Power dispatch switching operation safety checking method and device based on artificial intelligence

Abstract

The application discloses an artificial intelligence-based power dispatch switching operation safety checking method and device, relates to the technical field of power dispatch, and comprises the following steps: in response to a switching operation instruction, performing semantic analysis and feature extraction on the switching operation instruction based on a preset language processing model to obtain operation features; traversing associated devices corresponding to operation devices in the operation features, and performing associated checking on state conversion of the operation devices and the associated devices according to a preset conversion rule to output a first checking result; obtaining continuous information and consistent information of the operation features according to an operation sequence corresponding to the switching operation instruction, performing dynamic consistency checking according to the continuous information and the consistent information to output a second checking result; and outputting a power dispatch switching operation safety checking result based on the first checking result and the second checking result. The application has the beneficial effect of improving the reliability and accuracy of power dispatch switching operation safety checking through a double-checking mechanism.

Description

Technical Field

[0001] This application relates to the field of power dispatching technology, and in particular to a method and equipment for safety verification of power dispatching switching operations based on artificial intelligence. Background Technology

[0002] Power dispatching switching operations are a core component in ensuring the safe and stable operation of the power grid. The standardization of these operations directly impacts the safety of grid equipment, personnel, and power supply reliability. Switching operations must strictly adhere to preset operating rules, requiring precise control over the state transitions of operating equipment, operating timing, and equipment relationships. Therefore, safety verification of switching operations has become one of the key technologies in power dispatching automation systems.

[0003] Related power dispatching switching operation safety verification technologies mitigate the risk of misoperation during switching operations through single-dimensional or limited-dimensional verification. For example, some solutions use preset grid equipment topology relationships to interlock and verify the state transitions of operating equipment, ensuring that the operation conforms to the grid wiring logic; other solutions use fixed operation timing templates to verify the execution order of switching operations, avoiding timing violations.

[0004] However, using a pre-defined power grid equipment topology to lock out the state transitions of operating equipment ignores the close electrical connections, logical relationships, and scheduling synchronization relationships in the power system. This results in the verification results not matching the actual operating conditions, leading to low accuracy and reliability, and posing safety risks. Summary of the Invention

[0005] This application addresses the technical problem of low accuracy and reliability in the safety verification of power dispatching switching operations in related technologies. It provides a method and device for safety verification of power dispatching switching operations based on artificial intelligence. By utilizing the first verification result of the device-associated state verification and the second verification result of the operation sequence time dynamic consistency verification, the final safety verification result is output. The reliability and accuracy of the safety verification of power dispatching switching operations are improved through the dual verification mechanism.

[0006] To achieve the above technical objectives, this application provides a technical solution: a power dispatching switching operation safety verification method based on artificial intelligence, comprising the following steps: responding to a switching operation command, performing semantic parsing and feature extraction on the switching operation command based on a preset language processing model to obtain operation features; traversing the associated devices corresponding to the operating devices in the operation features, performing association verification on the state transitions of the operating devices and associated devices according to preset conversion rules, and outputting a first verification result; obtaining continuous information and consistency information of the operation features according to the operation sequence corresponding to the switching operation command, performing dynamic consistency verification based on the continuous information and consistency information, and outputting a second verification result; and outputting the power dispatching switching operation safety verification result based on the first verification result and the second verification result.

[0007] Furthermore, the construction process of the preset language processing model includes: acquiring historical switching operation instructions and their corresponding historical operation behavior data; for each operation behavior, extracting historical switching operation instructions containing that operation behavior from the historical operation behavior data, and constructing an instruction set; acquiring common instruction text information in the instruction set, establishing a text behavior association relationship between the common instruction text information and the corresponding operation behavior, using the common instruction text information as keywords, and constructing a preset language processing model based on the text behavior association relationship, a preset regular expression, and keywords.

[0008] Furthermore, the step of obtaining common instruction text information in the instruction set, establishing a text behavior association between common instruction text information and corresponding operation behavior, and using common instruction text information as keywords to construct a preset language processing model based on text behavior association, preset regular expressions, and keywords includes: taking instruction text information in the instruction set whose frequency of occurrence is greater than a preset common threshold as common instruction text information; establishing a text behavior association between common instruction text information and corresponding operation behavior; taking historical switching operation instructions in the instruction set that do not have common instruction text information as instructions to be confirmed; retrieving all common instruction text information corresponding to the instructions to be confirmed, combining the common instruction text information, determining the differential instruction text information in the instructions to be confirmed, constructing a consensus relationship with the differential instruction text information and the corresponding common instruction text information; and constructing a preset language processing model based on text behavior association, preset regular expressions, consensus relationship, and keywords using common instruction text information as keywords.

[0009] Furthermore, the step of obtaining common instruction text information in the instruction set, establishing text behavior associations between common instruction text information and corresponding operational behaviors, and constructing a preset language processing model based on text behavior associations, preset regular expressions, and keywords, includes: taking the instruction text information with the highest frequency of occurrence in the instruction set as common instruction text information; establishing a first text behavior association between common instruction text information and corresponding operational behaviors; taking instruction text information in the instruction set that differs from common instruction text information as differing instruction text information; constructing a difference association based on the differing instruction text information and the corresponding historical switching operation instruction scenarios; constructing a text behavior association based on the first text behavior association and the difference association; and constructing a preset language processing model based on text behavior associations, preset regular expressions, and keywords, using common instruction text information as keywords.

[0010] Furthermore, the process of traversing the associated devices corresponding to the operating device in the operation features, and verifying the state transitions of the operating device and associated devices according to preset conversion rules, and outputting the first verification result includes: traversing the topological relationship of the current operating device, and taking devices with a forced interlocking relationship with the operating device as associated devices of the operating device; performing initial state verification based on the real-time operating status of the operating device, associated devices, and the initial state of the operating device in the operation features; obtaining the state transition association between the operating device and associated devices based on the initial state and target state of the operating device in the operation features, and performing conversion verification on the state transition association according to preset conversion rules; and outputting the first verification result with the initial state verification result and the conversion verification result.

[0011] Furthermore, the step of obtaining continuous and consistent information of operation features based on the operation sequence corresponding to the switching operation command, performing dynamic consistency verification based on the continuous and consistent information, and outputting a second verification result includes: determining the sequential execution relationship of operations based on the execution steps in the switching operation command, as the operation sequence; performing time-sequence regularization on the operation features according to the operation sequence to obtain continuous and consistent information; wherein, the continuous information includes at least the state transition connection time sequence of each operating device, and the consistent information includes at least the synchronous state transition time sequence of different operating devices; performing time-sequence verification based on the continuous information, consistent information, and the transition delay threshold of each operating device, and outputting a second verification result.

[0012] Furthermore, the associated devices corresponding to the operating devices in the traversal operation features are used to perform association verification on the state transitions of the operating devices and associated devices according to preset conversion rules, and the first verification result is output. This includes: obtaining historical cooperating action devices of the operating devices as candidate associated devices based on historical switching operation data, and obtaining the coordination probability between the candidate associated devices and the operating devices; constructing the association influence relationship between the coordination probability and the topological relationship based on the topological relationship between the operating devices and the candidate associated devices; obtaining the target associated device and its coordination probability based on the operating devices, candidate associated devices, association influence relationship and current topological relationship in the operation features; taking the target associated device that does not belong to any operating device in the operation features as the action device to be confirmed, and retrieving all coordination probabilities of the action device to be confirmed; calculating the first operation influence based on all coordination probabilities of the action device to be confirmed, and calculating the second operation influence based on the redundant actions of the action device to be confirmed; and performing association verification on the state transitions of the operating devices and the target associated devices based on the minimum operation influence among the first operation influence and the second operation influence, and outputting the first verification result.

[0013] Furthermore, the step of calculating the first operational impact based on all the coordination probabilities of the devices to be confirmed, and calculating the second operational impact based on the redundant actions of the devices to be confirmed, includes: calculating the first operational impact based on all the coordination probabilities of the devices to be confirmed and the importance of the devices corresponding to the coordination probabilities; obtaining power grid fluctuation data based on the state transitions of the devices to be confirmed, and calculating the second operational impact based on the degree of influence of the power grid fluctuation data on the power grid fluctuations.

[0014] Furthermore, the process of verifying the state transitions of the operating equipment and the target associated equipment based on the minimum operational impact of the first and second operational impacts, and outputting the first verification result, includes: normalizing the first and second operational impacts respectively and mapping them to the power grid operation risk dimension; selecting the operational impact of the minimum power grid operation risk dimension; if the operational impact of the minimum power grid operation risk dimension is the first operational impact, then it is determined that the equipment to be confirmed does not perform a coordinated action; if the operational impact of the minimum power grid operation risk dimension is the second operational impact, then it is determined that the equipment to be confirmed performs a coordinated action. The first verification result includes the missing information on the coordination of the equipment to be confirmed.

[0015] Another technical solution provided in this application is an artificial intelligence-based power dispatching switching operation safety verification device, used to implement the method described above, including: an instruction parsing module, which, in response to a switching operation instruction, performs semantic parsing and feature extraction on the switching operation instruction based on a preset language processing model to obtain operation features; a first verification module, which traverses the associated devices corresponding to the operation devices in the operation features, performs association verification on the state transitions of the operation devices and associated devices according to preset conversion rules, and outputs a first verification result; a second verification module, which obtains continuous information and consistency information of the operation features according to the operation sequence corresponding to the switching operation instruction, performs dynamic consistency verification based on the continuous information and consistency information, and outputs a second verification result; and a result output module, which outputs the power dispatching switching operation safety verification result with the first verification result and the second verification result.

[0016] The beneficial effects of this application are as follows: 1. By utilizing a preset language processing model to extract operational features from switching operation commands, the adaptability of the verification is improved. Furthermore, the state transitions of the operating equipment and its associated equipment are verified according to preset conversion rules, and the linkage constraints between equipment are used to achieve associated verification, improving the comprehensiveness of the verification. Simultaneously, the continuity and consistency of information in the operation sequence corresponding to the switching operation command are used to perform dynamic consistency verification of the operational features, achieving dynamic verification throughout the entire process. Therefore, by using the first verification result of the equipment associated state verification and the second verification result of the dynamic consistency verification of the operation sequence timing, the final safety verification result is output. This dual verification mechanism improves the reliability and accuracy of the safety verification of power dispatching switching operations.

[0017] 2. Based on the set of all instructions to be confirmed, obtain all common instruction text information corresponding to it, and combine all common instruction text information. The part of the instruction to be confirmed that differs from the combined common instruction text is the difference instruction text information. Associate the difference instruction text information with the common instruction text information in the same position after combination to form a consensus relationship. There is no need to build an additional thesaurus. Moreover, since the consensus relationship is built based on the synonyms of actual switching operations, it is more in line with the actual scenario than the thesaurus built only based on word synonyms. Thus, the differences in terminology are reflected in the preset language processing model, improving the ability to identify and parse synonymous and near-synonymous instructions, and enhancing the adaptability and robustness of the model.

[0018] 3. Perform timing verification based on whether the sequential state transition connection sequence of each operating device and the synchronous transition sequence of different operating devices meet the transition delay threshold of each operating device. This enables joint verification of the action timing and transition delay between multiple operating devices, avoiding electrical shocks, equipment damage, and malicious misoperations caused by equipment action without delay or timing conflicts, thereby improving the safety and stability of switching operations.

[0019] 4. Based on historical switching operation data, extract historically coordinated action devices with synchronized / sequential execution records with the operating equipment. These devices are designated as candidate associated devices. The frequency of joint execution between the operating equipment and each candidate associated device is statistically analyzed to calculate the coordination probability. Combining this with the actual power grid topology relationship between the operating equipment and the candidate associated devices, a correlation influence relationship between the coordination probability and the topology relationship is constructed. This makes the identification of associated devices more consistent with the actual power grid structure, avoiding mismatches caused by relying solely on historical frequencies and improving the accuracy of association identification. Furthermore, for devices to be confirmed in non-operational characteristics, the minimum operational impact in the dual-operation influence is used as a preset conversion rule, taking into account historical operating habits and redundancy risk avoidance, thereby improving the accuracy of switching operation verification. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the safety verification method for power dispatching switching operations based on artificial intelligence, as described in this application.

[0021] Figure 2 This is a schematic diagram of the collaborative verification process of the AI-based power dispatching switching operation safety verification method of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] like Figure 1 As shown, the safety verification method for power dispatching switching operations based on artificial intelligence includes the following steps: In response to switching operation commands, semantic parsing and feature extraction of the switching operation commands are performed based on a preset language processing model to obtain operation features; Traverse the associated devices corresponding to the operating devices in the operation features, perform association verification on the state transition of the operating devices and associated devices according to the preset transformation rules, and output the first verification result; The continuous and consistent information of the operation characteristics are obtained according to the operation sequence corresponding to the switching operation command. Dynamic consistency verification is performed based on the continuous and consistent information, and the second verification result is output. Output the safety verification result of power dispatching switching operation based on the first verification result and the second verification result.

[0024] In this embodiment, a preset language processing model is used to extract operational features from switching operation commands, improving the verification adaptability. Furthermore, the state transitions of the operating equipment and its associated equipment are verified according to preset conversion rules, and the linkage constraints between equipment are used to achieve correlation verification, improving the comprehensiveness of the verification. Simultaneously, the continuity and consistency of information in the operation sequence corresponding to the switching operation command are used to perform dynamic consistency verification of the operational features, achieving dynamic verification throughout the entire process. Therefore, the final safety verification result is output using the first verification result of the equipment-associated state verification and the second verification result of the dynamic consistency verification of the operation sequence timing. This dual verification mechanism improves the reliability and accuracy of the safety verification of power dispatching switching operations.

[0025] In some embodiments, a preset language processing model can be constructed based on the artificial intelligence agent. For example, a pre-trained language model based on the Transformer (self-attention mechanism) architecture can be used as a basis to obtain historical switching operation instructions and corresponding operation behaviors, perform model training, and obtain the preset language processing model.

[0026] In other embodiments, semantic parsing and feature extraction are performed on the switching operation instructions based on a preset language processing model to obtain operation features, including: Acquire historical switching operation instructions and their corresponding historical operation behavior data; For each type of operation, extract historical switching operation instructions containing that operation from historical operation data and construct an instruction set; Obtain common instruction text information from the instruction set, establish text behavior association between common instruction text information and corresponding operation behavior, use common instruction text information as keywords, and construct a preset language processing model based on text behavior association, preset regular expressions, and keywords.

[0027] Historical switching operation instructions and historical operation behavior data can be directly obtained from systems such as power dispatch automation systems, operation ticket management systems, and dispatch instruction recording systems. Historical operation behavior data includes at least the operation action, the operation object, the operation equipment, and the voltage level.

[0028] The historical switching operation instructions and historical operation behavior data are first preprocessed to remove duplicate and invalid data, and a mapping table between the two is established.

[0029] In one scenario, each operational action, each operational object, each operational device, each voltage level, and each operational result corresponds to a specific operational behavior. For example, a historical switching operation command might be "Open -10kV -Chengbei Line -101 - Circuit Breaker," which is included in sets of commands corresponding to the "Open" operation behavior, the "10kV" operation behavior, the "Chengbei Line" operation behavior, and the "101 Circuit Breaker" operation behavior. The switching operation commands are segmented using a dedicated word segmentation dictionary for the power industry, and common command text information within the same command set is extracted. Since the same command set contains the same operational behavior, this common command text information corresponds to that same operational behavior. This establishes a text-behavior association relationship, and the common command text information is used as keywords. Based on the text-behavior association relationship, a pre-defined regular expression, and the keywords, a pre-defined language processing model is constructed.

[0030] In this scenario, a separate instruction set is constructed for each individual operation to ensure that common instruction text corresponds to the operation, facilitating the subsequent extraction of operation features based on the instruction text. Understandably, stop words are removed from the instruction set before extracting common instruction text information to avoid interference from stop words.

[0031] In practical applications, differences in terminology may lead to a lack of complete commonality in the text information of instruction sets. Therefore, it is necessary to obtain the common instruction text information from the instruction set, establish the text behavior association between the common instruction text information and the corresponding operation behavior, use the common instruction text information as keywords, and construct a preset language processing model based on the text behavior association, preset regular expressions, and keywords, including: Instruction text information whose frequency in the instruction set is greater than a preset commonality threshold is considered as common instruction text information; Establish a text behavior association between common instruction text information and corresponding operation behaviors; Historical switching operation instructions that do not have common instruction text information in the instruction set are regarded as instructions to be confirmed; Retrieve all common instruction text information corresponding to the instruction to be confirmed, combine the common instruction text information, determine the difference instruction text information in the instruction to be confirmed, and construct a consensus relationship with the difference instruction text information and the corresponding common instruction text information. Using common instruction text information as keywords, a pre-defined language processing model is constructed based on text behavior relationships, pre-defined regular expressions, common meaning relationships, and keywords.

[0032] The preset common threshold can be adjusted according to actual business data. In this embodiment, the preset common threshold is set to 80%. One instruction set corresponds to one common instruction text information. Historical switching operation instructions in that instruction set that do not contain their corresponding common instruction text information are considered as instructions to be confirmed. It is understood that if the same instruction text information appears multiple times in a historical switching operation instruction, it is only counted as one occurrence; the frequency of occurrence is only calculated based on the number of times the instruction text information appears in different historical switching operation instructions.

[0033] In this scenario, common command text information is first extracted by setting a preset common threshold. Then, historical switching operation commands that do not possess common command text information are filtered out as commands requiring secondary semantic confirmation. Since a historical switching operation command may exist in different command sets, and different command sets correspond to different common command text information, all common command text information corresponding to the command to be confirmed is obtained based on all command sets to which the command to be confirmed belongs. All common command text information is then combined. The part of the command to be confirmed that differs from the combined common command text is the difference command text information. The difference command text information is associated with the common command text information in the same position after combination to form a consensus relationship. There is no need to build an additional thesaurus. Moreover, since the consensus relationship is built based on the synonyms of actual switching operations, it is more in line with the actual scenario than a thesaurus built only based on word synonyms. Thus, the differences in terminology are reflected in the preset language processing model, improving the ability to identify and parse synonymous and near-synonymous commands, and enhancing the model's adaptability and robustness.

[0034] In this process, while filtering out instructions to be confirmed, common instruction text information that is not present in the instructions to be confirmed is recorded. If the recorded common instruction text information is different from the common instruction text information in the same position after combination, no commonality relationship is established, and it is filtered out as erroneous data to avoid the influence of abnormal instructions on the commonality relationship and improve the recognition accuracy. In this embodiment, common instruction text information is combined using preset regular expressions.

[0035] In response to a switching operation command, a preset language processing model is invoked. The switching operation command is first identified based on keywords and preset regular expressions to preliminarily determine the corresponding operation behavior. If there is text information in the switching operation command that is not covered by keywords and preset regular expressions, the commonality relation is invoked to map the uncovered text information to the corresponding commonality command text information for secondary semantic recognition, thereby improving the recognition accuracy of the switching operation command.

[0036] Understandably, if there is still no corresponding difference instruction text information in the consensus relationship, an abnormal prompt will be issued for the switching operation instruction.

[0037] In another scenario, an operational behavior includes at least a combination of operational actions, operational objects, operational equipment, and voltage levels. It is understood that the operational behavior here corresponds to a dimension of historical operational behavior data; if the historical operational behavior data also includes other dimensions, the operational behavior also includes those other dimensions.

[0038] At this point, historical switching operation instructions with the same operational behavior are grouped into one instruction set. Each historical switching operation instruction corresponds to only one instruction set. Common instruction text information in the instruction set is obtained, and a text behavior association relationship is established between the common instruction text information and the corresponding operation behavior. Using the common instruction text information as keywords, a preset language processing model is constructed based on the text behavior association relationship, preset regular expressions, and keywords, including: The instruction text information with the highest frequency of occurrence in the instruction set is taken as the common instruction text information; Establish a first text behavior association between common instruction text information and corresponding operation behavior; The instruction text information that differs from the common instruction text information in the instruction set is regarded as the differential instruction text information. A differential association is constructed based on the differential instruction text information and the corresponding historical switching operation instruction scenarios; Text behavior relationships are constructed based on the first text behavior relationship and the difference relationship; Using common instruction text information as keywords, a pre-defined language processing model is constructed based on text behavior relationships, pre-defined regular expressions, and keywords.

[0039] In this scenario, instruction text information that differs from the common instruction text information in the instruction set is designated as differential instruction text information. Both the common and differential instruction text information express the same complete switching operation sequence. This establishes a correlation between the differential instruction text information and the switching operation instruction scenario. Furthermore, the switching operation instruction scenario retrieves the differential instruction text information for identification, improving scenario adaptability and preventing incorrectly interpreted switching operation instructions with textual errors from being mistakenly interpreted as semantic deviations and proceeding with normal switching operations. In this embodiment, the instruction source interface can be used to distinguish switching operation instruction scenarios. In other embodiments, pre-defined switching control areas can be used to distinguish switching operation instruction scenarios.

[0040] In response to a switching operation command, a preset language processing model is invoked. The switching operation command is first identified based on keywords and preset regular expressions. If the switching operation command matches the preset regular expression and contains keywords, the first text behavior association is invoked to determine the corresponding operation behavior. If the switching operation command matches the preset regular expression but does not contain keywords, the difference association is invoked to obtain the difference command text information and perform secondary semantic recognition to improve the recognition accuracy of the switching operation command.

[0041] Understandably, if the switching operation command still does not contain the corresponding difference command text information, an error message will be displayed for the switching operation command.

[0042] In another scenario, semantic parsing and feature extraction of switching operation commands are performed simultaneously based on the two scenarios mentioned above to obtain operation features.

[0043] Once the corresponding operation behavior is determined according to the preset language processing model, the corresponding operation device feature information and operation action feature information are extracted as operation features. In this embodiment, the operation features include at least the operation device, operation device type, operation device initial state, operation device target state, and operation type.

[0044] In this embodiment, the associated devices corresponding to the operating device in the operation features are traversed, and the state transitions of the operating device and associated devices are verified according to preset conversion rules. The first verification result is output, including: Traverse the topology of the current operating device and identify devices that have a mandatory interlocking relationship with the current operating device as associated devices of the current operating device. Initial state verification is performed based on the real-time operating status of the operating equipment and associated equipment, as well as the initial state of the operating equipment in the operating characteristics. Based on the initial state and target state of the operating device in the operating characteristics, obtain the state transition association between the operating device and the associated device, and perform conversion verification on the state transition association according to the preset conversion rules; Output the first verification result based on the initial state verification result and the transformed verification result.

[0045] Based on the topological relationships of the operating equipment, the system matches the other equipment affected by its actions. As associated equipment, the system cross-checks the real-time operating status of the operating equipment and associated equipment with the initial state of the operating equipment to ensure that the switching operation command corresponds to the actual state of the operating equipment and associated equipment, thus ensuring the adaptability of the switching operation command to the current equipment state. Simultaneously, based on the initial and target states of the operating equipment, the system establishes state transition relationships between equipment and combines preset transition rules to achieve compliance verification of linked transitions, rather than just verifying the state change of a single device. This ensures that the switching operation command conforms to the actual operating logic of the power system, improving the safety of switching operations.

[0046] Specifically, each operating device in the operating characteristics is taken as the current operating device in turn, and its topological relationship is traversed to obtain the associated devices of that operating device. At this time, there are two possibilities for the associated devices: those belonging to the operating characteristics and those not belonging to the operating characteristics.

[0047] In the initial state verification: When the associated device belongs to the operating device in the operating feature, retrieve the initial state of the operating device corresponding to the current operating device and the initial state of the operating device corresponding to the associated device in the operating feature, and verify the initial state with its real-time running state respectively. When the associated device is not an operating device in the operating features, the initial state of the operating device corresponding to the current operating device in the operating features is retrieved and the real-time operating state of the current operating device is used for the first initial state verification. The forced interlocking relationship is retrieved and the second initial state verification is performed based on the forced interlocking relationship and the real-time operating state of the associated device.

[0048] At this point, the initial state verification result is one or more of the following: the initial state of the current operating device is abnormal, the initial state of the associated device is abnormal, the initial state of both the current operating device and the associated device is abnormal, and the initial state of both the current operating device and the associated device is normal.

[0049] During conversion verification: When the associated device belongs to the operation device in the operation feature, the initial state and target state of the operation device corresponding to the current operation device and the associated device are retrieved from the operation feature. A first state transition is obtained based on the initial state and target state of the operation device corresponding to the current operation device. A second state transition is obtained based on the initial state and target state of the operation device corresponding to the associated device. A conversion verification is performed based on whether the first state transition and the second state transition meet the preset conversion rules. When the associated device is not an operating device in the operating features, the initial state and target state of the current operating device in the operating features are retrieved. The first state transition is obtained based on the initial state and target state of the current operating device. The transition is verified based on whether the first state transition and the real-time operating state of the associated device meet the preset transition rules.

[0050] At this point, the conversion verification result is one or more of the following: first state transition error, second state transition error, first state transition and second state transition mismatch, first state transition and second state transition matching normally, and first state transition normal. It is understandable that since the real-time operating status of the associated device has already been verified in the initial state verification, if the first state transition and the real-time operating status of the associated device do not conform to the preset conversion rules during the conversion verification, then the first state transition error is output as the conversion verification result.

[0051] In this embodiment, the preset conversion rules are set based on the forced locking relationship and the operation type conversion rules. The conversion verification is performed based on whether the first state conversion and the second state conversion conform to the preset conversion rules: If the first state transition and the corresponding operation type do not conform to the operation type transition rules in the preset transition rules, then output "First state transition error"; If the second state transition and the corresponding operation type do not conform to the operation type transition rules in the preset transition rules, then output a second state transition error. If the action relationship between the first state transition and the second state transition does not conform to the mandatory locking relationship in the preset transition rules, then output "matching error between the first state transition and the second state transition"; If the action relationship between the first state transition and the second state transition conforms to the forced interlocking relationship in the preset transition rules, then the output will show that the first state transition and the second state transition are matched normally.

[0052] The conversion verification is performed based on whether the real-time operating status of the first state transition and the associated device conforms to the preset conversion rules. If the first state transition and the corresponding operation type do not conform to the operation type transition rules in the preset transition rules, then output "First state transition error"; If the first state transition and the real-time operating status of the associated device do not conform to the mandatory interlocking relationship in the preset transition rules, then an error in the first state transition will be output. If the first state transition and the real-time operating status of the associated device conform to the mandatory interlocking relationship in the preset transition rules, then the first state transition is output as normal.

[0053] Among them, the operation type conversion rules are pre-configured according to the legal state conversion direction corresponding to various types of operating equipment. For example, in switch equipment, the operation type conversion rule is: the state conversion from closing to opening state corresponds to the operation type of switch opening operation.

[0054] The first verification result, output based on the initial state verification result and the transformed verification result, includes: When the initial state verification result is any one of the following: the initial state of the current operating device is abnormal, the initial state of the associated device is abnormal, or the initial state of both the current operating device and the associated device is abnormal, the first verification result is output based on the initial state verification result. When the initial state verification result indicates that the current operating device and associated devices are in normal initial states, the conversion verification result is obtained, and the first verification result is output based on the conversion verification result and the initial state verification result.

[0055] In this scenario, an initial state check is performed first. If any abnormal result exists in the initial state check, the first check result is output directly based on the initial state check result. In this case, the first check result includes the abnormal result from the initial state check result. If no abnormal result exists in the initial state check result, a transformation check is performed. The first check result is output based on both the transformation check result and the initial state check result. In this case, the first check result includes the normal result from both the transformation check result and the initial state check result, or it includes the abnormal result from both the transformation check result and the normal result from the initial state check result.

[0056] Understandably, when any abnormal result exists in the first verification result, an abnormal alarm for the switching operation command is immediately output, and the abnormality is located based on the abnormal result.

[0057] Based on the operation sequence corresponding to the switching operation command, continuous and consistent information of the operation characteristics are obtained. Dynamic consistency verification is performed based on the continuous and consistent information, and the second verification result is output, including: The sequential execution relationship of the operations is determined based on the execution steps in the switching operation instruction, and this sequence is used as the operation sequence. The operation features are time-ordered according to the operation sequence to obtain continuous information and consistent information; among them, the continuous information includes at least the state transition connection sequence of each operation device, and the consistent information includes at least the synchronous state transition sequence of different operation devices. Based on continuous information, consistent information, and the conversion delay threshold of each operating device, a timing check is performed, and a second check result is output.

[0058] In this embodiment, the transition delay threshold is set based on the historical action time of each operating device. Timing verification is performed based on whether the sequential state transition connection sequence of each operating device and the synchronous transition sequence of different operating devices meet the transition delay threshold of each operating device. This achieves joint verification of the action sequence and transition delay between multiple operating devices, avoiding electrical shocks, equipment damage, and malicious misoperations caused by lack of device action delay or timing conflicts, thereby improving the safety and stability of switching operations.

[0059] It is understandable that formal switching operation tickets contain timing information, but there may still be scattered dispatch instructions, temporary single-item instructions, and non-operation ticket instructions that may not have timing information. Therefore, when the switching operation instruction has timing information, the state transition connection timing and synchronous state transition timing are obtained based on the timing information; when the switching operation instruction does not have timing information, the state transition connection timing and synchronous state transition timing are calculated based on the information transmission delay.

[0060] The second verification result includes the timing conflict result and the timing compliance result.

[0061] In some cases, the power dispatching switching operation safety verification results are output based on the first verification result and the second verification result, including: If the first verification result contains an abnormal result, the power dispatching switching operation safety verification result will be output based on the first verification result. If the first verification result does not contain any abnormal results, the power dispatching switching operation safety verification result will be output synchronously based on the first verification result and the second verification result.

[0062] In other cases, the power dispatching switching operation safety verification results are output based on both the first and second verification results, including: Simultaneously acquire the first verification result and the second verification result; If the first verification result does not contain any abnormal results and the second verification result does not contain any timing conflict results, then the safety verification result of the power dispatch switching operation is passed; If the first verification result contains an abnormal result or the second verification result contains a timing conflict result, the power dispatch switching operation safety verification result will be "failed", and all abnormal results and timing conflict results will be summarized and output.

[0063] like Figure 2 As shown, in Embodiment 2 of this application, the difference from Embodiment 1 is that the associated devices corresponding to the operating device in the operation features are traversed, and the state transitions of the operating device and associated devices are verified according to preset conversion rules. The first verification result is output, including: Based on historical switching operation data, obtain historical cooperative action devices of the operating equipment as candidate associated devices, and obtain the coordination probability between the candidate associated devices and the operating equipment; By combining the topological relationships between operating devices and candidate associated devices, a correlation and influence relationship between the cooperation probability and the topological relationships is constructed. Based on the operational features, including the operating equipment, candidate associated equipment, associated influence relationships, and current topological relationships, obtain the target associated equipment and its collaborative probability. Using a target associated device that does not belong to any of the operation features as the action device to be confirmed, retrieve all the cooperative probabilities of the action device to be confirmed. The first operational impact is calculated based on all the cooperative probabilities of the devices to be confirmed, and the second operational impact is calculated based on the redundant actions of the devices to be confirmed. The state transitions of the operating device and the target associated device are correlated and verified using the minimum operational impact of the first operational impact and the second operational impact, and the first verification result is output.

[0064] In this embodiment, historically coordinated action devices with synchronous / sequential execution records with the operating equipment are extracted from historical switching operation data and designated as candidate associated devices. The frequency of common execution between the operating equipment and each candidate associated device is statistically analyzed to calculate the coordination probability. Combining this with the actual power grid topology relationship between the operating equipment and the candidate associated devices, a correlation influence relationship between the coordination probability and the topology relationship is constructed. This makes the determination of associated devices more consistent with the actual power grid structure, avoiding mismatches caused by relying solely on historical frequencies and improving the accuracy of association identification. Furthermore, for devices to be confirmed in non-operational features, the minimum operational influence in the dual operational influence is used as a preset conversion rule, taking into account historical operating habits and redundancy risk avoidance, thereby improving the accuracy of switching operation verification.

[0065] Specifically, the topology should include at least electrical distance, number of intermediate components, and equipment type. Historical switching operation data should include at least historical switching operation records, operation tickets, and sequential control operation logs.

[0066] Specifically, the frequency of joint execution of operating equipment and candidate associated equipment in historical switching operations is used as the coordination probability. Based on the influence of topological changes on the coordination probability, an association influence relationship is constructed.

[0067] It is understandable that when pre-constructing candidate associated devices, candidate associated device types can be constructed based on device types. When the target associated device is determined, the candidate associated device types corresponding to the device types of the operating device are retrieved, and the candidate associated devices in the current topology are selected based on the retrieved candidate associated device types.

[0068] Furthermore, based on the candidate associated devices corresponding to the operating devices in the operational characteristics, the association influence relationship is retrieved, and the coordination probability of each candidate associated device is obtained according to the current topology. In this embodiment, candidate associated devices with a coordination probability of 0 are filtered out, and the remaining candidate associated devices are used as target associated devices. In other embodiments, candidate associated devices with a preset coordination probability threshold are filtered out. The preset coordination probability threshold is set based on the coordination probability with the highest randomness in participating in the switching operation within a preset period.

[0069] The first operational impact is calculated based on the total coordination probability of the device to be confirmed, and the second operational impact is calculated based on the redundant actions of the device to be confirmed, including: The first operational impact is calculated based on the total coordination probability of the devices to be confirmed and the importance of the operating devices corresponding to the coordination probabilities; The power grid fluctuation data is obtained based on the state transition of the device to be confirmed, and the impact of the second operation is calculated based on the degree of influence of the power grid fluctuation data on the power grid fluctuation.

[0070] One device requiring confirmation of an action may correspond to multiple operating devices. Therefore, the first operational impact is calculated based on the importance weight of the operating devices and the weighted average of all coordination probabilities. In this embodiment, the importance weights of the busbar, main transformer, line, and disconnector decrease in that order.

[0071] Simultaneously, grid fluctuation data during the state transition of the device to be verified is calculated, and the second operational impact is obtained based on the degree of influence of the grid fluctuation data on grid security. In this embodiment, the degree of influence of the grid fluctuation data on grid fluctuation can be obtained based on the ratio of the grid fluctuation data to the grid fluctuation threshold.

[0072] The state transitions of the operating equipment and the target associated equipment are correlated and verified using the minimum operational impact from the first operational impact and the second operational impact. The first verification result is output, including: The impacts of the first and second operations are normalized and mapped to the power grid operation risk dimension. Select the operational impact based on the dimension of minimum power grid operation risk; If the operational impact of the minimum power grid operation risk dimension is the first operational impact, then it is determined that the equipment to be confirmed will not perform the coordinated action; If the operational impact of the minimum power grid operation risk dimension is the second operational impact, then it is determined that the equipment to be confirmed will perform a coordinated action, and the first verification result includes the missing information on the coordinated action of the equipment to be confirmed.

[0073] In this embodiment, the power grid operation risk dimension can be set according to the actual power grid operation requirements. For example, in a power grid with better fluctuation absorption, the power grid operation risk dimension mapped by the second operation impact is lower than that in a power grid with poor fluctuation absorption.

[0074] Understandably, in practical applications, collaborative action verification, initial state verification, and transformation verification can be performed simultaneously. The first verification result includes the initial state verification result, the transformation verification result, and whether there is any missing collaborative action of the device to be confirmed.

[0075] In this embodiment, the difference from Embodiment 1 is that, after the safety verification result of the output power dispatching switching operation is passed, the following is executed: The system acquires real-time action delays of the operating equipment, corrects the state transition timing and synchronization timing based on these delays, and updates the transition delay thresholds. Real-time action delay acquisition and dynamic timing correction are introduced in the execution phase after safety verification, adapting the state transition timing from theoretical settings to the actual execution timing of the equipment. This avoids unexpected overlap of operation timings during actual operation and improves the accuracy of switching operation timing control. Simultaneously, the transition delay thresholds are iteratively updated based on actual action delays, achieving adaptive optimization of the verification thresholds and ensuring that subsequent safety verifications better reflect the actual operating conditions of the equipment on-site.

[0076] It is understandable that the state transition timing and synchronous state transition timing are corrected only when the delay in the operation of the device causes the subsequent state transition connection timing or synchronous state transition timing to overlap.

[0077] As a third embodiment of this application, an artificial intelligence-based power dispatching switching operation safety verification device is connected to the power dispatching system and includes: The instruction parsing module responds to switching operation instructions by performing semantic parsing and feature extraction on the switching operation instructions based on a preset language processing model to obtain operation features. The first verification module is used to traverse the associated devices corresponding to the operation devices in the operation features, perform association verification on the state transitions of the operation devices and associated devices according to the preset transformation rules, and output the first verification result. The second verification module is used to obtain continuous information and consistency information of operation characteristics according to the operation sequence corresponding to the switching operation instruction, perform dynamic consistency verification based on the continuous information and consistency information, and output the second verification result. The result output module is used to output the safety verification results of power dispatching switching operations based on the first verification result and the second verification result.

[0078] In this embodiment, the instruction parsing module is connected to the power dispatching system, obtains the switching operation instructions output by the power dispatching system, and outputs the operation characteristics to the first verification module and the second verification module. The first verification module and the second verification module are connected to each other and are synchronously connected to the result output module.

[0079] The specific embodiments described above are preferred embodiments of the power dispatching switching operation safety verification method and equipment based on artificial intelligence in this application, and are not intended to limit the specific implementation scope of this application. The scope of this application includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A safety verification method for power dispatching switching operations based on artificial intelligence, characterized in that: Includes the following steps: In response to switching operation commands, semantic parsing and feature extraction of the switching operation commands are performed based on a preset language processing model to obtain operation features; Traverse the associated devices corresponding to the operating devices in the operation features, perform association verification on the state transition of the operating devices and associated devices according to the preset transformation rules, and output the first verification result; The continuous and consistent information of the operation characteristics are obtained according to the operation sequence corresponding to the switching operation command. Dynamic consistency verification is performed based on the continuous and consistent information, and the second verification result is output. Output the safety verification result of power dispatching switching operation based on the first verification result and the second verification result; The associated devices corresponding to the operating devices in the traversal operation features are used to perform association verification on the state transitions of the operating devices and associated devices according to preset transformation rules, and the first verification result is output, including: Based on historical switching operation data, obtain historical cooperative action devices of the operating equipment as candidate associated devices, and obtain the coordination probability between the candidate associated devices and the operating equipment; By combining the topological relationships between operating devices and candidate associated devices, the correlation and influence relationship between the cooperation probability and the topological relationships are constructed. Based on the operational features, including the operating equipment, candidate associated equipment, associated influence relationships, and current topological relationships, obtain the target associated equipment and its collaborative probability. Using a target associated device that does not belong to any of the operation features as the action device to be confirmed, retrieve all the cooperative probabilities of the action device to be confirmed. The first operational impact is calculated based on all the cooperative probabilities of the devices to be confirmed, and the second operational impact is calculated based on the redundant actions of the devices to be confirmed. The state transitions of the operating device and the target associated device are correlated and verified using the minimum operational impact of the first operational impact and the second operational impact, and the first verification result is output.

2. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 1, characterized in that: The construction process of the preset language processing model includes: Acquire historical switching operation instructions and their corresponding historical operation behavior data; For each type of operation, extract historical switching operation instructions containing that operation from historical operation data and construct an instruction set; Obtain common instruction text information from the instruction set, establish text behavior association between common instruction text information and corresponding operation behavior, use common instruction text information as keywords, and construct a preset language processing model based on text behavior association, preset regular expressions, and keywords.

3. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 2, characterized in that: The process of acquiring common instruction text information from the instruction set, establishing text behavior associations between common instruction text information and corresponding operation behaviors, and constructing a preset language processing model based on text behavior associations, preset regular expressions, and keywords, includes: Instruction text information whose frequency in the instruction set is greater than a preset commonality threshold is considered as common instruction text information; Establish a text behavior association between common instruction text information and corresponding operation behaviors; Historical switching operation instructions that do not have common instruction text information in the instruction set are regarded as instructions to be confirmed; Retrieve all common instruction text information corresponding to the instruction to be confirmed, combine the common instruction text information, determine the difference instruction text information in the instruction to be confirmed, and construct a consensus relationship with the difference instruction text information and the corresponding common instruction text information. Using common instruction text information as keywords, a pre-defined language processing model is constructed based on text behavior relationships, pre-defined regular expressions, common meaning relationships, and keywords.

4. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 2, characterized in that: The process of acquiring common instruction text information from the instruction set, establishing text behavior associations between common instruction text information and corresponding operation behaviors, and constructing a preset language processing model based on text behavior associations, preset regular expressions, and keywords, includes: The instruction text information with the highest frequency of occurrence in the instruction set is taken as the common instruction text information; Establish a first text behavior association between common instruction text information and corresponding operation behavior; The instruction text information that differs from the common instruction text information in the instruction set is regarded as the differential instruction text information. A differential association is constructed based on the differential instruction text information and the corresponding historical switching operation instruction scenarios; Text behavior relationships are constructed based on the first text behavior relationship and the difference relationship; Using common instruction text information as keywords, a pre-defined language processing model is constructed based on text behavior relationships, pre-defined regular expressions, and keywords.

5. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 1, characterized in that: The step of obtaining continuous and consistent information of operation features based on the operation sequence corresponding to the switching operation command, performing dynamic consistency verification based on the continuous and consistent information, and outputting a second verification result includes: The sequential execution relationship of the operations is determined based on the execution steps in the switching operation instruction, and this sequence is used as the operation sequence. The operation features are time-ordered according to the operation sequence to obtain continuous information and consistent information; among them, the continuous information includes at least the state transition connection sequence of each operation device, and the consistent information includes at least the synchronous state transition sequence of different operation devices. Based on continuous information, consistent information, and the conversion delay threshold of each operating device, a timing check is performed, and a second check result is output.

6. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 1, characterized in that: The step of calculating the first operational impact based on all cooperative probabilities of the device to be confirmed, and calculating the second operational impact based on the redundant actions of the device to be confirmed, includes: The first operational impact is calculated based on the total coordination probability of the devices to be confirmed and the importance of the operating devices corresponding to the coordination probabilities; The power grid fluctuation data is obtained based on the state transition of the device to be confirmed, and the impact of the second operation is calculated based on the degree of influence of the power grid fluctuation data on the power grid fluctuation.

7. The method for safety verification of power dispatching switching operations based on artificial intelligence as described in claim 6, characterized in that: The process of performing a correlation verification on the state transition of the operating device and the target associated device based on the minimum operational influence among the first and second operational influences, and outputting the first verification result includes: The impacts of the first and second operations are normalized and mapped to the power grid operation risk dimension. Select the operational impact based on the dimension of minimum power grid operation risk; If the operational impact of the minimum power grid operation risk dimension is the first operational impact, then it is determined that the equipment to be confirmed will not perform the coordinated action; If the operational impact of the minimum power grid operation risk dimension is the second operational impact, then it is determined that the equipment to be confirmed will perform a coordinated action, and the first verification result includes the missing information on the coordinated action of the equipment to be confirmed.

8. A power dispatching switching operation safety verification device based on artificial intelligence, used to implement the method as described in any one of claims 1 to 7, characterized in that: include: The instruction parsing module responds to switching operation instructions by performing semantic parsing and feature extraction on the switching operation instructions based on a preset language processing model to obtain operation features. The first verification module is used to traverse the associated devices corresponding to the operation devices in the operation features, perform association verification on the state transitions of the operation devices and associated devices according to the preset transformation rules, and output the first verification result. The second verification module is used to obtain continuous information and consistency information of operation characteristics according to the operation sequence corresponding to the switching operation instruction, perform dynamic consistency verification based on the continuous information and consistency information, and output the second verification result. The result output module is used to output the safety verification results of power dispatching switching operations based on the first verification result and the second verification result.

Images