Two-ticket management system of new energy power station

Abstract

The invention discloses a two-ticket management system of a new energy power station, relates to the technical field of safety management, and solves the technical problem that the prior art cannot assist in generating work tickets and cannot count the duty performance of related personnel according to collected work tickets and operation tickets. The system comprises a two-ticket making module used for making two tickets; the initial verification module is used for verifying whether the work ticket and the operation ticket generated by the two-ticket manufacturing module meet the specification or not; the checking completion module is used for checking the content manually filled in the pictures of the two tickets after the uploaded task is finished, and if errors are found through checking, the two corresponding tickets are labeled and fed back to a specified user side; the counting module is used for counting the checking result of the initial checking module, completing the checking result of the checking module, scoring the duty performance conditions of related personnel and calculating the duty performance scores of all workers after the checking cycle is ended; the method can assist the power station to practically implement the two-ticket system.

Description

Technical Field

[0001] The present invention relates to the technical field of safety management, and in particular to a two-ticket management system for a new energy power station. Background Art

[0002] Two-ticket management is a widely used management method to ensure operational safety, personal safety and equipment safety in power plants. The two tickets refer to work tickets and operation tickets. The work ticket is confirmed by the work ticket issuer, the work supervisor and the work permit issuer in sequence, and is used to design the specific content of the maintenance work and confirm the necessity and safety of the work; the operation ticket is used to clarify the specific operation process of a single work, and is generally used to ensure the safety of operations with a wide range of impact and high operational risks. Two-ticket management is of great significance to the safety of daily maintenance and repair of power plants through multi-person confirmation and supervision at each level.

[0003] In existing research, there are some intelligent operation ticket generation systems for operation tickets whose contents are easy to solidify, but there is no auxiliary generation system for work tickets that require comprehensive consideration at all levels. At the same time, electronic operation tickets and work tickets are not applicable in some scenarios. In addition, the two-ticket management used in existing power stations mostly remains in the stage of manual design, filling, confirmation, and acceptance. The process requires a lot of manpower and simple manual confirmation cannot identify some errors well. Although the existing two-ticket process is set up for multi-person confirmation, the specific performance of personnel at all levels cannot be collected and counted, and it is impossible to judge whether personnel at all levels conscientiously implement the two-ticket rules and regulations to ensure the safety of personnel and equipment. This may sometimes lead to the two-ticket system failing to achieve the original safety management effect. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention provides a two-ticket management system for a new energy power station, which solves the technical problem that the prior art cannot assist in generating work tickets and cannot count the performance of relevant users based on the collected work tickets and operation tickets.

[0005] A two-ticket management system for a new energy power station, including: a two-ticket production module, an initial verification module, a completion inspection module and a statistics module

[0006] The two-ticket making module includes a template selection unit, a project selection unit, and an item recommendation unit;

[0007] The template selection unit is used to determine the types of two tickets to be produced from the template library and determine the corresponding templates;

[0008] The project selection unit is used to determine the project to be set from the project library according to the two ticket types;

[0009] The item recommendation unit is used to determine the items that need to be operated from the item library according to the set items;

[0010] The initial verification module is used to verify whether the work ticket and operation ticket generated by the two-ticket production module comply with regulations;

[0011] The completion check module is used to check the manually filled content in the two uploaded task completion pictures. If errors are found during the check, the corresponding two tickets will be labeled and fed back to the designated user end.

[0012] The statistical module is used to count the verification results of the initial verification module and the inspection results of the completion inspection module, score the performance of the relevant personnel involved, and calculate the performance scores of all staff members after the accounting cycle ends;

[0013] The user end is used for users to make, view and download the corresponding two tickets.

[0014] Furthermore, the template library and the project library are defined according to relevant work specifications; the matter library is produced by an information extraction model, and the production process includes:

[0015] All recent work ticket and operation ticket images are collected, preprocessed and stored separately, and then the information extraction model is used to extract information from the work ticket images and operation ticket images to obtain the work ticket dataset and operation ticket dataset, wherein the preprocessing includes grayscale processing, geometric transformation and graphic enhancement of the images in sequence.

[0016] Further processing of the extracted information includes: extracting each piece of data into two sections, the first section stores the template and task name, the last section stores the project name and the matters corresponding to each project, and then classifying and storing the data according to the first section, wherein the data with the same first section remains unchanged, and the data in the last section is integrated into a list, and then the matters corresponding to the same project name are integrated, and the duplicate and missing data are deleted, thereby obtaining a matter library that reflects the mapping relationship between template-task name and multiple project names-specific matters.

[0017] Furthermore, the workflow of the initial verification module includes: marking error-prone items in the project library according to statistical data, treating the remaining items as regular items, and first verifying whether the error-prone items are correct during verification. If the error-prone items are incorrect, the corresponding two tickets will be returned to the two-ticket production module. If the error-prone items are correct, the regular items will be verified again. If the regular items are verified to be correct, the corresponding two tickets will be pushed to the designated user. If the regular items are incorrect, the corresponding two tickets will be returned to the two-ticket production module. In actual application, if the number of errors in a regular item reaches a threshold within a set period, it will be marked as an error-prone item; if an error-prone item has not made any error within a set period, it will be marked as a regular item.

[0018] Furthermore, the verification process includes:

[0019] Feature extraction: Collect a certain number of two-ticket documents, and automatically identify and extract the features of the two-ticket type, task name, project name, and specific matters corresponding to each project. By analyzing the features, distinguish them according to the project name and list multiple two-ticket types-task names-project names-specific matters corresponding to the projects as identification data blocks.

[0020] Model training: To identify data blocks, corresponding labels are added as data sets to train and optimize the neural network. The data sets include positive samples and negative samples. The positive samples include two samples with correct labels, and the negative samples include two samples with error labels. The error labels are divided into work tickets and operation tickets. The error labels of work tickets include work tasks, unclear locations, failure to use double names in accordance with regulations, non-compliant working hours, failure to analyze dangerous points or analysis is not specific, non-standard or unclear terminology, and inconsistent numbering. The error labels of operation tickets include inconsistency between operation items and task names, failure to use double names in accordance with regulations, non-compliant working hours, and incorrect order of operation items. The trained neural network verifies the accuracy of the items in the two tickets and outputs the verification results based on the accuracy of the items.

[0021] Furthermore, the verification process also includes data updating: as time goes by, the verification standards of some projects will continue to change. The initial verification module will feed back the verification results to the two-ticket production module for confirmation. If the two-ticket production module does not agree with the verification results, the relevant data will be collected in real time and stored in the update database. When the amount of data in the updated database reaches the threshold, the model will be retrained to ensure the accuracy of the model verification.

[0022] Furthermore, the inspection process of completing the inspection module includes:

[0023] Several images of two compliant tickets and two non-compliant tickets are collected, and then the two ticket images are annotated and labeled accordingly as a data set. The compliant label is correct, and the non-compliant labels are divided into work tickets and operation tickets. The non-compliant labels for work tickets include: non-compliant signatures or seals, multiple modifications, non-compliant ticks, etc.; the non-compliant labels for operation tickets include: multiple modifications, operation contents are not processed one by one, non-compliant signatures or seals, etc.

[0024] The dataset was used to train the CNN model to obtain an image recognition and classification model to check the final two batches of pictures.

[0025] Furthermore, the scoring process of the statistical module includes:

[0026] An initial performance score is set for each registered user. When the initial verification module detects an error, it will be fed back to the two-ticket production module. If the two-ticket production module confirms that the error is established, the statistical module will increase the error index F of the user interacting with the two-ticket production module by 1. If the two-ticket production module confirms that the error is not established, the two tickets will be sent back to the initial verification module for re-examination. When the initial verification module verifies that there is no error, the statistical module will increase the correct index T of the user interacting with the two-ticket production module by 1.

[0027] When the completion inspection module detects an error, it will be fed back to the user who uploaded the two tickets. If the user confirms that the error is established, the statistics module will notify the relevant users in the ticket of the identification of this error, and increase the error index F of the relevant users by 1. If the user confirms that the error is not established, it will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that the error is established, the statistics module will notify the relevant users in the ticket of the identification of this error, and increase the error index F of the relevant users by 1; if the completion inspection module checks that there is no error, the correct result will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that there is no error, the correct index T of the relevant user will be increased by 1; if it is a work ticket, the relevant users are the work person in charge and the work permit holder, and if it is an operation ticket, the relevant users are the operator and supervisor. Specifically, the relevant personnel are called relevant users in the system.

[0028] Furthermore, the performance score calculation process is as follows:

[0029]

[0030] In the formula, S 1 is the initial performance score, θ is the identity weight, α i is the responsibility weight of the user when the i-th error index is increased by 1, β j The responsibility weight when the j-th correct index of the user is increased by 1.

[0031] Among them, the identity weight is positively correlated with the work authority possessed by the user, and the responsibility weight is positively correlated with the work responsibilities when making and executing the two tickets.

[0032] Furthermore, if the initial verification module finds an error, the responsibility weights of the relevant users who made the two tickets are added together as 1, and the respective weight ratios are increased or decreased according to the specific identities of the users; if the completion check module finds an error, the responsibility weights of the relevant users who executed the two tickets are added together as 1, and the respective weight ratios are increased or decreased according to the different execution contents.

[0033] Furthermore, the user side includes multiple user types, and different user types have corresponding permissions according to different responsibilities. Specifically, the responsibilities include work ticket issuer, work supervisor, work permit issuer, operator, shift leader, supervisor, and operator.

[0034] The beneficial effects of the present invention include:

[0035] The two-ticket production module is set up through the template library and project library. It first solidifies the specific format and project type of the two tickets, and then extracts and integrates the specific operation items corresponding to the tasks in the work tickets and operation tickets of the corresponding units through the design of the item library. When the relevant personnel are making the two tickets, specific items are recommended to facilitate faster and more accurate two-ticket production.

[0036] The initial verification module is used to verify whether the contents of the two tickets prepared by the relevant person in charge comply with relevant regulations, and then the completion inspection module is used to check whether the contents of the two tickets after execution are correct. In this way, artificial intelligence is used to assist in confirming the accuracy of the two tickets prepared and the compliance of the two tickets completed, ensuring the smooth operation of the two-ticket system from two levels.

[0037] Finally, the verification and inspection results of the initial verification module and the completion inspection module are counted through the statistical module, so as to evaluate the performance of relevant staff involved in the two-vote system based on objective data, so as to criticize and improve the staff who do not seriously implement the two-vote system, and promote the staff to effectively implement the two-vote system. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 The figure is a structural diagram of a two-ticket management system for a new energy power station. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0040] The content of the work ticket generally includes relevant personnel (work team members, issuer, person in charge, permit holder), work tasks (work content, name of maintenance equipment), work location, safety measures (switches and switches that should be pulled and have been pulled; installation of ground wires), number, working time, and task name;

[0041] The content of the operation ticket generally includes the task name, operation items, operation sequence, time (command operation time, end time, each operation time), and relevant personnel (shift leader, guardian, operator).

[0042] A two-ticket management system for a new energy power station, comprising:

[0043] Two-ticket production module, including template selection unit, project selection unit, and matter recommendation unit;

[0044] The template selection unit is used to determine the types of two tickets to be produced from the template library and determine the corresponding templates;

[0045] The project selection unit is used to determine the project to be set from the project library according to the two ticket types;

[0046] The item recommendation unit is used to determine the items that need to be operated from the item library according to the set items;

[0047] The initial verification module is used to verify whether the work ticket and operation ticket generated by the two-ticket production module comply with the regulations;

[0048] The completion check module is used to check the manually filled content in the two tickets of the uploaded task completion pictures. If errors are found during the check, the corresponding two tickets will be labeled and fed back to the designated user end.

[0049] The statistics module is used to count the verification results of the initial verification module and the inspection results of the completion inspection module, score the performance of the relevant personnel involved, and calculate the performance scores of all staff members after the accounting cycle ends.

[0050] Specifically, during the operation, relevant staff interact with the two-ticket production module through the user terminal, first determine the type of the two tickets, that is, whether to produce a work ticket or an operation ticket, and then the template selection unit recommends the corresponding template, and then fills in the corresponding project type in the template through the project selection unit according to the type of the two tickets, and finally determines the specific operation items of each project through the item recommendation unit according to the project type. If the required items are not covered in the item recommendation unit, you can add them by directly creating a new one in the item recommendation unit.

[0051] After the two tickets produced by the two-ticket production module are verified and qualified by the initial verification module, the staff who execute the two tickets can view and download them on the user end, print them out, take photos and upload them to the completion inspection module after execution, and the completion inspection module will check and archive them.

[0052] In another embodiment, the template library and the project library are defined according to relevant work specifications; specifically, the template library includes a work ticket template and an operation ticket template, the work ticket template includes a task name, working time, and number, and the operation ticket module includes an operation task name, a command operation time, and an end time; the project library includes two types of project sub-libraries for storing work ticket-related items and operation ticket-related items respectively, the project sub-library storing work tickets includes work team members, work content, maintenance equipment name, work location, safety measures, switches and knife switches that should be pulled and have been pulled, ground wires that need to be installed, work ticket issuers, persons in charge, and permit holders; the project sub-library storing operation tickets includes operation items, time for each operation, shift leader, guardian, and operator.

[0053] The event database is produced by an information extraction model, and the production process includes:

[0054] All recent work ticket and operation ticket images are collected, preprocessed and stored separately, and then the information extraction model is used to extract information from the work ticket images and operation ticket images to obtain the work ticket dataset and operation ticket dataset, wherein the preprocessing includes grayscale processing, geometric transformation and graphic enhancement of the images in sequence.

[0055] Further processing of the extracted information includes: extracting each piece of data into two sections, the first section stores the template and task name, the last section stores the project name and the matters corresponding to each project, and then classifying and storing the data according to the first section, wherein the data with the same first section remains unchanged, and the data in the last section is integrated into a list, and then the matters corresponding to the same project name are integrated, and the duplicate and missing data are deleted, so as to obtain a matter library that reflects the mapping relationship between template-task name and multiple project names-specific matters.

[0056] Specifically, the information extraction model is an LMDX model trained with a specific data set, which is a labeled work ticket training set and an operation ticket training set. The work ticket training set includes the area where the content involved in the template in each work ticket image is located, the area where each item is located, and the corresponding text labels, that is, the content of each item that can be obtained from the image; the same is true for the operation ticket training set.

[0057] In another embodiment, an initial verification module is used to verify whether the work tickets and operation tickets generated by the two-ticket production module comply with regulations; specifically, easy-to-error items in the project library are marked according to statistical data, and the remaining items are used as regular items. During verification, first verify whether the easy-to-error items are correct. If the easy-to-error items are incorrect, the corresponding two tickets are returned to the two-ticket production module. If the easy-to-error items are correct, the regular items are verified. If the regular items are verified to be correct, the corresponding two tickets are pushed to the designated user. If the regular items are incorrect, the corresponding two tickets are returned to the two-ticket production module. In actual application, if a regular item has an error number of times reaching a threshold within a set period, it will be marked as an easy-to-error item; if an easy-to-error item has no errors within a set period, it will be marked as a regular item.

[0058] In another embodiment, the specific verification process is as follows:

[0059] Feature extraction: Collect a certain number of two-ticket documents, and automatically identify and extract the features of the two-ticket type, task name, project name, and specific matters corresponding to each project. By analyzing the features, distinguish them according to the project name and list multiple two-ticket types-task names-project names-specific matters corresponding to the projects as identification data blocks.

[0060] Model training: To identify data blocks, corresponding labels are added as data sets to train and optimize the neural network. The data sets include positive samples and negative samples. The positive samples include two samples with correct labels, and the negative samples include two samples with error labels. The error labels are divided into two parts: work tickets and operation tickets. The error labels of work tickets include work tasks, unclear locations, failure to use double names in accordance with regulations, non-compliant working hours, failure to analyze dangerous points or analysis is unclear, non-standard or unclear terminology, and inconsistent numbering. The error labels of operation tickets include inconsistency between operation items and task names, failure to use double names in accordance with regulations, non-compliant working hours, and incorrect order of operation items. The trained neural network verifies the accuracy of items in the two tickets and outputs verification results based on the accuracy of the items.

[0061] Data update: As time goes by, the verification standards of some projects will continue to change. The initial verification module will feed back the verification results to the two-ticket production module for confirmation. If the two-ticket production module does not agree with the verification results, it will collect relevant data in real time and store it in the update database. When the amount of data in the update database reaches the threshold, the model will be retrained to ensure the accuracy of the model verification.

[0062] In another embodiment, the completion inspection module checks the manually filled content in the two uploaded task completion pictures. If errors are found during the inspection, the corresponding two tickets will be labeled and fed back to the designated user end.

[0063] The specific inspection process is as follows:

[0064] Several images of two compliant tickets and two non-compliant tickets are collected, and then the two ticket images are annotated and labeled accordingly as a data set. The compliant label is correct, and the non-compliant labels are divided into work tickets and operation tickets. The non-compliant labels for work tickets include: non-compliant signatures or seals, multiple modifications, non-compliant ticks, etc.; the non-compliant labels for operation tickets include: multiple modifications, operation contents are not processed one by one, non-compliant signatures or seals, etc.

[0065] The dataset was used to train the CNN model to obtain an image recognition and classification model to check the final two batches of pictures.

[0066] In another embodiment, the statistical module counts the verification results of the initial verification module and the inspection results of the completion inspection module, scores the performance of the relevant personnel involved, and calculates the performance scores after the accounting cycle ends. The performance of all staff members can be displayed in real time through the performance scores.

[0067] The specific scoring process is as follows:

[0068] An initial performance score is set for each registered user. When the initial verification module detects an error, it will be fed back to the two-ticket production module. If the two-ticket production module confirms that the error is established, the statistical module will increase the error index F of the user interacting with the two-ticket production module by 1. If the two-ticket production module confirms that the error is not established, the two tickets will be sent back to the initial verification module for re-examination. When the initial verification module verifies that there is no error, the statistical module will increase the correct index T of the user interacting with the two-ticket production module by 1.

[0069] When the completion inspection module detects an error, it will be fed back to the user who uploaded the two tickets. If the user confirms that the error is established, the statistical module will notify the relevant users in the ticket of the error and increase the error index F of the relevant users by 1. If the user confirms that the error is not established, it will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that the error is established, the statistical module will notify the relevant users in the ticket of the error and increase the error index F of the relevant users by 1. If the completion inspection module checks that there is no error, the correct result will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that there is no error, the correct index T of the relevant user will be increased by 1. If it is a work ticket, the relevant users are the work supervisor and the work permit holder. If it is an operation ticket, the relevant users are the operator and the supervisor. Specifically, the relevant personnel are called relevant users in the system.

[0070] In another embodiment, the error index F and the correct index T of all users in the accounting period content are uniformly calculated to obtain the performance score, and the performance score calculation process is:

[0071]

[0072] In the formula, S 1 is the initial performance score, θ is the identity weight, α i is the responsibility weight of the user when the i-th error index is increased by 1, β j The responsibility weight when the j-th correct index of the user is increased by 1.

[0073] The identity weight is related to the work permissions the user has. The more work permissions the user has, the greater the identity weight. For example, if the user has the permissions of the work ticket issuer, the work supervisor, and the work permit issuer, the identity weight is 1. If the user only has the permission of the work permit issuer, the identity weight is 0.6.

[0074] In another embodiment, the responsibility weight is related to the work responsibility when making and executing the two tickets. If the initial verification module finds an error, the responsibility weights of the relevant users who made the two tickets are added to 1, and the respective weight ratios are increased or decreased according to the specific identities of the users. For example, if the work ticket is incorrect, the user who makes and confirms the work ticket is the work ticket issuer, the work supervisor, and the work permit issuer, then the responsibility weights of the work ticket issuer, the work supervisor, and the work permit issuer are 0.5, 0.3, and 0.2 respectively. If the operation ticket is incorrect, the responsibility weights of the shift leader and the on-duty supervisor who made the operation ticket are 0.5 respectively; if the completion inspection module finds an error, the responsibility weights of the relevant users who execute the two tickets are added to 1, and the respective weight ratios are increased or decreased according to the different execution contents. For example, if the work ticket is incorrect, the responsibility weights of the work supervisor and the operator who execute the work ticket are 0.6 and 0.4 respectively; if the operation ticket is incorrect, the responsibility weights of the operator and the supervisor who execute the operation ticket are both 0.5.

[0075] In another embodiment, the user terminal is used for users to create, view and download corresponding work tickets and operation tickets. Specifically, each user has corresponding permissions according to different responsibilities, including work ticket issuer, work permit issuer, work supervisor, operator, shift supervisor, supervisor, operator, etc.;

[0076] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A two-ticket management system for a new energy power station, characterized in that: include: Two ticket making modules, initial verification module, completion inspection module and statistics module, The two-ticket making module includes a template selection unit, a project selection unit, and an item recommendation unit; The template selection unit is used to determine the types of two tickets to be produced from the template library and determine the corresponding templates; The project selection unit is used to determine the project to be set from the project library according to the two ticket types; The item recommendation unit is used to determine the items that need to be operated from the item library according to the set items; The initial verification module is used to verify whether the work ticket and operation ticket generated by the two-ticket production module comply with regulations; The completion check module is used to check the manually filled content in the two tickets of the uploaded task completion pictures. If errors are found during the check, the corresponding two tickets will be labeled and feedback will be given; The statistical module is used to count the verification results of the initial verification module and the inspection results of the completion inspection module, score the performance of the relevant personnel involved, and calculate the performance scores of all staff members after the accounting cycle ends; The user end is used for users to make, view and download the corresponding two tickets.

2. According to claim 1, a two-ticket management system for a new energy power station is characterized in that: The template library and project library are defined according to relevant work specifications; The event database is produced by an information extraction model, and the production process includes: Collect all recent work ticket and operation ticket images, pre-process and store them respectively, and then use the information extraction model to extract information from the work ticket images and operation ticket images to obtain the work ticket dataset and operation ticket dataset, wherein the pre-processing includes grayscale processing, geometric transformation and graphic enhancement of the images in sequence; Further processing of the extracted information includes: extracting each piece of data into two sections, the first section stores the template and task name, the last section stores the project name and the matters corresponding to each project, and then classifying and storing the data according to the first section, wherein the data with the same first section remains unchanged, and the data in the last section is integrated into a list, and then the matters corresponding to the same project name are integrated, and the duplicate and missing data are deleted, thereby obtaining a matter library that reflects the mapping relationship between template-task name and multiple project names-specific matters.

3. According to the two-ticket management system of a new energy power station according to claim 1, it is characterized in that: The workflow of the initial verification module includes: marking error-prone items in the project library according to statistical data, treating the remaining items as regular items, and first verifying whether the error-prone items are correct during verification. If the error-prone items are incorrect, the corresponding two tickets will be returned to the two-ticket production module. If the error-prone items are correct, the regular items will be verified again. If the regular items are verified to be correct, the corresponding two tickets will be pushed to the designated user. If the regular items are incorrect, the corresponding two tickets will be returned to the two-ticket production module. In actual application, if a regular item has an error number of times reaching a threshold within a set period, it will be marked as an error-prone item; if an error-prone item has not made any error within a set period, it will be marked as a regular item.

4. According to the two-ticket management system of a new energy power station according to claim 1, it is characterized in that: The verification process includes: Feature extraction: Collect a certain number of two-ticket documents, and automatically identify and extract the features of the two-ticket type, task name, project name, and specific items corresponding to each project. By analyzing the features, distinguish them according to the project name and list multiple two-ticket types-task names-project names-specific items corresponding to the projects as identification data blocks; Model training: To identify data blocks, corresponding labels are added as data sets to train and optimize the neural network. The data sets include positive samples and negative samples. The positive samples include two samples with correct labels, and the negative samples include two samples with error labels. The error labels are divided into work tickets and operation tickets. The error labels of work tickets include work tasks, unclear locations, failure to use double names in accordance with regulations, non-compliant working hours, failure to analyze dangerous points or analysis is not specific, non-standard or unclear terminology, and inconsistent numbering. The error labels of operation tickets include inconsistency between operation items and task names, failure to use double names in accordance with regulations, non-compliant working hours, and incorrect order of operation items. The trained neural network verifies the accuracy of the items in the two tickets and outputs the verification results based on the accuracy of the items.

5. According to the two-ticket management system of a new energy power station according to claim 1, it is characterized in that: The verification process also includes data updating: as time goes by, the verification standards of some projects will continue to change. The initial verification module will feed back the verification results to the two-ticket production module for confirmation. If the two-ticket production module does not agree with the verification results, the relevant data will be collected in real time and stored in the update database. When the amount of data in the updated database reaches the threshold, the model will be retrained to ensure the accuracy of the model verification.

6. A two-ticket management system for a new energy power station according to claim 1, characterized in that: The inspection process of completing the inspection module includes: Collect several images of two compliant tickets and two non-compliant tickets, then annotate the two tickets and add corresponding labels as a data set. The compliant label is correct, and the non-compliant labels are divided into work tickets and operation tickets. The non-compliant labels of work tickets include: non-compliant signature or seal, multiple modifications, and non-compliant ticks. The non-compliant operation ticket labels include: multiple modifications, operation content not processed one by one, and non-compliant signature or seal. The dataset was used to train the CNN model to obtain an image recognition and classification model to check the final two batches of pictures.

7. A two-ticket management system for a new energy power station according to claim 1, characterized in that: The scoring process of the statistical module includes: An initial performance score is set for each registered user. When the initial verification module verifies an error, it is fed back to the two-ticket production module. If the two-ticket production module confirms that the error is established, the statistical module will increase the error index F of the user interacting with the two-ticket production module by 1. If the two-ticket production module confirms that the error is not established, the two tickets will be sent back to the initial verification module for re-examination. When the initial verification module verifies that there is no error, the statistical module will increase the correct index T of the user interacting with the two-ticket production module by 1; When the completion inspection module detects an error, it will be fed back to the user who uploaded the two tickets. If the user confirms that the error is established, the statistical module will notify the relevant users in the ticket of the error and increase the error index F of the relevant users by 1. If the user confirms that the error is not established, it will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that the error is established, the statistical module will notify the relevant users in the ticket of the error and increase the error index F of the relevant users by 1. If the completion inspection module checks that there is no error, the correct result will be handed over to the relevant person in charge for review. If the relevant person in charge confirms that there is no error, the correct index T of the relevant user will be increased by 1. If it is a work ticket, the relevant users are the work supervisor and the work permit holder. If it is an operation ticket, the relevant users are the operator and the supervisor.

8. A two-ticket management system for a new energy power station according to claim 7, characterized in that: The performance score calculation process is as follows: In the formula, S1 is the initial performance score, θ is the identity weight, α i is the responsibility weight of the user when the i-th error index is increased by 1, β j The responsibility weight of the user when the j-th correct index is increased by 1; Among them, the identity weight is positively correlated with the work authority possessed by the user, and the responsibility weight is positively correlated with the work responsibilities when making and executing the two tickets.

9. A two-ticket management system for a new energy power station according to claim 8, characterized in that: If the initial verification module finds an error, the responsibility weights of the relevant users who made the two tickets will be added together as 1, and the respective weight ratios will be increased or decreased according to the specific identities of the users; if the completion check module finds an error, the responsibility weights of the relevant users who executed the two tickets will be added together as 1, and the respective weight ratios will be increased or decreased according to the different execution contents.

10. A two-ticket management system for a new energy power station according to claim 1, characterized in that: The user side includes multiple user types. Different user types have corresponding permissions according to different responsibilities. Specifically, the responsibilities include work ticket issuer, work supervisor, work permit issuer, operator, shift leader, supervisor, and operator.

Images