Electric power safety production risk digital management and control method based on three libraries and one closed loop

Abstract

The invention relates to an electric power safety production risk digital management and control method based on three libraries and one closed loop, and the method comprises the steps: building a database through collecting electric power safety production global data and corresponding safety problem types; constructing an artificial intelligence model by using the data stored in the database to learn the relationship between the safety production global data and the safety problem type; outputting related measures of different safety problem types through a measure set output model; and monitoring the completion progress of related measures, and after the measures are completed, repeatedly monitoring safety problems and measure completion conditions to form a closed-loop monitoring loop. The problems that in the prior art, manual patrol is relied on, most of the patrol is post patrol, the safety production risk cannot be monitored and pre-controlled in real time, the execution condition of rectification measures after safety patrol is difficult to track in real time, and closed-loop management means are lacked, so that the safety management and control efficiency is low are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power production safety, and relates to a digital management and control method for electric power production safety risks based on three databases and one closed loop. Background Art

[0002] Electricity production involves high voltage, complex equipment and multiple operating links. The potential risk of safety accidents often affects the safety of employees and the public. At the same time, with the continuous growth of global electricity demand and the rapid development of technology, the power system has become more complex and the requirements for safety management have become more stringent. Relevant laws, regulations and industry standards of various countries have put forward clear requirements for the safety responsibilities of power companies, forcing companies to establish a sound power production safety management system.

[0003] However, the existing power safety management and control relies on manual inspections, most of which are post-inspections, which cannot monitor and pre-control production safety risks in real time. In addition, it is difficult to track the implementation of rectification measures in real time after safety inspections, and there is a lack of closed-loop management methods, which leads to the problem of low safety management efficiency. In this context, digital transformation has become an inevitable trend to improve management efficiency and risk control capabilities. By introducing new technologies such as big data and the Internet of Things, power companies are expected to monitor and analyze safety risks in real time and realize timely identification and handling of risks. Summary of the invention

[0004] The present invention provides a digital management and control method for power production safety risks based on three databases and one closed loop, which aims to realize safety inspections, real-time monitoring of indicators, and risk pre-control through a database, a problem database, a measure database, and rectification closed-loop monitoring, so as to promote quality improvement and efficiency enhancement, and reduce the burden on grassroots units, and solve the problems that the existing technology relies on manual inspections, most of which are post-inspections, and cannot monitor and pre-control safety production risks in real time, and it is difficult to track the implementation of rectification measures in real time after the safety inspection, and there is a lack of closed-loop management means, which leads to the problem of low safety management and control efficiency.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] The present application provides a method for digital management and control of power production safety risks based on three databases and one closed loop, comprising the following steps:

[0007] S1. Build a database: Build a database by collecting global data on power safety production and corresponding safety problem types;

[0008] S2. Build a problem database: Use the data stored in the database to build an artificial intelligence model to learn the relationship between the global data of production safety and the types of safety issues, which is used to diagnose safety issues in power production;

[0009] S3. Build a measure library: Output relevant measures of different security problem types through a measure set output model, wherein the measure set output model is trained with the security problem type and the measure receiving object as input and the corresponding relevant measures as output;

[0010] S4. Retrieve the global data of power production safety monitored in real time through the database, and use the problem database to diagnose the types of safety problems in power production in real time;

[0011] S5. When a security problem is detected, the security problem type is input into the measure set output model of the measure library, and the relevant measures are output and sent according to the measure receiving object;

[0012] S6. Monitor the progress of the relevant measures. After the measures are completed, repeat steps S4-S6 to form a closed-loop monitoring circuit.

[0013] Furthermore, in step S1, the database also manages the collected data, including:

[0014] Group data samples of the same security issue type together and create categories for management;

[0015] Create indexes based on key fields of global data on power safety production in each category;

[0016] Determine the permission level according to the importance of each category, and set the permission level allowed for different user identities;

[0017] Regularly check and update the data in the database, and automatically back it up and store it in the cloud.

[0018] Furthermore, in step S2, the artificial intelligence model includes the following construction steps:

[0019] T1. Identify the main influencing variables of the global safety production data in each safety issue type;

[0020] T2. Determine the data characteristics of the main influencing variables when each type of security problem occurs;

[0021] T3. Train a deep learning model using the data features as independent variables and the security issue type as dependent variable.

[0022] Furthermore, in step T1, the identification of the main influencing variables of the global data of production safety in each type of safety issue includes the following steps:

[0023] T11. Randomly select several global production safety data variables as features and the types of safety issues as labels to build a classification tree model;

[0024] T12, repeat step T11 to obtain a classification tree model of different global production safety data variable combinations;

[0025] T13. Use the classification tree model to predict the type of security issues in the new data set;

[0026] T14. Sort the variables of global safety production data according to the accuracy of the prediction of safety problem types;

[0027] T15. Determine the first N global safety production data variables in order as the main influencing variables of this safety issue type.

[0028] Furthermore, in step T2, determining the data features of the main influencing variables when each type of security problem occurs includes the following steps:

[0029] T21. Divide the data set into two data subsets according to whether each type of security issue occurs;

[0030] T22. Select a data subset where the security issue type occurs, and use the numerical range or data type of the main influencing variable of the data subset as the data feature of the security issue type.

[0031] Furthermore, in step S3, the measure set output model includes the following construction steps:

[0032] Obtaining a set of measures that have been successfully resolved for each type of security issue in a historical period, wherein the set of measures includes the type of security issue, an object receiving the measures, and related measures;

[0033] Classify the measure set according to the measure receiving object to obtain the measure subset;

[0034] The security issue types in the measure subset are taken as input, the relevant measures are taken as target outputs, and the measure set output model is trained.

[0035] Furthermore, in step S6, the closed-loop monitoring circuit stores the global data on power production safety, types of safety issues and related measures monitored during the process into a database after the safety issue being monitored is completely resolved.

[0036] Further, in step S6, the closed-loop monitoring circuit, when the closed-loop monitoring circuit is executed M times, if the safety problem is still not solved, reports to the system and stops the power production activities; the value of M depends on the severity of the type of safety problem.

[0037] Furthermore, the global safety production data includes management process data, power grid operation data, equipment operation and maintenance data, operation control data, violation discovery data and accident event data, etc.

[0038] Furthermore, the types of safety issues include power grid risks, equipment risks, operational risks, network security risks, occupational health risks and public safety risks.

[0039] Beneficial effects of the present invention:

[0040] By collecting global data on power production safety and corresponding types of safety issues, a database is constructed; using the data stored in the database, an artificial intelligence model is constructed to learn the relationship between global data on production safety and types of safety issues; the relevant measures for different types of safety issues are output through the measure set output model; the completion progress of relevant measures is monitored, and after the measures are completed, the safety issues and the completion of measures are repeatedly monitored to form a closed-loop monitoring loop. This solves the problem that the existing technology relies on manual inspections, most of which are post-inspections, and cannot monitor and pre-control production safety risks in real time. It is also difficult to track the implementation of rectification measures in real time after safety inspections, and there is a lack of closed-loop management methods, which leads to low safety management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0042] Figure 1 The present invention is a flowchart of a method for digitally controlling power production safety risks based on three databases and one closed loop.

[0043] Figure 2 This is a flow chart for identifying the main influencing variables of global production safety data in each safety issue type in one embodiment of the present invention.

[0044] Figure 3 This is a step of constructing a measure set output model in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0046] See also Figure 1-Figure 3 The present application provides a method for digital management and control of power production safety risks based on three databases and one closed loop, comprising the following steps:

[0047] S1. Build a database: Build a database by collecting global data on power safety production and corresponding safety problem types;

[0048] In this embodiment, the global data of safe production can be automatically monitored and collected in real time through the data interface with the existing business process management system to ensure that the safe production management of power and the unsafe state can be prevented and controlled. The following are some types of global data of safe production that can be automatically monitored and collected:

[0049] 1) Manage process data

[0050] Production management data: power grid, equipment records, electricity users, etc.

[0051] Business process data: procurement and supply, warehousing and logistics, organizational structure, job responsibilities, fixed assets, etc.;

[0052] Project management data: investment plan, project progress, project settlement, etc.

[0053] 2) Grid operation data: commissioning plan, baseline risk assessment, basic question risk assessment, repeated power outages of main and distribution network equipment, power outage process monitoring, etc.

[0054] 3) Equipment operation and maintenance data: equipment commissioning, equipment risk assessment, equipment operation and maintenance strategy, equipment defects, equipment hidden dangers, equipment countermeasures, equipment maintenance, equipment retirement, etc.

[0055] 4) Operation control data: operation tasks, operation plans, operation risk assessments, on-site survey records, operation entities, personnel qualifications, online and offline inspection data, etc.

[0056] 5) Violation discovery data: violation discovery records, violation penalty results, violation closed-loop rectification status, etc.

[0057] 6) Accident data: accident category, accident level, accident loss, accident cause and responsibility classification, etc.

[0058] Therefore, further, the global safety production data includes management process data, power grid operation data, equipment operation and maintenance data, operation control data, violation discovery data and accident event data, etc.

[0059] Furthermore, the types of safety issues include power grid risks, equipment risks, operational risks, network security risks, occupational health risks and public safety risks.

[0060] In this embodiment, the global safety production data automatically monitored can correspond to a variety of safety problem types. The following are some common safety problem types:

[0061] 1) Grid risks

[0062] Inadequate planning and construction management: unreasonable power grid planning and design, failure to resolve risks at the source of planning and design, failure to dynamically adjust power grid planning based on typical fault events, and out-of-control progress of key construction projects;

[0063] Inadequate risk assessment management: failure to conduct baseline risk assessment in a standardized manner, failure to conduct basic risk assessment in a timely manner, and failure to issue risk control measures in a timely manner;

[0064] Inadequate dispatching and operation management: unreasonable grid operation mode, unreasonable maintenance arrangement, power outage plan not implemented, equipment operation and maintenance not carried out according to grid risk dynamics, etc.

[0065] 2) Equipment risks

[0066] Inadequate management of equipment commissioning: equipment procurement is not timely, equipment installation process is not strictly controlled, and handover acceptance is not carried out in a standardized manner;

[0067] Inadequate equipment risk assessment: risk assessment is not carried out dynamically, and equipment risk assessment results are not released in a timely manner;

[0068] Equipment risk control is not in place: equipment operation and maintenance strategies are not formulated in a differentiated manner, equipment operation and maintenance are not carried out in a standardized manner, equipment defect and hidden danger inspections are not carried out in a standardized manner, equipment maintenance management is unreasonable, and equipment retirement is not carried out in a standardized manner.

[0069] 3) Operational risks

[0070] Inadequate operation plan management: Inadequate communication and coordination of operation plans, unreasonable annual plan formulation, unreasonable coordination and control of monthly plans, unreasonable adjustment of weekly plans, and non-standard implementation of daily plans, etc.

[0071] Risk assessment management is not in place: on-site investigation is not carried out, benchmark operation risk assessment is inaccurate, risk assessment is not carried out dynamically, etc.;

[0072] Inadequate operation preparation management: unreasonable personnel preparation, unreasonable tool preparation, unreasonable document preparation, unreasonable material preparation, etc.;

[0073] Inadequate operation process management: risk control measures not implemented, operation process supervision not carried out, etc.;

[0074] Inadequate management of operational violations: serious violations occur repeatedly, violations are not rectified as required, etc.

[0075] 4) Cybersecurity risks

[0076] Inadequate planning and construction management: irregular control of development and implementation, lax acceptance control, and insufficient preparation before commissioning;

[0077] Inadequate operation and scheduling management: the operation mode is not included in the management and control, the operation mode is not implemented properly, and the operation monitoring is not in place, etc.

[0078] 5) Occupational health risks

[0079] Inadequate occupational health risk assessment: Inadequate hazard identification, inadequate implementation of control measures, etc.

[0080] Inadequate labor protection management: unreasonable allocation of labor protection equipment, irregular use of personal protective equipment, etc.

[0081] 6) Public safety risks

[0082] Inadequate hazard investigation and control: Inadequate environmental pollution prevention and control of operating equipment, inadequate risk control of important crossing hazards, inadequate control of flooding hazards in low-lying areas, inadequate control of leakage hazards in low-voltage equipment, etc.;

[0083] The joint management and control between the government and enterprises is not in place: the responsibility for reporting hidden dangers is not fulfilled in a timely manner, and the responsibilities for the safe management of power facilities are not clear.

[0084] Furthermore, in step S1, the database also manages the collected data, including:

[0085] Group data samples of the same security issue type together and create categories for management;

[0086] In this embodiment, grouping data samples of the same security problem type together and classifying them is an important strategy to improve data processing efficiency and security management level. First, by effectively classifying the data, relevant information can be quickly retrieved and accessed, reducing the time required for searching, thereby improving work efficiency. Secondly, classification management makes data analysis more convenient, can identify potential security risks and problem patterns, and provide strong support for decision-making. Classification standards can be set based on security problem types, event severity, location and time of occurrence, etc. to ensure the systematicness and accuracy of data. Specific steps include data collection, preliminary filtering, building a classification system, data entry and marking, etc. In addition, according to new security issues and management needs, the classification system is dynamically adjusted and optimized to ensure that it always conforms to the actual situation. This classification management method not only improves the efficiency of data utilization, but also enhances the ability of enterprises to identify and respond to security risks, which helps to form a safer and more efficient production environment.

[0087] Create indexes based on key fields of global data on power safety production in each category;

[0088] In this embodiment, the key fields of the global data of power production safety refer to data items that are crucial in the process of safety management and event processing, including event ID, event type, occurrence time, occurrence location, scope of impact, severity, processing status, and responsible person. The accurate and efficient management of these key fields is crucial to improving the level of safe production in the power industry. By creating indexes for these key fields, the query speed and data retrieval efficiency of the database can be significantly improved, and the retrieval time in the case of large amounts of data can be effectively reduced. The index is equivalent to the "directory" of the database, which enables the system to quickly locate the required information and avoid the performance bottleneck caused by line-by-line search. In addition, with the increase in data volume and changes in business, regular maintenance and optimization of the index is also essential to ensure the efficiency of the system in actual operation. Through efficient data indexing, power enterprise managers can quickly obtain key information and make decisions in a timely manner, so as to better respond to various safety hazards and production challenges, and ultimately achieve a dual improvement in safety and efficiency. Reasonable indexing of key fields not only helps to improve data query efficiency, but also helps to ensure power production safety and provide strong support for the sustainable development of enterprises.

[0089] Determine the permission level according to the importance of each category, and set the permission level allowed for different user identities;

[0090] Regularly check and update the data in the database, and automatically back it up and store it in the cloud.

[0091] In this embodiment, regularly checking and updating the data in the database and automatically backing it up and storing it in the cloud are key measures to ensure data integrity and security. First, regularly checking the database helps to discover and fix potential errors, vulnerabilities or data inconsistencies, and ensure the accuracy and reliability of the data. This process usually includes data verification, integrity checking, and performance optimization. Secondly, automatic backup is one of the core means of data protection. Storing database backups in the cloud can effectively prevent data loss or damage, because cloud backups provide a highly reliable storage solution and usually have data recovery and version control functions. The automated backup process not only reduces the risk of human error, but also ensures the timeliness and consistency of backups, ensuring that data can be quickly restored in the event of a system failure or disaster. In summary, regular data checks and automatic cloud backups can significantly improve the security and reliability of data management, reduce the risk of data loss or damage, and ensure business continuity and long-term stable use of data.

[0092] S2. Build a problem database: Use the data stored in the database to build an artificial intelligence model to learn the relationship between the global data of production safety and the types of safety issues, which is used to diagnose safety issues in power production;

[0093] In this embodiment, an artificial intelligence model is constructed to learn the relationship between the global data of power production safety and the types of safety issues, aiming to improve safety management efficiency and reduce accident risks. Power production involves multiple links and complex equipment. Traditional safety management often relies on manual experience and lacks foresight and real-time performance. With the help of a large amount of historical data in the database, the AI ​​model can analyze potential safety hazards, identify abnormal situations, and provide data support for timely decision-making. Through real-time monitoring, enterprises can quickly diagnose safety issues, reduce reaction time, and improve response capabilities, thereby effectively preventing accidents. In addition, this intelligent system can also optimize resource allocation and concentrate on dealing with the most urgent safety hazards, thereby improving management efficiency and reducing operating costs. At the same time, through the feedback mechanism, enterprises can continuously optimize the model, form closed-loop management, and continuously improve the level of safety production. In the face of increasingly stringent safety supervision requirements, the use of AI technology can also help enterprises achieve compliance and reduce legal risks and economic losses. In summary, building this model not only helps to ensure the safety of power production, but also promotes the transformation of enterprises to data-driven safety management and improves overall governance capabilities.

[0094] Furthermore, the artificial intelligence model includes the following construction steps:

[0095] T1. Identify the main influencing variables of the global safety production data in each safety issue type;

[0096] In this embodiment, since the global data of safe production covers multiple monitoring dimensions, these dimensions are closely related to the occurrence of various safety issues, and the reasons behind them are different. Therefore, when facing different types of safety issues, it is necessary to deeply analyze and identify the main influencing variables in the global data of safe production. This process can not only help us clearly understand the causes of various safety issues, but also significantly improve the accuracy and work efficiency of modeling. Specifically, by systematically screening and analyzing various types of data, we can focus on those factors that have the greatest impact on safety issues. On this basis, building a more accurate model will not only help identify potential safety hazards, but also formulate more effective preventive measures, thereby improving the overall safety management level. This methodology is not only applicable to current safe production practices, but also provides strong support and reference for future safety management.

[0097] Furthermore, the identification of the main influencing variables of the global data of production safety in each type of safety problem includes the following steps:

[0098] T11. Randomly select several global production safety data variables as features and the types of safety issues as labels to build a classification tree model;

[0099] In this embodiment, the classification tree model is a decision tree algorithm widely used in classification tasks, which uses a tree structure to classify data. The basic idea is to gradually divide the data set into subsets in a recursive way, and each segmentation is performed based on the value of a certain feature, thereby forming internal nodes and leaf nodes. The internal nodes represent the selected features, and the leaf nodes correspond to the final classification labels. When the model is constructed, the best feature is first selected at the root node for segmentation, and the commonly used information gain or Gini index measures the segmentation effect, and then this process is repeated on each subset until the preset stop condition is reached. The advantage of the classification tree model is that it is simple to understand and easy to explain; at the same time, it can handle nonlinear relationships and has low requirements for data preprocessing.

[0100] T12, repeat step T11 to obtain a classification tree model of different global production safety data variable combinations;

[0101] T13. Use the classification tree model to predict the type of security issues in the new data set;

[0102] T14. Sort the variables of global safety production data according to the accuracy of the prediction of safety problem types;

[0103] In this embodiment, the higher the accuracy of the combined prediction of the global data variables of production safety, the higher the ranking. There are many methods to evaluate the prediction accuracy of the classification tree model, and the combined use can provide a more comprehensive understanding of the performance of the model. First, the confusion matrix is ​​the most intuitive evaluation tool. It shows the relationship between the true category and the predicted category, and can clearly see the performance of the model in each category and the misclassification. Secondly, the classification accuracy represents the proportion of correctly predicted samples to the total samples. It is a simple and effective performance indicator, but it may be misleading when the categories are unbalanced. In order to comprehensively evaluate the model, F1-score is a very important indicator. It takes into account the precision and recall rate. The macro-average and micro-average F1-scores reflect the comprehensive performance of the model from different perspectives. In addition, the precision and recall rates provide a clear performance evaluation for each category, which can reveal the model's ability to identify positive and negative samples. The AUC-ROC curve is used to measure the performance of the model at different thresholds. The closer the AUC value is to 1, the stronger the classification ability of the model is, especially suitable for dealing with situations with multiple classes. Finally, Log loss is an indicator for evaluating the accuracy of the model's prediction probability for each class. The smaller the value, the more ideal the model. In summary, by applying these evaluation methods, the effect of the classification model can be comprehensively analyzed, thus providing an important basis for further optimization.

[0104] Among them, the calculation formula of F1-score is:

[0105]

[0106] Where F1 is the F1 value; P is the precision; S is the recall rate.

[0107] T15. Determine the first N global safety production data variables in order as the main influencing variables of this safety issue type.

[0108] In this embodiment, the value range of N depends on the accuracy of the model prediction. When the accuracy is not significantly improved after the N+1th variable is added to the model, the variables after N are no longer listed as the main influencing variables.

[0109] T2. Determine the data characteristics of the main influencing variables when each type of security problem occurs;

[0110] Furthermore, the determination of the data features of the main influencing variables when each type of security problem occurs includes the following steps:

[0111] T21. Divide the data set into two data subsets according to whether each type of security issue occurs;

[0112] T22. Select a data subset where the security issue type occurs, and use the numerical range or data type of the main influencing variable of the data subset as the data feature of the security issue type.

[0113] It should be noted that, in this embodiment, since the global data variables of production safety include continuous variables and categorical variables, the data features that mainly affect the variables should be the numerical range or data type.

[0114] T3. Train a deep learning model using the data features as independent variables and the security issue type as dependent variable.

[0115] In this embodiment, the deep learning model is used to train the relationship between the characteristics of the main influencing variables of the global safety production data and the corresponding safety problem types, and is used for the diagnosis of safety problems in power production. Common deep learning models include random forest models, artificial neural network models, etc.

[0116] S3. Build a measure library: Output relevant measures of different security problem types through a measure set output model, wherein the measure set output model is trained with the security problem type and the measure receiving object as input and the corresponding relevant measures as output;

[0117] In this embodiment, due to the different perspectives on security issues, the security processing measures required by different recipients of measures (such as the difference between the execution layer and the management layer) are also different. Specifically, the execution layer usually focuses on specific security technologies and systems. For example, after discovering the problem of repeated power outages in a short period of time, the security management problem is located. If the power outage plan demand is not reported in time, the execution layer will be pushed with a list of recommended monthly equipment power outage maintenance test plans. In contrast, the management layer pays more attention to the formulation and implementation of safety policies, including safety training, resource allocation and risk assessment. After discovering the problem of repeated power outages in a short period of time, the management layer will be reminded to scientifically and reasonably arrange the maintenance plan according to the equipment status, importance, heavy overload and other information. They need to ensure the effectiveness and sustainability of the overall safety management of the enterprise. Therefore, the management layer will formulate a long-term safety strategy to enhance the safety culture and emergency response capabilities of the organization. This hierarchical focus enables safety management to be comprehensively improved from both technical and management dimensions. Through the effective combination of technical measures and management strategies, enterprises can better respond to security challenges and ensure the safety of their operations. This embodiment uses both the security issue type and the measure receiving object as input variables. For different measure receiving objects, the measure output model will output different measures to deal with the security issue.

[0118] Furthermore, the measure set output model includes the following construction steps:

[0119] Obtaining a set of measures that have been successfully resolved for each type of security issue in a historical period, wherein the set of measures includes the type of security issue, an object receiving the measures, and related measures;

[0120] In this embodiment, the measures that can be used for training must be relevant processing methods that can successfully solve security problems in previous processing experience, so that the trained model can help accurately output relevant measures and solve security problems.

[0121] Classify the measure set according to the measure receiving object to obtain the measure subset;

[0122] The security issue types in the measure subset are taken as input, the relevant measures are taken as target outputs, and the measure set output model is trained.

[0123] S4. Retrieve the global data of power production safety monitored in real time through the database, and use the problem database to diagnose the types of safety problems in power production in real time;

[0124] S5. When a security problem is detected, the security problem type is input into the measure set output model of the measure library, and the relevant measures are output and sent according to the measure receiving object;

[0125] S6. Monitor the progress of the relevant measures. After the measures are completed, repeat steps S4-S6 to form a closed-loop monitoring circuit.

[0126] Furthermore, in step S6, the closed-loop monitoring circuit stores the global data on power production safety, types of safety issues and related measures monitored during the process into a database after the safety issue being monitored is completely resolved.

[0127] Further, in step S6, the closed-loop monitoring circuit, when the closed-loop monitoring circuit is executed M times, if the safety problem is still not solved, reports to the system and stops the power production activities; the value of M depends on the severity of the type of safety problem.

[0128] That is, the more serious the security problem type is, the smaller the value of M is.

[0129] In this embodiment, in order to effectively monitor the progress of the completion of relevant measures and continuously monitor the global data of production safety after the measures are completed, so as to achieve problem diagnosis and form a closed-loop monitoring circuit, the following specific steps can be taken.

[0130] First, develop a detailed plan for monitoring the progress of measures. It is necessary to clarify the goals, responsible persons, and completion time of each safety measure. During the implementation of measures, project management tools and software can be used to track progress and manage tasks. When the measures are completed, arrange for dedicated personnel to conduct effect evaluation to ensure that each measure achieves the expected effect and record relevant data for subsequent analysis.

[0131] Once the measures are implemented, the continuous monitoring and problem diagnosis phase begins. Data analysis tools are used to analyze the collected data on global production safety in real time, including trend analysis, abnormal alarms, and data visualization, to promptly identify potential safety hazards. In addition, machine learning and artificial intelligence technologies are used to conduct in-depth learning of historical data, identify potential risk patterns, and provide early warnings.

[0132] In the process of problem diagnosis, we can also combine data from multiple sources, such as employee feedback, equipment test data, and external environmental changes, to comprehensively evaluate possible sources of problems. We can formulate targeted corrective measures based on the diagnosis results and incorporate them into the closed-loop monitoring system. At the same time, we can use the data management platform to track the implementation of corrective measures to ensure that the problem is effectively resolved.

[0133] Finally, the key to forming a closed-loop monitoring loop is to integrate the feedback mechanism into the monitoring system. After each problem diagnosis, the corrective measures and their effects should be recorded in the system to form a complete safety management file. These files should be reviewed regularly to analyze the effectiveness and sustainability of the measures, so as to continuously optimize the safety management process. Integrate various types of data and feedback to form a circular monitoring system to achieve dynamic optimization of safety management.

[0134] After the safety issue is processed, the real-time monitored global safety data and the corresponding safety issue types will be stored in the database. This process involves organizing, classifying and archiving the data collected during the inspection for subsequent analysis and tracking. By establishing a systematic data storage mechanism, it can be ensured that the processing process and current status of all safety issues are effectively recorded. After the data is stored, the system will continue to monitor the processed global safety data in real time. This means that the monitoring system will continue to observe the safety status of the facility and collect new data related to the situation after the previous problem is processed to confirm whether the measures taken are effective. Through this continuous monitoring, potential safety hazards or problems can be quickly identified.

[0135] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A digital control method for power production safety risk based on three databases and one closed loop, characterized in that: The following steps are involved: S1. Build a database: Build a database by collecting global data on power safety production and corresponding safety problem types; S2. Build a problem database: Use the data stored in the database to build an artificial intelligence model to learn the relationship between the global data of production safety and the types of safety issues, which is used to diagnose safety issues in power production; S3. Build a measure library: Output relevant measures of different security problem types through a measure set output model, wherein the measure set output model is trained with the security problem type and the measure receiving object as input and the corresponding relevant measures as output; S4. Retrieve the global data of power production safety monitored in real time through the database, and use the problem database to diagnose the types of safety problems in power production in real time; S5. When a security problem is detected, the security problem type is input into the measure set output model of the measure library, and the relevant measures are output and sent according to the measure receiving object; S6. Monitor the progress of the relevant measures. After the measures are completed, repeat steps S4-S6 to form a closed-loop monitoring circuit.

2. According to claim 1, a digital control method for power production safety risk based on three databases and one closed loop is characterized by: In step S1, the database also manages the collected data, including: Group data samples of the same security issue type together and create categories for management; Create indexes based on key fields of global data on power safety production in each category; Determine the permission level according to the importance of each category, and set the permission level allowed for different user identities; Regularly check and update the data in the database, and automatically back it up and store it in the cloud.

3. According to claim 1, a digital control method for power production safety risk based on three databases and one closed loop is characterized in that: In step S2, the artificial intelligence model includes the following construction steps: T1. Identify the main influencing variables of the global safety production data in each safety issue type; T2. Determine the data characteristics of the main influencing variables when each type of security problem occurs; T3. Train a deep learning model using the data features as independent variables and the security issue type as dependent variable.

4. According to claim 3, a digital control method for power production safety risk based on three databases and one closed loop is characterized in that: In step T1, the identification of the main influencing variables of the global data of production safety in each type of safety problem includes the following steps: T11. Randomly select several global production safety data variables as features and the types of safety issues as labels to build a classification tree model; T12, repeat step T11 to obtain a classification tree model of different global production safety data variable combinations; T13. Use the classification tree model to predict the type of security issues in the new data set; T14. Sort the variables of global safety production data according to the accuracy of the prediction of safety problem types; T15. Determine the first N global safety production data variables in order as the main influencing variables of this safety issue type.

5. According to claim 3, a digital control method for power production safety risk based on three databases and one closed loop is characterized in that: In step T2, the determination of the data features of the main influencing variables when each type of security problem occurs includes the following steps: T21. Divide the data set into two data subsets according to whether each type of security issue occurs; T22. Select a data subset where the security issue type occurs, and use the numerical range or data type of the main influencing variable of the data subset as the data feature of the security issue type.

6. The method for digital control of power production safety risks based on three databases and one closed loop according to claim 1 is characterized by: In step S3, the measure set output model includes the following construction steps: Obtaining a set of measures that have been successfully resolved for each type of security issue in a historical period, wherein the set of measures includes the type of security issue, an object receiving the measures, and related measures; Classify the measure set according to the measure receiving object to obtain the measure subset; The security issue types in the measure subset are taken as input, the relevant measures are taken as target outputs, and the measure set output model is trained.

7. According to claim 1, a digital control method for power production safety risk based on three databases and one closed loop is characterized by: In step S6, the closed-loop monitoring circuit stores the global data on power production safety, the types of safety issues and related measures monitored during the process into a database after the safety issue being monitored is completely resolved.

8. The method for digital control of power production safety risks based on three databases and one closed loop according to claim 1 is characterized by: In step S6, the closed-loop monitoring circuit, when the closed-loop monitoring circuit is executed M times, if the safety problem is still not solved, reports to the system and stops the power production activities; the value of M depends on the severity of the type of safety problem.

9. The method for digital control of power production safety risks based on three databases and one closed loop according to claim 1 is characterized by: The global safety production data includes management process data, power grid operation data, equipment operation and maintenance data, operation control data, violation discovery data and accident event data, etc.

10. The method for digital control of power production safety risks based on three databases and one closed loop according to claim 1 is characterized by: The types of safety issues include power grid risks, equipment risks, operational risks, network security risks, occupational health risks and public safety risks.

Images