A high temperature superconducting cable monitoring system and method

Abstract

The present invention relates to the technical field of superconducting cable monitoring, and proposes a high-temperature superconducting cable monitoring system and method. The system and method can realize early prediction of cable quench faults through intelligent learning algorithms, and can remove and repair high-risk operating cables in advance to avoid damage to cables caused by major faults, reduce the operating cost of power transmission, and realize online real-time monitoring of the operating status of superconducting cables, thereby improving the safety and stability of system operation.

Description

Technical Field

[0001] The present disclosure relates to the technical field related to superconducting cable monitoring, and in particular, to a high-temperature superconducting cable monitoring system and method. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] For high-temperature superconducting cables, the main body of transmitting electric energy is the superconducting tape. In order to make it work normally, it is necessary to prevent the superconducting tape from quenching and changing from the unimpeded superconducting state to the normal state with larger resistance. The causes of quenching of high-temperature superconducting cables include: defects in the superconducting tape itself, failure of the superconducting cable cryogenic system that cannot effectively cool the liquid nitrogen, and power grid operation failure. Among them, the influence of superconducting tape characteristics, including the unevenness of the superconducting tape, will concentrate the hot spots in a fixed area during the current rise. If this hot spot is not effectively contained, it will lead to the occurrence of local quenching. In addition, the damaged points of the tape and the joints between the tapes are potential hot spots. When these hot spots generate heat accumulation during operation, long-term operation will also cause local quenching. If the tape has a local quench, the quench area will continue to expand, the temperature of the superconducting tape will rise rapidly, the superconductor structure will be destroyed, and ultimately affect the normal operation of the superconducting cable.

[0004] The inventors found in their research that the existing detection methods are not efficient in detecting local quenching. Generally, faults can only be detected after the entire cable has quenched. Post-fault detection has a lag, and the probability of superconducting cable damage is high, and the stability of power grid operation using superconducting cables cannot be guaranteed. Summary of the invention

[0005] In order to solve the above problems, the present disclosure proposes a high-temperature superconducting cable monitoring system and method, which can realize early prediction of cable quench faults through intelligent learning algorithms, and remove and repair high-risk operating cables in advance to avoid damage to the cables caused by major faults, reduce the operating costs of power transmission, and realize online real-time monitoring of the operating status of superconducting cables, thereby improving the safety and stability of system operation.

[0006] In order to achieve the above objectives, the present disclosure adopts the following technical solutions:

[0007] One or more embodiments provide a high temperature superconducting cable monitoring method, comprising the following steps:

[0008] Obtaining the operating data of superconducting cables, including electrical quantity operating data and non-electrical quantity operating data;

[0009] Based on the acquired electrical quantity signal, the current difference and current phase difference between the conductive layer and the shielding layer are calculated;

[0010] A deep learning network model is constructed, and the calculated current difference and current phase difference between the conductive layer and the shielding layer, voltage value, and non-electrical quantity operation data are taken as input and transmitted to the deep learning network model to obtain the probability of superconducting cable quenching.

[0011] One or more embodiments provide a high temperature superconducting cable monitoring system, comprising:

[0012] An acquisition module is configured to acquire operation data of the superconducting cable, including electrical quantity operation data and non-electrical quantity operation data;

[0013] Current data processing module: configured to calculate the current difference and current phase difference between the conductive layer and the shielding layer according to the acquired electrical quantity signal;

[0014] Probability prediction module: It is configured to build a deep learning network model, and transmit the calculated current difference and current phase difference between the conductive layer and the shielding layer, voltage value, and non-electrical quantity operation data as input to the deep learning network model to obtain the probability of superconducting cable quenching.

[0015] One or more embodiments provide a high temperature superconducting cable monitoring system, including a monitoring terminal, an electrical quantity detection device and a non-electrical quantity detection device communicatively connected to the monitoring terminal;

[0016] The monitoring terminal is configured as the steps of the above-mentioned high-temperature superconducting cable monitoring method.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] In the present disclosure, based on the deep learning network model, the key factors of quench are learned, and the subtle changes of the detection signal can be identified. The multi-dimensional data of quench detection is fused and processed through the deep learning network. Before the occurrence of the whole section quench, according to the probability of identification, the corresponding protection action can be performed before the superconducting cable quench is likely to occur, thereby reducing the probability of superconducting cable damage, thereby ensuring the stable and safe operation of the superconducting cable.

[0019] The advantages of the present disclosure and the advantages of additional aspects will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The exemplary embodiments of the present disclosure and the description thereof are used to explain the present disclosure but do not constitute a limitation of the present disclosure.

[0021] Figure 1 is a schematic diagram of the structure of a high-temperature superconducting cable according to Embodiment 1 of the present disclosure;

[0022] Figure 2 is a schematic structural diagram of a high-temperature superconducting cable monitoring system according to Embodiment 1 of the present disclosure;

[0023] Figure 3 It is a flow chart of the high-temperature superconducting cable monitoring method of Example 2 of the present disclosure. DETAILED DESCRIPTION

[0024] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0025] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0026] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof. It should be noted that, in the absence of conflict, the various embodiments in the present disclosure and the features in the embodiments can be combined with each other. The embodiments will be described in detail below in conjunction with the accompanying drawings.

[0027] Example 1

[0028] In the technical solutions disclosed in one or more embodiments, Figures 1 to 3 As shown, a high-temperature superconducting cable monitoring system includes: a monitoring terminal, an electrical quantity detection device and a non-electrical quantity detection device connected to the monitoring terminal for communication. The monitoring terminal obtains the electrical quantity and non-electrical quantity detected by the high-temperature superconducting cable, and predicts based on a deep learning network model to obtain the probability of superconducting cable quenching.

[0029] The structure of superconducting cable is as follows Figure 1 As shown, it is a multi-layer composite material structure, which includes, from the inside to the outside, a liquid nitrogen flow channel, a stainless steel bellows skeleton, a three-phase conductive layer formed by a high-temperature superconducting phase conductor layer separated by an insulating spacer, a shielding layer, a liquid nitrogen return flow channel, a protective layer and a low-temperature thermostat. Liquid nitrogen (LN2) is input from a low-temperature refrigeration system and flows inside the liquid nitrogen flow channel.

[0030] The three-phase conductive layer includes phase A, phase B and phase C, and an electrical insulation layer and a semi-conductive layer are arranged between each phase. The cryogenic refrigeration system includes a refrigerator and a circulating pump, which can provide liquid nitrogen that meets the cryogenic requirements for the superconducting cable.

[0031] In some embodiments, the non-electrical quantity detection device includes a device for detecting the temperature, pressure and flow rate of liquid nitrogen.

[0032] Optionally, the liquid nitrogen temperature detection device includes temperature sensors arranged at both ends of the superconducting cable. Specifically, they can be used to collect temperature signals at the input and output ends of the low-temperature refrigeration system at the inlet and outlet of the liquid nitrogen in the superconducting cable; temperature sensors can also be respectively arranged at both ends of the three-phase conductive layer of the superconducting cable to collect temperature changes during the operation of the three-phase conductive layer.

[0033] Optionally, the pressure detection device includes a pressure sensor arranged at the inlet and outlet of the liquid nitrogen of the superconducting cable.

[0034] Optionally, the flow detection device includes a flow meter arranged at the inlet and outlet of the liquid nitrogen of the superconducting cable.

[0035] In some embodiments, the electrical quantity detection device includes a device for detecting current and voltage signals of each layer of the superconducting cable.

[0036] Optionally, the current detection device can use a current transformer CT, which is set at both ends of each phase conductive layer to measure the current values ​​at both ends respectively; the current of the superconducting cable shielding layer is also detected and set at the shielding layer lead-out end of the superconducting cable terminal.

[0037] Optionally, the voltage detection device may adopt a voltage transformer PT, which may collect the voltage of each phase conductive layer.

[0038] In a further technical solution, the monitoring terminal is configured to perform a high-temperature superconducting cable monitoring method, and predicts based on a deep learning network model to obtain the probability of superconducting cable quenching, including the following steps:

[0039] Step 1, obtaining the operating data of the superconducting cable, including electrical quantity operating data and non-electrical quantity operating data;

[0040] Among them, the electrical quantity operation data includes the current signals of each layer of the superconducting cable and the three-phase voltage signals; the non-electrical quantity data may include the flow, temperature and pressure signals of liquid nitrogen.

[0041] Step 2: Calculate the current difference and current phase difference between the conductive layer and the shielding layer based on the acquired electrical quantity signal;

[0042] Step 3: Build a deep learning network model, take the calculated current difference and current phase difference between the conductive layer and the shielding layer, voltage value, and non-electrical operation data as input, transmit them to the deep learning network model, and obtain the probability of superconducting cable quenching.

[0043] When a full-section quench occurs, existing detection methods can effectively detect the occurrence of a fault; however, local quench is not easy to detect. In this embodiment, based on a deep learning network model, key factors of quench are learned, and subtle changes in detection signals can be identified. The multi-dimensional data of quench detection is fused and processed through the deep learning network. Before a full-section quench occurs, according to the probability of identification, corresponding protection actions can be performed before a superconducting cable quench is likely to occur, thereby reducing the probability of damage to the superconducting cable, thereby ensuring the stable and safe operation of the superconducting cable.

[0044] The superconducting cable has a shielding layer, and the current difference quench detection method can be used to judge the quench. Before the fault, the current of the conductor layer and the superconducting shielding layer are basically equal. After the fault, the current of the superconducting shielding layer is smaller than the current of the conductor layer.

[0045] Specifically, the current difference between the conductive layer and the shielding layer, assuming that the conductive layer current is I1 and the superconducting shielding layer current is I2. When the cable is operating normally, I1 and I2 are basically equal; when the cable is overcurrent, the shielding layer will be quenched, the shielding layer impedance becomes larger, and the induced current value is smaller than the conductive layer current. If a local fault occurs, the current difference exceeds the set value I A That is, a serious quench fault occurs.

[0046] The current difference between the conductive layer and the shielding layer is calculated as: |I1-I2| A

[0047] Specifically, when the superconducting cable is operating normally, the phase difference between the conductive layer current I1 and the superconducting shielding layer I2 is basically stable at about 180°. After the superconducting cable is quenched, the shielding layer will be quenched, generating a resistive component, which will cause the phase difference between the conductive layer current I1 and the superconducting shielding layer I2 to change. The formula for judging whether the superconducting cable is quenched based on the current phase difference is:

[0048] |Φ1-Φ2|-180°<Φd

[0049] Among them, Φ1 is the initial phase angle of the current in the conductive layer, Φ2 is the initial phase angle of the current in the superconducting shielding layer, and Φd is the phase difference detection setting threshold.

[0050] In this embodiment, the quench fault in the cable can be identified by the current signal based on the change of the current magnitude and the current phase. However, by simply using numerical comparison, the quench fault can only be judged after the fault develops to a certain extent, such as after the quench area exceeds the set value. The fault cannot be identified in the early stage of the fault development, and the hysteresis is serious. The high-temperature superconducting cable is a new type of cable with high preparation cost and serious loss after damage. Therefore, it is necessary to predict the quench fault of the superconducting cable in advance.

[0051] ​After a local quench occurs, the operating parameter data of the superconducting cable begins to change gradually. By integrating the multi-dimensional data of the cable that can be used to determine the quench, including temperature, flow, pressure, current difference, phase change, and voltage, the quench fault can be identified earlier and maintenance can be carried out in advance.

[0052] Optionally, the deep learning network model can be a support vector machine, a random forest model, and a neural network model. In this embodiment, a BP neural network can be used to construct a deep learning network model. After feature extraction of the acquired data, the probability of superconducting cable quenching can be predicted based on the BP neural network.

[0053] Furthermore, before making predictions through the deep learning network model, the deep learning network model is trained, including the following steps:

[0054] Step S1, obtaining historical data of superconducting cable operation and performing preprocessing, wherein the historical data includes normal operation data and fault data;

[0055] The preprocessing includes preprocessing the singular values ​​of the signal, processing missing values, and normalizing the signal using a predictive mean filtering algorithm.

[0056] Step S2: convert the acquired historical data into feature vectors, abstract feature attributes from the historical interaction data, and generate a multi-dimensional feature vector set;

[0057] The process of vectorizing the data may adopt methods in the prior art, which will not be described in detail here.

[0058] Step S3: Based on the feature vector, cluster the historical interaction data according to the faults to obtain multiple clusters;

[0059] Step S4: for the clustering results of the quench fault, identify the corresponding abnormal characteristic values, and determine the average number N of abnormal characteristic values ​​in the fault clustering results as the number of fault characteristic values;

[0060] Among them, abnormal data means exceeding the normal operating value, and the percentage of data fluctuation exceeding the normal value is set as the abnormal value;

[0061] Step S5: for each cluster, the ratio of the number of abnormal characteristic values ​​to the number of fault characteristic values ​​is defined as the cluster risk probability, that is, the risk probability of a quench fault is predicted based on the data in the cluster, and the risk probability is used as a label;

[0062] Specifically, the abnormal characteristic values ​​in each cluster X are identified, and the number m of abnormal characteristic values ​​is determined; then the probability of a quench fault occurring in the operating state corresponding to cluster X is:

[0063]

[0064] Step S6: Taking the feature vector of each cluster as input, obtaining the risk probability of each cluster as output, and inputting it into the deep learning network model for training;

[0065] Step S7: Calculate the cross entropy loss between the risk probability predicted by the model and the actual risk, adjust the parameters of the deep learning network model until the prediction accuracy requirement is met, and obtain the trained deep learning network model.

[0066] During the model's usage phase, fault prediction is performed on superconducting cables, and the data to be processed is vectorized. The feature vector obtained after processing is input into the trained deep learning model for prediction and identification.

[0067] In this embodiment, a corresponding relationship between abnormal values ​​and the probability of fault occurrence is established, and the relationship is learned through a deep learning network model. The intelligent learning algorithm can be used to predict cable quench faults in advance, and high-risk operating cables can be removed and repaired in advance to avoid damage to the cables caused by major faults, reduce the operating cost of power transmission, and realize online real-time monitoring of the operating status of superconducting cables, thereby improving the safety and stability of system operation.

[0068] Example 2

[0069] Based on Example 1, this embodiment provides a high-temperature superconducting cable monitoring method, such as Figure 3 As shown, the following steps are included:

[0070] Step 1, obtaining the operating data of the superconducting cable, including electrical quantity operating data and non-electrical quantity operating data;

[0071] Among them, the electrical quantity operation data includes the current signals of each layer of the superconducting cable and the three-phase voltage signals; the non-electrical quantity data may include the flow, temperature and pressure signals of liquid nitrogen.

[0072] Step 2: Calculate the current difference and current phase difference between the conductive layer and the shielding layer based on the acquired electrical quantity signal;

[0073] Step 3: Build a deep learning network model, take the calculated current difference and current phase difference between the conductive layer and the shielding layer, voltage value, and non-electrical operation data as input, transmit them to the deep learning network model, and obtain the probability of superconducting cable quenching.

[0074] When a full-section quench occurs, existing detection methods can effectively detect the occurrence of a fault; however, local quench is not easy to detect. In this embodiment, based on a deep learning network model, key factors of quench are learned, and subtle changes in detection signals can be identified. The multi-dimensional data of quench detection is fused and processed through the deep learning network. Before a full-section quench occurs, according to the probability of identification, corresponding protection actions can be performed before a superconducting cable quench is likely to occur, thereby reducing the probability of damage to the superconducting cable, thereby ensuring the stable and safe operation of the superconducting cable.

[0075] The superconducting cable has a shielding layer, and the current difference quench detection method can be used to judge the quench. Before the fault, the current of the conductor layer and the superconducting shielding layer are basically equal. After the fault, the current of the superconducting shielding layer is smaller than the current of the conductor layer.

[0076] Specifically, the current difference between the conductive layer and the shielding layer, assuming that the conductive layer current is I1 and the superconducting shielding layer current is I2. When the cable is operating normally, I1 and I2 are basically equal; when the cable is overcurrent, the shielding layer will be quenched, the shielding layer impedance becomes larger, and the induced current value is smaller than the conductive layer current. If a local fault occurs, the current difference exceeds the set value I A That is, a serious quench fault occurs.

[0077] The current difference between the conductive layer and the shielding layer is calculated as: |I1-I2| A

[0078] Specifically, when the superconducting cable is operating normally, the phase difference between the conductive layer current I1 and the superconducting shielding layer I2 is basically stable at about 180°. After the superconducting cable is quenched, the shielding layer will be quenched, generating a resistive component, which will cause the phase difference between the conductive layer current I1 and the superconducting shielding layer I2 to change. The formula for judging whether the superconducting cable is quenched based on the current phase difference is:

[0079] |Φ1-Φ2|-180°<Φd

[0080] Among them, Φ1 is the initial phase angle of the current in the conductive layer, Φ2 is the initial phase angle of the current in the superconducting shielding layer, and Φd is the phase difference detection setting threshold.

[0081] In this embodiment, the quench fault in the cable can be identified by the current signal based on the change of the current magnitude and the current phase. However, by simply using numerical comparison, the quench fault can only be judged after the fault develops to a certain extent, such as after the quench area exceeds the set value. The fault cannot be identified in the early stage of the fault development, and the hysteresis is serious. The high-temperature superconducting cable is a new type of cable with high preparation cost and serious loss after damage. Therefore, it is necessary to predict the quench fault of the superconducting cable in advance.

[0082] ​After a local quench occurs, the operating parameter data of the superconducting cable begins to change gradually. By integrating the multi-dimensional data of the cable that can be used to determine the quench, including temperature, flow, pressure, current difference, phase change, and voltage, the quench fault can be identified earlier and maintenance can be carried out in advance.

[0083] Optionally, the deep learning network model can be a support vector machine, a random forest model, and a neural network model. In this embodiment, a BP neural network can be used to construct a deep learning network model. After feature extraction of the acquired data, the probability of superconducting cable quenching can be predicted based on the BP neural network.

[0084] Furthermore, before making predictions through the deep learning network model, the deep learning network model is trained, including the following steps:

[0085] Step S1, obtaining historical data of superconducting cable operation and performing preprocessing, wherein the historical data includes normal operation data and fault data;

[0086] The preprocessing includes preprocessing the singular values ​​of the signal, processing missing values, and normalizing the signal using a predictive mean filtering algorithm.

[0087] Step S2: convert the acquired historical data into feature vectors, abstract feature attributes from the historical interaction data, and generate a multi-dimensional feature vector set;

[0088] The process of vectorizing the data may adopt methods in the prior art, which will not be described in detail here.

[0089] Step S3: Based on the feature vector, cluster the historical interaction data according to the faults to obtain multiple clusters;

[0090] Step S4: for the clustering results of the quench fault, identify the corresponding abnormal characteristic values, and determine the average number N of abnormal characteristic values ​​in the fault clustering results as the number of fault characteristic values;

[0091] Among them, abnormal data means exceeding the normal operating value, and the percentage of data fluctuation exceeding the normal value is set as the abnormal value;

[0092] Step S5: for each cluster, the ratio of the number of abnormal characteristic values ​​to the number of fault characteristic values ​​is defined as the cluster risk probability, that is, the risk probability of a quench fault is predicted based on the data in the cluster, and the risk probability is used as a label;

[0093] Specifically, the abnormal characteristic values ​​in each cluster X are identified, and the number m of abnormal characteristic values ​​is determined; then the probability of a quench fault occurring in the operating state corresponding to cluster X is:

[0094]

[0095] Step S6: Taking the feature vector of each cluster as input, obtaining the risk probability of each cluster as output, and inputting it into the deep learning network model for training;

[0096] Step S7: Calculate the cross entropy loss between the risk probability predicted by the model and the actual risk, adjust the parameters of the deep learning network model until the prediction accuracy requirement is met, and obtain the trained deep learning network model.

[0097] Example 3

[0098] Based on Example 1, this embodiment provides a high-temperature superconducting cable monitoring system, including:

[0099] An acquisition module is configured to acquire operation data of the superconducting cable, including electrical quantity operation data and non-electrical quantity operation data;

[0100] Current data processing module: configured to calculate the current difference and current phase difference between the conductive layer and the shielding layer according to the acquired electrical quantity signal;

[0101] Probability prediction module: It is configured to build a deep learning network model, and transmit the calculated current difference and current phase difference between the conductive layer and the shielding layer, voltage value, and non-electrical quantity operation data as input to the deep learning network model to obtain the probability of superconducting cable quenching.

[0102] It should be noted here that the various modules in this embodiment correspond one-to-one to the various steps in Example 1, and the specific implementation process is the same, which will not be repeated here.

[0103] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

[0104] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A high temperature superconducting cable monitoring method, characterized in that: The steps include: Obtaining the operating data of superconducting cables, including electrical quantity operating data and non-electrical quantity operating data; Based on the acquired electrical quantity signal, the current difference and current phase difference between the conductive layer and the shielding layer are calculated; A deep learning network model is constructed, and the calculated current difference and current phase difference between the conductive layer and the shielding layer, three-phase voltage signal, and non-electrical operation data are used as input and transmitted to the deep learning network model to obtain the probability of superconducting cable quenching. Training the deep learning network model includes the following steps: Obtain historical data of superconducting cable operation and perform preprocessing; Convert the acquired historical data into feature vectors, abstract feature attributes from the historical interaction data, and generate a multi-dimensional feature vector set; Based on the feature vector, the historical interaction data is clustered according to the faults to obtain multiple clusters; According to the clustering results of the quench fault, the corresponding abnormal characteristic values ​​are identified, and the average number of abnormal characteristic values ​​in the fault clustering results is determined as the number of fault characteristic values; For each cluster, the ratio of the number of abnormal feature values ​​to the number of fault feature values ​​is defined as the cluster risk probability, and the risk probability is used as the label; The feature vector of each cluster is used as input, and the risk probability of each cluster is obtained as output, which is input into the deep learning network model for training; Calculate the predicted loss function and adjust the parameters of the deep learning network model until the prediction accuracy requirements are met to obtain the trained deep learning network model.

2. A high temperature superconducting cable monitoring method according to claim 1, characterized in that: The electrical quantity operation data includes the current signals of each layer of the superconducting cable and the three-phase voltage signals; the non-electrical quantity data includes the flow, temperature and pressure signals of liquid nitrogen.

3. A high temperature superconducting cable monitoring method according to claim 1, characterized in that: Abnormal data refers to data that exceeds normal operating values, and the percentage of data fluctuation that exceeds the normal value is considered abnormal.

4. A high temperature superconducting cable monitoring method according to claim 1, characterized in that: The loss function is specifically the cross entropy loss between the risk probability predicted by the model and the actual risk.

5. A high-temperature superconducting cable monitoring system, characterized in that: include: An acquisition module is configured to acquire operation data of the superconducting cable, including electrical quantity operation data and non-electrical quantity operation data; Current data processing module: configured to calculate the current difference and current phase difference between the conductive layer and the shielding layer according to the acquired electrical quantity signal; Probability prediction module: It is configured to build a deep learning network model, and transmit the calculated current difference and current phase difference between the conductive layer and the shielding layer, the three-phase voltage signal, and the non-electrical quantity operation data as input to the deep learning network model to obtain the probability of superconducting cable quenching; Training the deep learning network model includes the following steps: Obtain historical data of superconducting cable operation and perform preprocessing; Convert the acquired historical data into feature vectors, abstract feature attributes from the historical interaction data, and generate a multi-dimensional feature vector set; Based on the feature vector, the historical interaction data is clustered according to the faults to obtain multiple clusters; According to the clustering results of the quench fault, the corresponding abnormal characteristic values ​​are identified, and the average number of abnormal characteristic values ​​in the fault clustering results is determined as the number of fault characteristic values; For each cluster, the ratio of the number of abnormal feature values ​​to the number of fault feature values ​​is defined as the cluster risk probability, and the risk probability is used as the label; The feature vector of each cluster is used as input, and the risk probability of each cluster is obtained as output, which is input into the deep learning network model for training; Calculate the predicted loss function and adjust the parameters of the deep learning network model until the prediction accuracy requirements are met to obtain the trained deep learning network model.

6. A high temperature superconducting cable monitoring system, characterized in that: It includes a monitoring terminal, an electrical quantity detection device and a non-electrical quantity detection device which are communicatively connected to the monitoring terminal; The monitoring terminal is configured to execute the steps of a high-temperature superconducting cable monitoring method as described in any one of claims 1 to 4.

7. A high temperature superconducting cable monitoring system as claimed in claim 6, characterized in that: The non-electrical quantity detection device includes a liquid nitrogen temperature detection device, including temperature sensors arranged at both ends of the superconducting cable, at the inlet and outlet of the liquid nitrogen of the superconducting cable, and temperature sensors arranged at both ends of the three-phase conductive layer of the superconducting cable.

8. A high temperature superconducting cable monitoring system as claimed in claim 6, characterized in that: Non-electrical quantity detection devices, including pressure detection devices and flow detection devices; The pressure detection device includes a pressure sensor arranged at the inlet and outlet of the liquid nitrogen of the superconducting cable; The flow detection device comprises a flow meter arranged at the inlet and outlet of the liquid nitrogen of the superconducting cable.

9. A high temperature superconducting cable monitoring system as claimed in claim 6, characterized in that: The electrical quantity detection device includes a device for detecting current and voltage signals of each layer of the superconducting cable; The current detection device adopts a current transformer, which is arranged at both ends of the conductive layer of each phase and at the shielding layer lead-out end of the superconducting cable terminal; The voltage detection device adopts a voltage transformer.

Images