Detection method and device for electric power safety production and storage medium

Abstract

The invention discloses an electric power safety production detection method and device and a storage medium, and the method comprises the steps: carrying out the time sequence operation parameter change analysis and normalized operation evolution of log data, and obtaining an operation state feature; performing safety production structure evolution on the operation state characteristics based on the log data to construct an electric power safety production knowledge graph; performing frame-by-frame image dynamic optical flow tracking on the video data to obtain production behavior characteristics of dynamic production behaviors of operators; performing dynamic behavior visual identification on the production behavior features and performing dynamic operation process feature analysis to generate dynamic operation process feature data; generating a production behavior risk quantized value for the dynamic operation process feature data of the operator based on the electric power safety knowledge graph; and performing emergency response decision based on the production behavior risk quantized value so as to optimize the electric power safety production knowledge graph. According to the embodiment of the invention, the capability of preventing and dealing with complex electric power safety accidents is improved.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a detection method, equipment and storage medium for safe production of electric power. Background Art

[0002] With the rapid development of the power industry and the continuous application of intelligent technology, power production safety has become one of the important links to ensure the stable operation of the power grid and the safety of people's lives and property. In the process of power production, transmission, distribution and operation of related equipment, various emergencies and hidden dangers occur from time to time, bringing challenges to power safety management.

[0003] Traditional power production safety management methods mostly rely on manual inspections, on-site operations and experience accumulation, but these methods have problems such as long response time, missing information, and low risk prediction capabilities, which makes the ability to prevent and respond to complex power safety accidents relatively low. Summary of the invention

[0004] In view of this, the present invention provides a detection method, device and storage medium for power safety production, so as to improve the ability to prevent and respond to complex power safety accidents.

[0005] A first aspect of the present invention provides a method for detecting safe production of electric power, comprising:

[0006] Obtaining log data from historical monitoring of power production behavior, analyzing the time series operation parameter changes and normalized operation evolution of the log data, and obtaining normalized operation status characteristics of power equipment;

[0007] Based on the log data, the operating status characteristics are subjected to a safety production structure evolution to construct a knowledge graph of power safety production;

[0008] Acquire video data for real-time monitoring of power production behavior, perform frame-by-frame dynamic optical flow tracking on the video data, and obtain production behavior characteristics of the dynamic production behavior of operators;

[0009] Performing dynamic behavior visual recognition on the production behavior characteristics and performing dynamic operation process characteristic analysis to generate dynamic operation process characteristic data of the operator;

[0010] Based on the power safety knowledge graph, a quantitative assessment of production safety risks is performed on the dynamic operation process characteristic data of the operator to generate a quantitative value of production behavior risk;

[0011] Emergency response decisions are made based on the quantified value of the production behavior risk to optimize the power production safety knowledge graph.

[0012] A second aspect of the present invention provides a detection device for safe production of electric power, comprising:

[0013] An operation status feature extraction module is used to obtain log data from historical monitoring of power production behavior, analyze the time series operation parameter changes and normalized operation evolution of the log data, and obtain the normalized operation status features of the power equipment;

[0014] A knowledge graph construction module, used for performing safety production structure evolution on the operating status characteristics based on the log data to construct a power safety production knowledge graph;

[0015] The production behavior feature extraction module is used to obtain video data for real-time monitoring of power production behavior, perform frame-by-frame dynamic optical flow tracking on the video data, and obtain production behavior features of the dynamic production behavior of operators;

[0016] An operation process characteristic data generation module, used for performing dynamic behavior visual recognition on the production behavior characteristics and performing dynamic operation process characteristic analysis to generate dynamic operation process characteristic data of the operator;

[0017] A risk quantification value calculation module, used to perform a safety production risk quantification assessment on the dynamic operation process characteristic data of the operator based on the power safety knowledge graph to generate a production behavior risk quantification value;

[0018] The knowledge graph optimization module is used to make emergency response decisions based on the quantified value of the production behavior risk so as to optimize the power production safety knowledge graph.

[0019] A third aspect of the present invention provides an electronic device, the electronic device comprising:

[0020] at least one processor; and

[0021] a memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the detection method for safe power production as described in the first aspect above.

[0023] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for detecting safe production of electric power as described in the first aspect above.

[0024] A fifth aspect of the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the detection method for safe power production as described in the first aspect above.

[0025] In this embodiment, the log data of historical monitoring of power production behavior is obtained, and the log data is analyzed for time series operation parameter changes and normalized operation evolution to obtain the normalized operation status characteristics of the power equipment; the operation status characteristics are evolved for safe production structure based on the log data to construct the knowledge graph of safe production of power; the video data of real-time monitoring of power production behavior is obtained, and the video data is tracked for frame-by-frame dynamic optical flow to obtain the production behavior characteristics of the dynamic production behavior of the operator; the production behavior characteristics are visually identified for dynamic behavior and analyzed for dynamic operation process characteristics to generate the dynamic operation process characteristic data of the operator; the dynamic operation process characteristic data of the operator is quantitatively evaluated for safe production risk based on the power safety knowledge graph to generate the quantitative value of production behavior risk; emergency response decisions are made based on the quantitative value of production behavior risk to optimize the knowledge graph of safe production of power. This embodiment continuously optimizes the knowledge graph of safe production of power based on real-time production business iteration, so that the knowledge graph of safe production of power continuously adapts to the development of business, accumulates a large amount of information, accurately predicts potential production risks, and quickly provides intelligent production decisions, which helps to improve the efficiency and accuracy of accident response, and improve the ability to prevent and respond to complex power safety accidents.

[0026] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 This is a flow chart of a method for detecting safe production of electric power provided in Embodiment 1 of the present invention.

[0029] Figure 2 It is a structural schematic diagram of a detection device for electric power production safety provided in Embodiment 2 of the present invention.

[0030] Figure 3 It is a structural schematic diagram of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can cover sequential implementations other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] Embodiment 1

[0034] See also Figure 1 , shows a flow chart of a method for detecting power safety production provided by the first embodiment of the present invention, the method can be executed by a detection device for power safety production, the detection device for power safety production can be implemented in the form of hardware and / or software, and the detection device for power safety production can be configured in an electronic device. Figure 1 As shown, the method includes:

[0035] Step 101: Obtain log data of historical monitoring of power production behavior, analyze the time series operation parameter changes and normalized operation evolution of the log data, and obtain normalized operation status characteristics of the power equipment.

[0036] In this embodiment, log data of historical monitoring of power production behavior can be obtained from the server, and the log data can be analyzed for time series operating parameter changes and normalized operating evolution to obtain normalized operating status characteristics of the power equipment.

[0037] Generally, log data is stored in a database or data warehouse of a power production management system, and is extracted therefrom using SQL (Structured Query Language) queries or data extraction tools (such as Python's Pandas library).

[0038] The log data includes information such as timestamp, device ID, operating parameters (such as voltage, power, temperature, etc.). If there is other relevant data (such as equipment maintenance records, fault records, etc.), the other relevant data is integrated with the original log data to form a comprehensive log data.

[0039] Clean the log data, remove duplicate records, process missing values, and format the data into a unified structure (such as CSV or DataFrame format) to ensure the integrity and consistency of the log data. Use data processing tools (such as Pandas) to load the cleaned log data, set the timestamp as the index of the log data, and ensure that the log data is arranged in chronological order to facilitate time series analysis.

[0040] Select appropriate time series analysis methods, such as time series decomposition, moving average, or autoregressive integrated moving average (ARIMA) model, etc., to analyze the changing trends of operating parameters, perform time series change analysis on each operating parameter, calculate the statistical characteristics of the parameters such as mean, variance, maximum, minimum, etc., and identify the changing patterns and trends.

[0041] Select a suitable normalization method, such as Z-score normalization or Min-Max scaling, to standardize the operating parameter change data, perform normalization on the operating parameter change data, ensure that the data of different parameters are compared under the same dimension, extract the operating status characteristics of the power equipment from the normalized data, such as the normal operating range, operating stability, etc., and further analyze the extracted operating status characteristics to identify the normalized operating status characteristics of the power equipment, such as normal operating conditions, abnormal fluctuations, etc.

[0042] The operating status characteristics are organized into data sets and stored in databases or files to provide a basis for subsequent monitoring and management.

[0043] In one embodiment of the present invention, step 101 may include the following steps:

[0044] Step 1011: Filter abnormal outlier data points from the log data.

[0045] In actual applications, the log data of historical monitoring of power production behavior provides a lot of behavioral information, which can be used to analyze the evolution and change trends of power production business.

[0046] Using machine learning methods (such as Isolation Forest and DBSCAN) to detect abnormal outlier data points and filter out abnormal outlier data points from log data can help reduce interference and make the data more representative, thereby effectively improving the quality of log data and ensuring the accuracy and reliability of subsequent analysis.

[0047] Step 1012: If filtering is completed, the log data is segmented into time series windows to obtain multiple log segments.

[0048] Defining the size of the time series window (such as 1 hour, 1 day), as well as the overlapping strategy (such as no overlap or partial overlap), using programming tools (such as Python's Pandas library) to split the log data into time series windows, and saving the log fragments in each time series window as independent data structures (such as DataFrame) will help you understand the characteristics and dynamic changes of the log data in more detail, better observe the time dependency between multiple log fragments, and discover periodic changes or trends.

[0049] Step 1013: Identify a business device entity from multiple log fragments.

[0050] Use regular expressions, string matching or machine learning models (such as text classifiers) to identify the equipment information in each log segment, add tags to the equipment data in each time window segment, record the equipment type, ID and related features, so as to identify business equipment entities in multiple log segments, and save the marked business equipment entities to a new data structure for subsequent analysis. This helps to associate log data with specific power production equipment to achieve more detailed analysis and monitoring. Marking the business equipment entities of power equipment provides a basis for subsequent parameter analysis.

[0051] Step 1014: Perform time-series operation parameter change analysis on multiple business device entities to generate operation parameter change data.

[0052] Use time series analysis methods (such as differential method and sliding window analysis) to analyze the changes in the operating parameters of each business equipment entity, identify the rising, falling or fluctuating trend of the parameters, generate the operating parameter change data of each business equipment entity, including statistical indicators (such as mean value, standard deviation) and change trends (such as increase or decrease), save the operating parameter change data of each device to a new file or database, and form a result set for subsequent analysis.

[0053] Step 1015: Perform normalized operation evolution on the operation parameter change data to obtain normalized operation status characteristics of the power equipment.

[0054] Select an appropriate normalization method, such as Z-score standardization, Min-Max normalization or box plot method, to normalize the operating parameter change data of each business equipment entity. By analyzing the normalized operating parameter change data and extracting the normalized operating status characteristics of the power equipment, including the normal range, abnormal threshold, etc., the normalized operating status characteristics are saved as a report or data file to facilitate subsequent decision support and business optimization.

[0055] Step 102: Evolve the safety production structure of the operating status characteristics based on the log data to construct a knowledge graph of power safety production.

[0056] In practical applications, based on log data, the operating status characteristics of power equipment can be used to evolve the safety production structure, thereby constructing a knowledge graph, which is recorded as the power safety production knowledge graph.

[0057] Among them, knowledge graph, as a way of building relationships between entities, can realize the integration and analysis of multi-dimensional information. In the field of power safety production, knowledge graph can not only systematically integrate various types of power equipment, production links, management specifications and other information, but also reveal potential safety hazards and predict failures or accidents through reasoning and association analysis, thereby providing accurate decision-making support for power safety production.

[0058] In one embodiment of the present invention, step 102 may include the following steps:

[0059] Step 1021: Perform job record analysis on multiple log segments to extract job record data.

[0060] In this embodiment, the job records of multiple log segments can be analyzed according to various tags in the log segments to understand the specific work conditions and operation records in each log segment. Extracting the job record data of multiple log segments helps to track the workflow and identify potential security risks and problems.

[0061] Ensure that multiple log fragments have been cleaned and structured to contain necessary fields (such as timestamp, job type, equipment ID, etc.), use programming languages ​​(such as Python) and data processing libraries (such as Pandas) to load log fragments one by one, extract data related to specific power equipment or job types in the job records, format the data, ensure that each record contains all necessary information, and save the extracted job record data to a new structured file (such as CSV or database) in preparation for subsequent analysis.

[0062] Step 1022: Perform a safety production behavior analysis on the operation record data to extract safety production behavior data.

[0063] In this embodiment, key features of safe production behaviors, such as operation records, maintenance records, and exception handling records, are determined, and data analysis tools (such as Python's Scikit-learn) are used to analyze the job record data. Behaviors related to safe production are identified based on the key features, and safe production behavior labels are added to each job record to form a data set of safe production behaviors. The extracted safe production behavior data is saved in a new data structure (such as DataFrame) to characterize the patterns and characteristics of safe production behaviors, which helps to evaluate the operators' compliance with behavioral codes and safety awareness.

[0064] Step 1023: extracting behavior timestamps and business processing personnel data from the production safety behavior data.

[0065] In this embodiment, behavior timestamps (start and end timestamps of the job) and business processing personnel data (such as operator ID, role, etc.) can be extracted from the production safety behavior data, and the timestamps and business processing personnel information can be integrated into a new data set, which helps to record the operation time and behavior information of the operators and facilitates subsequent analysis and behavior feature modeling.

[0066] Step 1024: Perform multi-period production behavior feature analysis based on the behavior timestamp and business processing personnel data to generate production behavior features for different time periods.

[0067] Ensure that each record contains behavior identification, timestamp and business processing personnel data. Divide the records into multiple time periods (such as by hour, by shift) according to the timestamp, generate an independent data set for each time period, and use statistical analysis methods (such as mean, frequency distribution) to analyze the production behavior characteristics of each time period. Identify the behavior patterns in different time periods, which helps to understand the changes and trends in production behavior patterns and identify potential safety hazards and improvement opportunities.

[0068] Step 1025: Based on the production behavior characteristics in different time periods, the operating status characteristics are evolved into a safe production structure to construct a knowledge graph of power safety production.

[0069] In the specific implementation, Python's Matplotlib or Seaborn is used for data visualization. According to the production behavior characteristics of different periods, the data related to safe production (such as equipment status, operation behavior, and personnel behavior) are integrated to build a knowledge graph of power safety production. The nodes (such as equipment, personnel, and behavior) and edges (such as operation relationships and influence relationships) in the knowledge graph are defined to build the structure of the knowledge graph. The integrated data is imported and a knowledge graph is generated using a knowledge graph construction tool (such as Neo4j and GraphDB). This is recorded as a knowledge graph of power safety production, which helps to systematically record safety information and support risk assessment and decision making.

[0070] In one embodiment of the present invention, step 1025 may further include the following steps:

[0071] Step 10251: Identify key behavior time nodes based on production behavior characteristics in different time periods.

[0072] In this embodiment, it is ensured that the production behavior characteristics of different time periods have been cleaned and structured, including timestamps and behavior categories, etc., and statistical analysis methods (such as cluster analysis) or machine learning algorithms (such as decision trees) are used to identify key behavior time nodes. For example, key nodes are identified by analyzing the frequency and importance of behaviors, which helps to determine the important time points in the production process. Extracting key behavior time nodes helps focus on key events and provides a basis for subsequent analysis.

[0073] Step 10252: Set pre- and post-trigger conditions for key behavior time nodes.

[0074] In this embodiment, pre- and post-trigger conditions are set for key behavior time nodes, and the pre- and post-trigger conditions include pre-trigger conditions (i.e., behaviors before key behavior time nodes) and post-trigger conditions (i.e., behaviors after key behavior time nodes). Analyzing the pre- and post-conditions of key time nodes helps to understand the background and triggering conditions of key events in these key behavior time nodes, and obtaining the pre- and post-trigger conditions of key behavior time nodes helps to establish the behavior triggering logic.

[0075] Step 10253: Based on the pre- and post-trigger conditions of the key behavior nodes, the production behavior characteristics of different time periods are subjected to up- and down-operation behavior dependency analysis to obtain production behavior time series dependency data.

[0076] Use association rule mining algorithms (such as Apriori or FP-Growth) to extract pre- and post-conditions, and use Python's mlxtend library to implement it. Based on the pre- and post-conditions of key behavior time nodes, analyze the dependencies between context operations. Use graph theory methods to construct a directed graph to represent the dependencies between operations, record the dependencies, and generate production behavior timing dependency data. Production behavior timing dependency data is the timing dependency of production behaviors, which can reveal the association and influence between operational behaviors and help understand the order and logic of operational behaviors.

[0077] Step 10254: Mining the safety production standard logic based on the production behavior timing dependency data.

[0078] Ensure that each record contains relevant key behavior time nodes and production behavior timing dependency data, use classification algorithms such as logistic regression or decision trees to analyze each record, extract logical rules that meet production safety standards, record them as production safety standard logic, and save the production safety standard logic as a rule library, which will help standardize operational behaviors, establish standard operating procedures for production safety, and improve work efficiency and safety.

[0079] Step 10255: Perform topological structure association analysis on production behavior characteristics in different time periods according to the operating status characteristics to extract topological structure data.

[0080] In this embodiment, a topological structure association analysis is performed on the production behavior characteristics of different time periods according to the operating status characteristics, so as to construct topological structure data. The topological structure data is a normalized topological structure association between operating behavior and power equipment. Each node represents a device or behavior, and each edge represents the relationship between the two. It reveals the relationship between the operating behavior and the operating status of the power equipment, which helps to understand the impact and effect of the operating behavior on the operation of the power equipment.

[0081] Step 10256: Evolve the safety production graph structure of the topological structure data based on the safety production standard logic to construct a power safety production knowledge graph.

[0082] In this embodiment, a graph analysis tool (such as NetworkX) is used to analyze the topology data, extract the association data between behavior and equipment, combine the safety production standard logic and the behavior-equipment normalization topology data, build a preliminary power safety production knowledge graph, and use a graph database (such as Neo4j) to import the standard logic and topology association data to build a comprehensive power safety production knowledge graph.

[0083] Step 103: Acquire video data for real-time monitoring of power production behavior, perform frame-by-frame dynamic optical flow tracking on the video data, and obtain production behavior characteristics of the dynamic production behavior of the operators.

[0084] In practical applications, real-time video data for monitoring power production behavior can be obtained from cameras deployed in the power production area. The deployment location of the camera should allow the camera to cover the power production area. Configure the network connection of the camera, use protocols such as RTSP (Real-time Streaming Protocol) to transmit video data in real time in the form of streams, and store the video data on a local server or cloud storage for subsequent analysis. Regularly check the quality of the video data to ensure that the picture is clear and free of interference, so as to facilitate subsequent image processing and analysis. Select a suitable image processing library (such as OpenCV, FFmpeg) to process the video data, supporting frame-by-frame analysis and optical flow calculation.

[0085] The video data contains multiple frames of image data. Use OpenCV and other tools to extract the video frame by frame to generate an image sequence. Preprocess the image data of each frame in the video data, including resizing, graying, and removing noise, to improve the accuracy of subsequent optical flow calculations.

[0086] Select a suitable optical flow calculation algorithm, such as the Lucas-Kanade method or the Farneback method, to calculate the motion vector between consecutive frames, use OpenCV's calcOpticalFlowFarneback or calcOpticalFlowPyrLK function to perform dynamic optical flow calculation, extract the motion optical flow information, draw the motion trajectory of the operator based on the calculated optical flow information, superimpose arrows or trajectory lines on the original image data to display the motion direction and speed, and extract the operator's dynamic behavior trajectory features from the optical flow data, such as motion path, speed, acceleration, etc. These features are used for subsequent behavior analysis to identify the production behavior characteristics of the operator's dynamic production behavior.

[0087] In one embodiment of the present invention, step 103 may include the following steps:

[0088] Step 1031: Filter the video data.

[0089] In this embodiment, a Gaussian filtering algorithm, a bilateral filtering algorithm, a spatial filtering algorithm, or the like may be used to filter each frame of image data in the video data to improve the clarity of the image data, thereby improving the quality of the video data.

[0090] Step 1032: If the filtering process is completed, the video data is subjected to visual recognition of the operator, and the image frame where the operator is located is segmented.

[0091] In this embodiment, target detection algorithms such as YOLO and SSD can be used to target workers and perform visual identification of workers on video data, thereby segmenting the image frame where the workers are located, saving each detected image frame as an independent image or area, and recording the position, timestamp and related information of each image frame to form structured data, which is helpful to accurately locate and track workers, providing basic information for subsequent analysis.

[0092] Step 1033: perform frame-by-frame dynamic optical flow tracking on the image frame to extract motion vector change data.

[0093] In this embodiment, an optical flow algorithm (such as Lucas-Kanade or Farneback) can be used to perform dynamic optical flow tracking on the image frame frame by frame, extract motion vector change data, capture the movement trajectory and changes of the operator, and help understand the dynamic behavior of the operator.

[0094] Step 1034: Perform dynamic behavior trajectory analysis on the motion vector change data to obtain production behavior characteristics of the operator's dynamic behavior.

[0095] In this embodiment, data analysis tools (such as Python's Pandas and NumPy) are used to perform dynamic behavior trajectory analysis on the motion vector change data, and the dynamic behavior trajectory characteristics of the operators are extracted from the motion vector data, including speed, acceleration, and movement direction, etc., to identify the behavior patterns of the operators (such as walking, pausing, operating, etc.), thereby obtaining the production behavior characteristics of the dynamic behavior of the operators, which is helpful to evaluate the standardization and efficiency of the operation and discover potential safety issues.

[0096] Step 104: Perform dynamic behavior visual recognition on the production behavior characteristics and perform dynamic operation process characteristic analysis to generate dynamic operation process characteristic data of the operator.

[0097] In practical applications, dynamic behavior visual recognition is performed on production behavior characteristics and dynamic operation process characteristics analysis is performed to generate dynamic operation process characteristic data of operators.

[0098] Use data processing libraries (such as Pandas) to load previously extracted production behavior features to ensure data availability, clean production behavior features, remove outliers and perform normalization to facilitate subsequent model training and analysis, select a suitable deep learning model for dynamic behavior recognition, for example, convolutional neural network (CNN) combined with long short-term memory network (LSTM) to process time series data, etc., apply the model to the processing of real-time production behavior features, identify the dynamic behavior of operators, generate corresponding behavior labels, analyze the operating procedures of operators based on the identified dynamic behavior labels, and extract corresponding dynamic operating procedure feature data, such as operation type, operation duration, operation frequency, etc.

[0099] Afterwards, a clustering algorithm (such as K-Means or DBSCAN) is used to classify the dynamic operation process feature data, identify different process patterns and behavioral characteristics, conduct statistical analysis on the frequency and duration of different operation processes, and generate relevant statistical charts for visualization and subsequent decision support.

[0100] The dynamic operation process characteristic data is organized into a structured data set, including information such as operation type, duration, frequency, etc., and marked with corresponding behavior labels. The organized dynamic operation process characteristic data is saved in the database or file system for subsequent query and analysis, and analysis reports are written to summarize the dynamic operation process characteristic data of the operators and provide them to the management for decision support.

[0101] In one embodiment of the present invention, step 104 may include the following steps:

[0102] Step 1041: Perform dynamic behavior visual recognition on production behavior characteristics to obtain power production behavior data of operators.

[0103] Use deep learning models (such as LSTM, CNN-RNN or other time series models) to perform dynamic behavior visual recognition of production behavior characteristics, input the dynamic behavior trajectory feature data of operators into the selected deep learning model for inference, identify the specific behaviors of operators in the power production process (such as operating equipment, checking equipment, etc.), and store the identified power production behavior data (including behavior type, timestamp, etc.) in a structured database

[0104] Step 1042: Calculate the dynamic behavior duration of the power production behavior data.

[0105] In this embodiment, the power production behavior data of the operators is calculated through dynamic behavior duration, that is, the duration of different behaviors of the operators in the production process is quantified to obtain the dynamic behavior duration, which helps to understand the time spent by the operators in various tasks and helps to evaluate work efficiency and standard operating procedures.

[0106] Specifically, the start time and end time are extracted from the operator's power production behavior data. For each behavior, its duration is calculated as the dynamic behavior duration. Dynamic behavior duration = end time - start time. The dynamic behavior duration is associated with the behavior data and saved as structured data.

[0107] Step 1043: quantify the behavior interval of the power production behavior data to obtain the production behavior interval value.

[0108] The behavior intervals of operators' power production behavior data are quantified, the time intervals between different production behaviors are measured, and the production behavior interval values ​​are obtained, which helps to determine the transition time between different tasks of operators and provide a basis for optimizing the operation process and reducing waiting time.

[0109] The so-called behavior interval usually refers to the time difference between two consecutive behaviors. The behavior data of the operators is processed, the time difference between each two consecutive behaviors is calculated, and the production behavior interval value is obtained.

[0110] Step 1044: Calculate the sequential characteristics of the operator's behavior based on the production behavior interval value.

[0111] Selecting appropriate sequence analysis methods, such as Markov models or sequence pattern mining, and conducting sequential analysis of multiple operating behaviors based on production behavior interval values ​​and behavior types will help reveal the sequential characteristics of operator behaviors and help understand the association and sequence between different operating behaviors, including the frequency and sequence relationship of behaviors, and provide a reference for formulating operating specifications and training.

[0112] Step 1045: Use the dynamic behavior duration and behavior sequence characteristics to construct the dynamic operation process characteristic data of the operator.

[0113] In this embodiment, a suitable analysis method (such as cluster analysis or principal component analysis) is selected to perform correlation analysis on the dynamic behavior duration and behavior sequence characteristics. The dynamic behavior duration and behavior sequence characteristics are combined to generate dynamic operation process characteristic data of the operators, which helps to identify the operation modes and behavior patterns of the operators and provide support for safe production management and risk control.

[0114] Step 105: Conduct a quantitative assessment of production safety risks on the dynamic operation process characteristic data of operators based on the power safety knowledge graph to generate a quantitative value of production behavior risk.

[0115] Based on the power safety knowledge graph, the dynamic operation process characteristic data of the operators are used to conduct a quantitative assessment of the production safety risks, thereby generating a quantitative value of the production behavior risk and intuitively displaying the risks of the production behavior.

[0116] Use data processing tools (such as Pandas) to load dynamic operation process characteristic data from the database or file, such as operation type, duration, frequency and other information, clean the dynamic operation process characteristic data, process missing values ​​and outliers, and standardize or normalize to ensure data quality and consistency.

[0117] According to the power safety knowledge graph, determine the risk assessment indicators related to the operating behavior, such as operation complexity, equipment status, historical failure rate, etc., select appropriate risk assessment models, such as the analytic hierarchy process (AHP), fuzzy logic model, machine learning model, etc., input the dynamic operation process characteristic data into the trained risk assessment model, calculate the risk score of each operation, and for each operation behavior, combine the risk scores of multiple evaluation indicators to calculate the quantitative value of its production behavior risk.

[0118] Considering the impact of different indicators on risk, the risk quantification value is normalized so that it is within a unified range to facilitate subsequent comparison and analysis. The calculated production behavior risk quantification value is integrated with the corresponding operation characteristic data to form a complete data set. The generated risk quantification value data is saved in a database or file to ensure the convenience of subsequent query and analysis. A detailed risk assessment report is generated to explain the production behavior risk quantification value of each operation and provide corresponding risk management suggestions to support decision-making.

[0119] In one embodiment of the present invention, step 105 may include the following steps:

[0120] Step 1051: Analyze the production safety operation rules on the dynamic operation process feature data based on the power safety knowledge graph to extract irregular operation behavior data.

[0121] In this embodiment, it is ensured that the dynamic operation process characteristic data of the operators in the power safety knowledge graph have been sorted and connected with the relevant safety operation rules. The association rule mining algorithm (such as Apriori or FP-Growth) is used to analyze the dynamic operation process characteristic data based on the power safety knowledge graph to perform safe production operation rule analysis, extract irregular operation behavior data, identify operation behaviors that do not comply with safety operation procedures or potential risks, record these irregular operation behavior data, and organize the identified safety hazard data into a report containing information such as hazard type, frequency of occurrence and consequences. This helps to identify violations of operation rules and improve the ability to identify potential safety risks.

[0122] Step 1052: Analyze the operation behavior data for potential safety hazards in production, and generate operation behavior potential safety hazards data.

[0123] In this embodiment, a risk propagation model is constructed using graph theory or network analysis methods to describe the relationship between different hidden dangers. The irregular operation behavior data is analyzed for production safety hazards, which helps to discover potential safety problems and behavioral anomalies, which are recorded as operation behavior safety hazard data, providing a basis for formulating improvement measures and strengthening training.

[0124] Step 1053: Mining the diffusible paths of security risks in the operational behavior security hazard data to extract the diffusible paths of risks.

[0125] In this embodiment, a path mining algorithm (such as the Dijkstra algorithm or the A* algorithm) is used to mine the diffusible paths of security risks in the operational behavior security hazard data, extract the diffusible paths of risks, and thus identify the potential paths for the propagation of security risks, which helps to understand the propagation mechanism and key influencing factors of security risks.

[0126] Step 1054: perform risk propagation prediction on the risk diffusible path to obtain safety risk propagation prediction data.

[0127] In this embodiment, a suitable risk propagation prediction model (such as an SIR model or an Agent-Based model) is selected, and modeling is performed according to historical data and hidden danger characteristics. Based on the constructed model, risk propagation is simulated for all risk diffusible paths to obtain safety risk propagation prediction data, which helps predict potential safety risk propagation situations, helps to identify safety issues in advance, identifies potential high-risk areas and key risk factors, and takes measures to reduce the possibility of risk occurrence.

[0128] Step 1055: Conduct a quantitative assessment of production safety risks on the safety risk propagation prediction data to generate a quantitative value of production behavior risk.

[0129] In practical applications, it is helpful to determine the indicators used for quantitative assessment of safety risks, such as the probability of risk occurrence, the degree of potential consequences, controllability, etc., to conduct quantitative assessment of production safety risks on the safety risk propagation prediction data, and to calculate the quantitative value of production behavior risk for each risk path, which will help to quantitatively assess the severity of potential risks and provide a basis for formulating risk management strategies.

[0130] Step 106: Make emergency response decisions based on the quantified value of production behavior risk to optimize the power production safety knowledge graph.

[0131] In practical applications, decisions can be made in the dimension of emergency response based on the quantitative value of production behavior risks, thereby optimizing the knowledge graph of power production safety.

[0132] Use data processing tools (such as Pandas) to load the production behavior risk quantification value data from the database or file, clean the production behavior risk quantification value data, process missing values ​​and outliers, and ensure the accuracy and consistency of the data for subsequent analysis.

[0133] Set risk thresholds based on industry standards and historical data to distinguish different levels of risk (such as low, medium, and high risk), and formulate corresponding emergency response strategies for different risk levels. For example, low risk: routine monitoring, stay vigilant; medium risk: strengthen monitoring,

[0134] Develop temporary operating procedures; High risk: Immediately launch the emergency plan and mobilize relevant resources. After the emergency response is implemented, record feedback data, including response effects, execution efficiency, and processing results, to evaluate the effectiveness of the emergency response strategy, analyze feedback data, identify deficiencies in the emergency response strategy, and optimize the emergency response strategy using statistical analysis or machine learning methods (such as regression analysis). Based on the feedback analysis results, adjust and optimize the emergency response strategy to improve the timeliness and effectiveness of future responses.

[0135] According to the optimized emergency response strategy, a real-time monitoring system is built. Data stream processing technology (such as Apache Kafka) is combined with intelligent decision-making models (such as reinforcement learning) to realize real-time emergency response decisions. According to the optimized emergency response strategy and decision results, the power safety production knowledge graph is updated, new nodes and edges are added to reflect the latest knowledge and experience, and the emergency response results and optimized strategy data are mapped to the power safety production knowledge graph to form a relationship between emergency response and safety production. The integrity and accuracy of the power safety production knowledge graph are regularly verified to ensure that it reflects the latest safety production knowledge, and it is updated and optimized when necessary.

[0136] In one embodiment of the present invention, step 106 may include the following steps:

[0137] Step 1061: Make an emergency response decision based on the quantified value of the production behavior risk to obtain a risk behavior emergency response strategy.

[0138] In this embodiment, the quantitative values ​​of production behavior risks are collected and organized to ensure that the data is complete and accurate. Methods such as decision trees, fuzzy logic, or expert systems are used to determine emergency response strategies based on the quantitative values ​​of production behavior risks. Corresponding response measures are set for different quantitative values ​​of production behavior risks. For example, high risk: immediate shutdown for inspection, medium risk: increase monitoring frequency, low risk: regular inspection, and the emergency response strategies are recorded as documents or database entries.

[0139] By presetting decision rules for application responses and making emergency response decisions based on the quantitative value of production behavior risk, it is helpful to formulate targeted emergency response strategies and obtain risk behavior emergency response strategies to improve the efficiency and accuracy of responding to emergencies and risk situations.

[0140] Step 1062: Perform iterative learning on the risk behavior emergency response strategy to obtain an iterative learning representation of the response strategy.

[0141] In this embodiment, corresponding operations are implemented according to the formulated emergency response strategy, the execution results and actual effects are recorded, and data related to the strategy execution are collected, including response time, effect evaluation, subsequent risk occurrence, etc. The execution results are analyzed using machine learning algorithms (such as Q-learning or reinforcement learning), the response strategy is adjusted and optimized, and an iterative learning representation of the response strategy is generated, which helps to continuously optimize the response strategy. The generation of iterative learning representation of the response strategy learns from experience and improves the intelligence level of emergency response.

[0142] Step 1063: Perform multi-scenario migration optimization on the iterative learning representation of the response strategy to obtain a migration optimized iterative learning representation.

[0143] In this embodiment, the effects and improvement suggestions of the emergency response strategy are recorded, different application scenarios are identified, the similarities and differences between these scenarios and existing strategies are analyzed, and transfer learning algorithms (such as domain adaptation, model fine-tuning, etc.) are applied to migrate the response strategy to the new scenario. The iterative learning representation of the response strategy is optimized for multi-scenario migration to obtain a migration-optimized iterative learning representation, which makes the emergency response more adaptable, helps to apply experience to different situations, and improves the quality and flexibility of decision-making.

[0144] Step 1064: Use transfer optimization iterative learning representation to make intelligent response decisions on the power safety production knowledge graph to optimize the power safety production knowledge graph.

[0145] In this embodiment, transfer optimization iterative learning representation is used to make intelligent response decisions on the power safety production knowledge graph, thereby optimizing the power safety production knowledge graph and improving the intelligence level of decision-making. Constructing the power safety production knowledge graph helps to integrate and apply various optimization measures and improve the efficiency and level of safety production management.

[0146] This process mainly involves the reconstruction of the power safety production knowledge graph, integrating the optimized emergency response strategies and learning results into the power safety production knowledge graph to form a more accurate decision support system, and combining graph analysis and machine learning techniques to optimize the intelligent response decision-making capabilities of the power safety production knowledge graph to ensure that emergency response strategies can be effectively implemented in new scenarios.

[0147] Set up a monitoring system to collect safety data in the power production process (such as equipment status, operation records, etc.) in real time, use the optimized safety production knowledge graph to analyze the real-time data, identify potential safety hazards and risks, and quickly execute emergency response decisions based on real-time analysis results to ensure the safety of the power production process, collect real-time analysis and emergency response results, and continuously update and optimize the power production safety knowledge graph and emergency response strategy.

[0148] In this embodiment, the log data of the historical monitoring of power production behavior is obtained, and the time series operation parameter change analysis and normal operation evolution of the log data are performed to obtain the normal operation status characteristics of the power equipment; the safe production structure evolution of the operation status characteristics is performed based on the log data to construct the power safety production knowledge map; the video data of the real-time monitoring of power production behavior is obtained, and the video data is subjected to frame-by-frame dynamic optical flow tracking to obtain the production behavior characteristics of the dynamic production behavior of the operator; the production behavior characteristics are subjected to dynamic behavior visual recognition and dynamic operation process characteristic analysis to generate the dynamic operation process characteristic data of the operator; the dynamic operation process characteristic data of the operator is subjected to a safe production risk quantitative assessment based on the power safety knowledge map to generate a production behavior risk quantitative value; emergency response decisions are made based on the production behavior risk quantitative value to optimize the power safety production knowledge map. This embodiment continuously optimizes the power safety production knowledge map based on the real-time production business iteration, so that the power safety production knowledge map continuously adapts to the development of the business, accumulates a large amount of information, accurately predicts potential production risks, and quickly provides intelligent production decisions, which helps to improve the efficiency and accuracy of accident response, and improve the prevention and response capabilities of complex power safety accidents.

[0149] Embodiment 2

[0150] See also Figure 2 , shows a schematic diagram of the structure of a detection device for power safety production provided by Embodiment 2 of the present invention. Figure 2 As shown, the device comprises:

[0151] The operation status feature extraction module 201 is used to obtain the log data of the historical monitoring of the power production behavior, analyze the time series operation parameter changes and the normalized operation evolution of the log data, and obtain the normalized operation status characteristics of the power equipment;

[0152] A knowledge graph construction module 202 is used to perform safety production structure evolution on the operating status characteristics based on the log data to construct a power safety production knowledge graph;

[0153] The production behavior feature extraction module 203 is used to obtain video data for real-time monitoring of power production behavior, and perform frame-by-frame dynamic optical flow tracking on the video data to obtain production behavior features of the dynamic production behavior of operators;

[0154] The operation process characteristic data generating module 204 is used to perform dynamic behavior visual recognition on the production behavior characteristics and perform dynamic operation process characteristic analysis to generate dynamic operation process characteristic data of the operator;

[0155] A risk quantification value calculation module 205 is used to perform a safety production risk quantification assessment on the dynamic operation process characteristic data of the operator based on the power safety knowledge graph to generate a production behavior risk quantification value;

[0156] The knowledge graph optimization module 206 is used to make emergency response decisions based on the quantified value of the production behavior risk to optimize the power production safety knowledge graph.

[0157] In one embodiment of the present invention, the operating status feature extraction module 201 is further used for:

[0158] filtering abnormal outlier data points from the log data;

[0159] If filtering is completed, the log data is segmented into time series windows to obtain multiple log segments;

[0160] identifying a business device entity from a plurality of the log fragments;

[0161] Performing time-series operation parameter change analysis on a plurality of the service equipment entities to generate operation parameter change data;

[0162] The operating parameter change data is subjected to normalized operating evolution to obtain normalized operating state characteristics of the power equipment.

[0163] In one embodiment of the present invention, the knowledge graph construction module 202 is also used to:

[0164] Performing operation record analysis on the plurality of log segments to extract operation record data;

[0165] Performing a production safety behavior analysis on the operation record data to extract production safety behavior data;

[0166] Extracting behavior timestamps and business processing personnel data from the safety production behavior data;

[0167] Perform multi-period production behavior feature analysis based on the behavior timestamp and the business processing personnel data to generate production behavior features for different time periods;

[0168] The safety production structure evolution of the operating status characteristics is performed based on the production behavior characteristics in different time periods to construct a knowledge graph of power safety production.

[0169] In one embodiment of the present invention, the knowledge graph construction module 202 is also used to:

[0170] Identifying key behavior time nodes for the production behavior characteristics in different time periods;

[0171] Setting pre- and post-trigger conditions for the key behavior time nodes;

[0172] Based on the pre- and post-trigger conditions of the key behavior nodes, up- and down-operation behavior dependency analysis is performed on the production behavior characteristics in different time periods to obtain production behavior time series dependency data;

[0173] The production behavior timing relies on data mining safety production standard logic;

[0174] Performing topological structure association analysis on the production behavior characteristics in different time periods according to the operating state characteristics to extract topological structure data;

[0175] Based on the safety production standard logic, the topological structure data is evolved into a safety production graph structure to construct a power safety production knowledge graph.

[0176] In one embodiment of the present invention, the production behavior feature extraction module 203 is further used for:

[0177] Performing filtering processing on the video data;

[0178] If the filtering process is completed, the video data is subjected to visual recognition of the operator, and the image frame where the operator is located is segmented;

[0179] Performing frame-by-frame dynamic optical flow tracking on the image frame to extract motion vector change data;

[0180] The motion vector change data is subjected to dynamic behavior trajectory analysis to obtain production behavior characteristics of the operator's dynamic behavior.

[0181] In one embodiment of the present invention, the operation process characteristic data generating module 204 is further used for:

[0182] Performing dynamic behavior visual recognition on the production behavior characteristics to obtain power production behavior data of the operator;

[0183] Calculating dynamic behavior duration for the power production behavior data;

[0184] quantifying the behavior interval of the power production behavior data to obtain a production behavior interval value;

[0185] Calculating the sequential characteristics of the operator's behavior based on the production behavior interval value;

[0186] The dynamic behavior duration and the behavior sequence characteristics are used to construct the dynamic operation process characteristic data of the operator.

[0187] In one embodiment of the present invention, the risk quantification value calculation module 205 is further used to:

[0188] Based on the power safety knowledge graph, the dynamic operation process feature data is analyzed for production safety operation rules to extract irregular operation behavior data;

[0189] Performing a production safety hazard analysis on the operation behavior data to generate operation behavior safety hazard data;

[0190] Mining the diffusible paths of security risks on the security potential hazard data of the operation behavior to extract the diffusible paths of risks;

[0191] Perform risk propagation prediction on the risk diffusible path to obtain safety risk propagation prediction data;

[0192] A quantitative assessment of production safety risks is performed on the safety risk propagation prediction data to generate a quantitative value of production behavior risk.

[0193] In one embodiment of the present invention, the knowledge graph optimization module 206 is further used to:

[0194] Make emergency response decisions based on the production behavior risk quantification value to obtain risk behavior emergency response strategies;

[0195] Performing iterative learning of strategy execution on the risk behavior emergency response strategy to obtain an iterative learning representation of the response strategy;

[0196] Performing multi-scenario migration optimization on the iterative learning representation of the response strategy to obtain a migration optimized iterative learning representation;

[0197] The migration optimization iterative learning representation is used to make intelligent response decisions on the electric power production safety knowledge graph to optimize the electric power production safety knowledge graph.

[0198] The detection device for safe production of electric power provided in the embodiment of the present invention can execute the detection method for safe production of electric power provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the detection method for safe production of electric power.

[0199] Embodiment 3

[0200] See also Figure 3, showing a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0201] like Figure 3 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0202] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0203] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the detection method for safe production of electric power.

[0204] In some embodiments, the detection method for safe production of electric power may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the detection method for safe production of electric power described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the detection method for safe production of electric power in any other appropriate manner (e.g., by means of firmware).

[0205] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0206] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0207] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0208] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0209] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0210] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0211] Embodiment 4

[0212] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the detection method for safe power production provided by any embodiment of the present invention is implemented.

[0213] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0214] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0215] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A detection method for power production safety, characterized in that: include: Obtaining log data from historical monitoring of power production behavior, analyzing the time series operation parameter changes and normalized operation evolution of the log data, and obtaining normalized operation status characteristics of power equipment; Based on the log data, the operating status characteristics are subjected to a safety production structure evolution to construct a knowledge graph of power safety production; Acquire video data for real-time monitoring of power production behavior, perform frame-by-frame dynamic optical flow tracking on the video data, and obtain production behavior characteristics of the dynamic production behavior of operators; Performing dynamic behavior visual recognition on the production behavior characteristics and performing dynamic operation process characteristic analysis to generate dynamic operation process characteristic data of the operator; Based on the power safety knowledge graph, a quantitative assessment of production safety risks is performed on the dynamic operation process characteristic data of the operator to generate a quantitative value of production behavior risk; Emergency response decisions are made based on the quantified value of the production behavior risk to optimize the power production safety knowledge graph.

2. The method according to claim 1, characterized in that The time series operation parameter change analysis and normalized operation evolution of the log data are performed to obtain the normalized operation status characteristics of the power equipment, including: filtering abnormal outlier data points from the log data; If filtering is completed, the log data is segmented into time series windows to obtain multiple log segments; identifying a business device entity from a plurality of the log fragments; Performing time-series operation parameter change analysis on a plurality of the service equipment entities to generate operation parameter change data; The operating parameter change data is subjected to normalized operating evolution to obtain normalized operating state characteristics of the power equipment.

3. The method according to claim 2, characterized in that The step of performing a safety production structure evolution on the operating status characteristics based on the log data to construct a power safety production knowledge graph includes: Performing operation record analysis on the plurality of log segments to extract operation record data; Performing a production safety behavior analysis on the operation record data to extract production safety behavior data; Extracting behavior timestamps and business processing personnel data from the safety production behavior data; Perform multi-period production behavior feature analysis based on the behavior timestamp and the business processing personnel data to generate production behavior features for different time periods; The safety production structure evolution of the operating status characteristics is performed based on the production behavior characteristics in different time periods to construct a knowledge graph of power safety production.

4. The method according to claim 3, characterized in that The step of performing a safety production structure evolution on the operation status characteristics based on the production behavior characteristics in different time periods to construct a knowledge graph of power safety production includes: Identifying key behavior time nodes for the production behavior characteristics in different time periods; Setting pre- and post-trigger conditions for the key behavior time nodes; Based on the pre- and post-trigger conditions of the key behavior nodes, up- and down-operation behavior dependency analysis is performed on the production behavior characteristics in different time periods to obtain production behavior time series dependency data; The production behavior timing depends on data mining safety production standard logic; Performing topological structure association analysis on the production behavior characteristics in different time periods according to the operating state characteristics to extract topological structure data; Based on the safety production standard logic, the topological structure data is evolved into a safety production graph structure to construct a power safety production knowledge graph.

5. The method according to any one of claims 1 to 4, characterized in that The step of performing frame-by-frame dynamic optical flow tracking on the video data to obtain production behavior characteristics of the dynamic production behavior of the operator includes: Performing filtering processing on the video data; If the filtering process is completed, the video data is subjected to visual recognition of the operator, and the image frame where the operator is located is segmented; Performing frame-by-frame dynamic optical flow tracking on the image frame to extract motion vector change data; The dynamic behavior trajectory analysis is performed on the motion vector change data to obtain the production behavior characteristics of the operator's dynamic behavior.

6. The method according to any one of claims 1 to 4, characterized in that The dynamic behavior visual recognition of the production behavior characteristics and the dynamic operation process characteristic analysis to generate the dynamic operation process characteristic data of the operator include: Performing dynamic behavior visual recognition on the production behavior characteristics to obtain power production behavior data of the operator; Calculating dynamic behavior duration for the power production behavior data; quantifying the behavior interval of the power production behavior data to obtain a production behavior interval value; Calculating the sequential characteristics of the operator's behavior based on the production behavior interval value; The dynamic behavior duration and the behavior sequence characteristics are used to construct the dynamic operation process characteristic data of the operator.

7. The method according to any one of claims 1 to 4, characterized in that The step of performing a quantitative assessment of production safety risks on the dynamic operation process characteristic data of the operator based on the power safety knowledge graph to generate a quantitative value of production behavior risk includes: Based on the power safety knowledge graph, the dynamic operation process feature data is analyzed for production safety operation rules to extract irregular operation behavior data; Performing a production safety hazard analysis on the operation behavior data to generate operation behavior safety hazard data; Mining the diffusible paths of security risks on the security potential hazard data of the operation behavior to extract the diffusible paths of risks; Perform risk propagation prediction on the risk diffusible path to obtain safety risk propagation prediction data; A quantitative assessment of production safety risks is performed on the safety risk propagation prediction data to generate a quantitative value of production behavior risk.

8. The method according to any one of claims 1 to 4, characterized in that The emergency response decision is made based on the production behavior risk quantification value to optimize the power safety production knowledge graph, including: Make emergency response decisions based on the production behavior risk quantification value to obtain risk behavior emergency response strategies; Performing iterative learning of strategy execution on the risk behavior emergency response strategy to obtain an iterative learning representation of the response strategy; Performing multi-scenario migration optimization on the iterative learning representation of the response strategy to obtain a migration optimized iterative learning representation; The migration optimization iterative learning representation is used to make intelligent response decisions on the electric power production safety knowledge graph to optimize the electric power production safety knowledge graph.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the detection method for safe power production according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the detection method for safe power production according to any one of claims 1 to 8 is implemented.

Images