Relay protection hidden fault identification method and device

Abstract

The invention relates to the technical field of power grid monitoring, and discloses a relay protection hidden fault identification method and device, and the method comprises the steps: carrying out the real-time monitoring of a relay protection system, and judging whether the relay protection system exceeds a preset dynamic threshold value or not through verification data; if yes, determining a relay protection fault condition; fusing multi-source data according to the relay protection fault condition to obtain fused data; identifying the hidden fault of the relay protection system according to the fusion data; and determining a maintenance scheme of the relay protection system according to the hidden fault identification result. Through cooperative application of dynamic threshold adjustment and multi-source data fusion, the hidden fault identification method integrating online monitoring, intelligent identification and prediction is constructed, the accuracy and timeliness of hidden fault identification are improved, and safe operation of a power grid is guaranteed.

Description

Technical Field

[0001] This application relates to the field of power grid monitoring technology, specifically to a method and device for identifying latent faults in relay protection. Background Technology

[0002] Power system relay protection is a crucial line of defense and an important component for ensuring the safe and stable operation of the power system, preventing the escalation of faults, and limiting the scope of equipment damage. However, some latent faults often exist in relay protection devices and related secondary circuits, which are difficult to detect in a timely manner using conventional monitoring methods. These faults usually do not show any trace under normal system operation and are not easily noticed or detected by operators. However, once the system encounters disturbances or changes in specific operating modes, they may be triggered, causing the protection devices to malfunction or fail to operate, resulting in the failure of the originally designed power system protection measures. Such failures may not only trigger chain reactions, causing local or even larger-scale power outages, but also have serious negative impacts and potential risks on the overall stability, power supply reliability, and operational safety of the power system. Therefore, in-depth analysis and research on latent faults in relay protection are of great significance, and have significant theoretical value and urgent practical significance for improving the power system's security defense capabilities and ensuring reliable power supply from the grid.

[0003] However, most existing technologies are limited to offline analysis of single-type data (such as electrical quantity data) or simple alarms based on fixed thresholds, lacking the ability to deeply integrate and dynamically analyze multi-source heterogeneous monitoring data. Specifically, they have the following main limitations:

[0004] (1) Insufficient data utilization: Existing methods usually rely only on traditional electrical quantity information (such as current and voltage) or single status monitoring signals for analysis, failing to make full use of the rich operating status information of the relay protection device itself (such as temperature, power supply voltage, port light intensity, etc.) as well as environmental parameters (such as ambient temperature and humidity), historical maintenance records and other multi-dimensional data sources for collaborative analysis. This results in a single dimension of perception of latent faults, making it difficult to capture early and weak fault symptoms.

[0005] (2) Rigid threshold setting: Static thresholds are commonly used for limit judgment. Static thresholds cannot adapt to changes in equipment operating conditions (such as load fluctuations and changes in ambient temperature) and equipment aging process, which can easily lead to false alarms (normal fluctuations are misjudged as abnormal) or missed alarms (actual abnormalities fail to trigger alarms due to improper threshold settings), reducing the accuracy and reliability of hidden fault identification.

[0006] (3) Limited analytical dimensions: The analysis of monitoring data is often limited to simple limit checks, lacking in-depth mining and correlation analysis of dynamic characteristics such as data change trends (e.g., slow degradation) and sudden change characteristics (e.g., instantaneous anomalies). This makes it difficult for the system to distinguish between normal fluctuations and early signals of potential faults, and also makes it unable to effectively identify complex and coupled fault modes.

[0007] (4) Lack of dynamic response and prediction capabilities: Existing technologies focus on post-event analysis or periodic inspections, lacking the ability to conduct online monitoring and identify latent faults in real time based on multi-source data. At the same time, there is a lack of effective means to predict when latent faults may occur, making it difficult to adjust maintenance strategies in a timely manner to prevent them from happening.

[0008] (5) Limited level of intelligence: The identification of hidden faults relies on human experience and judgment, which is inefficient and inconsistent. Summary of the Invention

[0009] This application aims to at least address the technical problems in related technologies, such as insufficient data utilization, rigid threshold settings, single analysis dimensions, and lack of dynamic response and prediction capabilities, which lead to low accuracy and slow response speed in the detection of latent faults in relay protection.

[0010] To address the aforementioned technical problems, embodiments of this application provide a method for identifying latent faults in relay protection, comprising:

[0011] The relay protection system is monitored in real time, and the verification data is used to determine whether the relay protection system exceeds the preset dynamic threshold.

[0012] If so, determine the fault condition of the relay protection;

[0013] Based on the relay protection fault conditions, multi-source data are fused to obtain fused data;

[0014] The latent faults of the relay protection system are identified based on the fused data;

[0015] Based on the results of the latent fault identification, a maintenance plan for the relay protection system is determined.

[0016] In some embodiments, the preset dynamic threshold includes at least one of an over-limit dynamic threshold, a sudden change dynamic threshold, and a trend change dynamic threshold.

[0017] Determining whether the relay protection system exceeds a preset dynamic threshold by verifying data includes:

[0018] Determine whether the current monitoring value of the relay protection system exceeds the over-limit dynamic threshold, so as to determine whether the relay protection system has exceeded the limit;

[0019] Determine whether the difference between the current monitoring value and the monitoring value at the previous historical storage point of the relay protection system exceeds the sudden change dynamic threshold, so as to determine whether the relay protection system has undergone a sudden change;

[0020] To determine whether a trend change has occurred in the relay protection system, it is necessary to determine whether the difference between the average monitoring value of the most recent cycle and the average monitoring value of the previous cycle exceeds the trend change dynamic threshold.

[0021] In some embodiments, the method further includes:

[0022] The preset dynamic threshold is corrected based on environmental parameters.

[0023] In some embodiments, the relay protection fault conditions include only exceeding the limit, only abrupt change, only a trend change, both exceeding the limit and abrupt change, both exceeding the limit and a trend change, both abrupt change and a trend change, and both exceeding the limit, abrupt change and a trend change.

[0024] In some embodiments, identifying latent faults in the relay protection system based on the fused data includes:

[0025] Normalize the data from each multi-source source;

[0026] Construct a one-dimensional multi-source data matrix based on the normalized multi-source data;

[0027] Weight matrices are constructed according to the different relay protection fault scenarios.

[0028] The one-dimensional multi-source data matrix and the weight matrix are input into a preset relay protection latent fault identification model to identify latent faults.

[0029] In some embodiments, determining the maintenance plan for the relay protection system based on the results of latent fault identification includes:

[0030] If a hidden fault is identified, it will be checked and repaired manually.

[0031] If no latent fault is identified, the time point at which the relay protection system may experience a latent fault is predicted, and the next maintenance schedule for the relay protection system is determined based on the prediction results.

[0032] In some embodiments, predicting the time points at which the relay protection system may experience latent faults includes:

[0033] Obtain multi-source data corresponding to the relay protection fault conditions;

[0034] The multi-source data is input into a preset CNNs-LSTM prediction model to predict the time points at which the relay protection system may experience latent faults; wherein, the CNNs-LSTM prediction model includes multiple CNNs, each corresponding to input data under different relay protection fault scenarios.

[0035] In some embodiments, determining the next maintenance schedule time for the relay protection system based on the prediction results includes:

[0036] Compare the predicted time points with the scheduled maintenance time points;

[0037] If the predicted time point is earlier than the scheduled maintenance time point, the scheduled maintenance time point will be adjusted to be earlier than the predicted time point; otherwise, the scheduled maintenance time point will not be changed.

[0038] In some embodiments, manual verification and maintenance include:

[0039] If manual verification confirms the existence of a latent fault, the relay protection system shall be repaired.

[0040] Otherwise, the manual verification record will be sent to the relay protection latent fault identification model for optimization.

[0041] This application also provides a relay protection latent fault identification device, including:

[0042] The monitoring module is configured to monitor the relay protection system in real time and determine whether the relay protection system exceeds a preset dynamic threshold by verifying the data.

[0043] If the module is configured to be active, determine the relay protection fault conditions.

[0044] The multi-source fusion module is configured to fuse multi-source data according to the relay protection fault conditions to obtain fused data;

[0045] The identification module is configured to identify latent faults in the relay protection system based on the fused data;

[0046] The maintenance module is configured to determine the maintenance plan for the relay protection system based on the results of hidden fault identification.

[0047] This application also provides an electronic device, which includes at least a processor and a memory. The memory stores a computer program, and the processor implements the above-described relay protection latent fault identification method when executing the computer program in the memory.

[0048] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying latent faults in relay protection.

[0049] The relay protection latent fault identification method and apparatus provided in this application embodiment monitors the relay protection system in real time and determines whether the relay protection system exceeds a preset dynamic threshold through verification data; if so, it determines the relay protection fault condition; it fuses multi-source data according to the relay protection fault condition to obtain fused data; it identifies latent faults in the relay protection system based on the fused data; and it determines the maintenance plan for the relay protection system based on the latent fault identification result. This application constructs a relay protection latent fault identification method integrating online monitoring, intelligent identification, and prediction through the synergistic application of dynamic threshold adjustment and multi-source data fusion, significantly improving the detection accuracy and response speed of relay protection latent faults and ensuring the safe operation of the power grid. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of a relay protection latent fault identification method according to an embodiment of this application;

[0052] Figure 2 This is another flowchart of the relay protection latent fault identification method according to an embodiment of this application;

[0053] Figure 3 This is a schematic diagram of the CNNs-LSTM prediction model structure of the relay protection latent fault identification method according to an embodiment of this application;

[0054] Figure 4 This is a schematic diagram of the structure of the relay protection latent fault identification device according to an embodiment of this application. Detailed Implementation

[0055] Various embodiments and features of this application are described herein with reference to the accompanying drawings.

[0056] It should be understood that various modifications can be made to the embodiments described herein. Therefore, the above description should not be considered as limiting, but merely as an example of embodiments. Other modifications within the scope and spirit of this application will be apparent to those skilled in the art.

[0057] The accompanying drawings, which are included in and form part of this specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0058] These and other features of this application will become apparent from the following description of preferred forms of embodiments given as non-limiting examples, with reference to the accompanying drawings.

[0059] It should also be understood that although this application has been described with reference to some specific examples, those skilled in the art can certainly implement many other equivalent forms of this application, which have the features described in the claims and are therefore all within the scope of protection defined herein.

[0060] The above and other aspects, features and advantages of this application will become more apparent when taken in conjunction with the accompanying drawings and in view of the following detailed description.

[0061] Specific embodiments of this application are described thereafter with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of this application, which can be implemented in various ways. Well-known and / or repeated functions and structures are not described in detail to avoid unnecessary or redundant details that could obscure the application. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as the basis and representative basis for the claims to teach those skilled in the art to use this application in a variety of substantially any suitable detailed structures.

[0062] This specification may use the phrases “in one embodiment,” “in another embodiment,” “in yet another embodiment,” or “in other embodiments,” all of which may refer to one or more of the same or different embodiments according to this application.

[0063] Example 1

[0064] Figure 1 and Figure 2 A flowchart illustrating a relay protection latent fault identification method according to an embodiment of this application is shown. Figure 1 As shown in the embodiments of this application, the relay protection latent fault identification method includes:

[0065] S101: Real-time monitoring of the relay protection system, and determination of whether the relay protection system exceeds the preset dynamic threshold through verification data.

[0066] First, the status of the relay protection system is monitored in real time to obtain the monitoring data of the relay protection system. Then, the obtained monitoring data of the relay protection system is used as verification data and compared and judged based on a preset dynamic threshold.

[0067] Optionally, in step S101, the preset dynamic threshold includes at least one of an over-limit dynamic threshold, a sudden change dynamic threshold, and a trend change dynamic threshold.

[0068] Determining whether the relay protection system exceeds a preset dynamic threshold by verifying data includes:

[0069] S201: Determine whether the current monitoring value of the relay protection system exceeds the over-limit dynamic threshold, so as to determine whether the relay protection system has exceeded the limit;

[0070] S202: Determine whether the difference between the current monitoring value of the relay protection system and the monitoring value of the previous historical storage point exceeds the sudden change dynamic threshold, so as to determine whether the relay protection system has a sudden change;

[0071] S203: Determine whether the difference between the average value of the most recent cycle monitoring and the average value of the previous cycle monitoring of the relay protection system exceeds the trend change dynamic threshold, so as to determine whether the relay protection system has undergone a trend change.

[0072] Among these, "exceeding the limit" refers to the instantaneous value of a measured parameter (such as current, voltage, or device temperature) exceeding the dynamic threshold set for that parameter at the current monitoring moment. "Sudden change" focuses on the drastic nature of parameter changes, indicating that the difference between the current monitored value and the previously recorded historical monitored value exceeds the dynamic threshold set for this type of change. "Trend change" focuses on the overall trend of a parameter's change over a period of time, rather than a single point; specifically defined as: the difference between the average monitoring value of the most recent monitoring period and the average monitoring value of the previous monitoring period exceeding the dynamic threshold set for this trend change.

[0073] Specifically, if the judgment result satisfies at least one of the above-mentioned over-limit, sudden change, and trend change, the relay protection system exceeds the preset dynamic threshold and needs to be carried out for subsequent maintenance.

[0074] S102: If so, determine the relay protection fault condition.

[0075] If the relay protection system exceeds the preset dynamic threshold in this step, it is necessary to further determine the relay protection fault condition.

[0076] Optionally, in step S102, the relay protection fault conditions include only exceeding the limit, only abrupt change, only a trend change, both exceeding the limit and abrupt change, both exceeding the limit and a trend change, both abrupt change and a trend change, and both exceeding the limit, abrupt change and a trend change.

[0077] Specifically, this step categorizes complex monitoring phenomena into seven specific scenarios, each corresponding to a unique maintenance plan.

[0078] S103: Based on the relay protection fault conditions, the multi-source data are fused to obtain fused data.

[0079] Specifically, the seven relay protection fault scenarios are denoted as Scenarios 1 to 7, as shown in Table 1. The multi-source data required for the seven relay protection fault scenarios differs, and is divided into general data and characteristic data:

[0080] Table 1. Multi-source data for different relay protection fault scenarios

[0081]

[0082] General data refers to data that is needed in any situation, including electrical quantity data, condition monitoring data, and historical data. Characteristic data refers to data that is needed in specific situations, specifically different combinations of general data.

[0083] In this step, multi-source data under different relay protection fault conditions are obtained according to Table 1. By matching the data with specific feature combinations, a data foundation is laid for the subsequent accurate identification of different relay protection fault conditions.

[0084] S104: Identify latent faults in the relay protection system based on the fused data.

[0085] In this step, the obtained fused data needs to be processed and identified using a pre-defined relay protection latent fault identification model to obtain the latent fault identification results. The relay protection latent fault identification model is a convolutional neural network model, which provides a powerful tool for intelligent identification of relay protection latent faults through a hierarchical automatic feature extraction mechanism.

[0086] S105: Based on the results of the latent fault identification, determine the maintenance plan for the relay protection system.

[0087] In this step, based on the results of the latent fault identification, i.e. whether a latent fault has been identified, the maintenance plan for the relay protection system is determined, and differentiated maintenance decisions are made.

[0088] The relay protection latent fault identification method provided in this application embodiment monitors the relay protection system in real time and determines whether the relay protection system exceeds a preset dynamic threshold through verification data. If so, the relay protection fault situation is determined. Multi-source data is fused according to the relay protection fault situation to obtain fused data. Latent faults in the relay protection system are identified based on the fused data. Based on the latent fault identification results, a maintenance plan for the relay protection system is determined. Through the synergistic application of dynamic threshold adjustment and multi-source data fusion, a relay protection latent fault identification method integrating online monitoring, intelligent identification, and prediction is constructed, significantly improving the detection accuracy and response speed of relay protection latent faults and ensuring the safe operation of the power grid.

[0089] In addition, this application improves the intelligence level of latent fault identification by automatically adapting to different fault characteristic patterns (such as single limit exceedance, sudden change, trend change and their combination), fusing multi-source data, and using intelligent analysis models (such as deep learning) to identify latent faults.

[0090] In some embodiments, step S101, the method further includes:

[0091] S1011: Correct the preset dynamic threshold based on environmental parameters.

[0092] Specifically, the dynamic threshold is dynamically corrected by incorporating environmental parameters (such as temperature and humidity), and the calculation formula is as follows:

[0093]

[0094] in, and These represent the thresholds before and after dynamic correction, respectively. Indicates weight, This represents a function representing changes in environmental parameters. This represents the i-th environment parameter.

[0095] In this embodiment, when the ambient temperature rises, the dynamic threshold will also be dynamically adjusted accordingly, which effectively prevents false alarms in hot weather and greatly improves the accuracy of the judgment.

[0096] In some embodiments, step S104, identifying latent faults in the relay protection system based on the fused data, includes:

[0097] S1041: Normalize the data from each multi-source source;

[0098] S1042: Construct a one-dimensional multi-source data matrix based on the normalized multi-source data;

[0099] S1043: Construct weight matrices according to the different relay protection fault conditions;

[0100] S1044: Input the one-dimensional multi-source data matrix and the weight matrix into a preset relay protection latent fault identification model to identify latent faults.

[0101] Specifically, the "min-max" normalization method is first used to normalize the multi-source data. The "min-max" normalization method linearly transforms the data to a specific range, such as the interval [0, 1].

[0102] Secondly, the normalized multi-source data is constructed into a one-dimensional multi-source data matrix:

[0103]

[0104] in, Represents a one-dimensional multi-source data matrix. This represents the nth normalized multi-source data.

[0105] Then, based on different relay protection fault scenarios, fixed-size [structures] are constructed respectively. Weight matrix:

[0106]

[0107] in, Indicates the weight magnitude. This indicates the amount of multi-source data under different circumstances. This is a fixed value set by an individual.

[0108] Finally, the one-dimensional multi-source data matrix and weight matrix The input is fed into the relay protection latent fault identification model of the convolutional neural network, where the input of the latent fault identification model is... The output indicates whether a hidden fault exists.

[0109] In this embodiment, by introducing a weight matrix, key features are emphasized according to the specific relay protection fault situation, and deep learning models such as convolutional neural networks are used to automatically learn complex fault feature patterns, thereby significantly improving the ability to identify hidden faults.

[0110] In some embodiments, step S105, determining the maintenance plan for the relay protection system based on the latent fault identification result, includes:

[0111] S1051: If a latent fault is identified, it shall be checked and repaired manually;

[0112] S1052: If no latent fault is identified, the time point at which the relay protection system may develop a latent fault is predicted, and the next maintenance schedule time point of the relay protection system is determined based on the prediction results.

[0113] Specifically, based on the results of latent fault identification, if a latent fault is identified, further manual judgment is required to determine whether a latent fault exists and to carry out repairs, thereby achieving closed-loop optimization. If no latent fault is identified, it is necessary to predict the time points at which the relay protection system may experience latent faults, and determine the next maintenance schedule for the relay protection system based on the prediction results, thereby achieving predictive maintenance and improving the intelligence level and reliability of the relay protection system.

[0114] In some embodiments, step S1052, predicting the time points at which the relay protection system may experience latent faults, includes:

[0115] S301: Obtain multi-source data corresponding to the relay protection fault conditions;

[0116] S302: Input the multi-source data into a preset CNNs-LSTM prediction model to predict the time points when the relay protection system may experience latent faults; wherein, the CNNs-LSTM prediction model includes multiple CNNs, each CNN corresponding to input data under different relay protection fault scenarios.

[0117] The CNNs-LSTM prediction model, also known as the Convolutional Long Short-Term Memory Network (often simply referred to as the CNN-LSTM model), is an advanced deep learning hybrid architecture that cleverly combines the advantages of convolutional neural networks in spatial feature extraction with the strengths of long short-term memory networks in time series modeling. It is very suitable for handling time series data prediction tasks that have both local correlations and long-term dependencies.

[0118] In this embodiment, firstly, multi-source time series data corresponding to different relay protection fault conditions are obtained, as shown in Table 2:

[0119] Table 2. Multi-source time series data for different relay protection fault scenarios

[0120]

[0121] Then, the acquired multi-source time series data is normalized using the "min-max" normalization method. Finally, the multi-source time series data is input into a pre-defined CNNs-LSTM prediction model to predict the time points at which latent faults may occur in the relay protection system.

[0122] Among them, such as Figure 3As shown, the CNNs-LSTM prediction model structurally comprises a parallelized feature extraction front-end and a sequence prediction back-end. At the input end, three CNN modules act as feature extractors, each adapted to different relay protection fault scenarios, ensuring that various types of input data can be effectively encoded. Within the feature space, these CNNs map heterogeneous input data into feature matrices of the same size, laying the foundation for subsequent fusion analysis. At the output end, the LSTM model learns the evolutionary patterns of these deeply refined feature sequences and ultimately achieves accurate prediction of the timing of future latent faults.

[0123] In some embodiments, step S1052, determining the next maintenance schedule time for the relay protection system based on the prediction results, includes:

[0124] S401: Compare the predicted time point with the maintenance schedule time point;

[0125] S402: If the predicted result time point is earlier than the maintenance schedule time point, then the maintenance schedule time point is adjusted to be earlier than the predicted result time point; otherwise, the maintenance schedule time point is not changed.

[0126] In this embodiment, the predicted time point is compared with the maintenance schedule time point. If the predicted time point is earlier than the maintenance schedule time point, it indicates that there is a risk of "unplanned" failure of the relay protection device. The original maintenance plan cannot cover this risk and intervention must be carried out in advance. In this case, the maintenance schedule time point needs to be adjusted to be earlier than the predicted time point. Otherwise, it indicates that the original maintenance plan is sufficient and timely, and the relay protection device is very likely to remain stable before the next planned maintenance. In this case, the maintenance schedule time point will not be changed.

[0127] The prediction result time point refers to the future time point predicted by the CNNs-LSTM prediction model for the possible latent faults of the relay protection device. The maintenance schedule time point refers to the next maintenance time pre-set according to established procedures (such as the annual maintenance plan).

[0128] In some embodiments, step S1051, manual verification and maintenance, includes:

[0129] S501: If manual verification confirms the existence of a latent fault, then the relay protection system shall be repaired;

[0130] S502: Otherwise, send the manual verification record to the relay protection hidden fault identification model for optimization.

[0131] In this embodiment, if manual verification determines the existence of a latent fault, the relay protection system is repaired to restore its normal function and directly ensure power grid safety. If manual verification fails to detect a latent fault, indicating an error in the latent fault identification, the complete verification record is used as an optimization dataset and fed back to the latent fault identification model for optimization. Furthermore, if manual verification fails to detect a latent fault, the process proceeds to step S1052 to predict the latent fault's time point.

[0132] In summary, this application integrates heterogeneous information from multiple sources, including electrical quantities, condition monitoring data (such as temperature, power supply voltage, and port light intensity), and historical maintenance records, to achieve in-depth collaborative analysis of multi-dimensional data. A dynamic threshold mechanism is introduced to correct thresholds for exceeding limits, sudden changes, and trend changes in real time based on environmental parameters, effectively avoiding false alarms or missed alarms caused by static threshold settings and improving identification accuracy and reliability. Based on seven latent fault scenarios (including exceeding limits, sudden changes, trend changes, and their combinations), multi-dimensional dynamic features are mined and correlated to accurately distinguish between normal fluctuations and potential fault modes. Simultaneously, the system possesses online monitoring and real-time identification capabilities, and combines a CNNs-LSTM prediction model to accurately predict the occurrence time of latent faults, thereby optimizing maintenance strategies. The method proposed in this application significantly improves the detection accuracy, response speed, and prediction capability of latent faults in relay protection, providing strong support for the safe and stable operation of power systems.

[0133] Example 2

[0134] Figure 4 This is a schematic diagram of the structure of a relay protection latent fault identification device according to an embodiment of this application. Figure 4 As shown in the figure, this application provides a relay protection latent fault identification device, including:

[0135] Monitoring module 10 is configured to monitor the relay protection system in real time and determine whether the relay protection system exceeds a preset dynamic threshold by verifying the data.

[0136] If module 20 is configured to be set to "yes", determine the relay protection fault condition.

[0137] The fusion module 30 is configured to fuse multi-source data according to the relay protection fault conditions to obtain fused data;

[0138] The identification module 40 is configured to identify latent faults in the relay protection system based on the fused data.

[0139] The maintenance module 50 is configured to determine the maintenance plan for the relay protection system based on the results of hidden fault identification.

[0140] In some embodiments, the monitoring module 10 is further configured to:

[0141] Determining whether the relay protection system exceeds a preset dynamic threshold by verifying data includes:

[0142] Determine whether the current monitoring value of the relay protection system exceeds the over-limit dynamic threshold, so as to determine whether the relay protection system has exceeded the limit;

[0143] Determine whether the difference between the current monitoring value and the monitoring value at the previous historical storage point of the relay protection system exceeds the sudden change dynamic threshold, so as to determine whether the relay protection system has undergone a sudden change;

[0144] To determine whether a trend change has occurred in the relay protection system, it is necessary to determine whether the difference between the average monitoring value of the most recent cycle and the average monitoring value of the previous cycle exceeds the trend change dynamic threshold.

[0145] In some embodiments, the monitoring module 10 is further configured to:

[0146] The preset dynamic threshold is corrected based on environmental parameters.

[0147] In some embodiments, the determining module 20 is further configured to:

[0148] The relay protection fault scenarios include only exceeding the limit, only abrupt change, only trend change, both exceeding the limit and abrupt change, both exceeding the limit and trend change, both abrupt change and trend change, and both exceeding the limit, abrupt change and trend change.

[0149] In some embodiments, the identification module 40 is further configured to:

[0150] Normalize the data from each multi-source source;

[0151] Construct a one-dimensional multi-source data matrix based on the normalized multi-source data;

[0152] Weight matrices are constructed according to the different relay protection fault scenarios.

[0153] The one-dimensional multi-source data matrix and the weight matrix are input into a preset relay protection latent fault identification model to identify latent faults.

[0154] In some embodiments, the maintenance module 50 is further configured to:

[0155] If a hidden fault is identified, it will be checked and repaired manually.

[0156] If no latent fault is identified, the time point at which the relay protection system may experience a latent fault is predicted, and the next maintenance schedule for the relay protection system is determined based on the prediction results.

[0157] In some embodiments, the maintenance module 50 is further configured to:

[0158] Obtain multi-source data corresponding to the relay protection fault conditions;

[0159] The multi-source data is input into a preset CNNs-LSTM prediction model to predict the time points at which the relay protection system may experience latent faults; wherein, the CNNs-LSTM prediction model includes multiple CNNs, each corresponding to input data under different relay protection fault scenarios.

[0160] In some embodiments, the maintenance module 50 is further configured to:

[0161] Compare the predicted time points with the scheduled maintenance time points;

[0162] If the predicted time point is earlier than the scheduled maintenance time point, the scheduled maintenance time point will be adjusted to be earlier than the predicted time point; otherwise, the scheduled maintenance time point will not be changed.

[0163] In some embodiments, the maintenance module 50 is further configured to:

[0164] If manual verification confirms the existence of a latent fault, the relay protection system shall be repaired.

[0165] Otherwise, the manual verification record will be sent to the relay protection latent fault identification model for optimization.

[0166] The relay protection latent fault identification device provided in this application corresponds to the relay protection latent fault identification method in the above embodiments. Any option in the relay protection latent fault identification method embodiments is also applicable to the embodiments of the relay protection latent fault identification device, and will not be repeated here.

[0167] Example 3

[0168] This application also provides an electronic device, which includes at least a memory and a processor. The memory stores a computer program, and the processor implements the above-described relay protection latent fault identification method when executing the computer program in the memory.

[0169] In some embodiments, the processor executing a computer program may be a processing device that includes one or more general-purpose processing devices, such as a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), etc. More specifically, the processor may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor that runs other instruction sets, or a processor that runs a combination of instruction sets. The processor may also be one or more special-purpose processing devices, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), system-on-a-chip (SoCs), etc.

[0170] The memory may be a read-only memory (ROM), random access memory (RAM), phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), electrically erasable programmable read-only memory (EEPROM), other types of random access memory (RAM), flash drives or other forms of flash memory, cache, registers, static memory, optical disc read-only memory (CD-ROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape cassette or other magnetic storage devices, or any other possible non-transitory medium used to store information or instructions that can be accessed by computer equipment.

[0171] The electronic devices in this application embodiment may include, but are not limited to, fixed terminal devices such as servers, desktop computers, and digital TVs, as well as mobile terminal devices such as in-vehicle devices (e.g., head-up displays), handheld devices (e.g., mobile phones, tablets, etc.), and wearable devices (e.g., smartwatches, smart bracelets, etc.).

[0172] Example 4

[0173] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying latent faults in relay protection.

[0174] The computer-readable storage medium in this application embodiment can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. In this application embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device; for example, it can be the aforementioned memory.

[0175] The computer programs of embodiments of this application can be organized into one or more computer-executable components or modules. Various aspects of this application can be implemented with any number and combination of such components or modules. For example, aspects of this application are not limited to the specific computer-executable instructions or specific components or modules shown in the drawings and described herein. Other embodiments may include different computer-executable instructions or components having more or fewer functions than those shown and described herein.

[0176] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for identifying latent faults in relay protection, characterized in that, include: The relay protection system is monitored in real time, and the verification data is used to determine whether the relay protection system exceeds the preset dynamic threshold. If so, determine the fault condition of the relay protection; Based on the relay protection fault conditions, multi-source data are fused to obtain fused data; The latent faults of the relay protection system are identified based on the fused data; Based on the results of the latent fault identification, a maintenance plan for the relay protection system is determined.

2. The relay protection latent fault identification method according to claim 1, characterized in that, The preset dynamic threshold includes at least one of the following: an over-limit dynamic threshold, a sudden change dynamic threshold, and a trend change dynamic threshold. Determining whether the relay protection system exceeds a preset dynamic threshold by verifying data includes: Determine whether the current monitoring value of the relay protection system exceeds the over-limit dynamic threshold, so as to determine whether the relay protection system has exceeded the limit; Determine whether the difference between the current monitoring value and the monitoring value at the previous historical storage point of the relay protection system exceeds the sudden change dynamic threshold, so as to determine whether the relay protection system has undergone a sudden change; To determine whether a trend change has occurred in the relay protection system, it is necessary to determine whether the difference between the average monitoring value of the most recent cycle and the average monitoring value of the previous cycle exceeds the trend change dynamic threshold.

3. The relay protection latent fault identification method according to claim 1, characterized in that, The method further includes: The preset dynamic threshold is corrected based on environmental parameters.

4. The relay protection latent fault identification method according to claim 1, characterized in that, The relay protection fault scenarios include only exceeding the limit, only abrupt change, only trend change, both exceeding the limit and abrupt change, both exceeding the limit and trend change, both abrupt change and trend change, and both exceeding the limit, abrupt change and trend change.

5. The relay protection latent fault identification method according to claim 1, characterized in that, Identifying latent faults in the relay protection system based on the fused data includes: Normalize the data from each multi-source source; Construct a one-dimensional multi-source data matrix based on the normalized multi-source data; Weight matrices are constructed according to the different relay protection fault scenarios. The one-dimensional multi-source data matrix and the weight matrix are input into a preset relay protection latent fault identification model to identify latent faults.

6. The relay protection latent fault identification method according to claim 1, characterized in that, Based on the results of latent fault identification, a maintenance plan for the relay protection system is determined, including: If a hidden fault is identified, it will be checked and repaired manually. If no latent fault is identified, the time point at which the relay protection system may experience a latent fault is predicted, and the next maintenance schedule for the relay protection system is determined based on the prediction results.

7. The relay protection latent fault identification method according to claim 6, characterized in that, Predicting the time points at which latent faults may occur in the relay protection system, including: Obtain multi-source data corresponding to the relay protection fault conditions; The multi-source data is input into a preset CNNs-LSTM prediction model to predict the time points at which the relay protection system may experience latent faults; wherein, the CNNs-LSTM prediction model includes multiple CNNs, each corresponding to input data under different relay protection fault scenarios.

8. The relay protection latent fault identification method according to claim 6, characterized in that, The timing of the next maintenance schedule for the relay protection system is determined based on the prediction results, including: Compare the predicted time points with the scheduled maintenance time points; If the predicted time point is earlier than the scheduled maintenance time point, the scheduled maintenance time point will be adjusted to be earlier than the predicted time point; otherwise, the scheduled maintenance time point will not be changed.

9. The relay protection latent fault identification method according to claim 6, characterized in that, Manual verification and repair include: If manual verification confirms the existence of a latent fault, the relay protection system shall be repaired. Otherwise, the manual verification record will be sent to the relay protection latent fault identification model for optimization.

10. A relay protection latent fault identification device, characterized in that, include: The monitoring module is configured to monitor the relay protection system in real time and determine whether the relay protection system exceeds a preset dynamic threshold by verifying the data. If the module is configured to be active, determine the relay protection fault conditions. The multi-source fusion module is configured to fuse multi-source data according to the relay protection fault conditions to obtain fused data; The identification module is configured to identify latent faults in the relay protection system based on the fused data; The maintenance module is configured to determine the maintenance plan for the relay protection system based on the results of hidden fault identification.

Images