 , Figure 5 The corresponding relationship between the sensor and the sample signal during actual detection is shown:

[0049] During actual detection, sensor A and sensor C can be selected to collect signal data of the cable. Signal data 1 collected by sensor A contains sensor A label information, and signal data 3 collected by sensor C contains sensor C label information. It can be known that AdaBoosting strong classifier a is trained using sample signal 1 containing sensor A label information, and AdaBoosting strong classifier b is also trained using sample signal 1 containing sensor A label information. Then, based on the selected sensor A, AdaBoosting strong classifier a and AdaBoosting strong classifier b can be used as target sensors.

[0050] At the same time, the AdaBoosting strong classifier c is trained using the sample signal 3 containing the sensor C label. Then, according to the selected sensor B, the AdaBoosting strong classifier c can be used as the target sensor.

[0051] In summary, AdaBoosting strong classifier a, AdaBoosting strong classifier b and AdaBoosting strong classifier c are determined as target classifiers.

[0052] S3: Input the signal data into each target classifier and output each initial detection result.

[0053] After the target classifier is determined in step S2, the signal data collected in step S1 is input into the target classifier, and the target classifier outputs an initial detection result.

[0054] S4: setting a weight vector for each of the initial detection results, inputting each of the initial detection results after the weight vector is set into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

[0055] The Bagging classifier is pre-trained, and the initial detection result obtained in step S2 is set as a weight vector, which is then input into the Bagging classifier to output the final detection result. The final detection result can be used to determine whether partial discharge occurs in the cable.

[0056] The detection method of partial discharge of cables provided in the embodiment of the present application first collects the signal data of the cable containing the sensor tag information through one or more sensors, then selects the target classifier from multiple trained AdaBoosting strong classifiers based on the signal data, inputs the signal data into the target classifier and outputs the initial detection result, finally sets the weight vector of the initial detection result and inputs it into the trained Bagging classifier and outputs the final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable. This method can use one or more sensors to collect the signal data of the cable during actual detection, and then use the trained AdaBoosting classifier and Bagging classifier to detect the signal data, so as to more accurately determine whether partial discharge occurs in the cable and improve the detection accuracy.

[0057] Optionally, the present application embodiment provides an optional training process for multiple AdaBoosting strong classifiers, such as Figure 6 As shown, including:

[0058] S21: Acquire a sample signal of the cable, where the sample signal is collected by a plurality of sensors, wherein the sample signal includes a partial discharge signal and a non-abnormal signal, where the non-abnormal signal is a signal generated by the cable in a normal working state.

[0059] S22: Label the sample signal.

[0060] S23: Randomly select different numbers of sample signals according to different selection orders to construct multiple training sets.

[0061] S24: using multiple training sets to train the AdaBoosting classifiers respectively, so as to obtain multiple AdaBoosting strong classifiers.

[0062] In an optional embodiment, the training process of multiple AdaBoosting strong classifiers may include:

[0063] Select N types of sensors, use one of each sensor, and give a parameter i, so that i=1.

[0064] Select one sample signal from the sample signals collected by the N sensors, input this signal into a base classifier, and train it using the AdaBoosting algorithm to obtain a weak classifier. The weak classifier is used as an AdaBoosting strong classifier. Continue to select one signal from the sample signals collected by the N sensors and train it using the base classifier until all the sample signals collected by the N sensors are selected, and N AdaBoosting strong classifiers are obtained.

[0065] Let i = i + 1, and execute the above process of training AdaBoosting classifier until i = N, and finally get AdaBoosting strong classifiers, each of which outputs its own initial detection result.

[0066] Specifically, before the above step S21, the local discharge scenario and the normal working scenario of the cable can be artificially simulated first to generate sample signals, and the sample signals include local discharge signals and non-abnormal signals. A variety of sensors are arranged around the local discharge scenario and the normal working scenario. After these sensors collect the sample signals of the cable at the same time, the server can obtain the sample signals.

[0067] Optionally, before step S22, artificial noise may be added to the acquired sample signal, which may enrich the sample signal and improve the accuracy of detection.

[0068] Optionally, in step S22, the sample signal may be labeled in the following ways:

[0069] The partial discharge signal is labeled with the first label, or the abnormal signal is labeled with the second label, or the partial discharge signal is labeled with the first label and the non-abnormal signal is labeled with the second label.

[0070] In step S23, for ease of understanding, the specific process of selecting three sensors to collect sample signals is given below:

[0071] Select three sensors and call them A, B, and C. Use these three sensors to collect signal data of the cable, and collect sample signal 1, sample signal 2, and sample signal 3, hereinafter referred to as 1, 2, and 3. Randomly select different numbers of sample signals in different orders to construct multiple training sets, which may include:

[0072] 1, 2, 3, 12, 21, 13, 31, 23, 32, 123, a total of 10 training sets.

[0073] It should be noted that the above-mentioned 12 training sets and 21 training sets are different training sets, which means that the order of selecting sample signal 1 and sample signal 2 is different. Similarly, 13 training sets and 31 training sets, 23 training sets and 32 training sets are all different training sets.

[0074] Optionally, step S24 may include the following process:

[0075] Select sample signals according to the selection order of each sample signal in the training set.

[0076] According to the classification result output by the weak classifier corresponding to the previous sample signal, the AdaBoosting classifier is trained using the selected sample signal to obtain a weak classifier corresponding to the selected sample signal; after the weak classifier corresponding to each sample signal in the training set is obtained, the weak classifiers corresponding to each sample signal are linearly combined to obtain an AdaBoosting strong classifier.

[0077] For ease of understanding, a detailed embodiment is given. The AdaBoosting classifier is trained using the 10 training sets in step S23, taking the 123 data set as an example:

[0078] In the order of 123, sample signal 1, sample signal 2, and sample signal 3 are selected in turn.

[0079] The sample signal 1 is input into a base classifier and trained using the AdaBoosting algorithm to obtain the first weak classifier. If the sample signal 1 is a non-abnormal signal, but the first weak classifier classifies the sample signal 1 as a partial discharge signal, then the classification is wrong.

[0080] According to the classification result output by the first weak classifier, the sample signal 2 is input into the new base classifier for training to obtain the second weak classifier. The second weak classifier outputs the classification result.

[0081] According to the classification result output by the second weak classifier, the sample signal 3 is input into the new base classifier for training to obtain the third weak classifier.

[0082] Finally, the three weak classifiers mentioned above are linearly combined to obtain the AdaBoosting strong classifier and output the initial detection result of the sample.

[0083] In the above scheme, the AdaBoosting algorithm belongs to a kind of Boosting algorithm. The Boosting algorithm is a classification algorithm that combines weak classifiers by linear combination to obtain a strong classifier with greatly improved classification performance. The algorithm can transform some rough empirical rules into highly accurate prediction rules. The strong classifier classifies signal data through the majority voting mechanism of the weak classifier. Compared with the Boosting algorithm, the AdaBoosting algorithm can adaptively adjust the classification results of the obtained weak classifiers, so as to obtain a weak classifier with higher classification degree. Therefore, in the embodiment provided in the present application, the AdaBoosting algorithm can be used for training, so as to achieve more accurate classification results, i.e., the initial detection results of the samples.

[0084] Optionally, the embodiment of the present application provides an optional training process of a bagging classifier, including:

[0085] The initial detection results of all trained AdaBoosting strong classifiers are randomly sampled multiple times to obtain multiple sample sets, which are input into the base classifiers respectively and trained using the Bagging algorithm to obtain multiple trained Bagging classifiers. The Bagging (Boostrao aggregating) algorithm is a guided clustering algorithm, also known as the bagging algorithm. The Bagging algorithm can be combined with other classification algorithms and regression algorithms to improve its classification accuracy and stability and avoid overfitting.

[0086] Combine multiple trained Bagging classifiers, and the combined Bagging classifier outputs the final detection result of the sample.

[0087] In an optional embodiment, before the sample initial detection result is input into the trained Bagging classifier, the sample initial detection result can be set Dimensional weight vector Among them, a weight vector is composed of many dimensional data, and each dimension of the weight vector corresponds to the initial detection result of the sample output by each of the above AdaBoosting strong detectors. Then set the weight vector The initial detection results of the samples are then input into the Bagging classifier for training.

[0088] Specifically, the training process of the Bagging classifier can include the following steps:

[0089] The iteration number T of the weak learner algorithm and the weak classifier is preset, and the initial detection result output by the AdaBoosting strong classifier is used as input.

[0090] The initial detection results of the samples are randomly sampled with replacement t times, t = 1, 2, ..., T, and the base classifiers are trained using the Bagging algorithm to obtain multiple Bagging classifiers. All the Bagging classifiers are combined as a whole to obtain the final Bagging classifier and output the final detection results of the samples.

[0091] The following is a description of a cable partial discharge detection device provided in an embodiment of the present application. The cable partial discharge detection device described below and the cable partial discharge detection method described above can refer to each other.

[0092] Combination Figure 7 , the detection device of cable partial discharge is introduced, such as Figure 7 As shown, the detection device may include:

[0093] A signal data acquisition module, used to acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information;

[0094] A target classifier determination module, used to select at least one target classifier from a plurality of trained AdaBoosting strong classifiers according to the sensor tag information;

[0095] An initial detection result obtaining module, used for inputting the signal data into each target classifier and outputting each initial detection result;

[0096] The final detection result obtaining module is used to set a weight vector for each of the initial detection results, input each initial detection result after setting the weight vector into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

[0097] In an optional embodiment, the cable partial discharge detection device may further include:

[0098] A sample signal acquisition module is used to acquire a sample signal of the cable. The sample signal is collected by multiple sensors. The sample signal includes a partial discharge signal and a non-abnormal signal. The non-abnormal signal is a signal generated by the cable in a normal working state.

[0099] A sample signal labeling module, used for labeling sample signals;

[0100] A training set construction module, used for randomly selecting different numbers of sample signals according to different selection orders to construct multiple training sets;

[0101] The AdaBoosting classifier training module is used to respectively train the AdaBoosting classifier using multiple training sets to obtain multiple AdaBoosting strong classifiers.

[0102] In an optional embodiment, the cable partial discharge detection device of the present application may further include:

[0103] The artificial noise adding module is used to add artificial noise to the sample signal.

[0104] Furthermore, an embodiment of the present application provides a cable partial discharge detection system, including a sensor and a server. The specific implementation logic of these two units can refer to the cable partial discharge detection method provided in the method embodiment of the present application.

[0105] Furthermore, the embodiment of the present application provides a cable partial discharge detection device, such as a terminal: a mobile phone, a computer, etc. Optionally, Figure 8 The hardware structure diagram of the cable partial discharge detection device is shown in FIG. Figure 8 The hardware structure of the cable partial discharge detection device may include: at least one processor 01 , at least one communication interface 02 , at least one memory 03 and at least one communication bus 04 .

[0106] In the embodiment of the present application, the number of the processor 01 , the communication interface 02 , the memory 03 , and the communication bus 04 is at least one, and the processor 01 , the communication interface 02 , and the memory 03 communicate with each other through the communication bus 04 .

[0107] The processor 01 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0108] The memory 03 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory.

[0109] The memory stores a program, and the processor can call the program stored in the memory, where the program is used to execute the partial discharge detection method described in the method embodiment.

[0110] Optionally, the detailed functions and extended functions of the program may refer to the description of the partial discharge detection method in the method embodiment.

[0111] An embodiment of the present application further provides a storage medium, which can store a program suitable for execution by a processor, wherein the program is used to execute the partial discharge detection method described in the method embodiment.

[0112] Specifically, the storage medium may be a computer-readable storage medium, and the computer-readable storage medium may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk or a ROM.

[0113] Optionally, the detailed functions and extended functions of the program may refer to the description of the partial discharge detection method in the method embodiment.

[0114] In addition, the functional modules in the various embodiments of the present disclosure can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a live broadcast device, or a network device, etc.) to perform all or part of the steps of the method described in the various embodiments of the present disclosure.

[0115] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0116] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0117] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting partial discharge of a cable, characterized in that: include: Acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information; At least one target classifier is selected from a plurality of trained AdaBoosting strong classifiers according to the sensor tag information; the training process of the plurality of the AdaBoosting strong classifiers includes: Acquire a sample signal of the cable, wherein the sample signal is collected by multiple sensors, wherein the sample signal includes a partial discharge signal and a non-abnormal signal, wherein the non-abnormal signal is a signal generated by the cable in a normal working state; label the sample signal; randomly select different numbers of sample signals according to different selection orders to construct multiple training sets; use the multiple training sets to train AdaBoosting classifiers respectively to obtain multiple AdaBoosting strong classifiers; wherein, select Types of sensors, use one for each sensor, set parameters ,make ; From here Select one sample signal from the sample signals collected by the sensors, input this sample signal into a base classifier, train it using the AdaBoosting algorithm, get a weak classifier, and use the weak classifier as the AdaBoosting strong classifier; continue in this Select one sample signal from the sample signals collected by the sensors and train it using the base classifier until the sample signal is After selecting the sample signals collected by the sensors, we get AdaBoosting strong classifier; , execute the above process of training AdaBoosting strong classifier until , and finally get AdaBoosting strong classifiers, each of which outputs its own initial detection result; Inputting the signal data into each of the target classifiers and outputting each initial detection result; A weight vector is set for each of the initial detection results, and each of the initial detection results after the weight vector is set is input into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

2. The method according to claim 1, characterized in that The marking of the sample signal includes: Using a first label to mark the partial discharge signal, and / or, The non-abnormal signal is labeled with a second label.

3. The method according to claim 1, characterized in that Before marking the sample signal, the method further includes: Artificial noise is added to the sample signal.

4. The method according to claim 1, characterized in that: The process of training the AdaBoosting strong classifier using the training set includes: Selecting sample signals according to the selection order of the sample signals in the training set; According to the classification result output by the weak classifier corresponding to the previous sample signal, the AdaBoosting classifier is trained using the selected sample signal to obtain a weak classifier corresponding to the selected sample signal; after the weak classifier corresponding to each sample signal in the training set is obtained, the weak classifiers corresponding to each sample signal are linearly combined to obtain an AdaBoosting strong classifier.

5. The method according to claim 4, characterized in that The training process of the Bagging classifier includes: Setting a weight vector for the sample initial detection result output by the AdaBoosting strong classifier; Perform multiple random sampling on the initial detection results of the samples after setting the weight vector to obtain multiple sample sets; Inputting the sample sets into bagging classifiers for training respectively to obtain multiple trained bagging classifiers; Combine multiple trained Bagging classifiers, and the combined Bagging classifier outputs the final detection result of the sample.

6. A cable partial discharge detection device, characterized in that: include: A signal data acquisition module, used to acquire signal data of the cable, wherein the signal data is collected by one or more sensors and includes sensor tag information; The target classifier determination module is used to select at least one target classifier from multiple trained AdaBoosting strong classifiers according to the sensor tag information; the training process of multiple AdaBoosting strong classifiers includes: obtaining sample signals of the cable, the sample signals are collected by multiple sensors, wherein the sample signals include local discharge signals and non-abnormal signals, and the non-abnormal signals are signals generated by the cable in normal working state; marking the sample signals; randomly selecting different numbers of sample signals according to different selection orders to construct multiple training sets; using multiple training sets to train AdaBoosting classifiers respectively to obtain multiple AdaBoosting strong classifiers; wherein, selecting Types of sensors, use one for each sensor, set parameters ,make ; From here Select one sample signal from the sample signals collected by the sensors, input this sample signal into a base classifier, train it using the AdaBoosting algorithm, get a weak classifier, and use the weak classifier as the AdaBoosting strong classifier; continue in this Select one sample signal from the sample signals collected by the sensors and train it using the base classifier until the sample signal is After selecting the sample signals collected by the sensors, we get AdaBoosting strong classifier; , execute the above process of training AdaBoosting strong classifier until , and finally get AdaBoosting strong classifiers, each of which outputs its own initial detection result; An initial detection result obtaining module, used for inputting the signal data into each target classifier and outputting each initial detection result; The final detection result obtaining module is used to set a weight vector for each of the initial detection results, input each initial detection result after setting the weight vector into a trained Bagging classifier to output a final detection result, and the final detection result is used to determine whether partial discharge occurs in the cable.

7. A cable partial discharge detection system, characterized in that: include: A sensor, used for collecting signal data or sample signals of the cable; A server, used to execute the cable partial discharge detection method as described in any one of claims 1-5.

8. A cable partial discharge detection device, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the cable partial discharge detection method as described in any one of claims 1-5.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method for detecting partial discharge of a cable as described in any one of claims 1 to 5 is implemented.

Images