A deep learning-based operation ticket service logic intelligent control method and system

Abstract

The application discloses a kind of based on deep learning's operation ticket business logic intelligent control method and system, including the following steps: collecting operation ticket original service data set, executes pre-processing;Extract operation step unit, operation object unit and state constraint unit, and construct sequential dependence relationship and interlocking response relationship, generate operation ticket logic topology diagram;Operation ticket logic topology diagram is mapped to virtual execution environment, and executable digital twin ticket surface is constructed, and ticket surface execution state space is generated;Step-by-step rehearsal is executed on executable digital twin ticket surface, and deep learning state evolution prediction model is called, and ticket surface dynamic rehearsal sequence is generated;Generate operation ticket implicit logic risk result;Executable digital twin ticket surface is executed control correction, and operation ticket business logic intelligent control result is obtained.The application adopts deep learning rehearsal control method, realizes operation ticket logic intelligent checking and correction, with the advantages of high safety, accurate identification and control intelligence.

Description

Technical Field

[0001] This invention relates to the field of intelligent processing of operation tickets, and in particular to an intelligent control method and system for operation ticket business logic based on deep learning. Background Technology

[0002] In existing technologies, operation tickets are mainly used to standardize the management of switching operations, state transitions, and execution sequences of power equipment. They typically rely on manual preparation, manual review, and ticket verification based on fixed rules to complete pre-execution checks. Some solutions can perform individual verifications on ticket text, step sequence, or equipment status, but overall, they still mainly rely on static content verification and lack overall associative modeling of operation steps, operation objects, state constraints, and interlocking relationships.

[0003] However, existing technologies make it difficult to dynamically rehearse the business logic of the operation ticket before execution, and cannot identify implicit logical risks in advance by combining equipment state transitions, interlock responses, and constraint satisfaction. This can easily lead to problems such as logic breakpoints, unreachable states, interlock mismatches, and execution conflicts that are difficult to detect in a timely manner, thereby affecting the safety, accuracy, and control effectiveness of operation ticket execution. Summary of the Invention

[0004] One objective of this invention is to propose an intelligent control method and system for operation ticket business logic based on deep learning. This invention adopts a deep learning pre-simulation control method to realize intelligent verification and correction of operation ticket logic, and has the advantages of high security, accurate identification and intelligent control.

[0005] According to an embodiment of the present invention, a deep learning-based intelligent control method for operation ticket business logic includes the following steps: Collect the text data, structure data, equipment status data, and interlock constraint data of the operation ticket to be executed to form the original business dataset of the operation ticket, and perform preprocessing to generate a standardized business feature set of the operation ticket; Based on the standardized operation ticket business feature set, operation step units, operation object units, and state constraint units are extracted, and sequential dependency relationships and interlocking response relationships are constructed to generate an operation ticket logical topology diagram. The operation ticket logic topology diagram is mapped to the virtual execution environment to construct an executable digital twin ticket and generate the ticket execution state space; A step-by-step pre-play is performed on the executable digital twin ticket, and a deep learning state evolution prediction model is called during each pre-play process to predict the equipment state transition results, interlock response results and constraint satisfaction results in the current pre-play state, and generate a dynamic pre-play sequence of the ticket. Based on the dynamic pre-simulation sequence of the ticket, identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps, and generate implicit logical risk results for the operation ticket; Based on the implicit logic risk results of the operation ticket, the execution control of the executable digital twin ticket is modified, an optimized control scheme for the operation ticket is generated, and the intelligent control results of the operation ticket business logic are output.

[0006] Optionally, the generation of the original business dataset for the operation ticket specifically includes: Read the text data on the ticket corresponding to the operation to be executed, extract the operation instructions, condition descriptions and status confirmations for each operation step, and form a subset of the ticket text. Obtain the ticket structure data corresponding to the operation ticket to be executed, and extract the ticket structure subset; The device status data corresponding to the operation ticket to be executed is collected and associated to obtain the current status information and permission status information of each operation object at the current moment, forming a subset of device status; The interlock constraint data corresponding to each operation step is retrieved to obtain the interlock constraint content and execution permission content corresponding to each operation step, forming an interlock constraint subset. The subsets of ticket text, ticket structure, equipment status, and interlocking constraints are merged to generate the original business dataset of the operation ticket.

[0007] Optionally, the preprocessing includes field normalization, temporal alignment, semantic mapping, and object association.

[0008] Optionally, the generation of the operation ticket logic topology diagram specifically includes: Based on the standardized operation ticket business feature set, the standard expression data of operation steps and the corresponding relationship data between steps are analyzed to determine the step content, step identifier and step sequence of each operation step, and the step boundary and step affiliation of each operation step are determined to form a set of operation step units. Based on the standard expression data of operation objects and the correspondence data between steps and operation objects in the standardized operation ticket business feature set, the operation objects corresponding to each operation step are matched and labeled to determine the operation object identifier and object affiliation corresponding to each operation step, thus forming a set of operation object units. Based on the state constraint standard expression data and the correspondence data between steps and state constraints in the standardized operation ticket business feature set, the state conditions, sequence conditions and permission conditions corresponding to each operation step are constrained and analyzed to determine the constraint content and constraint attribution of each operation step, thus forming a set of state constraint units. Based on the set of operation step units and the correspondence data between steps, establish the sequential connection relationship and dependency triggering relationship between each operation step to form a sequential dependency relationship; Based on the set of operation step units, the set of operation object units, the set of state constraint units, and the current state data and permission state data, the association and determination of the current state satisfaction relationship and permission state satisfaction relationship corresponding to the execution of each operation step are performed to establish the interlock constraint transmission relationship and execution permission transmission relationship between each operation step, thus forming an interlock response relationship. The operation step unit set, operation object unit set, state constraint unit set, sequential dependency relationship and interlock response relationship are associated and organized to generate an operation ticket logic topology diagram.

[0009] Optionally, the generation of the ticket execution state space specifically includes: Read the set of operation step units, the set of operation object units, the set of state constraint units, the sequence dependency relationship and the interlock response relationship in the operation ticket logic topology diagram. According to the same operation step identifier and the same operation object identifier, perform mapping and splitting on each data to obtain the operation step mapping result, operation object mapping result, state constraint mapping result, sequence dependency mapping result and interlock response mapping result. Based on the operation step mapping results and operation object mapping results, the operation objects corresponding to each operation step are loaded into the corresponding operation step positions in the virtual execution environment, establishing a virtual execution correspondence between each operation step and each operation object, and forming a step object loading result; Based on the loading results of the step objects and the mapping results of the state constraints, the state constraints corresponding to each operation step are loaded into the corresponding operation step position. Combined with the sequential dependency mapping results and the interlocking response mapping results, virtual sequential inheritance relationships and virtual interlocking response relationships between each operation step are established to form the relationship loading results. Based on the relation loading results, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are synchronized and organized to form the ticket status organization result. Based on the ticket status organization results, an executable digital twin ticket is constructed in a virtual execution environment, which includes each operation step, each operation object, each state constraint, each sequential dependency relationship, and each interlocking response relationship. Based on the executable digital twin ticket, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are associated and arranged to generate the ticket execution state space.

[0010] Optionally, the virtual sequence succession relationship represents the relationship in which each operation step is connected and advanced sequentially according to the sequential dependency relationship in the virtual execution environment, and the virtual interlock response relationship represents the relationship in which each operation step is constrained and authorized according to the interlock response relationship in the virtual execution environment.

[0011] Optionally, the generation of the dynamic pre-show sequence of the ticket specifically includes: Read the executable digital twin ticket and the ticket execution state space, determine the step-by-step pre-play starting position according to the order of the operation steps, extract the corresponding execution state, current state, satisfied state and permitted state, and obtain the current pre-play state result; Based on the current pre-simulation results and the executable digital twin ticket, determine the corresponding operation objects, state constraint units, and interlocking response relationships in the current pre-simulation state; The current pre-simulation state results, operation objects, state constraint units, and interlocking response relationships are encoded to form the current step pre-simulation input feature vector. After normalization, the vector is input into a deep learning state evolution prediction model to predict the equipment state transition results, interlocking response results, and constraint satisfaction results under the current pre-simulation state. The equipment state transition results, interlock response results, and constraint satisfaction results are associated with the same operation step identifier to form the single-step pre-simulation result corresponding to the current operation step. Based on the single-step pre-simulation result corresponding to the current operation step, the execution state, current state, satisfied state and permitted state in the ticket execution state space are updated, and the updated execution state, current state, satisfied state and permitted state are used as the pre-simulation input for the operation step corresponding to the next sequential position to generate the next pre-simulation state result; The result of the next pre-simulation state is used as the pre-simulation input for the operation step corresponding to the next sequential position. The single-step pre-simulation result corresponding to the next operation step is generated and repeated according to the sequential position of the operation steps to obtain the set of single-step pre-simulation results corresponding to each operation step. The sets of single-step pre-simulation results corresponding to each operation step are linked and organized according to their sequential positions to generate a dynamic pre-simulation sequence for the ticket.

[0012] Optionally, the generation of the implicit logic risk result of the operation ticket specifically includes: Read the dynamic pre-simulation sequence of the ticket, and extract the equipment state transition results, interlocking response results and constraint satisfaction results corresponding to each operation step according to the operation step identifier to form a step pre-simulation result set; Based on the step pre-simulation result set, the corresponding comparison of the equipment state transition results between adjacent operation steps is performed to identify the discontinuous step connection positions between the equipment state transition result of the previous operation step and the execution requirements of the next operation step, and to obtain the logic breakpoint identification results. Based on the pre-simulation result set, the equipment state transition results and constraint satisfaction results corresponding to each operation step are jointly verified to identify the operation step where the equipment state transition results do not meet the corresponding state constraint requirements, and the state unreachable node identification results are obtained. Based on the pre-simulation result set, the interlocking response results and constraint satisfaction results corresponding to each operation step are verified, and the operation step positions where the interlocking response results are inconsistent with the corresponding permission conditions are identified to obtain the interlocking mismatch node identification results. Based on the step-by-step simulation result set, combined with the equipment state transition results, interlock response results and constraint satisfaction results between adjacent operation steps, the positions of operation steps with conflicting execution conditions in the same temporal progression path are identified, and the execution conflict node identification results are obtained. The step association path of risk status transmission to subsequent operation steps along the sequential position is also identified, and the risk propagation node identification results are obtained. The results of logical breakpoint identification, unreachable state node identification, interlock mismatch node identification, execution conflict node identification, and risk propagation node identification are associated and organized according to the operation step identifier and sequential position to generate implicit logical risk results for the operation ticket.

[0013] Optionally, the generation of the intelligent control result of the operation ticket business logic specifically includes: Read the implicit logical risk results and executable digital twin ticket of the operation ticket, and extract the logical breakpoint identification results, unreachable status node identification results, interlock mismatch node identification results, execution conflict node identification results and risk propagation node identification results according to the operation step identifier and sequence position; The execution order of operation steps with logical breakpoint identification results and execution conflict node identification results is rearranged to form a step rearrangement result; For the operation steps of identifying unreachable nodes and interlocking mismatch nodes, supplement the corresponding state conditions and permission conditions, and load the supplemented state conditions and permission conditions into the corresponding state constraint positions in the executable digital twin ticket to form the condition supplement result; Based on the step association path corresponding to the risk propagation node identification result, the interlocking response relationship in the executable digital twin ticket is corrected to form an interlocking that satisfies the correction result; Based on the results of step rearrangement, condition supplementation, and interlocking satisfaction correction, the operation steps, state constraints, sequential dependencies, and interlocking response relationships in the executable digital twin ticket are updated accordingly to form the execution path correction results. Based on the execution path correction results, an optimized control scheme for operation tickets is generated, which serves as the intelligent control result for the operation ticket business logic.

[0014] According to an embodiment of the present invention, a deep learning-based intelligent control system for operation ticket business logic includes: The raw business data construction module is used to collect the text data, structure data, equipment status data, and interlocking constraint data of the operation tickets to be executed, forming the raw business dataset of the operation tickets, and performing preprocessing to generate a standardized business feature set of the operation tickets; The logical topology construction module is used to extract operation step units, operation object units, and state constraint units based on the standardized operation ticket business feature set, and to construct sequential dependency relationships and interlocking response relationships to generate an operation ticket logical topology diagram. The digital twin ticket construction module is used to map the logic topology diagram of the operation ticket to the virtual execution environment, construct an executable digital twin ticket, and generate the ticket execution state space; The dynamic pre-show prediction module is used to perform step-by-step pre-shows on the executable digital twin ticket and call the deep learning state evolution prediction model during each pre-show process to predict the equipment state transition results, interlocking response results and constraint satisfaction results in the current pre-show state, and generate a dynamic pre-show sequence for the ticket. The hidden risk identification module is used to identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps based on the dynamic pre-play sequence of the ticket, and generate hidden logical risk results for the operation ticket. The control correction output module is used to perform control corrections on the executable digital twin ticket based on the implicit logic risk results of the operation ticket, generate an optimized control scheme for the operation ticket, and output the intelligent control results of the operation ticket business logic.

[0015] The beneficial effects of this invention are: Compared to existing operation ticket processing methods that primarily rely on manual review or static rule verification, this invention can uniformly organize and correlate the text data, structural data, equipment status data, and interlocking constraint data of the operation ticket, forming a standardized operation ticket business feature set. Furthermore, it constructs an operation ticket logical topology diagram, an executable digital twin ticket, and a ticket execution state space, thereby achieving a structured, executable, and dynamic expression of the operation ticket's business logic. This processing method goes beyond isolated verification of single ticket content or local step relationships; it enables systematic analysis of the operation ticket from a holistic perspective, considering operation steps, operation objects, state constraints, sequential dependencies, and interlocking response relationships, improving the completeness and consistency of the operation ticket's business logic expression.

[0016] Simultaneously, this invention performs a step-by-step pre-play on the executable digital twin ticket and, combined with a deep learning state evolution prediction model, predicts the equipment state transition results, interlock response results, and constraint satisfaction results during the advancement of each operation step. This enables the early detection of implicit logical risks that are difficult to identify directly using traditional methods before formal execution, including logical breakpoints, unreachable state nodes, interlock mismatch nodes, execution conflict nodes, and risk propagation nodes. Based on this, and according to the identified implicit logical risk results of the operation ticket, the executable digital twin ticket is modified for control, generating an optimized control scheme for the operation ticket and intelligent control results for the operation ticket's business logic. This transforms the operation ticket from a post-event problem discovery to a proactive prediction and early correction before execution. Therefore, this invention can effectively improve the accuracy of risk identification before operation ticket execution, the depth of business logic verification, and the pertinence of control correction, enhancing the safety, reliability, and intelligent control capabilities of the operation ticket execution process. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an intelligent control method for operation ticket business logic based on deep learning proposed in this invention; Figure 2 This diagram illustrates the implicit logic risk identification of operation tickets, a method for intelligent control of operation ticket business logic based on deep learning proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figure 1 and Figure 2 A deep learning-based intelligent control method for ticket operation business logic includes the following steps: Collect the text data, structure data, equipment status data, and interlock constraint data of the operation ticket to be executed to form the original business dataset of the operation ticket, and perform preprocessing to generate a standardized business feature set of the operation ticket; Based on the standardized operation ticket business feature set, operation step units, operation object units, and state constraint units are extracted, and sequential dependency relationships and interlocking response relationships are constructed to generate an operation ticket logical topology diagram. The operation ticket logic topology diagram is mapped to the virtual execution environment to construct an executable digital twin ticket and generate the ticket execution state space; A step-by-step pre-play is performed on the executable digital twin ticket, and a deep learning state evolution prediction model is called during each pre-play process to predict the equipment state transition results, interlock response results and constraint satisfaction results in the current pre-play state, and generate a dynamic pre-play sequence of the ticket. Based on the dynamic pre-simulation sequence of the ticket, identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps, and generate implicit logical risk results for the operation ticket; Based on the implicit logic risk results of the operation ticket, the execution control of the executable digital twin ticket is modified, an optimized control scheme for the operation ticket is generated, and the intelligent control results of the operation ticket business logic are output.

[0020] In this embodiment, the generation of the original business dataset for the operation ticket specifically includes: Read the text data on the ticket corresponding to the operation to be executed, extract the operation instructions, condition descriptions and status confirmations for each operation step, and form a subset of the ticket text. Obtain the ticket structure data corresponding to the operation ticket to be executed, and extract the step number information, operation object information, execution order information, condition constraint information and step association information corresponding to each operation step to form a ticket structure subset; Condition constraint information is used to characterize the state conditions, sequence conditions, and permission conditions that each operation step needs to meet when it is executed. Step association information is used to characterize the sequential connection, dependency triggering, or linkage correspondence between each operation step. The device status data corresponding to the operation ticket to be executed is collected and associated to obtain the current status information and permission status information of each operation object at the current moment, forming a subset of device status; Current status information is used to characterize the actual operating status of each operation object at the time of data collection, while permission status information is used to characterize whether each operation object has the qualification to perform the corresponding operation under the current interlock constraints and execution conditions. Based on the condition description and status confirmation content in the ticket text subset, and combined with the condition constraint information and step association information in the ticket structure subset, the interlock constraint data corresponding to each operation step is retrieved to obtain the interlock constraint content and execution permission content corresponding to each operation step, thus forming an interlock constraint subset. Interlocking constraints are used to characterize the equipment linkage and safety restriction relationships that each operation step is subject to during execution. Execution permission is used to characterize whether each operation step is allowed to be executed under the current state conditions and constraint conditions. The subsets of ticket text, ticket structure, equipment status, and interlocking constraints are merged according to the same operation ticket identifier, the same operation step identifier, and the same operation object identifier to generate the original business dataset of operation tickets.

[0021] In this embodiment, preprocessing includes field normalization, temporal alignment, semantic mapping, and object association. The standardized operation ticket business feature set is a data set formed after preprocessing the original operation ticket business dataset. It is used to extract operation step units, operation object units, and state constraint units, and to construct sequential dependencies and interlocking response relationships. This includes standard expression data for operation steps, standard expression data for operation objects, standard expression data for state constraints, correspondence data between steps, correspondence data between steps and operation objects, correspondence data between steps and state constraints, current state data, and permitted state data. Standard expression data for operation steps is data formed after normalizing and semantically unifying the step content, step identifier, and sequential position of each operation step in the operation ticket. It is used to characterize the standardized step information of each operation step. Standard expression data for operation objects refers to data formed after normalizing and semantically unifying the equipment object name, object identifier, and object affiliation corresponding to each operation step. It is used to characterize the operation object information corresponding to each operation step. Standard expression data for state constraints refers to the normalization of the state conditions, sequence conditions, and permitted conditions corresponding to the execution of each operation step. The data formed after standardization and semantic unification is used to represent the state constraint information of each operation step. The correspondence data between steps refers to the data formed after organizing the connection relationship and dependency trigger relationship between each operation step, and is used to represent the association relationship between each operation step. The correspondence data between steps and operation objects refers to the data formed after organizing the matching relationship between each operation step and its corresponding operation object, and is used to represent which operation object each operation step acts on. The correspondence data between steps and state constraints refers to the data formed after organizing the matching relationship between each operation step and its corresponding state constraints, and is used to represent which state constraints each operation step is subject to. The current state data is the data formed after standardizing the actual running state of each operation object at the time of collection, and is used to represent the current state information of each operation object. The permission state data is the data formed after standardizing whether each operation object has the qualification to perform operation under the current interlock constraints and execution conditions, and is used to represent the permission state information of each operation object. The dependency trigger relationship is used to represent the correspondence relationship that the execution of a certain operation step is premised on the completion of another operation step or the satisfaction of the corresponding conditions.

[0022] In this embodiment, the generation of the operation ticket logic topology diagram specifically includes: Based on the standardized operation ticket business feature set, the standard expression data of operation steps and the corresponding relationship data between steps are analyzed to determine the step content, step identifier and step sequence of each operation step, and the step boundary and step affiliation of each operation step are determined to form a set of operation step units. Step sequence is information used to characterize the sequential position of each operation step in the entire operation ticket execution process; Based on the standard expression data of operation objects and the correspondence data between steps and operation objects in the standardized operation ticket business feature set, the operation objects corresponding to each operation step are matched and labeled to determine the operation object identifier and object affiliation corresponding to each operation step, thus forming a set of operation object units. Based on the state constraint standard expression data and the correspondence data between steps and state constraints in the standardized operation ticket business feature set, the state conditions, sequence conditions and permission conditions corresponding to each operation step are constrained and analyzed to determine the constraint content and constraint attribution of each operation step, thus forming a set of state constraint units. The generation of the state constraint unit set specifically includes: reading the state constraint standard expression data and the correspondence data between steps and state constraints from the standardized operation ticket business feature set; extracting the state constraint standard expression data matching each operation step according to the step identifier corresponding to each operation step to obtain the state constraint candidate data corresponding to each operation step; based on the state constraint candidate data corresponding to each operation step, identifying and merging the data content representing the state meeting the requirements to obtain the state condition data corresponding to each operation step; identifying and merging the data content representing the execution sequence restrictions to obtain the sequence condition data corresponding to each operation step; identifying and merging the data content representing the execution qualification permission requirements to obtain the permission condition data corresponding to each operation step; then combining the state condition data, sequence condition data, and permission condition data corresponding to each operation step according to the same operation step identifier to form the constraint content data corresponding to each operation step; performing attribution matching on the constraint content data corresponding to each operation step according to the correspondence data between steps and state constraints to determine the operation step attribution corresponding to each set of constraint content data to obtain the constraint attribution data corresponding to each operation step; and associating and organizing the constraint content data and constraint attribution data corresponding to each operation step to form the state constraint unit set. Based on the set of operation step units and the correspondence data between steps, establish the sequential connection relationship and dependency triggering relationship between each operation step to form a sequential dependency relationship; The generation of sequential dependencies specifically includes: reading the set of operation step units and the correspondence data between steps; extracting the sequential order of each operation step unit according to its corresponding step identifier to obtain the result of the sequential arrangement of operation steps; identifying the sequential connection relationship between adjacent operation step units in the sequential arrangement of operation steps based on the correspondence data between steps, determining the preceding and following operation step units for each operation step unit, and obtaining the result of the sequential connection between steps; and then, based on the result of the sequential connection between steps, extracting the correspondence between each operation step unit that requires the completion of another operation step as a prerequisite for execution. After obtaining the step completion dependency results, the correspondence between each operation step unit with the condition being met as the execution premise is extracted to obtain the condition satisfaction dependency results. Then, the step completion dependency results and the condition satisfaction dependency results are merged according to the same operation step identifier to form the dependency trigger results corresponding to each operation step. On this basis, the step connection results are matched with the dependency trigger results, and the step connection paths and trigger inheritance paths between each operation step unit are associated and organized to obtain the relationship association results between each operation step. Based on the relationship association results, the sequential connection relationship and dependency trigger relationship between each operation step are established to form a sequential dependency relationship. Based on the set of operation step units, the set of operation object units, the set of state constraint units, and the current state data and permission state data, the association and determination of the current state satisfaction relationship and permission state satisfaction relationship corresponding to the execution of each operation step are performed to establish the interlock constraint transmission relationship and execution permission transmission relationship between each operation step, thus forming an interlock response relationship. The generation of interlocking response relationships specifically includes: reading the set of operation step units, the set of operation object units, the set of state constraint units, the current state data, and the permitted state data; matching the operation object units corresponding to each operation step unit according to the same operation step identifier and the same operation object identifier to obtain the operation object matching result corresponding to each operation step; based on the operation object matching result corresponding to each operation step, loading the current state data and permitted state data corresponding to each operation object into the corresponding operation step unit to obtain the state loading result corresponding to each operation step; then, according to the set of state constraint units, verifying the correspondence between the current state data in the state loading result corresponding to each operation step and the state conditions in the corresponding state constraint units to determine whether the current state satisfies the corresponding state constraint when each operation step is executed, obtaining the current state satisfaction result corresponding to each operation step, and verifying the state loading result corresponding to each operation step. The permission status data is verified item by item against the permission conditions in the corresponding state constraint unit to determine whether the permission status meets the corresponding state constraint when each operation step is executed, thus obtaining the permission status satisfaction result for each operation step. The current state satisfaction result and the permission status satisfaction result for each operation step are associated and organized according to the same operation step identifier to obtain the constraint satisfaction association result for each operation step. Based on this, according to the sequential position in the operation step unit set, the constraint satisfaction association result corresponding to the preceding operation step is passed and matched to the subsequent operation steps to determine the constraint inheritance relationship formed by the current state satisfaction result between each operation step, thus obtaining the interlock constraint transmission result. The permission inheritance relationship formed by the permission status satisfaction result between each operation step is also determined to obtain the execution permission transmission result. The interlock constraint transmission result and the execution permission transmission result are associated and organized according to the same operation step identifier to form an interlock response relationship. The operation step unit set, operation object unit set, state constraint unit set, sequential dependency relationship and interlock response relationship are associated and organized to generate an operation ticket logic topology diagram; An operation ticket logic topology diagram is a graphical logical organization result formed by structuring and associating each operation step, operation object, state constraint, and their logical relationships in an operation ticket. It represents the order in which each operation step is executed, which operation objects it acts on, which state constraints it is subject to, and how sequential dependencies and interlocking response relationships are formed between each step. The operation step unit, operation object unit, and state constraint unit constitute the node content in the operation ticket logic topology diagram, and the sequential dependencies and interlocking response relationships constitute the connection relationships in the operation ticket logic topology diagram. This transforms the static content of the operation ticket into a data structure for subsequent virtual execution environment mapping and digital twin pre-simulation.

[0023] In this embodiment, the generation of the ticket execution state space specifically includes: Read the set of operation step units, the set of operation object units, the set of state constraint units, the sequence dependency relationship and the interlock response relationship in the operation ticket logic topology diagram. According to the same operation step identifier and the same operation object identifier, perform mapping and splitting on each data to obtain the operation step mapping result, operation object mapping result, state constraint mapping result, sequence dependency mapping result and interlock response mapping result. Based on the operation step mapping results and operation object mapping results, the operation objects corresponding to each operation step are loaded into the corresponding operation step positions in the virtual execution environment, establishing a virtual execution correspondence between each operation step and each operation object, and forming a step object loading result; The virtual execution environment is a data-driven runtime environment used to carry out the simulation execution of the business logic of the operation ticket. This virtual execution environment is constructed based on the logic topology diagram of the operation ticket. It is used to load and organize the set of operation step units, the set of operation object units, the set of state constraint units, the sequential dependency relationship, and the interlocking response relationship. This allows each operation step to be virtually advanced under the combined action of the corresponding operation object, the corresponding state constraint, the corresponding sequential dependency relationship, and the corresponding interlocking response relationship. This represents the step execution state, the current state of the operation object, the state constraint satisfaction state, and the interlocking response permission state of the operation ticket before formal execution. It also provides the runtime foundation for constructing an executable digital twin ticket and generating the ticket execution state space. The virtual execution correspondence is the correspondence between each operation step and its corresponding operation object in the virtual execution environment, representing the loading correspondence and execution action relationship. Based on the loading results of the step objects and the mapping results of the state constraints, the state constraints corresponding to each operation step are loaded into the corresponding operation step position. Combined with the sequential dependency mapping results and the interlocking response mapping results, virtual sequential inheritance relationships and virtual interlocking response relationships between each operation step are established to form the relationship loading results. Based on the relation loading results, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are synchronized and organized to form the ticket status organization result. The generation of the ticket status organization result specifically includes: reading the relationship loading result, extracting the virtual execution correspondence, virtual sequence succession relationship, and virtual interlocking response relationship corresponding to each operation step, and obtaining the ticket relationship reading result; based on the ticket relationship reading result, locating the execution advancement position and execution succession position corresponding to each operation step, determining the execution status affiliation of each operation step in the virtual execution environment, and obtaining the operation step execution status positioning result; based on the operation step execution status positioning result, loading the current state of the operation object corresponding to each operation step into the corresponding virtual execution position, forming the operation object current state loading result; and based on the operation object current state loading result and the state corresponding to each operation step... Constraints are checked item by item for the state conditions, sequence conditions, and permission conditions corresponding to each operation step, to determine the satisfaction status of each state constraint at the current virtual execution position, and to obtain the state constraint satisfaction status result. Based on this, according to the virtual interlocking response relationship, the state constraint satisfaction status result corresponding to each operation step is transmitted and matched to the corresponding interlocking response position to determine the permission status of each interlocking response at the current virtual execution position, and to obtain the interlocking response permission status result. The operation step execution state positioning result, the current state loading result of the operation object, the state constraint satisfaction status result, and the interlocking response permission status result are synchronously associated and organized according to the same operation step identifier to form the ticket state organization result. Based on the ticket status organization results, an executable digital twin ticket is constructed in a virtual execution environment, which includes each operation step, each operation object, each state constraint, each sequential dependency relationship, and each interlocking response relationship. An executable digital twin ticket refers to a runnable ticket structure formed in a virtual execution environment by loading, associating, and organizing the set of operation step units, the set of operation object units, the set of state constraint units, the sequential dependency relationship, and the interlocking response relationship according to the operation ticket logical topology diagram. This executable digital twin ticket can synchronously represent the execution progress of each operation step, the state change process of each operation object, the satisfaction process of each state constraint, and the permitted change process of each interlocking response. This allows the operation ticket to be digitally rehearsed step by step before formal execution and serves as the execution carrier for generating the ticket execution state space and conducting subsequent state evolution prediction. Based on the executable digital twin ticket, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are associated and arranged to generate the ticket execution state space.

[0024] In this embodiment, the virtual sequential succession relationship represents the relationship in which each operation step is connected and advanced sequentially according to the sequential dependency relationship in the virtual execution environment, and the virtual interlocking response relationship represents the relationship in which each operation step is constrained and authorized according to the interlocking response relationship in the virtual execution environment.

[0025] In this embodiment, the generation of the dynamic pre-simulation sequence of the ticket specifically includes: Read the executable digital twin ticket and the ticket execution state space, determine the step-by-step pre-play starting position according to the order of the operation steps, extract the corresponding execution state, current state, satisfied state and permitted state, and obtain the current pre-play state result; The generation of the current rehearsal state result specifically includes: reading the executable digital twin ticket and the ticket execution state space; extracting all operation step identifiers from the executable digital twin ticket and the state data corresponding to each operation step identifier in the ticket execution state space to obtain the rehearsal object reading result; based on the rehearsal object reading result, sequentially searching according to the sequential position of each operation step in the executable digital twin ticket to determine the operation step currently participating in the step-by-step rehearsal, and obtaining the current operation step determination result; then, based on the current operation step determination result, extracting the operation step execution state corresponding to the current operation step in the ticket execution state space to obtain the execution state result, and extracting the current state of the operation object corresponding to the current operation step to obtain the state result; subsequently, extracting the state constraint satisfaction state corresponding to the current operation step to obtain the satisfaction state result, and extracting the interlocking response permission state corresponding to the current operation step to obtain the permission state result; and associating and organizing the execution state result, state result, satisfaction state result, and permission state result according to the same operation step identifier to form the current rehearsal state result. The execution status result, status result, satisfied status result, and permitted status result are status data extracted from the ticket execution state space and corresponding to the current operation step identifier. Among them, the execution status result is used to characterize the execution stage of the current operation step at the current rehearsal time, the status result is used to characterize the running state of the operation object corresponding to the current operation step at the current rehearsal time, the satisfied status result is used to characterize whether the state constraints corresponding to the current operation step meet the execution requirements at the current rehearsal time, and the permitted status result is used to characterize whether the interlocking response corresponding to the current operation step meets the execution permitted conditions at the current rehearsal time. In specific implementation, based on the current pre-simulation state results, the execution state, current state, satisfied state, and permitted state corresponding to the current operation step are extracted and mapped to corresponding state code values ​​to obtain the current pre-simulation state numerical results; the operation objects corresponding to the current pre-simulation state are categorized and identified, resulting in operation object numerical results; the state conditions, sequence conditions, and permitted conditions in the state constraint units corresponding to the current pre-simulation state are constrained, resulting in state constraint numerical results; and the interlocking constraint transmission relationship and execution permitted transmission relationship in the interlocking response relationship corresponding to the current pre-simulation state are relationally encoded, resulting in interlocking response numerical results; the current pre-simulation state numerical results, operation object numerical results, state constraint numerical results, and interlocking response numerical results are concatenated according to the same operation step identifier to form the current step pre-simulation input feature vector; based on this, the current step pre-simulation input feature vector is normalized to obtain a normalized pre-simulation input vector; the normalized pre-simulation input vector is input into a deep learning state evolution prediction model, which outputs device state transition results, interlocking response results, and constraint satisfied results; Based on the current pre-simulation results and the executable digital twin ticket, determine the corresponding operation objects, state constraint units, and interlocking response relationships in the current pre-simulation state; The current pre-simulation state results, operation objects, state constraint units, and interlocking response relationships are encoded to form the current step pre-simulation input feature vector. After normalization, the vector is input into a deep learning state evolution prediction model to predict the equipment state transition results, interlocking response results, and constraint satisfaction results under the current pre-simulation state. The deep learning state evolution prediction model consists of a graph attention network, a gated recurrent unit, and a result output layer. The graph attention network is used to encode the features of the relationship structure between the operation object, the state constraint unit, and the interlocking response relationship. The gated recurrent unit is used to receive the feature encoding results and perform temporal evolution modeling in combination with the sequential progression of the operation steps, and output temporal prediction features. The result output layer is used to output the prediction results based on the temporal prediction features. During the training of the deep learning state evolution prediction model, historical operation ticket samples, corresponding equipment state change records, interlock response records, and state constraint satisfaction records are collected. These samples are then merged according to the same operation ticket identifier, the same operation step identifier, and the same operation object identifier to form a historical training sample set. The historical training sample set undergoes preprocessing consistent with the current operation ticket business data to generate a historical standardized operation ticket business feature set. Based on this feature set, a historical operation ticket logical topology diagram, a historical executable digital twin ticket, and a historical ticket execution state space are constructed. From the historical ticket execution state space, the execution state, current state, satisfied state, and permitted state corresponding to each historical operation step are extracted in the order of operation steps. Combined with the corresponding operation object, state constraint unit, and interlock response relationship, historical pre-simulation input samples are formed. The equipment state transition results, interlock response results, and constraint satisfaction results corresponding to each historical operation step after actual execution are used as historical pre-simulation output labels. The historical pre-simulation input samples are quantized, encoded, and normalized to form a model. The training input vector is then fed into a deep learning state evolution prediction model consisting of a graph attention network encoding layer, a gated recurrent unit temporal prediction layer, and a result output layer. The result output layer uses a multilayer perceptron to output device state transition prediction vectors, interlocking response prediction vectors, and constraint satisfaction prediction vectors, respectively. A total loss function is constructed using the mean squared error loss between the device state transition prediction vector and the device state transition result, the cross-entropy loss between the interlocking response prediction vector and the interlocking response result, and the cross-entropy loss between the constraint satisfaction prediction vector and the constraint satisfaction result. The Adam optimizer is used to iteratively update the model parameters, with an initial learning rate of 0.001 and a batch size of 32. Validation loss is calculated on the validation sample set after each training epoch. When the validation loss decreases by less than 0.0001 for 10 consecutive training epochs, or when the training epochs reach 100, the model is considered to have reached convergence and training is stopped. The model parameters corresponding to the minimum validation loss are saved, resulting in the trained deep learning state evolution prediction model. The equipment state transition results, interlock response results, and constraint satisfaction results are associated with the same operation step identifier to form the single-step pre-simulation result corresponding to the current operation step. Based on the single-step pre-simulation result corresponding to the current operation step, the execution state, current state, satisfied state and permitted state in the ticket execution state space are updated, and the updated execution state, current state, satisfied state and permitted state are used as the pre-simulation input for the operation step corresponding to the next sequential position to generate the next pre-simulation state result; When updating the execution state, current state, satisfied state, and permitted state in the ticket execution state space, the single-step pre-simulation result and the ticket execution state space corresponding to the current operation step are read. The equipment state transition result, interlocking response result, and constraint satisfaction result corresponding to the current operation step are extracted from the single-step pre-simulation result to obtain the single-step pre-simulation result reading result. Based on the single-step pre-simulation result reading result, the execution state position of the operation step, the current state position of the operation object, the state constraint satisfaction state position, and the interlocking response permitted state position corresponding to the current operation step are located in the ticket execution state space to obtain the state update positioning result. Based on the equipment state transition result, the current state position of the operation object in the state update positioning result is replaced to form the current state update result. Finally, based on the constraint satisfaction result, the state is updated accordingly. The state constraints in the state update positioning result are satisfied, and state replacement is performed on the state position to form a satisfied state update result. Then, based on the interlocking response result, the interlocking response permission state position in the state update positioning result is replaced, and permission state update result is formed. Based on the execution progress of the current operation step in the single-step pre-rehearsal result, the state position of the operation step in the state update positioning result is updated in stages to form an execution state update result. On this basis, the current state update result, satisfied state update result, permission state update result, and execution state update result are associated and merged according to the same operation step identifier to obtain the single-step state update result. Based on the single-step state update result, the corresponding state position in the ticket execution state space is written back as a whole to complete the update of the execution state, current state, satisfied state, and permission state. The result of the next pre-simulation state is used as the pre-simulation input for the operation step corresponding to the next sequential position. The single-step pre-simulation result corresponding to the next operation step is generated and repeated according to the sequential position of the operation steps to obtain the set of single-step pre-simulation results corresponding to each operation step. The sets of single-step pre-simulation results corresponding to each operation step are linked and organized according to their sequential positions to generate a dynamic pre-simulation sequence for the ticket.

[0026] In this embodiment, the generation of implicit logic risk results for the operation ticket specifically includes: Read the dynamic pre-simulation sequence of the ticket, and extract the equipment state transition results, interlocking response results and constraint satisfaction results corresponding to each operation step according to the operation step identifier to form a step pre-simulation result set; Based on the step pre-simulation result set, the corresponding comparison of the equipment state transition results between adjacent operation steps is performed to identify the discontinuous step connection positions between the equipment state transition result of the previous operation step and the execution requirements of the next operation step, and to obtain the logic breakpoint identification results. Based on the pre-simulation result set, the equipment state transition results and constraint satisfaction results corresponding to each operation step are jointly verified to identify the operation step where the equipment state transition results do not meet the corresponding state constraint requirements, and the state unreachable node identification results are obtained. Based on the pre-simulation result set, the interlocking response results and constraint satisfaction results corresponding to each operation step are verified, and the operation step positions where the interlocking response results are inconsistent with the corresponding permission conditions are identified to obtain the interlocking mismatch node identification results. The generation of interlock mismatch node identification results specifically includes: reading the step pre-rehearsal result set, extracting the interlock response results and constraint satisfaction results corresponding to each operation step according to the same operation step identifier, and obtaining the interlock constraint extraction results; based on the interlock constraint extraction results, parsing the interlock response permission status in the interlock response results corresponding to each operation step, obtaining the interlock permission parsing results corresponding to each operation step, and parsing the permission condition satisfaction status in the constraint satisfaction results corresponding to each operation step, obtaining the permission condition parsing results corresponding to each operation step; then, according to the same operation step identifier, matching the interlock permission parsing results and permission condition parsing results corresponding to each operation step to form the permission correspondence verification results corresponding to each operation step; subsequently, performing consistency judgment on the interlock response permission status and permission condition satisfaction status in the permission correspondence verification results, identifying the operation step positions where the interlock response permission status and permission condition satisfaction status are inconsistent, and obtaining the interlock mismatch position identification results; and associating and organizing the interlock mismatch position identification results according to the operation step identifier and sequential position to form the interlock mismatch node identification results. Based on the step-by-step simulation result set, combined with the equipment state transition results, interlock response results and constraint satisfaction results between adjacent operation steps, the positions of operation steps with conflicting execution conditions in the same temporal progression path are identified, and the execution conflict node identification results are obtained. The step association path of risk status transmission to subsequent operation steps along the sequential position is also identified, and the risk propagation node identification results are obtained. The generation of conflict node identification results and risk propagation node identification results specifically includes: reading the step pre-rehearsal result set, extracting the corresponding equipment state transition results, interlock response results, and constraint satisfaction results between adjacent operation steps according to the operation step identifier and sequential position, and obtaining the adjacent step result extraction results; based on the adjacent step result extraction results, comparing the equipment state transition results of the previous operation step with the constraint satisfaction results of the next operation step in the same time sequence advancement path, identifying whether there is a conflict between the state conditions required for the execution of the next operation step and the state changes formed by the previous operation step, and obtaining the state conflict identification results; based on the adjacent step result extraction results, verifying the interlock response results and constraint satisfaction results between adjacent operation steps, identifying whether there is a conflict between the execution permission conditions of the next operation step and the interlock response results transmitted by the previous operation step, and obtaining the permission conflict identification results; and arranging the state conflict identification results and permission conflict identification results in the same time sequence. The execution paths are correlated and merged to identify the positions of operation steps with conflicting execution conditions within the same temporal execution path, thus obtaining the execution conflict position identification result. Based on this, according to the execution conflict position identification result, the equipment state transition results, interlock response results, and constraint satisfaction results corresponding to the execution conflict position are traced step by step along the sequential position to subsequent operation steps, identifying the risk state continuously transmitted from the execution conflict position to subsequent operation steps, thus obtaining the risk state transmission identification result. According to the risk state transmission identification result, the operation step identifiers, sequential positions, and preceding and following connections of the risk state are correlated and organized to determine the step association path of the risk state transmitted along the sequential position to subsequent operation steps, thus obtaining the risk propagation path identification result. The execution conflict position identification result is correlated and organized according to the operation step identifiers and sequential positions to form the execution conflict node identification result, and the risk propagation path identification result is correlated and organized according to the operation step identifiers and sequential positions to form the risk propagation node identification result. The results of logical breakpoint identification, unreachable state node identification, interlock mismatch node identification, execution conflict node identification, and risk propagation node identification are associated and organized according to the operation step identifier and sequential position to generate implicit logical risk results for the operation ticket.

[0027] In this embodiment, the generation of the intelligent control result of the operation ticket business logic specifically includes: Read the implicit logical risk results and executable digital twin ticket of the operation ticket, and extract the logical breakpoint identification results, unreachable status node identification results, interlock mismatch node identification results, execution conflict node identification results and risk propagation node identification results according to the operation step identifier and sequence position; The execution order of operation steps with logical breakpoint identification results and execution conflict node identification results is rearranged, and the connection position of the corresponding operation steps in the executable digital twin ticket is adjusted to form the step rearrangement result; For the operation steps of identifying unreachable nodes and interlocking mismatch nodes, supplement the corresponding state conditions and permission conditions, and load the supplemented state conditions and permission conditions into the corresponding state constraint positions in the executable digital twin ticket to form the condition supplement result; Based on the step association path corresponding to the risk propagation node identification result, the interlocking response relationship in the executable digital twin ticket is corrected to form an interlocking that satisfies the correction result; The generation of the interlocking satisfaction correction result specifically includes: reading the risk propagation node identification result and the executable digital twin ticket; extracting the step association path corresponding to the risk propagation node identification result according to the operation step identifier and sequential position; locating the interlocking response relationship corresponding to the step association path in the executable digital twin ticket to obtain the interlocking response location result; based on the interlocking response location result, extracting the interlocking constraint transmission relationship and execution permission transmission relationship that constitute the interlocking response relationship between the preceding and following operation steps on the step association path segment by segment to obtain the interlocking response segmentation result; then, based on the interlocking response segmentation result, verifying the interlocking constraint transmission relationship between each adjacent operation step segment by segment, identifying the position of the interlocking constraint transmission relationship that is inconsistent with the risk propagation node identification result, obtaining the interlocking constraint transmission relationship correction location result, and correcting the relationship between each adjacent operation step segment by segment. The execution permission transmission relationship between the nodes is verified segment by segment. Positions of execution permission transmission relationships that are inconsistent with the risk propagation node identification results are identified, resulting in corrected execution permission transmission relationship positioning results. Subsequently, based on the corrected positioning results of the interlocking constraint transmission relationship, the interlocking constraint transmission relationships in the corresponding interlocking response relationships are replaced and updated, forming corrected interlocking constraint transmission relationship results. Then, based on the corrected positioning results of the execution permission transmission relationship, the execution permission transmission relationships in the corresponding interlocking response relationships are replaced and updated, forming corrected execution permission transmission relationship results. The corrected interlocking constraint transmission relationship results and the corrected execution permission transmission relationship results are associated and merged according to the same associated path, resulting in corrected interlocking response results. Finally, the corrected interlocking response results are written back to the corresponding interlocking response relationship position in the executable digital twin ticket, forming corrected interlocking satisfaction results. Based on the results of step rearrangement, condition supplementation, and interlocking satisfaction correction, the operation steps, state constraints, sequential dependencies, and interlocking response relationships in the executable digital twin ticket are updated accordingly to form the execution path correction results. Based on the execution path correction results, an optimized control scheme for the operation ticket is generated. This scheme serves as the intelligent control result for the operation ticket's business logic. The intelligent control result for the operation ticket's business logic is a data result output based on the optimized control scheme for the operation ticket. It represents the final business logic processing result of the operation ticket after control correction. This includes the step rearrangement content, condition supplementation content, interlock satisfaction correction content, and execution path correction content for the corresponding operation steps.

[0028] A deep learning-based intelligent control system for ticket operation logic includes: The raw business data construction module is used to collect the text data, structure data, equipment status data, and interlocking constraint data of the operation tickets to be executed, forming the raw business dataset of the operation tickets, and performing preprocessing to generate a standardized business feature set of the operation tickets; The logical topology construction module is used to extract operation step units, operation object units, and state constraint units based on the standardized operation ticket business feature set, and to construct sequential dependency relationships and interlocking response relationships to generate an operation ticket logical topology diagram. The digital twin ticket construction module is used to map the logic topology diagram of the operation ticket to the virtual execution environment, construct an executable digital twin ticket, and generate the ticket execution state space; The dynamic pre-show prediction module is used to perform step-by-step pre-shows on the executable digital twin ticket and call the deep learning state evolution prediction model during each pre-show process to predict the equipment state transition results, interlocking response results and constraint satisfaction results in the current pre-show state, and generate a dynamic pre-show sequence for the ticket. The hidden risk identification module is used to identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps based on the dynamic pre-play sequence of the ticket, and generate hidden logical risk results for the operation ticket. The control correction output module is used to perform control corrections on the executable digital twin ticket based on the implicit logic risk results of the operation ticket, generate an optimized control scheme for the operation ticket, and output the intelligent control results of the operation ticket business logic.

[0029] Example 1: To verify the feasibility of this invention in practice, it was applied to a pre-execution safety verification scenario for operation tickets at a substation in a certain region. In this scenario, the daily operations involve frequent equipment switching, tightly linked operation steps, and some steps requiring interlocking constraints and status permissions before execution can continue. Existing methods primarily rely on operators manually reviewing the ticket content and using experience to judge the rationality of the step sequence, the equipment status, and the correctness of interlocking relationships. While this approach can address visible errors, it often fails to detect hidden business logic problems in step connections, status transmission, and interlocking responses in a timely manner. This can easily lead to logical breakpoints, unreachable states, interlocking mismatches, and execution conflicts before formal execution.

[0030] In this embodiment, the text data, structure data, equipment status data, and interlocking constraint data of the operation ticket to be executed are first collected. The relevant content is then processed through field normalization, temporal alignment, semantic mapping, and object association to form a standardized operation ticket business feature set. Subsequently, based on the standardized data content, operation step units, operation object units, and state constraint units are extracted. Sequential dependencies between operation steps and interlocking response relationships between steps are established, thereby forming an operation ticket logical topology diagram. Based on this logical topology diagram, an executable digital twin ticket is constructed in a virtual execution environment, and a ticket execution state space is generated. This transforms the originally static ticket content into a digital execution carrier that can be progressively advanced, dynamically updated, and correlated for judgment.

[0031] In practical applications, before issuing formal operation instructions, operators import the operation ticket to be reviewed into the intelligent control system corresponding to this invention. The system first reads the current execution status, the current status of the operation object, the state constraint satisfaction status, and the interlock response permission status step by step according to the operation ticket, and inputs these status contents into a deep learning state evolution prediction model. This model combines the association structure between operation objects, the restriction relationship between state constraints, and the sequential relationship of operation steps to pre-predict the possible equipment state transition results, interlock response results, and constraint satisfaction results of subsequent steps, thereby forming a complete dynamic pre-prediction sequence of the ticket. Based on this pre-prediction sequence, the system then identifies whether there are interruptions in the connection between each step, whether a certain step cannot meet the predetermined state requirements, whether the interlock permission is consistent with the constraint conditions, whether there are conflicting execution conditions between successive steps, and whether the risk status continues to be transmitted along the subsequent path, and outputs the corresponding implicit logical risk results of the operation ticket.

[0032] Once the system identifies a relevant risk, it does not stop at the level of prompting, but further combines the executable digital twin ticket to make targeted corrections to the execution sequence, status conditions, permission conditions, and interlocking response relationships of the problematic steps, generate an operation ticket optimization control scheme, and then feed the corrected execution path back to the operators for review.

[0033] To verify the performance of the present invention, it was compared with the traditional method. The comparison results are shown in Table 1.

[0034] Table 1. Comparison of the overall performance of intelligent control methods and traditional verification methods for operation ticket business logic.

[0035] As shown in Table 1, the method of this invention significantly outperforms traditional verification methods in key indicators of operational ticket business logic verification. Specifically, the logic breakpoint identification rate increased from 81.6% to 94.4%, the interlocking mismatch node identification rate increased from 76.8% to 92.6%, and the execution conflict node identification rate increased from 74.5% to 90.8%. This indicates that the invention can not only detect surface-level problems on the ticket but also, by combining operational steps, operational objects, state constraints, sequential dependencies, and interlocking response relationships, more deeply identify logical problems hidden in the process of step connection and interlocking transmission. In particular, the improvement of 16.3% in execution conflict node identification demonstrates that the invention, through executable digital twin tickets and dynamic pre-playback sequences, can detect inconsistencies between preceding and following steps that are difficult to expose directly using traditional methods at an earlier stage.

[0036] Meanwhile, the overall accuracy rate of operation ticket verification improved to 96.2%, and the first-pass rate after correction increased to 94.3%, indicating that the present invention not only improves risk identification capabilities but also enhances the effectiveness of subsequent control corrections. This is because the present invention does not remain at the static review level; instead, it first constructs a logical topology diagram of the operation ticket, then forms an executable digital twin ticket in a virtual execution environment, and combines it with a deep learning state evolution prediction model to jointly pre-simulate equipment state transition results, interlock response results, and constraint satisfaction results. Based on this, the system can target identified implicit logical risks and make targeted corrections to the step sequence, state conditions, and interlock response relationships. Therefore, the corrected operation ticket is more likely to meet actual execution requirements, ultimately demonstrating higher verification accuracy and a higher first-pass rate.

[0037] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent control of ticket operation business logic based on deep learning, characterized in that, Includes the following steps: Collect the text data, structure data, equipment status data, and interlock constraint data of the operation ticket to be executed to form the original business dataset of the operation ticket, and perform preprocessing to generate a standardized business feature set of the operation ticket; Based on the standardized operation ticket business feature set, operation step units, operation object units, and state constraint units are extracted, and sequential dependency relationships and interlocking response relationships are constructed to generate an operation ticket logical topology diagram. The operation ticket logic topology diagram is mapped to the virtual execution environment to construct an executable digital twin ticket and generate the ticket execution state space; A step-by-step pre-play is performed on the executable digital twin ticket, and a deep learning state evolution prediction model is called during each pre-play process to predict the equipment state transition results, interlock response results and constraint satisfaction results in the current pre-play state, and generate a dynamic pre-play sequence of the ticket. Based on the dynamic pre-simulation sequence of the ticket, identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps, and generate implicit logical risk results for the operation ticket; Based on the implicit logic risk results of the operation ticket, the execution control of the executable digital twin ticket is modified, an optimized control scheme for the operation ticket is generated, and the intelligent control results of the operation ticket business logic are output.

2. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the original business dataset for the operation ticket specifically includes: Read the text data on the ticket corresponding to the operation to be executed, extract the operation instructions, condition descriptions and status confirmations for each operation step, and form a subset of the ticket text. Obtain the ticket structure data corresponding to the operation ticket to be executed, and extract the ticket structure subset; The device status data corresponding to the operation ticket to be executed is collected and associated to obtain the current status information and permission status information of each operation object at the current moment, forming a subset of device status; The interlock constraint data corresponding to each operation step is retrieved to obtain the interlock constraint content and execution permission content corresponding to each operation step, forming an interlock constraint subset. The subsets of ticket text, ticket structure, equipment status, and interlocking constraints are merged to generate the original business dataset of the operation ticket.

3. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The preprocessing includes field normalization, temporal alignment, semantic mapping, and object association.

4. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the operation ticket logic topology diagram specifically includes: Based on the standardized operation ticket business feature set, the standard expression data of operation steps and the corresponding relationship data between steps are analyzed to determine the step content, step identifier and step sequence of each operation step, and the step boundary and step affiliation of each operation step are determined to form a set of operation step units. Based on the standard expression data of operation objects and the correspondence data between steps and operation objects in the standardized operation ticket business feature set, the operation objects corresponding to each operation step are matched and labeled to determine the operation object identifier and object affiliation corresponding to each operation step, thus forming a set of operation object units. Based on the state constraint standard expression data and the correspondence data between steps and state constraints in the standardized operation ticket business feature set, the state conditions, sequence conditions and permission conditions corresponding to each operation step are constrained and analyzed to determine the constraint content and constraint attribution of each operation step, thus forming a set of state constraint units. Based on the set of operation step units and the correspondence data between steps, establish the sequential connection relationship and dependency triggering relationship between each operation step to form a sequential dependency relationship; Based on the set of operation step units, the set of operation object units, the set of state constraint units, and the current state data and permission state data, the association and determination of the current state satisfaction relationship and permission state satisfaction relationship corresponding to the execution of each operation step are performed to establish the interlock constraint transmission relationship and execution permission transmission relationship between each operation step, thus forming an interlock response relationship. The operation step unit set, operation object unit set, state constraint unit set, sequential dependency relationship and interlock response relationship are associated and organized to generate an operation ticket logic topology diagram.

5. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the ticket execution state space specifically includes: Read the set of operation step units, the set of operation object units, the set of state constraint units, the sequence dependency relationship and the interlock response relationship in the operation ticket logic topology diagram. According to the same operation step identifier and the same operation object identifier, perform mapping and splitting on each data to obtain the operation step mapping result, operation object mapping result, state constraint mapping result, sequence dependency mapping result and interlock response mapping result. Based on the operation step mapping results and operation object mapping results, the operation objects corresponding to each operation step are loaded into the corresponding operation step positions in the virtual execution environment, establishing a virtual execution correspondence between each operation step and each operation object, and forming a step object loading result; Based on the loading results of the step objects and the mapping results of the state constraints, the state constraints corresponding to each operation step are loaded into the corresponding operation step position. Combined with the sequential dependency mapping results and the interlocking response mapping results, virtual sequential inheritance relationships and virtual interlocking response relationships between each operation step are established to form the relationship loading results. Based on the relation loading results, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are synchronized and organized to form the ticket status organization result. Based on the ticket status organization results, an executable digital twin ticket is constructed in a virtual execution environment, which includes each operation step, each operation object, each state constraint, each sequential dependency relationship, and each interlocking response relationship. Based on the executable digital twin ticket, the execution status of each operation step, the current status of each operation object, the satisfaction status of each state constraint, and the permission status of each interlocking response are associated and arranged to generate the ticket execution state space.

6. The intelligent control method for operation ticket business logic based on deep learning according to claim 5, characterized in that, The virtual sequential succession relationship represents the relationship in which each operation step is connected and advanced sequentially according to the sequential dependency relationship in the virtual execution environment. The virtual interlocking response relationship represents the relationship in which each operation step is constrained and authorized according to the interlocking response relationship in the virtual execution environment.

7. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the dynamic pre-simulation sequence for the ticket specifically includes: Read the executable digital twin ticket and the ticket execution state space, determine the step-by-step pre-play starting position according to the order of the operation steps, extract the corresponding execution state, current state, satisfied state and permitted state, and obtain the current pre-play state result; Based on the current pre-simulation results and the executable digital twin ticket, determine the corresponding operation objects, state constraint units, and interlocking response relationships in the current pre-simulation state; The current pre-simulation state results, operation objects, state constraint units, and interlocking response relationships are encoded to form the current step pre-simulation input feature vector. After normalization, the vector is input into a deep learning state evolution prediction model to predict the equipment state transition results, interlocking response results, and constraint satisfaction results under the current pre-simulation state. The equipment state transition results, interlock response results, and constraint satisfaction results are associated with the same operation step identifier to form the single-step pre-simulation result corresponding to the current operation step. Based on the single-step pre-simulation result corresponding to the current operation step, the execution state, current state, satisfied state and permitted state in the ticket execution state space are updated, and the updated execution state, current state, satisfied state and permitted state are used as the pre-simulation input for the operation step corresponding to the next sequential position to generate the next pre-simulation state result; The result of the next pre-simulation state is used as the pre-simulation input for the operation step corresponding to the next sequential position. The single-step pre-simulation result corresponding to the next operation step is generated and repeated according to the sequential position of the operation steps to obtain the set of single-step pre-simulation results corresponding to each operation step. The sets of single-step pre-simulation results corresponding to each operation step are linked and organized according to their sequential positions to generate a dynamic pre-simulation sequence for the ticket.

8. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the implicit logic risk result of the operation ticket specifically includes: Read the dynamic pre-simulation sequence of the ticket, and extract the equipment state transition results, interlocking response results and constraint satisfaction results corresponding to each operation step according to the operation step identifier to form a step pre-simulation result set; Based on the step pre-simulation result set, the corresponding comparison of the equipment state transition results between adjacent operation steps is performed to identify the discontinuous step connection positions between the equipment state transition result of the previous operation step and the execution requirements of the next operation step, and to obtain the logic breakpoint identification results. Based on the pre-simulation result set, the equipment state transition results and constraint satisfaction results corresponding to each operation step are jointly verified to identify the operation step where the equipment state transition results do not meet the corresponding state constraint requirements, and the state unreachable node identification results are obtained. Based on the pre-simulation result set, the interlocking response results and constraint satisfaction results corresponding to each operation step are verified, and the operation step positions where the interlocking response results are inconsistent with the corresponding permission conditions are identified to obtain the interlocking mismatch node identification results. Based on the step-by-step simulation result set, combined with the equipment state transition results, interlock response results and constraint satisfaction results between adjacent operation steps, the positions of operation steps with conflicting execution conditions in the same temporal progression path are identified, and the execution conflict node identification results are obtained. The step association path of risk status transmission to subsequent operation steps along the sequential position is also identified, and the risk propagation node identification results are obtained. The results of logical breakpoint identification, unreachable state node identification, interlock mismatch node identification, execution conflict node identification, and risk propagation node identification are associated and organized according to the operation step identifier and sequential position to generate implicit logical risk results for the operation ticket.

9. The intelligent control method for operation ticket business logic based on deep learning according to claim 1, characterized in that, The generation of the intelligent control result of the operation ticket business logic specifically includes: Read the implicit logical risk results and executable digital twin ticket of the operation ticket, and extract the logical breakpoint identification results, unreachable status node identification results, interlock mismatch node identification results, execution conflict node identification results and risk propagation node identification results according to the operation step identifier and sequence position; The execution order of operation steps with logical breakpoint identification results and execution conflict node identification results is rearranged to form a step rearrangement result; For the operation steps of identifying unreachable nodes and interlocking mismatch nodes, supplement the corresponding state conditions and permission conditions, and load the supplemented state conditions and permission conditions into the corresponding state constraint positions in the executable digital twin ticket to form the condition supplement result; Based on the step association path corresponding to the risk propagation node identification result, the interlocking response relationship in the executable digital twin ticket is corrected to form an interlocking that satisfies the correction result; Based on the results of step rearrangement, condition supplementation, and interlocking satisfaction correction, the operation steps, state constraints, sequential dependencies, and interlocking response relationships in the executable digital twin ticket are updated accordingly to form the execution path correction results. Based on the execution path correction results, an optimized control scheme for operation tickets is generated, which serves as the intelligent control result for the operation ticket business logic.

10. A deep learning-based intelligent control system for ticket operation logic, executing the deep learning-based intelligent control method for ticket operation logic as described in any one of claims 1 to 9, characterized in that, include: The raw business data construction module is used to collect the text data, structure data, equipment status data, and interlocking constraint data of the operation tickets to be executed, forming the raw business dataset of the operation tickets, and performing preprocessing to generate a standardized business feature set of the operation tickets; The logical topology construction module is used to extract operation step units, operation object units, and state constraint units based on the standardized operation ticket business feature set, and to construct sequential dependency relationships and interlocking response relationships to generate an operation ticket logical topology diagram. The digital twin ticket construction module is used to map the logic topology diagram of the operation ticket to the virtual execution environment, construct an executable digital twin ticket, and generate the ticket execution state space; The dynamic pre-show prediction module is used to perform step-by-step pre-shows on the executable digital twin ticket and call the deep learning state evolution prediction model during each pre-show process to predict the equipment state transition results, interlocking response results and constraint satisfaction results in the current pre-show state, and generate a dynamic pre-show sequence for the ticket. The hidden risk identification module is used to identify logical breakpoints, unreachable state nodes, interlocking mismatch nodes, execution conflict nodes, and risk propagation nodes between operation steps based on the dynamic pre-play sequence of the ticket, and generate hidden logical risk results for the operation ticket. The control correction output module is used to perform control corrections on the executable digital twin ticket based on the implicit logic risk results of the operation ticket, generate an optimized control scheme for the operation ticket, and output the intelligent control results of the operation ticket business logic.

Images