Method and device for judging reliability of power grid monitoring alarm information

Abstract

The invention discloses a power grid monitoring alarm information reliability discrimination method and device, and relates to the technical field of power grid control, and the method comprises the steps: capturing new power grid monitoring real-time alarm information; matching the alarm information with the content of the maintenance plan of the data medium station on the day, and judging whether the alarm information is taken out by the maintenance work; matching the alarm information with historical mistakenly sent alarm information of the data intermediate station, calculating the similarity, and if the similarity is greater than a set threshold value, judging that the alarm information is a mistakenly sent signal; an event rule base is constructed, the alarm information is analyzed and judged, an event preliminary classification result is formed, and a training sample set is constructed for training and parameter learning to obtain an event analysis model; and performing data preprocessing on the alarm information, inputting the alarm information into the trained event analysis model to obtain an alarm information event result, and retraining the event analysis model according to an alarm information event and actual power grid event matching result. The method achieves the reliability judgment of the power grid monitoring alarm information, reduces the false monitoring probability of a worker, and improves the operation and maintenance efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid control, and in particular to a method and device for determining the reliability of power grid monitoring alarm information. Background Art

[0002] As the scale of the power grid gradually expands, the level of power grid intelligence continues to improve, and the digitalization capabilities of equipment information continue to increase, the number of alarm information connected to the centralized monitoring system for centralized monitoring has increased rapidly, and the average daily number of alarm information sent to the monitoring system has also increased.

[0003] At present, the power grid monitoring method is to manually identify and analyze each alarm information of the monitoring system one by one. When the monitoring personnel confirm that the signal is not a maintenance signal and is actually sent on site, the operating personnel are notified to come to the station for inspection. After the operating personnel arrive at the station, they will make a preliminary judgment on the signal again and contact the maintenance team for processing as needed. It can be seen that the inspection and confirmation of an alarm information often takes a lot of time, so it is of great significance to improve the reliability of monitoring alarm signals.

[0004] At present, the method for judging the reliability of monitoring alarm information is to manually judge and analyze each alarm information of the monitoring system one by one, but there are the following disadvantages:

[0005] (1) Some power grid equipment has the problem of occasional or erroneous inherent alarm information. Under the manual judgment method, it is impossible to quickly judge the signal reliability based on historical alarm information and disposal results.

[0006] (2) Personnel are often unable to comprehensively obtain the planned work information of the main and distribution networks for the day, and may misjudge some alarm information caused by work, resulting in mismonitoring behavior.

[0007] (3) The manual judgment method lacks the ability of information logical association and multi-dimensional analysis. In the judgment of information reliability, weather and event characteristics are not taken into comprehensive consideration, which reduces the reliability of information judgment.

[0008] (4) The manual judgment method is overly dependent on the skill level and knowledge reserve of the on-duty personnel. Different on-duty personnel may make different credibility judgments on the same alarm information, and the urgency of handling the alarm information will also be different, which may cause waste of resources or expand the accident. Summary of the invention

[0009] The technical problem to be solved by the present invention is to solve the deficiencies in the prior art and to design a method and device for determining the reliability of power grid monitoring alarm information.

[0010] The technical solution adopted by the present invention to solve its technical problem is:

[0011] In a first aspect, the present invention provides a method for determining the reliability of power grid monitoring alarm information, comprising:

[0012] Capture new real-time alarm information for power grid monitoring;

[0013] Match the alarm information with the maintenance plan of the day in the data center to determine whether it is an alarm caused by maintenance work;

[0014] Match the alarm information with the historical mis-sent alarm information of the data center and calculate the similarity. If the similarity is greater than the set threshold, it is judged as a mis-sent signal;

[0015] Build an event-based rule base and analyze and judge the alarm information to form a preliminary classification result of the event. Build a training sample set for training and parameter learning to obtain an event-based analysis model.

[0016] The alarm information is preprocessed and input into the trained event-based analysis model to obtain the alarm information event-based results. The event-based analysis model is retrained based on the matching results between the alarm information event-based results and the actual power grid events.

[0017] As a preferred embodiment of the present invention, the capturing of newly added power grid monitoring real-time alarm information specifically includes:

[0018] By using alarm linkage, when the monitoring system sends a signal, the automatic capture module is triggered to capture accidents, abnormalities, line crossings, displacements and notification information within the specified time (10 seconds).

[0019] As a preferred embodiment of the present invention, it also includes:

[0020] Establishing an expert system maintenance related information knowledge base, wherein establishing an expert system maintenance related information knowledge base specifically includes:

[0021] Use the TextRank model to extract keywords from real-time warning information;

[0022] Collect basic data of maintenance information;

[0023] Establish an expert system maintenance-related information knowledge base based on maintenance information basic data and keywords;

[0024] The keywords include: substation, time scale, interval time, voltage level, signal type, equipment, power outage status, and whether the alarm information is sent simultaneously;

[0025] The basic data of maintenance information includes: historical information of power grid monitoring alarms, historical maintenance operation records, voltage level, substation, equipment, whether there is a power outage, planned time, and signal type.

[0026] As a preferred embodiment of the present invention, the alarm information is matched with the maintenance plan content of the data center on the same day to determine whether it is the alarm information caused by the maintenance work; specifically, it includes:

[0027] A two-way matching algorithm is used for the keywords to search and match with the maintenance plan content of the day in the data center. When the equipment is matched successfully, it is determined whether the maintenance plan in the risk management and control platform is in the start state. If it is in the start state, the expert system is started to determine whether the alarm information is caused by the maintenance work. If the judgment is true, the reliability of the output alarm information is determined to be caused by the maintenance work; if the judgment is false, the next step of false signal judgment is carried out.

[0028] As a preferred embodiment of the present invention, the alarm information is matched with the historical mis-sent alarm information of the data center, and the similarity is calculated. If the similarity is greater than a set threshold, it is a mis-sent signal; specifically, it includes:

[0029] A complete matching algorithm is used for keywords to search and match with the historical false alarm information of the data center. If a match occurs, the N-gram model is used to convert the keywords into vectors, and the cosine similarity is used to calculate the similarity between the two. The calculation formula is:

[0030]

[0031] Where, “.” represents the vector dot product, and ||x|| represents the vector x=(x1, x2, x3, …, x n ) is defined as

[0032] Furthermore, the threshold is set to 0.8. When the similarity is greater than 0.8, the reliability of the output alarm information is determined to be a false signal. Otherwise, the next step of event analysis is performed.

[0033] As a preferred embodiment of the present invention, the event-based rule base is constructed and the alarm information is analyzed and determined to form a preliminary event classification result, and a training sample set is constructed for training and parameter learning to obtain an event-based analysis model; specifically, it includes:

[0034] Establish a standardized model library for monitoring information based on logical relationships;

[0035] Use the N-gram model to perform unsupervised training on monitoring alarm information, deeply mine the association between the previous and next phrases in a single piece of information, and generate high-dimensional word vectors containing semantic features;

[0036] The word vectors in a single monitoring alarm message are averaged and processed to output a monitoring alarm message sentence vector that can be recognized by a computer;

[0037] Analyze the standardized information objects associated with power grid events, combine time and space constraints, and build an event-based rule base;

[0038] Comprehensively use event feature rules and event rules to analyze and judge alarm information, form preliminary event classification results, and convert them into vectors;

[0039] Combined with the vector of the preliminary classification result, a training sample set is constructed;

[0040] The convolutional neural network (TextCNN) classification model is used to train the training sample set and learn parameters to obtain the event analysis model.

[0041] As a preferred embodiment of the present invention, the alarm information is preprocessed and input into a trained event analysis model to obtain an alarm information event result, and the event analysis model is retrained according to the matching result between the alarm information event and the actual power grid event, specifically including:

[0042] Perform data preprocessing on captured real-time monitoring alarm information;

[0043] Use the N-gram model to generate a feature vector containing signal features;

[0044] The vector representation is input into the trained model, and the alarm information event result is output;

[0045] The model is retrained based on the feature vectors generated by real-time monitoring alarm information and actual power grid events.

[0046] In a second aspect, the present invention provides a method for verifying the reliability of power grid monitoring alarm information, comprising the following steps:

[0047] Information capture module, captures new real-time alarm information of power grid monitoring;

[0048] The maintenance association matching module matches the alarm information with the maintenance plan content of the day in the data center to determine whether it is an alarm information caused by maintenance work;

[0049] The mis-sent signal matching module matches the alarm information with the historical mis-sent alarm information of the data center and calculates the similarity. If the similarity is greater than the set threshold, it is a mis-sent signal;

[0050] The event analysis module builds an event rule base and analyzes and determines the alarm information to form a preliminary classification result of the event, builds a training sample set for training and parameter learning to obtain an event analysis model;

[0051] The result output and feedback module performs data preprocessing on the alarm information and inputs it into the trained event-based analysis model to obtain the alarm information event-based results. The event-based analysis model is retrained according to the matching results between the alarm information event-based results and the actual power grid events.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The present invention constructs an expert system library of maintenance-related signals and equipment erroneous signaling; extracts basic objects according to the electrical function attributes of monitoring information to form event characteristics, and uses an artificial intelligence neural network model to establish a power grid event classification model; captures all alarm information sent by the monitoring system in the same time period through a program, and uses the expert system to match the work content of the middle station maintenance plan and the historical alarm information processing records; according to the matching results, uses the artificial neural network model to perform power grid event analysis, outputs the reliability results of the power grid monitoring alarm information, and provides the main basis for the analysis; the present invention realizes the reliability judgment of the power grid monitoring alarm information, reduces the probability of personnel mismonitoring, and improves the operation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A flow chart of a method for determining the reliability of power grid monitoring alarm information proposed by the present invention;

[0055] Figure 2 A flow chart of a specific embodiment of the present invention;

[0056] Figure 3 This is the logical structure diagram of the power grid monitoring information eventization proposed by the present invention;

[0057] Figure 4 This is a structural diagram of a power grid monitoring alarm information reliability determination device proposed by the present invention. DETAILED DESCRIPTION

[0058] The specific implementation of the present invention is described below in conjunction with the accompanying drawings and embodiments:

[0059] It should be noted that the structures, colors, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so that people familiar with this technology can understand and read them, and are not used to limit the conditions under which the present invention can be implemented. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0060] Unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present invention should be understood by people with ordinary skills in the field to which the present disclosure belongs. "First", "second" and similar words used in the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0062] Embodiment 1

[0063] like Figure 1 As shown, the present invention provides a method for determining the reliability of power grid monitoring alarm information, comprising:

[0064] Step 101, capturing newly added power grid monitoring real-time alarm information;

[0065] Step 102, matching the alarm information with the maintenance plan content of the day in the data center to determine whether it is an alarm information caused by the maintenance work;

[0066] Step 103, matching the alarm information with the historical mis-sent alarm information of the data center, and calculating the similarity. If the similarity is greater than a set threshold, it is a mis-sent signal;

[0067] Step 104, construct an event-based rule base and analyze and determine the alarm information to form a preliminary event classification result, construct a training sample set for training and parameter learning to obtain an event-based analysis model;

[0068] Step 105 , pre-process the alarm information and input it into the event-based analysis model to output the alarm information event-based result, and re-train the event-based analysis model according to the alarm information event-based result.

[0069] The present invention constructs an expert system library of maintenance-related signals and equipment erroneous signaling; extracts basic objects according to the electrical function attributes of monitoring information to form event characteristics, and uses an artificial intelligence neural network model to establish a power grid event classification model; captures all alarm information sent by the monitoring system in the same time period through a program, and uses the expert system to match the work content of the middle station maintenance plan and the historical alarm information processing records; according to the matching results, uses the artificial neural network model to perform power grid event analysis, outputs the reliability results of the power grid monitoring alarm information, and provides the main basis for the analysis; the present invention realizes the reliability judgment of the power grid monitoring alarm information, reduces the probability of personnel mismonitoring, and improves the operation and maintenance efficiency.

[0070] Among them, in step 101, the newly added power grid monitoring real-time alarm information is captured, specifically including:

[0071] By utilizing alarm linkage, when the monitoring system sends a signal, the automatic capture module is triggered to capture accidents, abnormalities, line crossings, displacements and notification information within the specified time.

[0072] The present invention establishes an expert system maintenance-related information knowledge base, performs maintenance, mis-sending and event-based possibility analysis on real-time alarm information of power grid monitoring, and outputs reliability results of power grid monitoring alarm information. The establishment of the expert system maintenance-related information knowledge base specifically includes:

[0073] Use the TextRank model to extract keywords from real-time warning information;

[0074] Collect basic data of maintenance information;

[0075] Establish an expert system maintenance-related information knowledge base based on maintenance information basic data and keywords;

[0076] The keywords include: substation, time scale, interval time, voltage level, signal type, equipment, power outage status, and whether the alarm information is sent simultaneously;

[0077] The basic data of maintenance information includes: historical information of power grid monitoring alarms, historical maintenance operation records, voltage level, substation, equipment, whether there is a power outage, planned time, and signal type.

[0078] By analyzing the real-time alarm information of power grid monitoring, the reliability of power grid monitoring alarm information can be judged, the probability of personnel mismonitoring can be reduced, and the operation and maintenance efficiency can be improved.

[0079] See also Figure 2 In step 102, the present invention matches the alarm information with the maintenance plan content of the day in the data center to determine whether it is an alarm information caused by the maintenance work; specifically including:

[0080] A two-way matching algorithm is used for the keywords to search and match with the maintenance plan content of the day in the data center. When the equipment is matched successfully, it is determined whether the maintenance plan in the risk management and control platform is in the start state. If it is in the start state, the expert system is started to determine whether the alarm information is caused by the maintenance work. If the judgment is true, the reliability of the output alarm information is determined to be caused by the maintenance work; if the judgment is false, the next step of false signal judgment is carried out.

[0081] In step 103, the present invention matches the alarm information with the historical mis-sent alarm information of the data center, and calculates the similarity. If the similarity is greater than a set threshold, it is a mis-sent signal; specifically, it includes:

[0082] A complete matching algorithm is used for keywords to search and match with the historical false alarm information of the data center. If a match occurs, the N-gram model is used to convert the keywords into vectors, and the cosine similarity is used to calculate the similarity between the two. The calculation formula is:

[0083]

[0084] Where, “.” represents the vector dot product, and ||x|| represents the vector x=(x1, x2, x3, …, x n ) is defined as

[0085] The threshold value set in the present invention is 0.8. When the similarity is greater than 0.8, the reliability of the output alarm information is determined to be a false signal. If there is no match or the similarity is less than 0.8, the next step of event analysis is performed.

[0086] See also Figure 3 In step 104, an event-based rule base is constructed and the alarm information is analyzed and determined to form a preliminary event classification result, and a training sample set is constructed for training and parameter learning to obtain an event-based analysis model; specifically, the following steps are included:

[0087] Establish a standardized model library for monitoring information based on logical relationships.

[0088] Use the N-gram model to perform unsupervised training on monitoring alarm information, deeply mine the association between the previous and next phrases in a single piece of information, and generate high-dimensional word vectors containing semantic features;

[0089] The word vectors in a single monitoring alarm message are averaged and processed to output a monitoring alarm message sentence vector that can be recognized by a computer;

[0090] Analyze the standardized information objects associated with power grid events, combine time and space constraints, and build an event-based rule base;

[0091] Comprehensively use event feature rules and event rules to analyze and judge alarm information, form preliminary event classification results, and convert them into vectors;

[0092] Combined with the vector of the preliminary classification result, a training sample set is constructed;

[0093] The convolutional neural network (TextCNN) classification model is used to train the training sample set and learn parameters to obtain the event analysis model.

[0094] In step 105, the alarm information is preprocessed and input into the trained event analysis model to obtain the alarm information event analysis result. The event analysis model is retrained according to the matching result between the alarm information event and the actual power grid event, which specifically includes:

[0095] Perform data preprocessing on captured real-time monitoring alarm information;

[0096] Use the N-gram model to generate a feature vector containing signal features;

[0097] The vector representation is input into the trained model, and the alarm information event result is output;

[0098] The model is retrained based on the feature vectors generated by real-time monitoring alarm information and actual power grid events.

[0099] The present invention performs data preprocessing on the real-time monitoring alarm information captured in step 101, and uses the N-gram model to perform unsupervised training on the monitoring alarm information, deeply mines the association relationship between the previous and next phrases in a single message, generates a high-dimensional word vector containing semantic features, and inputs the vector into the trained model, and outputs the reliability of the output alarm information as a certain power grid event, such as tripping, grounding, etc.; if the final output power grid event does not match the actual one, the model is retrained based on the feature vector generated by the real-time monitoring alarm information and the actual power grid event.

[0100] Embodiment 2

[0101] See also Figure 4 , a reliability determination device for power grid monitoring alarm information, comprising:

[0102] Information capture module 201, captures newly added power grid monitoring real-time alarm information;

[0103] The maintenance association matching module 202 matches the alarm information with the maintenance plan content of the data center on that day to determine whether it is the alarm information caused by the maintenance work;

[0104] The mis-sent signal matching module 203 matches the alarm information with the historical mis-sent alarm information of the data center and calculates the similarity. If the similarity is greater than a set threshold, it is a mis-sent signal;

[0105] The event analysis module 204 builds an event rule base and analyzes and determines the alarm information to form a preliminary event classification result, builds a training sample set for training and parameter learning to obtain an event analysis model;

[0106] The result output and feedback module 205 performs data preprocessing on the alarm information and inputs it into the trained event analysis model to obtain the alarm information event result, and retrains the event analysis model according to the matching result between the alarm information event and the actual power grid event.

[0107] The various variations and specific examples of a method for determining the reliability of power grid monitoring alarm information in the aforementioned embodiment are also applicable to a device for determining the reliability of power grid monitoring alarm information in the present embodiment. Through the aforementioned detailed description of a method for determining the reliability of power grid monitoring alarm information, those skilled in the art can clearly know a device for determining the reliability of power grid monitoring alarm information in the present embodiment. Therefore, for the sake of brevity of the specification, they will not be described in detail here.

[0108] The present invention constructs an expert system library and an artificial neural network model to analyze the possibility of maintenance, mis-sending, and eventing of real-time alarm information of power grid monitoring, outputs the reliability results of power grid monitoring alarm information, and provides the main basis for analysis, thereby realizing the reliability judgment of power grid monitoring alarm information, reducing the probability of personnel mis-monitoring, and improving operation and maintenance efficiency.

[0109] The patent of this invention adopts convolutional neural network to build and train the model. It should be pointed out that without changing the principle of the scheme, changing the model construction and training method should also be regarded as the scope of protection of this invention.

[0110] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0111] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0113] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for determining the reliability of power grid monitoring alarm information, characterized in that: include: Capture new real-time alarm information for power grid monitoring; Match the alarm information with the maintenance plan of the day in the data center to determine whether it is an alarm caused by maintenance work; Match the alarm information with the historical mis-sent alarm information of the data center and calculate the similarity. If the similarity is greater than the set threshold, it is judged as a mis-sent signal; Build an event-based rule base and analyze and judge the alarm information to form a preliminary classification result of the event. Build a training sample set for training and parameter learning to obtain an event-based analysis model. The alarm information is preprocessed and input into the trained event-based analysis model to obtain the alarm information event-based results. The event-based analysis model is retrained based on the matching results between the alarm information event-based results and the actual power grid events.

2. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: The capturing of newly added power grid monitoring real-time alarm information specifically includes: By utilizing alarm linkage, when the monitoring system sends a signal, the automatic capture module is triggered to capture accidents, abnormalities, line crossings, displacements and notification information within the specified time.

3. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: Also includes: Establishing an expert system maintenance-related information knowledge base, wherein establishing an expert system maintenance-related information knowledge base specifically includes: Use the TextRank model to extract keywords from real-time warning information; Collect basic data of maintenance information; Establish an expert system maintenance-related information knowledge base based on maintenance information basic data and keywords; The keywords include: substation, time scale, interval time, voltage level, signal type, equipment, power outage status, and whether the alarm information is sent simultaneously; The basic data of maintenance information includes: historical information of power grid monitoring alarms, historical maintenance operation records, voltage level, substation, equipment, whether there is a power outage, planned time, and signal type.

4. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: Match the alarm information with the maintenance plan of the day in the data center to determine whether it is an alarm caused by maintenance work; specifically, it includes: A two-way matching algorithm is used for the keywords to search and match with the maintenance plan content of the day in the data center. When the equipment is matched successfully, it is determined whether the maintenance plan in the risk management and control platform is in the start state. If it is in the start state, the expert system is started to determine whether the alarm information is caused by the maintenance work. If the judgment is true, the reliability of the output alarm information is determined to be caused by the maintenance work; if the judgment is false, the next step of false signal judgment is carried out.

5. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: The alarm information is matched with the historical mis-sent alarm information of the data center, and the similarity is calculated. If the similarity is greater than a set threshold, it is judged as a mis-sent signal; specifically, it includes: A complete matching algorithm is used for keywords to search and match with the historical false alarm information of the data center. If a match occurs, the N-gram model is used to convert the keywords into vectors, and the cosine similarity is used to calculate the similarity between the two. The calculation formula is: Among them, "." represents the vector dot product, ||x|| represents the vector x=(x1,x2,x3,…,x n ) is defined as 6. A method for determining the reliability of power grid monitoring alarm information according to claim 5, characterized in that: The threshold is set to 0.

8. When the similarity is greater than 0.8, the reliability of the output alarm information is determined to be a false signal. Otherwise, the next step of event analysis is performed.

7. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: The event-based rule base is constructed and the alarm information is analyzed and determined to form a preliminary event classification result, and a training sample set is constructed for training and parameter learning to obtain an event-based analysis model; specifically, the following steps are included: Establish a standardized model library for monitoring information based on logical relationships. Use the N-gram model to perform unsupervised training on monitoring alarm information, deeply mine the association between the previous and next phrases in a single piece of information, and generate high-dimensional word vectors containing semantic features; The word vectors in a single monitoring alarm message are averaged and processed to output a monitoring alarm message sentence vector that can be recognized by a computer; Analyze the standardized information objects associated with power grid events, combine time and space constraints, and build an event-based rule base; Comprehensively use event feature rules and event rules to analyze and judge alarm information, form preliminary event classification results, and convert them into vectors; Combined with the vector of the preliminary classification result, a training sample set is constructed; The convolutional neural network (TextCNN) classification model is used to train the training sample set and learn parameters to obtain the event analysis model.

8. A method for determining the reliability of power grid monitoring alarm information according to claim 1, characterized in that: The alarm information is preprocessed and input into the trained event analysis model to obtain the alarm information event analysis result, and the event analysis model is retrained according to the matching result between the alarm information event and the actual power grid event, specifically including: Perform data preprocessing on captured real-time monitoring alarm information; Use the N-gram model to generate a feature vector containing signal features; The vector representation is input into the trained model, and the alarm information event result is output; The model is retrained based on the feature vectors generated by real-time monitoring alarm information and actual power grid events.

9. A device for determining the reliability of power grid monitoring alarm information, using a method for determining the reliability of power grid monitoring alarm information as claimed in any one of claims 1 to 8, characterized in that: include: Information capture module, captures new real-time alarm information of power grid monitoring; The maintenance association matching module matches the alarm information with the maintenance plan content of the day in the data center to determine whether it is an alarm information caused by maintenance work; The mis-sent signal matching module matches the alarm information with the historical mis-sent alarm information of the data center and calculates the similarity. If the similarity is greater than the set threshold, it is judged as a mis-sent signal; The event analysis module builds an event rule base and analyzes and determines the alarm information to form a preliminary classification result of the event, builds a training sample set for training and parameter learning to obtain an event analysis model; The result output and feedback module performs data preprocessing on the alarm information and inputs it into the trained event-based analysis model to obtain the alarm information event-based results. The event-based analysis model is retrained according to the matching results between the alarm information event-based results and the actual power grid events.

Images