Switching operation generation verification method and system based on NFC and intelligent ticket forming model

Abstract

The invention belongs to the technical field of electric power automation and intellectualization, and more particularly relates to a switching operation generation verification method and system based on NFC and an intelligent ticket forming model, and the method specifically comprises the steps: enabling an REID induction module to make contact with an NFC patch corresponding to a lock, obtaining an NFC unique ID, verifying whether the NFC unique ID meets a preset triggering condition or not, and if yes, carrying out the verification; if the conditions are met, the collected NFC data information is transmitted back to the power transformation operation inspection mobile operation platform; the method comprises the following steps: establishing an equipment ledger and lockset ledger module by establishing a power equipment knowledge graph in a power transformation operation inspection mobile operation platform, generating a switching operation rule decision tree, and performing one-to-one matching on equipment ledger names and lockset ledger names by utilizing a corresponding relationship between electrical equipment and locksets provided by a five-prevention system; according to the power equipment knowledge graph and the switching operation rule decision tree, in combination with the historical ticket and the typical ticket of the business data, the GLM model is adopted to carry out the operation logic learning of the intelligent ticket forming expert model to complete the operation ticket automation, and the anti-misoperation risk safety management and control capability of the transformer substation is improved.

Description

Technical Field

[0001] The present invention belongs to the field of power automation and intelligent technology, and more specifically, to a switching operation generation and verification method and system based on NFC and intelligent ticketing model. Background Art

[0002] As one of the important daily tasks of substation, switching operation must strictly follow its technical principles and organizational measures due to the particularity of power production to ensure the safety of personnel and equipment. However, in the actual production process, there are many factors that affect the standardization and normalization of switching operation. Even if organizational and technical measures to ensure safety and dangerous point control measures are formulated, accidents such as mistaken entry into live intervals and misoperation of equipment still occur from time to time, which seriously threatens the safety of front-line operators and equipment assets.

[0003] At present, in the switching operation link, the operation control of the switching operation link is strengthened through the execution rehearsal of the operation ticket, the management of safety tools and equipment, and the association of the two tickets, including equipment inspection, maintenance management, equipment maintenance and testing, and certain application results have been achieved, laying a solid foundation for ensuring the safety of operators and the stable operation of the power grid. However, in the process of switching operation, there are still points to be improved, including the following aspects:

[0004] 1) The pre-operation lock verification prompt and confirmation has not been implemented on the mobile operation platform for operation and inspection, which poses certain safety hazards and also creates a control blind spot for the overall switching operation process.

[0005] 2 Operation tickets are usually generated based on pre-confirmed equipment status, but may not fully reflect the actual situation on site, including whether the equipment and locks on site are fully matched, changes in the status or location of the equipment, etc. When the operator scans the lock information and finds that it is inconsistent with the operation ticket information, the operation will be blocked, which may cause unnecessary delays and confusion.

[0006] 3) When conducting grassroots research, the substation operation and maintenance professionals pointed out that the preparation of switching operation tickets has a strong logical correlation. At present, the operation tickets are still manually produced, which is inefficient, the means of verifying the correctness of the operation tickets are single, and during the operation, personnel frequently travel back and forth between the main control room and the site to confirm the status of the equipment. These problems seriously affect the quality and efficiency of on-site work.

[0007] Therefore, the present invention proposes a method and system for generating and verifying switching operations based on NFC and intelligent ticketing models. Summary of the invention

[0008] In order to address the deficiencies in the prior art, the present invention provides a method and system for generating and verifying switching operations based on NFC and the intelligent ticketing model. Based on a new generation of centralized monitoring systems and substation anti-mistaken locking systems, the present invention gives full play to the advantages of data integration and sharing, and builds a digital switching operation application through new technologies such as equipment knowledge graphs, intelligent ticketing expert models, and status verification. This creates full-process control of switching operations with intelligent ticketing, anti-mistaken locking confirmation, and real-time verification of equipment status, thereby improving the substation's risk safety control capabilities to prevent misoperation.

[0009] The present invention adopts the following technical solution.

[0010] The present invention provides a switching operation generation verification method based on NFC and intelligent ticketing model, comprising the following steps:

[0011] S1: Use the REID sensing module of the substation operation and maintenance mobile operation terminal to contact the NFC patch corresponding to the lock to obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform;

[0012] S2: In the substation operation and inspection mobile operation platform, the equipment ledger and lock ledger modules are constructed by establishing a knowledge graph of power equipment, generating a decision tree for switching operation rules, and using the correspondence between electrical equipment and locks provided by the five-defense system to match the equipment ledger names with the lock ledger names one by one;

[0013] S3: According to step S2, the knowledge graph of power equipment and the decision tree of switching operation rules are obtained, combined with the business data historical tickets and typical tickets, and the GLM model is used to learn the operation logic of the intelligent ticket expert model to complete the training of automatic generation of operation tickets;

[0014] S4: The power professional speech recognition model converts the power operator's voice instructions into text, and combines it with the intelligent ticket expert model in step S3, using the converted text data as input to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements;

[0015] S5: Based on the successful NFC verification in step S1 to ensure that the on-site conditions are correct, the voice command is input on-site through the voice recognition module in step S4 and a real-time operation ticket is automatically generated, the lock information in the operation ticket is identified, and the lock name corresponding to the operation device is identified using the terminology of the operation ticket;

[0016] S6: According to step S5, confirm the lock information corresponding to the switching operation equipment, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, check and remind the correctness and sequential execution of the locks of the operation object, complete the business closed loop of the lock verification link before the operation, and then confirm that the operator correctly performs the switching operation content.

[0017] Preferably, the step S1 of verifying whether the preset trigger condition is met includes:

[0018] Determine whether the NFC signal is valid. If the lock ID is not detected, the strength of the NFC patch signal does not meet the system's set range, and the distance between the device and the NFC patch does not meet the minimum distance, the system will display an error prompt to notify the operator to recalibrate the position or check the device;

[0019] If the NFC signal is valid, the unique ID of the NFC patch is used to determine whether the NFC patch matches the lock and confirm whether the operating device and the lock are correctly bound;

[0020] If the device and the lock do not match, the system will issue an alarm and prohibit further operation. The operator needs to reconfirm the consistency of the on-site device and the lock.

[0021] Preferably, step S2 comprises the following steps:

[0022] S2.1: Obtain equipment topology information corpus from multiple data sources and perform data cleaning and preprocessing on the equipment topology information corpus to extract entity attribute data related to power equipment and correlation data between equipment;

[0023] S2.2: The entity attribute data and association data obtained in step S2.1 are integrated into a knowledge graph of electric power equipment in a unified format through data fusion;

[0024] S2.3: Based on the power equipment knowledge graph of step S2.2, the logic and rules of the switching operation are extracted to construct a switching operation rule decision tree that meets the actual operation requirements. The decision tree automatically derives the switching operation steps of various types of equipment under different states.

[0025] S2.4: Arrange the physical device attributes and status recorded in the power equipment knowledge graph of step S2.2 into an equipment ledger, and arrange the recorded lock entity attributes and status into a lock ledger;

[0026] S2.5: Based on step S2.4, each device information in the device ledger is matched with the corresponding lock in the lock ledger. This matching information will provide a basis for comparing the subsequent operation ticket execution information with the NFC acquisition information to ensure that the device and lock names in the operation ticket are consistent.

[0027] Preferably, after screening and preprocessing, fusing the multi-distance data into a unified knowledge graph of electric power equipment includes:

[0028] The relationship edges between device nodes are used as the basic information of the power equipment knowledge graph;

[0029] Each device node contains the attribute information of the device, including the type, status, location and specifications of the device;

[0030] The relationship edge attributes between device nodes include type, dependency, and operation order;

[0031] The types include control, connection, dependency and influence, which represent the nature of the relationship between devices. The dependency refers to the strength of the relationship between devices and whether there is a mandatory operation sequence between the two devices. The operation sequence describes the sequence of device operations.

[0032] Preferably, the step S3 specifically includes:

[0033] S3.1: Using the GLM model based on historical operation tickets, typical tickets and common operation record data, a supervised fine-tuning algorithm is used to fine-tune the data in the field of power switching operation tickets, and a human feedback reinforcement learning optimization model is used based on actual needs and complex scenarios;

[0034] S3.2: Based on the GLM model optimized in step S3.1, the intelligent ticket expert model is constructed using the device nodes and corresponding relationship edges of the power equipment knowledge graph, and the intelligent ticket expert model reasoning is trained through the device operation sequence, state changes and the relationship between devices;

[0035] S3.3: Based on the intelligent ticket expert model obtained in step S3.2, the switching operation steps of each device node are inferred using the switching operation rule decision tree, and the intelligent ticket expert model derives the switching operation ticket that meets the operation specifications according to the current status of the device, the dependency relationship between the devices and the operation sequence;

[0036] S3.4: The intelligent ticket expert model combined with the thinking chain algorithm gradually infers the switching operation ticket generation task obtained in step S3.3 and breaks the task into multiple small steps. Each step is inferred based on the equipment status, operation sequence and dependency relationship, and accurate operation ticket content is gradually generated.

[0037] S3.5: The current state of the equipment is confirmed in real time through knowledge graph query in the reasoning process in step S3.4 and the state is verified. If the equipment state is inconsistent with the reasoning step, the system will issue an alarm to readjust the operation steps;

[0038] S3.6: Based on the reasoning verification of step S3.5, combined with the reasoning results of the expert model, the information of the equipment knowledge graph and the switching operation rules, a switching operation ticket that complies with the power grid management regulations is automatically generated.

[0039] Preferably, the supervised fine-tuning algorithm in step S3.1 specifically includes:

[0040] Collect historical operation tickets, typical operation tickets and common operation record data from power equipment and manually annotate them to clarify the correct sequence of each operation step, the relationship between equipment, and the changes in equipment status;

[0041] The labeled historical data is provided as input to the GLM model for training. The standard supervised fine-tuning loss function is used to calculate the difference between the output of the analysis model and the true label. By minimizing the loss function, the model learns how to generate correct operation steps and related decisions based on the input data.

[0042] Preferably, the human feedback reinforcement learning optimization model in step S3.1 specifically includes:

[0043] Define a reinforcement learning environment suitable for power switching operations to simulate different scenarios during the switching operation, including device state changes, dependencies between devices, and operation sequences;

[0044] Define state space and action space, where the state refers to the current state of the device and the action refers to specific operation steps;

[0045] Incorporate industry experts into the feedback process. Experts provide feedback on each operation step through a scoring mechanism based on the operation steps output by the GLM model, including scoring the correctness, timing, and whether the dependencies of the operation steps meet the specifications.

[0046] These expert scoring data feedbacks are converted into reward signals for the training of reinforcement learning algorithms. The model is updated through the reinforcement learning algorithm. The reward signals of each operation step affect the weight adjustment of the model, so that the GLM model can make optimized decisions according to the expert feedback during the next operation.

[0047] Preferably, the step S4 specifically includes:

[0048] S4.1: Extract power professional terms and phonetic features related to the terms from the multi-source power knowledge base, and annotate all terms into a standardized format;

[0049] The power professional terms include station equipment, substation, operation specification terms and geographical location;

[0050] The knowledge base includes two-ticket management regulations, substation operation procedures, electrical equipment operation procedures, operation execution points, equipment wiring methods, operation matters, equipment operation rules, and equipment error prevention rules;

[0051] The speech features associated with the term include the frequency spectrum, pitch, and volume of the audio;

[0052] S4.2: manually input the power industry professional terms obtained in step S4.1 to generate a speech recognition library containing power industry professional terms;

[0053] S4.3: Manually record power industry terms and generate more than 100 hours of audio files, clean and annotate all recorded speech to ensure that each audio file corresponds to the correct terminology and instructions;

[0054] S4.4: Use the PEFT algorithm to fine-tune the electric power professional speech recognition model through artificial recording samples, adjust the model weight, adjust the model parameters according to the characteristics of the speech data, and complete the optimization of the electric power professional speech recognition model.

[0055] Preferably, the NFC patch ID is matched with the lock ledger. By reading the unique identifier on the NFC patch, the system searches for the corresponding lock information in the lock ledger. When the correspondence between the device and the lock is known, the system confirms whether the patch matches the lock recorded in the lock ledger through the NFC patch ID. It is valid when the NFC patch ID matches the lock ID. When the match is successful, the system records this relationship and provides the necessary basis for subsequent lock verification and operation ticket execution. If the match fails, the current operation is interrupted.

[0056] The present invention also provides a switching operation generation verification system based on NFC and intelligent ticket model, which runs the aforementioned switching operation generation verification method based on NFC and intelligent ticket model:

[0057] The NFC induction verification module is used to use the REID induction module of the substation operation and maintenance mobile operation terminal to perform contact induction with the NFC patch corresponding to the lock, obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform;

[0058] The equipment and lock matching module is used to build the equipment ledger and lock ledger module by establishing the knowledge graph of power equipment in the substation operation and inspection mobile operation platform, generate the decision tree of switching operation rules, and use the correspondence between electrical equipment and locks provided by the five-defense system to match the equipment ledger name with the lock ledger name one by one;

[0059] The automatic operation ticket generation module is used to combine the business data historical tickets and typical tickets based on the power equipment knowledge graph and the switching operation rule decision tree provided by the equipment and lock matching module, and use the GLM model to learn the operation logic of the intelligent ticket expert model to complete the training of automatic operation ticket generation;

[0060] The power professional voice recognition module is used to convert the voice instructions of power operators into text, and is combined with the intelligent ticket expert model of the automatic operation ticket generation module to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements using the converted text data as input;

[0061] The module corresponding to the operation ticket and the operation equipment lock is used to ensure the correctness of the on-site conditions based on the successful NFC verification of the NFC induction verification module. The voice command is input on-site through the power professional voice recognition module and the real-time operation ticket is automatically generated. The lock information in the operation ticket is identified, and the name of the lock corresponding to the operation equipment is identified using the operation ticket terminology specification, which provides a basis for the subsequent collection and verification of whether the lock is correct;

[0062] The switching execution module is used to confirm the corresponding lock information of the switching operation equipment according to the operation ticket and the corresponding module of the operating equipment lock, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, the correctness and sequential execution of the locks of the operation objects are verified and reminded, completing the business closed loop of the lock verification link before the operation, and then confirming that the operating personnel correctly executed the switching operation content.

[0063] Compared with the prior art, the beneficial effects of the present invention include at least:

[0064] 1. The anti-misoperation lock verification and control based on the switching operation of the substation operation and maintenance mobile operation platform is not only an important measure to fully respond to the State Grid headquarters' risk safety control of on-site operations, but also an effective exploration to further enrich the "five-level five-control" risk prevention and control system of production site operations. It also uses NFC patches as lock verification and confirmation carriers to complete the data comparison and confirmation of lock ledger data and substation operation and maintenance mobile operation ticket information, further clearing the blind spots of switching operation execution control, preventing operators from operating incorrectly or missing operation links, greatly improving the safety of on-site operators and equipment assets, and adding safety insurance for the stable operation of the power grid.

[0065] 2. By utilizing the cooperation of NFC and REID sensing modules, each operation step can be verified, the correctness of the on-site equipment and locks can be confirmed, and operational accidents caused by incorrect identification of locks or equipment can be avoided. By verifying the status of the locks in real time, it can be ensured that the operators will not perform incorrect operations, further improving the safety of on-site operations.

[0066] 4. By performing NFC verification first, the matching relationship between the device and the lock can be confirmed before the operation ticket is generated, avoiding incorrect device information or lock information when generating the operation ticket. This advance verification mechanism ensures that the lock and device in the operation ticket are more accurately matched, reducing the risk of inconsistent device and lock information.

[0067] 3. Combining the GLM model with the knowledge graph of power equipment and automatically generating operation tickets can reduce manual intervention and improve the accuracy of operations. At the same time, voice recognition technology enables operators to issue instructions in natural language to quickly complete operation tasks, further accelerate the switching operation process, and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a schematic diagram of the steps of a switching operation generation verification method based on NFC and smart ticketing model provided in accordance with an embodiment of the present invention;

[0069] Figure 2 According to an embodiment of the present invention, Figure 1 A logic-refined judgment flow chart of the verification method for the switching operation of NFC and the smart ticketing model;

[0070] Figure 3 It is a corresponding schematic diagram of the power equipment knowledge graph taking "xx line 26126 circuit breaker" as an example in a specific embodiment. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The described embodiments are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0072] like Figure 1 to Figure 2 As shown, embodiment 1 of the present invention provides a switching operation generation verification method based on NFC and smart ticketing model, comprising the following steps:

[0073] S1: Use the REID sensing module of the substation operation and maintenance mobile operation terminal to contact the NFC patch corresponding to the lock to obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform.

[0074] Preferably but not limitatively, the step S1 of verifying whether the preset trigger condition is met specifically includes:

[0075] Determine whether the NFC signal is valid. If the lock ID is not detected, the strength of the NFC patch signal does not meet the system's set range, and the distance between the device and the NFC patch does not meet the minimum distance, the system will display an error prompt to notify the operator to recalibrate the position or check the device;

[0076] If the NFC signal is valid, the unique ID of the NFC patch is used to determine whether the NFC patch matches the lock and confirm whether the operating device and the lock are correctly bound;

[0077] If the device and the lock do not match, the system will issue an alarm and prohibit further operation. The operator needs to reconfirm the consistency of the on-site device and the lock.

[0078] In the embodiment of the present invention, it is checked whether the strength of the NFC patch signal meets the set range of the system. Usually, the setting range of the patch signal strength usually depends on multiple factors, such as the specific application scenario of the device, the version of the NFC technology, and the system's requirements for signal stability and distance. Generally speaking, the unit of NFC signal strength is RSSI. In different devices and systems, the RSSI threshold may be different. The system setting range is generally between -60dBm and -80dBm. The specific value depends on the design and use requirements of the system. In some more stringent applications, the signal strength may be required to be above -70dBm to ensure the stability and security of communication.

[0079] S2: In the substation operation and maintenance mobile operation platform, the equipment ledger and lock ledger modules are constructed by establishing a knowledge graph of power equipment, and a decision tree for switching operation rules is generated. The correspondence between electrical equipment and locks provided by the five-defense system is used to match the equipment ledger names with the lock ledger names one by one, providing a basis for comparing the subsequent operation ticket execution information with the NFC acquired information.

[0080] In the embodiment of the present invention, the NFC patch ID is matched with the lock register. By reading the unique identifier (such as the AES key) on the NFC patch, the system will search for the corresponding lock information in the lock register. When the corresponding relationship between the device and the lock is known, the system confirms whether the patch matches the lock recorded in the lock register through the NFC patch ID. Only when the NFC patch ID matches the lock ID is it considered valid; when the match is successful, the system records this relationship and provides the necessary basis for subsequent lock verification and operation ticket execution; if the match fails, the current operation is interrupted.

[0081] Preferably but not limiting, the step S2 specifically comprises the following steps:

[0082] S2.1: Obtain equipment topology information corpus from multiple data sources and perform data cleaning and preprocessing on the equipment topology information corpus to extract entity attribute data related to power equipment and correlation data between equipment;

[0083] The device topology information corpus includes device information, error prevention rules, operating status and load;

[0084] The device information includes device records and topology information;

[0085] The error prevention rules include equipment error prevention relations;

[0086] The operating state includes switch state, knife switch state and pressure plate state;

[0087] The load load includes current, voltage and power;

[0088] It is further explained that the relevant entity attribute data includes device type, device ID, device status, device location and device parameters. The main goal of entity attribute extraction is to convert unstructured or semi-structured data into attributes in a structured form, so as to provide available data for subsequent relationship extraction and knowledge graph construction; at the same time, the association data between the devices includes identifying the physical connection relationship between devices, extracting the control relationship and identifying the operation sequence and dependency relationship. For example, the physical connection relationship between devices refers to which circuit breaker is connected to which bus, which switch controls which device, etc. The control relationship refers to the fact that the switch status of a certain device may control the operation of another device; the operation sequence and dependency relationship refer to the fact that some devices must be operated in a specific order, such as turning off the circuit breaker first and then performing the isolating switch operation. The core goal of extracting and processing the association data between devices is to ensure that the constructed knowledge graph is both complete and accurate, and to avoid erroneous information affecting subsequent operations and analysis.

[0089] The data sources selected in the embodiment of the present invention include equipment inventory management system, monitoring control and data acquisition system, five-defense system, equipment status online monitoring system, maintenance and repair records, engineering drawings and design documents, contracts and procurement records, standards and specification documents, historical fault and accident reports, external databases, Internet of Things devices or handheld devices, etc.

[0090] In the embodiment of the present invention, data cleaning and preprocessing include removing duplicate data, filling missing data and formatting; removing duplicate data means that there may be duplicate records in data sources such as equipment ledgers, topological relationships and switch states, and through deduplication operations, it is ensured that each device has only a unique identifier in the data; filling missing data means checking the missing fields in the data (such as equipment model, installation time, etc.), supplementing the missing data by interpolation or comparing with other data sources to complete the information; formatting means unifying the formats in different data sources to ensure the uniformity between the fields and avoid data processing errors caused by inconsistent formats.

[0091] S2.2: The entity attribute data and association data obtained in step S2.1 are integrated into a knowledge graph of electric power equipment in a unified format through data fusion.

[0092] After screening and preprocessing, the multi-distance data is fused to build a unified knowledge graph of power equipment;

[0093] The relationship edges between device nodes are included, and each node contains attribute information of the device, including the type, status, location and specifications of the device; the attributes of the relationship edges between the device nodes include type, dependency, and operation sequence; the type can be control, connection, dependence, and influence, etc., indicating the nature of the relationship between devices, and the dependency refers to the description of the strength of the relationship between devices (for example, high, medium, and low dependency), or indicates whether there is a mandatory operation sequence between the two devices, and the operation sequence describes the order of device operations, for example, one device must be operated before another device.

[0094] The power equipment knowledge graph is used to graphically represent various relationships between devices and is displayed through graphical tools (such as Gephi, Graphviz, etc.). The device nodes are the vertices in the graph, and the relationship edges are the connections between the vertices. Different edge types (control, dependency, connection, etc.) can be distinguished by different colors, line types, etc.

[0095] Figure 3 Taking "xx line 26126 circuit breaker" as an example, the specific presentation form of the knowledge graph is demonstrated, which includes various relationships related to circuit breakers, such as switch protection, starting failure pressure plate, load side relationship, I mother power side relationship, II mother power side relationship, etc., as well as related equipment entities such as disconnectors.

[0096] S2.3: Based on the power equipment knowledge graph of step S2.2, the logic and rules of the switching operation are extracted to construct a switching operation rule decision tree that meets the actual operation requirements. The decision tree automatically derives the switching operation steps of various types of equipment under different states.

[0097] S2.4: Arrange the physical device attributes and status recorded in the power equipment knowledge graph of step S2.2 into an equipment ledger, and arrange the recorded lock entity attributes and status into a lock ledger;

[0098] S2.5: Based on step S2.4, match each device information in the device ledger with the corresponding lock in the lock ledger. This matching information will provide a basis for comparing the subsequent operation ticket execution information with the NFC acquisition information to ensure that the device and lock names in the operation ticket are consistent.

[0099] S3: According to step S2, the knowledge graph of power equipment and the decision tree of switching operation rules are obtained, combined with the business data historical tickets and typical tickets, and the GLM model is used to learn the operation logic of the intelligent ticket expert model to complete the training of automatic generation of operation tickets.

[0100] Optionally but not limitingly, step S3 specifically includes:

[0101] S3.1: Using the GLM model through historical operation tickets, typical tickets and common operation record data, a supervised fine-tuning algorithm is used to fine-tune the data in the field of power switching operation tickets, and human feedback reinforcement learning optimization model is used according to actual needs and complex scenarios.

[0102] Optionally but not limitingly, the supervised fine-tuning algorithm in step S3.1 specifically includes:

[0103] Collect historical operation tickets, typical operation tickets and common operation record data from power equipment and manually annotate them to clarify the correct sequence of each operation step, the relationship between equipment, and the changes in equipment status;

[0104] The labeled historical data is provided as input to the GLM model for training. The standard supervised fine-tuning loss function is used to calculate the difference between the output of the analysis model and the true label. By minimizing the loss function, the model learns how to generate correct operation steps and related decisions based on the input data.

[0105] The human feedback reinforcement learning optimization model in step S3.1 specifically includes:

[0106] Define a reinforcement learning environment suitable for power switching operations to simulate different scenarios during the switching operation, including device state changes, dependencies between devices, and operation sequences;

[0107] Define state space and action space, where the state refers to the current state of the device and the action refers to specific operation steps;

[0108] Incorporate industry experts into the feedback process. Experts provide feedback on each operation step through a scoring mechanism based on the operation steps output by the model, including scoring the correctness, timing, and whether the dependencies of the operation steps meet the specifications.

[0109] These expert feedbacks are converted into reward signals for the training of reinforcement learning algorithms. The model is updated through the reinforcement learning algorithm. The reward signals of each operation step affect the weight adjustment of the model, so that the model can make more appropriate decisions based on the expert feedback during the next operation.

[0110] As expert feedback accumulates, the model is gradually improved to better adapt to the complex scenarios of power switching operations and can better handle abnormal situations, emergency operations or dependencies between complex equipment.

[0111] In the embodiment of the present invention, the state space refers to the current state of each device in the system, including the device type, device status (such as the switch status and fault status of switches and circuit breakers, etc.) and the dependencies between related devices and the characteristics of the current operating environment (such as power load, environmental status, etc.); the action space refers to the operation steps that the intelligent agent can perform, including specific switch control, status query, operation sequence adjustment, etc.

[0112] S3.2: Based on the GLM model optimized in step S3.1, the intelligent ticketing expert model is constructed using the device nodes and corresponding relationship edges of the power equipment knowledge graph, and the intelligent ticketing expert model reasoning is trained through the equipment operation sequence, state changes and the relationship between devices.

[0113] S3.3: Based on the intelligent ticket-making expert model obtained in step S3.2, the switching operation steps of each device node are inferred using the switching operation rule decision tree. According to the current status of the device, the dependency relationship between devices and the operation sequence, the intelligent ticket-making expert model derives the switching operation ticket that meets the operation specifications.

[0114] S3.4: The intelligent ticket expert model combined with the thinking chain algorithm gradually infers the switching operation ticket generation task obtained in step S3.3 and breaks the task into multiple small steps. Each step is inferred based on the equipment status, operation sequence and dependency relationship, and accurate operation ticket content is gradually generated.

[0115] S3.5: The reasoning process in step S3.4 is used to confirm the current status of the equipment in real time through knowledge graph query and verify the status. If the equipment status is inconsistent with the reasoning steps, the system will issue an alarm to readjust the operation steps.

[0116] S3.6: Based on the reasoning verification of step S3.5, combined with the reasoning results of the expert model, the information of the equipment knowledge graph and the switching operation rules, a switching operation ticket that complies with the power grid management regulations is automatically generated.

[0117] S4: The electric power professional speech recognition model converts the voice instructions of the electric power operator into text, and combines it with the intelligent ticket expert model in step S3.3, using the converted text data as input to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements.

[0118] Optionally but not limitingly, step S4 specifically includes:

[0119] S4.1: Extract power professional terms and phonetic features related to the terms from the multi-source power knowledge base, and annotate all terms into a standardized format;

[0120] The power professional terms include station equipment, substation, operation specification terms and geographical location;

[0121] The knowledge base includes two-ticket management regulations, substation operation procedures, electrical equipment operation procedures, operation execution points, equipment wiring methods, operation matters, equipment operation rules, and equipment error prevention rules;

[0122] The term relates to speech features including the frequency spectrum, pitch and volume of the audio.

[0123] S4.2: manually input the power industry professional terms obtained in step S4.1 to generate a speech recognition library containing power industry professional terms;

[0124] S4.3: Manually record power industry terms and generate more than 100 hours of audio files, clean and annotate all recorded speech to ensure that each audio file corresponds to the correct terminology and instructions;

[0125] S4.4: Use the PEFT algorithm to fine-tune the electric power professional speech recognition model through artificial recording samples, adjust the model weight, and adjust the model parameters according to the characteristics of the speech data to ensure its adaptability in the substation environment;

[0126] S5: Based on the successful NFC verification in step S1, ensure that the on-site conditions are correct. Enter the voice command on-site through the voice recognition module in step S4 and automatically generate a real-time operation ticket. Identify the lock information in the operation ticket, and use the operation ticket terminology to identify the lock name corresponding to the operating device, so as to provide a basis for subsequent collection and verification of whether the lock is correct.

[0127] In the implementation of the present invention, generating an operation ticket and identifying a lock with NFC are two parallel and mutually supportive tasks. Generating an operation ticket is mainly to ensure the standardization of operation steps and equipment, and NFC operation is to verify whether the lock information matches the equipment. If the generation of the operation ticket contains detailed information on the equipment and the lock, then generating the operation ticket after confirming the lock information may be more accurate.

[0128] S6: According to step S5, confirm the lock information corresponding to the switching operation equipment, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, check and remind the correctness and sequential execution of the locks of the operation object, complete the business closed loop of the lock verification link before the operation, and then confirm that the operator correctly performs the switching operation content.

[0129] Embodiment 2 of the present invention provides a switching operation generation verification system based on NFC and smart ticket model, and runs the switching operation generation verification method based on NFC and smart ticket model described in embodiment 1, including:

[0130] The NFC induction verification module is used to use the REID induction module of the substation operation and maintenance mobile operation terminal to perform contact induction with the NFC patch corresponding to the lock, obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform;

[0131] The equipment and lock matching module is used to build the equipment ledger and lock ledger module by establishing the knowledge graph of power equipment in the substation operation and inspection mobile operation platform, generate the decision tree of switching operation rules, and use the correspondence between electrical equipment and locks provided by the five-defense system to match the equipment ledger name with the lock ledger name one by one;

[0132] The automatic operation ticket generation module is used to combine the business data historical tickets and typical tickets based on the power equipment knowledge graph and the switching operation rule decision tree provided by the equipment and lock matching module, and use the GLM model to learn the operation logic of the intelligent ticket expert model to complete the training of automatic operation ticket generation;

[0133] The power professional voice recognition module is used to convert the voice instructions of power operators into text, and is combined with the intelligent ticket expert model of the automatic operation ticket generation module to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements using the converted text data as input;

[0134] The module corresponding to the operation ticket and the operation equipment lock is used to ensure the correctness of the on-site conditions based on the successful NFC verification of the NFC induction verification module. The voice command is input on-site through the power professional voice recognition module and the real-time operation ticket is automatically generated. The lock information in the operation ticket is identified, and the name of the lock corresponding to the operation equipment is identified using the operation ticket terminology specification, which provides a basis for the subsequent collection and verification of whether the lock is correct;

[0135] The switching execution module is used to confirm the corresponding lock information of the switching operation equipment according to the operation ticket and the corresponding module of the operating equipment lock, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, the correctness and sequential execution of the locks of the operation objects are verified and reminded, completing the business closed loop of the lock verification link before the operation, and then confirming that the operating personnel correctly executed the switching operation content.

[0136] Compared with the prior art, the beneficial effects of the present invention include at least:

[0137] 1. The anti-misoperation lock verification and control based on the switching operation of the substation operation and maintenance mobile operation platform is not only an important measure to fully respond to the State Grid headquarters' risk safety control of on-site operations, but also an effective exploration to further enrich the "five-level five-control" risk prevention and control system of production site operations. It also uses NFC patches as lock verification and confirmation carriers to complete the data comparison and confirmation of lock ledger data and substation operation and maintenance mobile operation ticket information, further clearing the blind spots of switching operation execution control, preventing operators from operating incorrectly or missing operation links, greatly improving the safety of on-site operators and equipment assets, and adding safety insurance for the stable operation of the power grid.

[0138] 2. By utilizing the cooperation of NFC and REID sensing modules, each operation step can be verified, the correctness of the on-site equipment and locks can be confirmed, and operational accidents caused by incorrect identification of locks or equipment can be avoided. By verifying the status of the locks in real time, it can be ensured that the operators will not perform incorrect operations, further improving the safety of on-site operations.

[0139] 4. By performing NFC verification first, the matching relationship between the device and the lock can be confirmed before the operation ticket is generated, avoiding incorrect device information or lock information when generating the operation ticket. This advance verification mechanism ensures that the lock and device in the operation ticket are more accurately matched, reducing the risk of inconsistent device and lock information.

[0140] 3. Combining the GLM model with the knowledge graph of power equipment and automatically generating operation tickets can reduce manual intervention and improve the accuracy of operations. At the same time, voice recognition technology enables operators to issue instructions in natural language to quickly complete operation tasks, further accelerate the switching operation process, and improve work efficiency.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A switching operation generation and verification method based on NFC and intelligent ticketing model, characterized in that: The following steps are involved: S1: Use the REID sensing module of the substation operation and maintenance mobile operation terminal to contact the NFC patch corresponding to the lock to obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform; S2: In the substation operation and inspection mobile operation platform, the equipment ledger and lock ledger modules are constructed by establishing a knowledge graph of power equipment, generating a decision tree for switching operation rules, and using the correspondence between electrical equipment and locks provided by the five-defense system to match the equipment ledger names with the lock ledger names one by one; S3: According to step S2, the knowledge graph of power equipment and the decision tree of switching operation rules are obtained, combined with the business data historical tickets and typical tickets, and the GLM model is used to learn the operation logic of the intelligent ticket expert model to complete the training of automatic generation of operation tickets; S4: The power professional speech recognition model converts the power operator's voice instructions into text, and combines it with the intelligent ticket expert model in step S3, using the converted text data as input to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements; S5: Based on the successful NFC verification in step S1 to ensure that the on-site conditions are correct, the voice command is input on-site through the voice recognition module in step S4 and a real-time operation ticket is automatically generated, the lock information in the operation ticket is identified, and the lock name corresponding to the operation device is identified using the terminology of the operation ticket; S6: According to step S5, confirm the lock information corresponding to the switching operation equipment, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, check and remind the correctness and sequential execution of the locks of the operation object, complete the business closed loop of the lock verification link before the operation, and then confirm that the operator correctly performs the switching operation content.

2. According to claim 1, a switching operation generation and verification method based on NFC and intelligent ticketing model is characterized in that: The step S1 of verifying whether the preset trigger condition is met includes: Determine whether the NFC signal is valid. If the lock ID is not detected, the strength of the NFC patch signal does not meet the system's set range, and the distance between the device and the NFC patch does not meet the minimum distance, the system will display an error prompt to notify the operator to recalibrate the position or check the device; If the NFC signal is valid, the unique ID of the NFC patch is used to determine whether the NFC patch matches the lock and confirm whether the operating device and the lock are correctly bound; If the device and the lock do not match, the system will issue an alarm and prohibit further operation. The operator needs to reconfirm the consistency of the on-site device and the lock.

3. According to claim 1, a switching operation generation and verification method based on NFC and intelligent ticketing model is characterized in that: Step S2 includes the following steps: S2.1: Obtain equipment topology information corpus from multiple data sources and perform data cleaning and preprocessing on the equipment topology information corpus to extract entity attribute data related to power equipment and correlation data between equipment; S2.2: The entity attribute data and association data obtained in step S2.1 are integrated into a knowledge graph of electric power equipment in a unified format through data fusion; S2.3: Based on the knowledge graph of power equipment in step S2.2, the logic and rules of the switching operation are extracted to construct a switching operation rule decision tree that meets the actual operation requirements. The decision tree automatically derives the switching operation steps of various types of equipment under different states; S2.4: Arrange the physical device attributes and status recorded in the power equipment knowledge graph of step S2.2 into an equipment ledger, and arrange the recorded lock entity attributes and status into a lock ledger; S2.5: Based on step S2.4, each device information in the device ledger is matched with the corresponding lock in the lock ledger. This matching information will provide a basis for comparing the subsequent operation ticket execution information with the NFC acquisition information to ensure that the device and lock names in the operation ticket are consistent.

4. According to claim 3, a method for generating and verifying a switching operation based on NFC and an intelligent ticketing model is characterized in that: After screening and preprocessing, the multi-distance data is fused to build a unified knowledge graph of power equipment including; The relationship edges between device nodes are used as the basic information of the power equipment knowledge graph; Each device node contains the attribute information of the device, including the type, status, location and specifications of the device; The relationship edge attributes between device nodes include type, dependency, and operation order; The types include control, connection, dependency and influence, which represent the nature of the relationship between devices. The dependency refers to the description of the strength of the relationship between devices and whether there is a mandatory operation sequence between the two devices. The operation sequence describes the sequence of device operations.

5. According to claim 1, a switching operation generation and verification method based on NFC and intelligent ticketing model is characterized in that: The step S3 specifically includes: S3.1: Using the GLM model through historical operation tickets, typical tickets and common operation record data, a supervised fine-tuning algorithm is used to fine-tune the data in the field of power switching operation tickets, and human feedback reinforcement learning optimization model is used according to actual needs and complex scenarios; S3.2: Based on the GLM model optimized in step S3.1, the intelligent ticketing expert model is constructed using the equipment nodes and corresponding relationship edges of the power equipment knowledge graph, and the intelligent ticketing expert model reasoning is trained through the equipment operation sequence, state changes and the relationship between equipment; S3.3: Based on the intelligent ticket expert model obtained in step S3.2, the switching operation steps of each device node are inferred using the switching operation rule decision tree, and the intelligent ticket expert model derives the switching operation ticket that meets the operation specifications according to the current status of the device, the dependency relationship between the devices and the operation sequence; S3.4: The intelligent ticket expert model combined with the thinking chain algorithm gradually infers the switching operation ticket generation task obtained in step S3.3 and breaks the task into multiple small steps. Each step is inferred based on the equipment status, operation sequence and dependency relationship, and the accurate operation ticket content is gradually generated; S3.5: The current state of the equipment is confirmed in real time through knowledge graph query in the reasoning process in step S3.4 and the state is verified. If the equipment state is inconsistent with the reasoning step, the system will issue an alarm to readjust the operation steps; S3.6: Based on the reasoning verification of step S3.5, combined with the reasoning results of the expert model, the information of the equipment knowledge graph and the switching operation rules, a switching operation ticket that complies with the power grid management regulations is automatically generated.

6. A method for generating and verifying a switching operation based on NFC and a smart ticketing model according to claim 5, characterized in that: The supervised fine-tuning algorithm in step S3.1 specifically includes: Collect historical operation tickets, typical operation tickets and common operation record data from power equipment and manually annotate them to clarify the correct sequence of each operation step, the relationship between equipment, and the changes in equipment status; The labeled historical data is provided as input to the GLM model for training. A standard supervised fine-tuning loss function is used to calculate the difference between the output of the analysis model and the true label. By minimizing the loss function, the model learns how to generate correct operation steps and related decisions based on the input data.

7. The method for generating and verifying a switching operation based on NFC and a smart ticketing model according to claim 5 is characterized in that: The human feedback reinforcement learning optimization model in step S3.1 specifically includes: Define a reinforcement learning environment suitable for power switching operations to simulate different scenarios during the switching operation, including device state changes, dependencies between devices, and operation sequences; Define state space and action space, where the state refers to the current state of the device and the action refers to specific operation steps; Incorporate industry experts into the feedback process. Experts provide feedback on each operation step through a scoring mechanism based on the operation steps output by the GLM model, including scoring the correctness, timing, and whether the dependencies of the operation steps meet the specifications. These expert scoring data feedbacks are converted into reward signals for the training of reinforcement learning algorithms. The model is updated through the reinforcement learning algorithm. The reward signals of each operation step affect the weight adjustment of the model, so that the GLM model can make optimized decisions according to the expert feedback during the next operation.

8. The method for generating and verifying a switching operation based on NFC and a smart ticketing model according to claim 1 is characterized in that: The step S4 specifically includes: S4.1: Extract power professional terms and phonetic features related to the terms from the multi-source power knowledge base, and annotate all terms into a standardized format; The power professional terms include station equipment, substation, operation specification terms and geographical location; The knowledge base includes two-ticket management regulations, substation operation procedures, electrical equipment operation procedures, operation execution points, equipment wiring methods, operation matters, equipment operation rules, and equipment error prevention rules; The speech features associated with the term include the frequency spectrum, pitch, and volume of the audio; S4.2: manually input the power industry professional terms obtained in step S4.1 to generate a speech recognition library containing power industry professional terms; S4.3: Manually record power industry terms and generate more than 100 hours of audio files, clean and annotate all recorded speech to ensure that each audio file corresponds to the correct terminology and instructions; S4.4: Use the PEFT algorithm to fine-tune the electric power professional speech recognition model through artificial recording samples, adjust the model weight, adjust the model parameters according to the characteristics of the speech data, and complete the optimization of the electric power professional speech recognition model.

9. The method for generating and verifying a switching operation based on NFC and a smart ticketing model according to claim 1 is characterized in that: By matching the NFC patch ID with the lock register, the system searches for the corresponding lock information in the lock register by reading the unique identifier on the NFC patch. When the correspondence between the device and the lock is known, the system confirms whether the patch matches the lock recorded in the lock register through the NFC patch ID. If the NFC patch ID matches the lock ID, it is valid. When the match is successful, the system records this relationship and provides the necessary basis for subsequent lock verification and operation ticket execution. If the match fails, the current operation is interrupted.

10. A switching operation generation verification system based on NFC and intelligent ticket model, running a switching operation generation verification method based on NFC and intelligent ticket model as claimed in any one of claims 1 to 9, characterized in that: The NFC induction verification module is used to use the REID induction module of the substation operation and maintenance mobile operation terminal to perform contact induction with the NFC patch corresponding to the lock, obtain the NFC unique ID and verify whether it meets the preset trigger conditions. If the conditions are met, the collected NFC data information will be transmitted back to the substation operation and maintenance mobile operation platform; The equipment and lock matching module is used to build the equipment ledger and lock ledger module by establishing the knowledge graph of power equipment in the substation operation and inspection mobile operation platform, generate the decision tree of switching operation rules, and use the correspondence between electrical equipment and locks provided by the five-defense system to match the equipment ledger name with the lock ledger name one by one; The automatic operation ticket generation module is used to combine the business data historical tickets and typical tickets based on the power equipment knowledge graph and the switching operation rule decision tree provided by the equipment and lock matching module, and use the GLM model to learn the operation logic of the intelligent ticket expert model to complete the training of automatic operation ticket generation; The power professional voice recognition module is used to convert the voice instructions of power operators into text, and is combined with the intelligent ticket expert model of the automatic operation ticket generation module to automatically generate operation ticket content that meets the equipment type, voltage level and operation requirements using the converted text data as input; The module corresponding to the operation ticket and the operation equipment lock is used to ensure the correctness of the on-site conditions based on the successful NFC verification of the NFC induction verification module. The voice command is input on-site through the power professional voice recognition module and the real-time operation ticket is automatically generated. The lock information in the operation ticket is identified, and the name of the lock corresponding to the operation equipment is identified using the operation ticket terminology specification, which provides a basis for the subsequent collection and verification of whether the lock is correct; The switching execution module is used to confirm the corresponding lock information of the switching operation equipment according to the operation ticket and the corresponding module of the operating equipment lock, and transmit the NFC patch data to the substation operation and maintenance mobile operation platform. Through the data returned by RFID sensing, the correctness and sequential execution of the locks of the operation objects are verified and reminded, completing the business closed loop of the lock verification link before the operation, and then confirming that the operating personnel correctly executed the switching operation content.

Images