Substation four-remote signal test system and method based on distributed monitoring

Abstract

The invention discloses a transformer substation four-remote signal test system and method based on distributed monitoring. The test system comprises a test host and an SCADA system. The test host comprises an AI model, a database, a signal simulation generator and a controller; the SCADA system comprises a main controller, an acquisition device and execution equipment; the test host realizes automatic learning and repairing of unknown faults based on an artificial intelligence model and a digital twinborn model, intelligently generates a fault repairing strategy according to known similar signal features, sends the fault repairing strategy to the digital twinborn model for rehearsal, and then compares the fault repairing strategy with an actual execution result for feedback. And the intelligent automation level and the unknown risk coping capability of the system are improved. In addition, the model precision is corrected and updated in combination with an execution feedback result, and the precision and the application range of the test system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power dispatching, and in particular to a substation four-remote signal testing system and method based on distributed monitoring. Background Art

[0002] With the current diversified development of energy types and the grid-connected operation of various small-scale renewable energy power stations, the existing substation energy management system can no longer meet the development needs of the power grid. After the construction of a new substation energy management system, it is necessary to undergo a long period of testing and adjustment, and the development and testing cycle is relatively long.

[0003] In addition, after the system is put into operation, as the system runs, since the system model cannot be updated in real time with the system, the system operation efficiency will be low, the operation safety cannot be guaranteed, and the ability to respond to unknown risk failures will be low.

[0004] For example, the invention patent with the patent publication number CN108918997A provides a method for testing a new system using a simulated four-remote signal generator. It only performs simulation testing through a four-remote signal simulation system and is put into operation after a short test. Although it shortens the test cycle, the number of fault simulations is limited, and system operation faults cannot be fully detected, posing a safety hazard to operation.

[0005] At present, artificial intelligence has developed rapidly and its intelligence has been greatly improved, but it is mostly used in security, robotics, autonomous driving and other fields. There is little research on its combined application in substation management systems, resulting in the current substation management system being less intelligent and unable to achieve highly intelligent automated operation.

[0006] Therefore, how to propose a substation four-remote signal testing system and method with a wide range of applications, high safety, and a high level of intelligent automation on the basis of shortening the R&D cycle is an urgent problem to be solved. Summary of the invention

[0007] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a substation four-remote signal test system and method based on distributed monitoring, which combines artificial intelligence technology and digital twin technology to realize rapid intelligent testing of the substation management system, and automatically tests and repairs unknown faults based on the discovery, thereby improving the intelligence and safety of the system. In addition, the model accuracy is corrected and updated in combination with the execution feedback results, thereby improving the accuracy and scope of application of the test system.

[0008] The present invention provides a substation four-remote signal test system based on distributed monitoring, the test system comprises a test host and a SCADA system; the test host comprises an AI model, a database, a signal simulation generator and a controller; the SCADA system comprises a main controller, a collection device, and an execution device; The acquisition device acquires the system telemetry signal and transmits it to the controller in the test host through the telesignal channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal has obvious abnormality. If it has abnormality, it issues a telemetry signal abnormality alarm; if it has no abnormality, it extracts the signal feature of the telemetry signal and searches it in the database; The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the controller uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator to simulate sending an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device to perform corresponding actions, and then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

[0009] Preferably, the test host further includes a digital twin model; When the controller determines that there is no signal feature similar to the telemetry signal in the database based on the retrieval results, the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview. The controller determines the first optimal fault repair strategy based on the preview results.

[0010] Preferably, after the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification and obtains the fault repair result. If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model. The AI ​​model generates a second set of candidate fault repair strategies again according to the fault repair result and the first set of candidate fault repair strategies, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview. The controller determines the second optimal fault repair strategy according to the preview result, and then sends the second optimal fault repair strategy to the main controller for analysis and verification again.

[0011] In order to solve the above technical problems, the present invention also provides a substation four-telemetry signal testing method based on distributed monitoring. The acquisition device in the SCADA system collects the system telemetry signal and transmits it to the controller in the test host through the telemetering channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal has obvious abnormalities. If there is an abnormality, an abnormal telemetry signal alarm is issued; if there is no abnormality, the signal characteristics of the telemetry signal are extracted and retrieved in the database in the test host. The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator in the test host to simulate and send an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, and the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device of the SCADA system to perform corresponding actions. Then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

[0012] Furthermore, the controller determines whether the telemetry signal belongs to an existing fault signal based on the retrieval result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, which specifically includes: setting a similarity threshold. If there is a signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an existing fault signal, and the fault type and fault repair strategy code stored in the database corresponding to the signal feature are extracted. The controller of the test host performs secondary confirmation on the extracted fault type and fault repair strategy code, and then controls the AI ​​model to extract the corresponding fault repair strategy.

[0013] Furthermore, if there is no signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an unknown fault signal, and the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview, and the controller determines the first optimal fault repair strategy based on the preview results.

[0014] Furthermore, after the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification to obtain the execution result of the execution device and the fault repair result; If the ratio of the execution result to the execution result previewed in the digital twin model is within the second preset range, the remote adjustment and remote control functions are verified to be normal; if the ratio of the execution result to the execution result previewed in the digital twin model is not within the second preset range, an abnormal alarm of the remote adjustment and remote control functions is issued; The controller further determines whether the fault has been repaired according to the fault repair result. If the fault has been repaired, the fault signal feature and the first optimal fault repair strategy are correspondingly stored in the database.

[0015] Furthermore, the controller further determines whether the fault is repaired according to the fault repair result, which specifically includes: If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model, and the AI ​​model generates a second set of candidate fault repair strategies again according to the fault repair result and the first set of candidate fault repair strategies, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview, and the controller determines the second best fault repair strategy according to the preview result, and then sends the second best fault repair strategy to the main controller again for analysis and verification; The controller further determines whether the fault has been repaired according to the fault repair result, and if the fault has been repaired, stores the fault signal feature and the second best fault repair strategy in the database accordingly; If the fault is still not repaired, the feedback execution step is repeated until the fault is repaired and the fault signal characteristics and the Nth best fault repair strategy are stored accordingly; where N is an integer greater than or equal to 2.

[0016] Furthermore, the controller is provided with an upper limit value N of the number of execution feedbacks. max , if the number of execution feedback is equal to N max If the fault is still not repaired, the test host sends an alarm signal to the management personnel for manual intervention.

[0017] Furthermore, the controller periodically updates the AI ​​model based on the number of feedback executions and the best fault repair strategies in history.

[0018] The beneficial effects of the present invention are: 1. After the test system is developed, it is first run in the test mode. While verifying the four remote signals, it also realizes the accumulation of fault repair strategies and the accumulation of system operation experience. After the system is verified to be correct, it enters the operation mode, shortening the test-investment development cycle; and in the operation mode, real-time testing is also carried out, using the test host and SCADA system for two-way verification to improve the safety of system operation; 2. Automatic learning and repair of unknown faults are realized based on artificial intelligence models and digital twin models. The fault repair strategy is first intelligently generated based on known similar signal characteristics, and then sent to the digital twin model for preview. Then, the strategy is compared and fed back with the actual execution results, so as to improve the system's intelligent automation level and ability to deal with unknown risks. 3. The artificial intelligence model and digital twin model also automatically learn and update based on the feedback results to form a closed-loop adjustment, thereby improving the accuracy and scope of application of the test system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is the schematic diagram of the substation four-remote signal test system based on distributed monitoring; Figure 2 This is a flow chart of the substation four-remote signal testing method based on distributed monitoring. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings. Example 1

[0021] The present invention provides a substation four-remote signal test system based on distributed monitoring, such as Figure 1As shown, The test system includes a test host and a SCADA system; the test host includes an AI model, a database, a signal simulation generator and a controller; the SCADA system includes a main controller, a collection device, and an execution device; The acquisition device acquires the system telemetry signal and transmits it to the controller in the test host through the telesignal channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal has obvious abnormality. If it has abnormality, it issues a telemetry signal abnormality alarm; if it has no abnormality, it extracts the signal feature of the telemetry signal and searches it in the database; The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the controller uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator to simulate sending an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device to perform corresponding actions, and then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

[0022] The test host is used to verify the telemetry signal, remote adjustment and remote control functions, and the SCADA system is used to verify the remote signal channel. The two-way multi-party verification can eliminate interference and make the system have higher test accuracy.

[0023] Preferably, the test host further includes a digital twin model; the digital twin model constructs corresponding virtual modules according to physical device components, and is constructed according to physical connection relationships to perform real-time simulation operation; When the controller determines that there is no signal feature similar to the telemetry signal in the database based on the retrieval results, the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview. The controller determines the first optimal fault repair strategy based on the preview results.

[0024] Preferably, after the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification and obtains the fault repair result. If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model. The AI ​​model generates a second set of candidate fault repair strategies again according to the fault repair result and the first set of candidate fault repair strategies, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview. The controller determines the second optimal fault repair strategy according to the preview result, and then sends the second optimal fault repair strategy to the main controller for analysis and verification again. Example 2

[0025] In order to solve the above technical problems, the present invention also provides a substation four-remote signal testing method based on distributed monitoring: the testing process is as follows: Figure 2 As shown, the acquisition device in the SCADA system collects the system telemetry signal and transmits it to the controller in the test host through the telesignaling channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal has obvious abnormalities. If there is an abnormality, an abnormal telemetry signal alarm is issued; if there is no abnormality, the signal characteristics of the telemetry signal are extracted and searched in the database in the test host; The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator in the test host to simulate and send an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, and the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device of the SCADA system to perform corresponding actions. Then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

[0026] The controller verifies whether the remote adjustment and remote control functions are normal according to the execution results, specifically including: the controller receives the execution feedback signal, parses the executed adjustment parameters and control instructions, and then compares them with the adjustment parameters and control instructions included in the fault repair strategy. If the comparison result deviation exceeds the threshold, it is determined that the remote adjustment and remote control functions are abnormal; if the comparison result deviation is less than or equal to the threshold, it is determined that the remote adjustment and remote control functions are normal.

[0027] Furthermore, the controller determines whether the telemetry signal belongs to an existing fault signal based on the retrieval result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, which specifically includes: setting a similarity threshold. If there is a signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an existing fault signal, and the fault type and fault repair strategy code stored in the database corresponding to the signal feature are extracted. The controller of the test host performs secondary confirmation on the extracted fault type and fault repair strategy code, and then controls the AI ​​model to extract the corresponding fault repair strategy.

[0028] Furthermore, if there is no signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an unknown fault signal, and the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview, and the controller determines the first optimal fault repair strategy based on the preview results.

[0029] The signal feature similarity is specifically set as follows: in the signal features, key point features are extracted to construct a key point feature matrix, and then the key point feature matrix stored in the database is traversed, and covariance calculations are performed with the extracted key point feature matrix respectively. If there is a covariance value in the calculation result that is less than or equal to the covariance value corresponding to the similarity threshold, it is determined that there is a signal feature whose signal feature similarity with the telemetry signal exceeds the similarity threshold, and then the signal is determined to be an existing fault signal.

[0030] The controller determines the first optimal fault repair strategy based on the preview results, specifically including: firstly judging the fault repair degree of the preview results, and selecting the fault repair strategy with the highest fault repair degree as the first fault repair strategy; if there are multiple preview results with the highest fault repair degree, further judging the system operation energy consumption, and selecting the fault repair strategy with the lowest system operation energy consumption as the first fault repair strategy.

[0031] Furthermore, after the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification to obtain the execution result of the execution device and the fault repair result; If the ratio of the execution result of the execution device to the execution result previewed in the digital twin model is within the second preset range, the remote adjustment and remote control functions are verified to be normal; if the ratio of the execution result of the execution device to the execution result previewed in the digital twin model is not within the second preset range, an abnormal alarm of the remote adjustment and remote control functions is issued; the ratio of the execution result of the execution device to the execution result previewed in the digital twin model is the ratio of a parameter value in the execution result of the execution device to the corresponding parameter value in the previewed execution result; Among them, the test host database also records the preview execution results and the execution results of the execution device in real time, and draws a ratio curve between the execution results of the execution device and the preview execution results in the digital twin model. If the ratio curve deviates from the standard curve for a long time, the digital twin model is updated according to the execution results of the execution device.

[0032] The controller further determines whether the fault has been repaired according to the fault repair result. If the fault has been repaired, the fault signal feature and the first optimal fault repair strategy are correspondingly stored in the database.

[0033] Furthermore, the controller further determines whether the fault is repaired according to the fault repair result, which specifically includes: If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model, and the AI ​​model generates a second set of candidate fault repair strategies again according to the fault repair result and the first set of candidate fault repair strategies, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview, and the controller determines the second best fault repair strategy according to the preview result, and then sends the second best fault repair strategy to the main controller again for analysis and verification; The controller further determines whether the fault has been repaired according to the fault repair result, and if the fault has been repaired, stores the fault signal feature and the second best fault repair strategy in the database accordingly; If the fault is still not repaired, the feedback execution step is repeated until the fault is repaired and the fault signal characteristics and the Nth best fault repair strategy are stored accordingly; where N is an integer greater than or equal to 2.

[0034] Furthermore, the controller is provided with an upper limit value N of the number of execution feedbacks. max , if the number of execution feedback is equal to N max If the fault is still not repaired, the test host sends an alarm signal to the management personnel for manual intervention.

[0035] Furthermore, the controller regularly updates the AI ​​model based on the number of feedback executions and the best fault repair strategies in history, and automatically learns and updates according to the feedback results, forming a closed-loop self-updating evolutionary model to continuously improve the accuracy of the test system.

[0036] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a ROM, a RAM, etc.

[0037] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A substation four-remote signal testing method based on distributed monitoring, characterized in that: The acquisition device in the SCADA system collects the system telemetry signal and transmits it to the controller in the test host through the telemetering channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal has obvious abnormalities. If there is an abnormality, an abnormal telemetry signal alarm is issued; If there is no abnormality, the signal characteristics of the telemetry signal are extracted and searched in the database in the test host; The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator in the test host to simulate and send an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, and the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device of the SCADA system to perform corresponding actions. Then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

2. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 1, characterized in that: The controller determines whether the telemetry signal is an existing fault signal based on the retrieval result. If it is an existing fault signal, the test host uses the AI ​​model to extract the corresponding fault repair strategy, which specifically includes: setting a similarity threshold. If there is a signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an existing fault signal, and the fault type and fault repair strategy code stored in the database corresponding to the signal feature are extracted. The controller of the test host performs secondary confirmation on the extracted fault type and fault repair strategy code, and then controls the AI ​​model to extract the corresponding fault repair strategy.

3. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 2, characterized in that: If there is no signal feature in the database whose similarity with the signal feature of the telemetry signal exceeds the similarity threshold, it is determined to be an unknown fault signal, and the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview. The controller determines the first optimal fault repair strategy based on the preview results.

4. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 3, characterized in that: After the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification to obtain the execution result of the execution device and the fault repair result; If the ratio of the execution result to the execution result previewed in the digital twin model is within the second preset range, it is verified that the remote adjustment and remote control functions are normal; If the ratio of the execution result to the execution result previewed in the digital twin model is not within the second preset range, an abnormal alarm of remote adjustment and remote control function is issued; The controller further determines whether the fault has been repaired according to the fault repair result. If the fault has been repaired, the fault signal feature and the first optimal fault repair strategy are correspondingly stored in the database.

5. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 4, characterized in that: The controller further determines whether the fault is repaired according to the fault repair result, specifically including: If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model, and the AI ​​model generates a second set of candidate fault repair strategies according to the fault repair result and the first set of candidate fault repair strategies again, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview, and the controller determines the second best fault repair strategy according to the preview result, and then sends the second best fault repair strategy to the main controller again for analysis and verification; The controller further determines whether the fault has been repaired according to the fault repair result, and if the fault has been repaired, stores the fault signal feature and the second best fault repair strategy in the database accordingly; If the fault is still not repaired, the feedback execution step is repeated until the fault is repaired and the fault signal characteristics and the Nth best fault repair strategy are stored accordingly; where N is an integer greater than or equal to 2.

6. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 5, characterized in that: The controller is set with an upper limit value N of the number of execution feedback times max , if the number of execution feedback is equal to N max If the fault is still not repaired, the test host sends an alarm signal to the management personnel for manual intervention.

7. A method for testing four remote signals of a substation based on distributed monitoring as claimed in claim 6, characterized in that: The controller periodically updates the AI ​​model based on the number of feedback executions and the best fault repair strategy in history.

8. A substation four-remote signal test system based on distributed monitoring, characterized in that: The test system includes a test host and a SCADA system; the test host includes an AI model, a database, a signal simulation generator and a controller; the SCADA system includes a main controller, a collection device, and an execution device; The acquisition device acquires the system telemetry signal and transmits it to the controller in the test host through the telemetering channel. After receiving the telemetry signal, the controller analyzes and determines whether the telemetry signal is obviously abnormal. If there is an abnormality, an abnormal telemetry signal alarm is issued; If there is no abnormality, extract the signal characteristics of the telemetry signal and search it in the database; The controller determines whether the telemetry signal is an existing fault signal according to the search result. If it is an existing fault signal, the controller uses the AI ​​model to extract the corresponding fault repair strategy, wherein the fault repair strategy includes remote adjustment and remote control signals; The controller controls the signal simulation generator to simulate sending an analog signal identical to the telemetry signal to the main controller of the SCADA system, and the main controller determines whether the received analog signal is within a first preset range, and if not, issues a telemetering channel abnormality alarm; If so, the controller sends a fault repair strategy to the main controller, the main controller parses the adjustment parameters and control instructions contained in the remote adjustment and remote control signals and sends them to the execution device to perform corresponding actions, and then the acquisition device collects the execution feedback signal and returns it to the controller of the test host. The controller verifies whether the remote adjustment and remote control functions are normal based on the execution results.

9. A substation four-remote signal test system based on distributed monitoring as claimed in claim 8, characterized in that: The test host also includes a digital twin model; When the controller determines that there is no signal feature similar to the telemetry signal in the database based on the retrieval results, the controller extracts the signal feature with the highest similarity to the signal feature of the telemetry signal in the database and sends it to the AI ​​model. The AI ​​model generates a first set of candidate fault repair strategies based on the signal feature with the highest similarity, and then sends multiple candidate fault repair strategies in the first set of candidate fault repair strategies to the digital twin model for preview. The controller determines the first optimal fault repair strategy based on the preview results.

10. A substation four-remote signal test system based on distributed monitoring as claimed in claim 9, characterized in that: After the controller determines the first optimal fault repair strategy according to the preview result, the controller sends the first optimal fault repair strategy to the main controller for analysis and verification and obtains the fault repair result. If the controller determines that the fault has not been repaired according to the fault repair result, the fault repair result is fed back to the AI ​​model. The AI ​​model generates a second set of candidate fault repair strategies again according to the fault repair result and the first set of candidate fault repair strategies, and then sends multiple candidate fault repair strategies in the second set of candidate fault repair strategies to the digital twin model for preview. The controller determines the second optimal fault repair strategy according to the preview result, and then sends the second optimal fault repair strategy to the main controller for analysis and verification again.

Images