A trip warning management system for substation equipment based on fault signal classification and recognition

Abstract

The present invention relates to the technical field of substation equipment management, and specifically discloses a substation equipment tripping warning management system based on fault signal classification and identification. The system comprises: a substation fault knowledge graph module, a collection data analysis module, a meteorological data module, a substation equipment tripping intelligent perception module, and a fault warning model module. The present invention ensures timely warning when the substation equipment trips, which is conducive to improving the fault handling efficiency and accuracy of the substation equipment. The present invention solves the need to accurately mine equipment tripping information from massive equipment operation data in order to efficiently carry out subsequent fault finding, integrates multiple information sources, conducts in-depth model mechanism research and data mining, and realizes intelligent perception of substation equipment tripping.

Description

Technical Field

[0001] The present invention relates to the technical field of substation equipment management, and in particular to a substation equipment tripping early warning management system based on fault signal classification and identification. Background Art

[0002] Substation equipment has the functions of converting, transmitting and distributing electric energy. The tripping of substation equipment not only brings losses to power supply companies, but also has a huge impact on residents' daily lives and the production of industrial enterprises, seriously affecting the customer experience of power supply companies. Therefore, whether the tripping of substation equipment can be accurately warned and the cause of the fault can be diagnosed in time will directly affect the reliability of power supply of the power grid and the safe operation of the system.

[0003] The existing technology can meet the current requirements to a certain extent in terms of real-time perception and accurate judgment of equipment tripping, but there are still certain defects, which are specifically reflected in:

[0004] (1) Existing technologies include research on intelligent perception of substation equipment tripping and fault diagnosis technology based on time series data analysis, but research on automatic data tracking and intelligent linkage based on abnormal defects of station equipment is still blank.

[0005] (2) Most existing technologies are limited to fault classification, and rarely involve the study of fault location. Most fault diagnosis methods are slightly lacking in fault location and are unable to accurately diagnose power transformer faults. In addition, most current technologies rely on massive amounts of equipment operating condition data for diagnosis, and are not linked to camera detection data and meteorological data for analysis, making it difficult to accurately identify the cause of equipment tripping failures. Summary of the invention

[0006] In order to overcome the shortcomings of the background technology, the present invention provides a substation tripping warning management system based on fault signal classification and identification, which can effectively solve the problems involved in the above background technology.

[0007] The technical solution adopted by the present invention is:

[0008] A substation equipment trip warning management system based on fault signal classification and identification, comprising: a substation fault knowledge graph module, a collection data analysis module, a meteorological data module, a video monitoring module and a fault warning model module; the substation fault knowledge graph module is connected with the collection data analysis module and the video monitoring module; the collection data analysis module, the video monitoring module and the meteorological data module are connected with the fault warning model module;

[0009] Substation fault knowledge graph module: By abstracting the physical structure of substation equipment, sorting out the attributes of existing substation equipment and historical fault analysis, constructing a substation fault knowledge graph based on a graph database, and establishing a substation fault entity, attribute, and relationship model;

[0010] Data collection and analysis module: Based on the real-time operating conditions of substation equipment, it can provide real-time perception of substation equipment operating data and abnormal warning reminders;

[0011] Meteorological data module: associates meteorological information with substation equipment, collects various weather conditions and forecasts disasters, and detects climate anomalies in substation equipment;

[0012] Video monitoring module: detects anomalies in monitoring videos through the spatiotemporal attention mechanism and performs environmental anomaly detection on substation equipment;

[0013] Fault warning model module: Calculate the equipment tripping probability based on the real-time operation data analysis results of the substation equipment, the video monitoring data of the substation equipment, and the meteorological data collection results, and perform analysis and warning.

[0014] Furthermore, the established substation fault entity, attribute, and relationship model includes: entity information, attribute information, and relationship definitions;

[0015] Among them, entity information includes: equipment name, fault event, substation, operation and maintenance unit and manufacturer;

[0016] Attribute information includes: equipment voltage level, physical ID, equipment type, equipment model, production date, fault time, fault location, fault cause, fault description, corrective measures and voltage level;

[0017] The relationship definitions include: the relationship between the fault event and the equipment is defined as: belongs to; the relationship between the equipment and the substation is defined as: belongs to; the relationship between the substation and the operation and maintenance unit is defined as: belongs to; the relationship between the equipment and the manufacturer is defined as: shipped from the factory.

[0018] Furthermore, the data collection and analysis module is based on the LSTM network to capture the long-term dependencies of the operating condition data of the substation equipment, and use the training model to directly predict the data at the next time point based on the historical data of the substation equipment time series to determine whether there are any abnormalities in the real-time operating condition data of the substation equipment.

[0019] Further, the meteorological information of the meteorological data module is associated with the substation equipment, including the following steps:

[0020] Firstly, the real-time monitoring data of the meteorological monitoring station and the meteorological forecast and warning information provided by the meteorological station are connected to the GIS platform;

[0021] Then the geographical location information of the meteorological monitoring station is verified, and the coordinates of the meteorological monitoring station are converted into the coordinates used by the power grid equipment;

[0022] Finally, the two are associated based on the meteorological monitoring station positioning information and the power grid facility positioning information after coordinate conversion.

[0023] Further, in the meteorological data module, the meteorological information associated with the substation equipment is determined based on the distance L between the power grid equipment and the meteorological monitoring station;

[0024] The calculation formula is as follows:

[0025] ,

[0026] Where d is the spherical distance calculated based on the longitude and latitude coordinates of the power grid equipment and the meteorological monitoring station; h is the height of the power grid equipment; , It is an effective influencing factor of meteorological information.

[0027] Furthermore, the video surveillance module associates the substation equipment with the camera installation configuration information and performs abnormal behavior detection based on the attention mechanism;

[0028] The video surveillance module specifically consists of three parts: encoder, attention module, and decoder; the encoder outputs encoding vectors based on continuous video frames; the attention module learns weight vectors through a fully connected neural network; the decoder is responsible for outputting the model prediction value at the next moment, and judging whether there is any abnormality in the video surveillance by comparing the predicted value with the actual value.

[0029] Furthermore, the encoder input is: a three-channel video frame of length T;

[0030] The encoder output is calculated as:

[0031] ,

[0032] in, is the output representation of the encoder when processing a sequence of input video frames; For the encoder, is a video frame of length T;

[0033] The encoder output is subjected to global maximum pooling and global average pooling respectively, and then sent to the attention module to calculate the weight;

[0034] The specific calculation formula is as follows:

[0035] ,

[0036] in, is the channel attention weight; is the activation function; It is a fully connected neural network; as well as Represent global maximum pooling and global average pooling respectively;

[0037] After the channel attention mechanism, the encoding vector is obtained:

[0038] ,

[0039] in, represents the encoding vector after channel attention weight adjustment; Represents the attention weight of the i-th channel; represents the feature representation of the i-th channel at time step t; c is the total number of channels;

[0040] The attention module also needs to calculate the spatial attention weight; the encoding vector Continue with global maximum pooling and global average pooling and then perform full connection to obtain vector ; Then use the convolution kernel to perform convolution operation to obtain the spatial attention weight;

[0041] The specific calculation formula is as follows:

[0042] ,

[0043] in, is the spatial attention weight; Indicates that convolution operation is performed through the convolution kernel; as well as Represent global maximum pooling and global average pooling respectively;

[0044] The final output of the spatial attention mechanism is:

[0045] ,

[0046] in, is the final spatial attention output vector; is the spatial attention weight; is the encoding vector The vector obtained by fully connecting after global maximum pooling and global average pooling;

[0047] The encoder output vector according to the spatial attention mechanism , predict the image at time t+T;

[0048] The specific calculation formula is as follows:

[0049] ,

[0050] in, represents the image predicted at time t+T; A function representing a decoder.

[0051] Furthermore, in the video monitoring module, the predicted value output by the decoder is compared with the actual value to determine whether there is any abnormality in the video monitoring; the video monitoring module compares the original image with the predicted image, and uses the peak signal-to-noise ratio to determine the difference between the original image and the predicted image. If the threshold is exceeded, an early warning is issued.

[0052] Furthermore, the fault warning model analyzes common tripping fault causes, including hardware failure, line failure, and tripping caused by external factors.

[0053] Furthermore, the fault warning model is built based on the Markov model; the model defines the state evolution of the substation equipment according to the real-time operating conditions of the substation equipment, the meteorological monitoring data associated with the substation equipment, and the video monitoring results associated with the substation equipment, which is defined as ; Indicates that the substation equipment trips. Indicates that the substation equipment is in good working condition; the transition between states is represented by the transition probability matrix A:

[0054] ,

[0055] in, is the transition probability matrix, which represents the transition probability between different states; Representation Matrix The elements in represent the probability of transitioning from state i to state j; N is the total number of states, and the dimension of matrix A is Represents all possible state combinations; represents the probability of transitioning from state 1 to state 1; represents the probability of transitioning from state N to state 1; represents the probability of transitioning from state N to state N;

[0056] Defining states is the failure threshold state, and the tripping probability of the substation equipment is The evaluation method is:

[0057] ,

[0058] in, is the specific value of the i-th state, used to compare with the failure threshold Make comparisons; is the probability distribution of the substation equipment status; Represents the transition probability matrix A, which represents the transition probability between device states after t time steps.

[0059] Compared with the prior art, the present invention has at least the following advantages or beneficial effects:

[0060] (1) The present invention summarizes the failures of substation equipment and uses knowledge graphs to analyze and mine the historical failure information of substation equipment in the existing network to assist in the analysis of substation equipment failures.

[0061] (2) The present invention adopts a time series data analysis method to identify abnormal data of the equipment based on the real-time operating conditions of the substation equipment in the data collection and analysis module, thereby providing a data basis for the substation equipment fault warning model.

[0062] (3) The present invention associates meteorological information with substation equipment in the meteorological data module, realizes the collection of various weather conditions and disaster forecasts, realizes climate anomaly detection of substation equipment, realizes real-time meteorological monitoring and early warning of substation equipment, line tripping early warning in areas with high correlation with lightning strikes, and rapid location of line fault points, providing a decision-making basis for maintenance arrangements, fault handling, risk analysis, etc. of substation equipment.

[0063] (4) The present invention performs environmental monitoring of substation equipment in the video monitoring module. By planning the camera point layout, the camera's collected data is associated with the substation equipment, thereby achieving real-time monitoring and inspection of the alarm equipment area, completing the automatic collection of important information, and providing data support for subsequent diagnosis and analysis of the cause of the fault.

[0064] (5) The present invention combines typical tripping cases in the fault warning model module, integrates the real-time operation data analysis results of the substation equipment, the video monitoring data of the substation equipment, and the meteorological data collection results to calculate the tripping probability of the equipment, and constructs a substation equipment tripping fault warning model, which further improves the recognition rate and success rate of fault warning and helps operation and maintenance personnel quickly locate the cause and location of the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0066] Figure 1 It is a schematic diagram of module connection of the present invention.

[0067] Figure 2 The present invention is a flowchart of the system implementation steps. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings of the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without creative work are within the scope of protection of the present invention. Example

[0069] This embodiment provides a substation trip warning management system based on fault signal classification and identification, referring to Figure 1 As shown, the substation equipment tripping warning management system based on fault signal classification and identification includes: a substation fault knowledge graph module, a collection data analysis module, a meteorological data module, a video monitoring module, and a fault warning model module.

[0070] Among them, the substation fault knowledge graph module is connected with the acquisition data analysis module and the video monitoring module; the acquisition data analysis module, the video monitoring module, the meteorological data module are connected with the fault warning model module.

[0071] (1) Substation fault knowledge graph module: By abstracting the physical structure of substation equipment, sorting out the attributes of existing substation equipment and historical fault analysis, a substation fault knowledge graph based on a graph database was constructed, and a substation fault entity, attribute, and relationship model was established.

[0072] In the substation fault entity, attribute, and relationship model, the entity information includes: equipment name, fault event, substation, operation and maintenance unit, and manufacturer.

[0073] The established substation fault entity, attribute, and relationship model includes: entity information, attribute information, and relationship definitions;

[0074] Among them, entity information includes: equipment name, fault event, substation, operation and maintenance unit and manufacturer;

[0075] Attribute information includes: equipment voltage level, physical ID, equipment type, equipment model, production date, fault time, fault location, fault cause, fault description, corrective measures and voltage level;

[0076] The relationship definitions include: the relationship between a fault event and equipment is defined as: belongs to; the relationship between equipment and substation is defined as: belongs to; the relationship between substation and operation and maintenance unit is defined as: belongs to; the relationship between equipment and manufacturer is defined as: shipped from the factory.

[0077] In the substation fault knowledge graph module, the substation fault entity, attribute, and relationship model are established based on the substation fault data, which is used for subsequent substation equipment fault classification and early warning.

[0078] (2) Data collection and analysis module: Based on the real-time operating conditions of the substation equipment, it realizes real-time perception and early warning of equipment tripping.

[0079] In this embodiment, the data collection and analysis module is based on the LSTM network to capture the long-term dependencies of the operating condition data of the substation equipment, and use the training model to directly predict the data at the next time point based on the historical data of the substation equipment time series to determine whether there are any abnormalities in the real-time operating condition data of the substation equipment.

[0080] In the data collection and analysis module, by comparing the predicted value of the operating condition data with the observed value, it is confirmed whether there is any abnormality in the operating condition data. If there is an abnormality, relevant warning information is sent, and the knowledge graph is used to mark the abnormal equipment and components to accurately determine the scope of subsequent fault inspections.

[0081] (3) Meteorological data module: associates meteorological information with substation equipment to collect various weather conditions and forecast disasters, and detect climate anomalies in substation equipment.

[0082] In this embodiment, the steps of associating meteorological information with substation equipment are as follows:

[0083] Firstly, the real-time monitoring data of the meteorological monitoring station and the meteorological forecast and warning information provided by the meteorological station are connected to the GIS platform;

[0084] Then the geographical location information of the meteorological monitoring station is verified, and the coordinates of the meteorological monitoring station are converted into the coordinates used by the power grid equipment;

[0085] Finally, the two are associated based on the meteorological monitoring station positioning information and the power grid facility positioning information after coordinate conversion.

[0086] It should be noted that the meteorological information associated with the substation equipment can be determined based on the distance between the power grid equipment and the meteorological monitoring station. Make judgments;

[0087] The calculation formula is as follows:

[0088] ,

[0089] Where d is the spherical distance calculated based on the longitude and latitude coordinates of the power grid equipment and the meteorological monitoring station; h is the height of the power grid equipment; , It is an effective influencing factor of meteorological information.

[0090] It should be noted that meteorological information includes temperature, humidity, wind speed and direction, lightning distribution, precipitation, and the changing trends of the above indicators, which are updated every ten minutes.

[0091] In the meteorological data module, meteorological disasters are associated with abnormalities in power grid equipment to establish more accurate meteorological risk warning indicators and warning threshold sets, which can realize real-time meteorological monitoring, warning and trend analysis of power grid facilities, and provide decision-making information for power grid new energy forecasting, planned maintenance arrangements, fault warning and disposal, risk analysis and other tasks.

[0092] (4) Video surveillance module: Through the spatiotemporal attention mechanism, anomalies in the surveillance video are detected to achieve anomaly detection of substation equipment.

[0093] In this embodiment, the video monitoring module is associated with the substation equipment according to the camera installation configuration information, and abnormal behavior detection is performed based on the attention mechanism.

[0094] The video surveillance module specifically consists of three parts: encoder, attention module, and decoder. The encoder is responsible for outputting encoding vectors based on continuous video frames. The attention module learns weight vectors through a fully connected neural network. The decoder is responsible for outputting the model prediction value at the next moment, and comparing the predicted value with the actual value to determine whether there is any abnormality in the video surveillance.

[0095] It should be noted that the method for associating and identifying the collected data of the camera with the substation equipment is based on the camera point control method and the tripping fault inspection rules.

[0096] It should be noted that the encoder input is: a three-channel video frame of length T;

[0097] The encoder output is calculated as:

[0098] ,

[0099] in, is the output representation of the encoder when processing a sequence of input video frames; For the encoder, is a video frame of length T.

[0100] The encoder output is subjected to global maximum pooling and global average pooling respectively, and then sent to the attention module to calculate the weight;

[0101] The specific calculation formula is as follows:

[0102] ,

[0103] in, is the channel attention weight; is the activation function; It is a fully connected neural network; as well as They represent global maximum pooling and global average pooling respectively.

[0104] After the channel attention mechanism, the encoding vector is obtained:

[0105] ,

[0106] in, represents the encoding vector after channel attention weight adjustment; Represents the attention weight of the i-th channel; represents the feature representation of the i-th channel at time step t; c is the total number of channels;

[0107] The attention module also needs to calculate the spatial attention weight; the encoding vector Continue with global maximum pooling and global average pooling and then perform full connection to obtain vector ; Then use the convolution kernel to perform convolution operation to obtain the spatial attention weight;

[0108] The specific calculation formula is as follows:

[0109] ,

[0110] in, is the spatial attention weight; Indicates that convolution operation is performed through the convolution kernel; as well as They represent global maximum pooling and global average pooling respectively.

[0111] The final output of the spatial attention mechanism is:

[0112] ,

[0113] in, is the final spatial attention output vector; is the spatial attention weight; is the encoding vector The vector obtained by fully connecting after global maximum pooling and global average pooling.

[0114] The encoder output vector according to the spatial attention mechanism , predict the image at time t+T;

[0115] The specific calculation formula is as follows:

[0116] ,

[0117] in, represents the image predicted at time t+T; A function representing a decoder.

[0118] Finally, the predicted value is compared with the actual value, that is, the original image is compared with the predicted image; the peak signal-to-noise ratio is used to determine the difference between the original image and the predicted image, and an early warning is issued if the threshold is exceeded.

[0119] In this embodiment, the environment of the substation equipment is monitored by a camera in the video monitoring module. When an abnormality occurs in the video environment monitoring or a user discovers an event, the abnormal situation or event information can be identified, and the abnormal information can be associated with the substation equipment that may be affected, thereby realizing video monitoring of the substation equipment.

[0120] (5) Fault warning model module: Calculate the equipment tripping probability based on the real-time operation data analysis results of the substation equipment, the video monitoring data of the substation equipment, and the meteorological data collection results, and perform analysis and warning.

[0121] Among them, the fault warning model analyzes common causes of tripping faults, including hardware fault tripping, line fault tripping, and tripping faults caused by external factors.

[0122] In this embodiment, the fault warning model is built based on the Markov model. The model defines the state evolution of the substation equipment according to the real-time operating conditions of the substation equipment, the meteorological monitoring data associated with the substation equipment, and the video monitoring results associated with the substation equipment. It is defined as .in, Indicates that the substation equipment trips; Indicates that the substation equipment is in a warning state and there is a certain risk of failure, but it has not yet completely tripped; Indicates that the substation equipment is in an intermediate state, for example, it is not functioning properly or has degraded performance, but is still within normal operating range and requires further monitoring or maintenance; Indicates that the substation equipment is in good working condition. The transition between states is represented by the transition probability matrix A:

[0123] ;

[0124] in, is the transition probability matrix, which represents the transition probability between different states; Representation Matrix The elements in represent the probability of transitioning from state i to state j; N is the total number of states, and the dimension of matrix A is Represents all possible state combinations; represents the probability of transitioning from state 1 to state 1; represents the probability of transitioning from state N to state 1; represents the probability of transitioning from state N to state N.

[0125] Defining states is the failure threshold state, and the tripping probability of the substation equipment is The evaluation method is:

[0126] ,

[0127] in, is the specific value of the i-th state, used to compare with the failure threshold Make comparisons; is the probability distribution of the substation equipment status; Represents the transition probability matrix A, which represents the transition probability between device states after t time steps.

[0128] In the fault warning model, the real-time operating conditions of the substation equipment, the meteorological monitoring data associated with the substation equipment, and the video monitoring results associated with the substation equipment are correlated to conduct a multi-dimensional analysis and warning of the possibility of substation equipment tripping.

[0129] In summary, the substation tripping warning management system provided in this embodiment improves the fault identification and response capabilities by establishing a comprehensive knowledge graph and an efficient data collection and analysis mechanism, and provides reliable support for ensuring the stable operation of the power system.

[0130] Example 2

[0131] Based on the substation tripping warning management system based on fault signal classification and identification provided in Example 1, this embodiment provides a substation tripping warning management method based on fault signal classification and identification.

[0132] like Figure 2 As shown, the substation tripping warning management method based on fault signal classification and identification includes:

[0133] Substation fault knowledge graph module: Build a knowledge graph in the field of substation faults, integrate data information on substation faults, and convert redundant data generated by the existing network into a rigorous knowledge graph;

[0134] Data collection and analysis module: Based on the real-time operating conditions of substation equipment, it realizes real-time perception of substation equipment operating data and abnormal early warning reminders;

[0135] Meteorological data module: associates meteorological information with substation equipment to collect various weather conditions and forecast disasters, and detects climate anomalies in substation equipment;

[0136] Video monitoring module: detects anomalies in monitoring videos through the spatiotemporal attention mechanism, and realizes environmental anomaly detection of substation equipment;

[0137] Fault warning model module: Calculate the equipment tripping probability based on the real-time operation data analysis results of the substation equipment, the video monitoring data of the substation equipment, and the meteorological data collection results, and perform analysis and warning.

[0138] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A trip warning management system for substation equipment based on fault signal classification and identification, characterized in that: include: Substation fault knowledge graph module, data collection analysis module, meteorological data module, video monitoring module and fault warning model module; The substation fault knowledge graph module is connected with the acquisition data analysis module and the video monitoring module; the acquisition data analysis module, the video monitoring module, the meteorological data module are connected with the fault warning model module; Substation fault knowledge graph module: By abstracting the physical structure of substation equipment, sorting out the attributes of existing substation equipment and historical fault analysis, constructing a substation fault knowledge graph based on a graph database, and establishing a substation fault entity, attribute, and relationship model; Data collection and analysis module: Based on the real-time operating conditions of substation equipment, it can provide real-time perception of substation equipment operating data and abnormal warning reminders; Meteorological data module: associates meteorological information with substation equipment, collects various weather conditions and forecasts disasters, and detects climate anomalies in substation equipment; Video monitoring module: detects anomalies in monitoring videos through the spatiotemporal attention mechanism and performs environmental anomaly detection on substation equipment; Fault warning model module: Calculate the equipment tripping probability based on the real-time operation data analysis results of the substation equipment, the video monitoring data of the substation equipment, and the meteorological data collection results, and perform analysis and warning; The fault warning model is built based on the Markov model; The model defines the state evolution of substation equipment according to the real-time operating conditions of substation equipment, meteorological monitoring data associated with substation equipment, and video monitoring results associated with substation equipment, which is defined as S(t)∈[S1,S2,S3...S N ]; S1 indicates that the substation equipment trips, S N Indicates that the substation equipment is in good working condition; the transition between states is represented by the transition probability matrix A: Among them, A is the transition probability matrix, which represents the transition probability between different states; [a ij ] represents the elements in matrix A, which represents the probability of transitioning from state i to state j; N is the total number of states, and the dimension of matrix A is N*N, which represents all possible state combinations; a 11 represents the probability of transitioning from state 1 to state 1; a N1 represents the probability of transitioning from state N to state 1; a NN represents the probability of transitioning from state N to state N; Define state S F is the failure threshold state, and the trip probability R of the substation equipment S The evaluation method is: Among them, S i is the specific value of the i-th state, used to compare with the failure threshold S F For comparison; π is the probability distribution of the substation equipment state; A t Represents the transition probability matrix A, which represents the transition probability between device states after t time steps.

2. The power substation tripping early warning management system based on fault signal classification and identification according to claim 1 is characterized in that: The established substation fault entity, attribute, and relationship model includes: entity information, attribute information, and relationship definitions; Among them, entity information includes: equipment name, fault event, substation, operation and maintenance unit and manufacturer; Attribute information includes: equipment voltage level, physical ID, equipment type, equipment model, production date, fault time, fault location, fault cause, fault description, corrective measures and voltage level; The relationship definitions include: the relationship between a fault event and equipment is defined as: belongs to; the relationship between equipment and substation is defined as: belongs to; the relationship between substation and operation and maintenance unit is defined as: belongs to; the relationship between equipment and manufacturer is defined as: shipped from the factory.

3. The power substation tripping early warning management system based on fault signal classification and identification according to claim 1 is characterized in that: The data collection and analysis module is based on the LSTM network to capture the long-term dependencies of the substation equipment operating condition data, and use the training model to directly predict the data at the next time point based on the historical data of the substation equipment time series to determine whether there are any abnormalities in the real-time operating condition data of the substation equipment.

4. The power substation tripping early warning management system based on fault signal classification and identification according to claim 1 is characterized in that: The meteorological information of the meteorological data module is associated with the substation equipment, including the following steps: Firstly, the real-time monitoring data of the meteorological monitoring station and the meteorological forecast and warning information provided by the meteorological station are connected to the GIS platform; Then the geographical location information of the meteorological monitoring station is verified, and the coordinates of the meteorological monitoring station are converted into the coordinates used by the power grid equipment; Finally, the two are associated based on the meteorological monitoring station positioning information and the power grid facility positioning information after coordinate conversion.

5. The system for early warning and management of tripping of substation equipment based on fault signal classification and identification according to claim 1 is characterized in that: In the meteorological data module, the meteorological information associated with the substation equipment is determined based on the distance L between the power grid equipment and the meteorological monitoring station; The calculation formula is as follows: L=α1*d+α2*h Among them, d is the spherical distance calculated based on the longitude and latitude coordinates of the power grid equipment and the meteorological monitoring station; h is the height of the power grid equipment; α1 and α2 are effective influencing factors of meteorological information.

6. The system for early warning and management of tripping of substation equipment based on fault signal classification and identification according to claim 1 is characterized in that: The video surveillance module associates the camera installation configuration information with the substation equipment and detects abnormal behavior based on the attention mechanism; The video surveillance module specifically consists of three parts: encoder, attention module, and decoder; the encoder outputs encoding vectors based on continuous video frames; the attention module learns weight vectors through a fully connected neural network; the decoder is responsible for outputting the model prediction value at the next moment, and judging whether there is any abnormality in the video surveillance by comparing the predicted value with the actual value.

7. The system for early warning and management of tripping of substation equipment based on fault signal classification and identification according to claim 6 is characterized in that: The encoder input is: a three-channel video frame of length T; The encoder output is calculated as: P t =Encoder(x t,t+1,...,t+T-1 ) Among them, P t is the output representation of the encoder when processing the input video frame sequence; Encoder is the encoder, x t,t+1,...,t+T-1 is a video frame of length T; The encoder output is subjected to global maximum pooling and global average pooling respectively, and then sent to the attention module to calculate the weight; The specific calculation formula is as follows: W c =Softmax(MLP(MaxPool(P t ))+MLP(AvgPool(P t ))) Among them, W c is the channel attention weight; Softmax is the activation function; MLP is a fully connected neural network; MaxPool and AvgPool represent global maximum pooling and global average pooling respectively; After the channel attention mechanism, the encoding vector is obtained: in, represents the encoding vector after channel attention weight adjustment; Represents the attention weight of the i-th channel; represents the feature representation of the i-th channel at time step t; c is the total number of channels; The attention module also needs to calculate the spatial attention weight; the encoding vector Continue with global maximum pooling and global average pooling and then perform full connection to obtain vector P C ; Then use the convolution kernel to perform convolution operation to obtain the spatial attention weight; The specific calculation formula is as follows: Among them, W s is the spatial attention weight; Cov represents the convolution operation through the convolution kernel; MaxPool and AvgPool represent the global maximum pooling and global average pooling respectively; The final output of the spatial attention mechanism is: P SC =W s ×P C Among them, P SC is the final spatial attention output vector; W s is the spatial attention weight; P C is the encoding vector The vector obtained by fully connecting after global maximum pooling and global average pooling; The encoder output vector P according to the spatial attention mechanism SC , predict the image at time t+T; The specific calculation formula is as follows: x t+T =Decoder(P SC ) Among them, x t+T Represents the image predicted at time t+T; Decoder represents the function of the decoder.

8. The system for early warning and management of tripping of substation equipment based on fault signal classification and identification according to claim 6 is characterized in that: In the video surveillance module, the predicted value output by the decoder is compared with the actual value to determine whether there is any abnormality in the video surveillance; The video monitoring module compares the original image with the predicted image, uses the peak signal-to-noise ratio to determine the difference between the original image and the predicted image, and issues an early warning if the difference exceeds the threshold.

9. The system for early warning and management of tripping of substation equipment based on fault signal classification and identification according to claim 1 is characterized in that: The fault warning model analyzes common tripping fault causes, including hardware failure, line failure, and tripping caused by external factors.

Images