Risk state sensing method and device for power grid safety production operation and equipment

Abstract

The invention discloses a risk state sensing method and device for power grid safety production operation and equipment, and relates to the technical field of power grid operation management. The method comprises the following steps: acquiring power grid production operation data in real time; processing the power grid production operation data based on the feature extraction model to obtain a comprehensive feature vector; processing the comprehensive feature vector based on a BiLSTM model to obtain a dynamic risk feature sequence; calculating an attention score corresponding to the risk feature of each time point in the dynamic risk feature sequence, weighting the corresponding risk feature based on the attention score of each time point, and generating a comprehensive risk perception feature vector; and determining a real-time risk state of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector. Potential risks in the power grid production scene can be recognized in time, the risk state of power grid production operation and equipment is comprehensively evaluated, and accurate perception of the global risk state of the power grid production scene is achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of power grid operation management, and in particular, to a method and device for perceiving the risk status of power grid safety production operations and equipment. Background Art

[0002] As the scale of power grids expands and their complexity increases, power grid production operations involve the operation and maintenance of a large number of equipment and personnel management, so the safety of power grid production operations faces multiple challenges.

[0003] In related technologies, power grid safety management systems are usually used for equipment operation and maintenance and personnel management. However, when dealing with complex operating scenarios, power grid safety management systems have the following limitations: 1. Risk identification is not timely; 2. It is unable to comprehensively assess the risk status of key operations and equipment in power grid safety production.

[0004] Therefore, how to timely identify potential risks in power grid production scenarios, comprehensively assess the risk status of power grid production operations and equipment, and achieve accurate perception of the global risk status of power grid production scenarios is an urgent problem to be solved. Summary of the invention

[0005] The present application provides a method and device for perceiving the risk status of power grid production safety operations and equipment, which can timely identify potential risks in power grid production scenarios, and can comprehensively assess the risk status of power grid production operations and equipment, thereby achieving accurate perception of the global risk status of power grid production scenarios.

[0006] In a first aspect, an embodiment of the present application provides a method for perceiving risk status of power grid safety production operations and equipment, the method comprising:

[0007] Real-time acquisition of power grid production operation data in the power grid operation management system; power grid production operation data includes: operation plan data, operator task allocation data, operation environment condition data and equipment operation and maintenance data;

[0008] The power grid production operation data is processed based on the feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0009] The comprehensive feature vector is processed based on the Bi-directional Long Short-Term Memory (BiLSTM) model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of power grid production operation data at different time points;

[0010] Calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, weight the corresponding risk feature based on the attention score at each time point, and generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0011] Based on the comprehensive risk perception feature vector, the real-time risk status of the power grid production operation data at the current moment is determined.

[0012] In a second aspect, the embodiment of the present application further provides a risk status perception device for power grid safety production operations and equipment, the device comprising:

[0013] An acquisition module is used to acquire power grid production operation data in the power grid operation management system in real time; the power grid production operation data includes: operation plan data, operation personnel task allocation data, operation environment condition data and equipment operation and maintenance data;

[0014] A first processing module is used to process the power grid production operation data based on a feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0015] The second processing module is used to process the comprehensive feature vector based on a bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of the power grid production operation data at different time points;

[0016] A calculation module, used to calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, weight the corresponding risk feature based on the attention score at each time point, and generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0017] A determination module is used to determine the real-time risk status of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0019] one or more processors;

[0020] a memory for storing one or more programs,

[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement the risk status perception method for power grid safety production operations and equipment described in any embodiment of the present application.

[0022] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for perceiving the risk status of power grid safety production operations and equipment as described in any embodiment of the present application is implemented.

[0023] The embodiment of the present application proposes a method and device for perceiving the risk status of power grid production safety operations and equipment, which obtains power grid production operation data in a power grid operation management system in real time; the power grid production operation data includes: operation plan data, operator task allocation data, operation environment condition data and equipment operation and maintenance data; the power grid production operation data is processed based on a feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; the comprehensive feature vector is processed based on a bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of power grid production operation data at different time points; the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence is calculated, and the corresponding risk feature is weighted based on the attention score at each time point to generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; based on the comprehensive risk perception feature vector, the real-time risk status of the power grid production operation data at the current moment is determined. That is to say, in the technical solution of the present application, a real-time dynamic risk status identification framework is constructed through a feature extraction model and a BiLSTM model. The feature extraction model uniformly characterizes the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; the BiLSTM model captures the time series characteristics of these multi-dimensional data, and realizes the accurate perception of the task execution status and the dynamic changes of the environment, so that potential risk factors can be identified in real time, and potential risks in the power grid production scene can be identified in time, which significantly improves the timeliness of emergency response. The attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence is calculated, and the corresponding risk feature is weighted based on the attention score at each time point, the contribution of the key features is strengthened, and a comprehensive risk perception feature vector is generated to comprehensively characterize the relationship between multi-dimensional data. Through the comprehensive risk perception vector, risk factors such as operation plan data, operator task allocation data, operation environment condition data and equipment operation and maintenance data are uniformly quantified and integrated for analysis, which overcomes the problem of data isolation in traditional technologies and realizes accurate perception and evaluation of the global risk status of the power grid production scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 One of the flow charts of the method for sensing the risk status of power grid safety production operations and equipment provided in one embodiment of the present application;

[0025] Figure 2 A second flow chart of a method for sensing the risk status of power grid safety production operations and equipment provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of the structure of a risk status perception device for power grid safety production operations and equipment provided in one embodiment of the present application;

[0027] Figure 4 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0028] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only the parts related to the present application, rather than all structures, are shown in the accompanying drawings.

[0029] First, the relevant knowledge involved in the embodiments of the present application is introduced.

[0030] As the scale and complexity of power grids increase, the safety of power grid operations faces multiple challenges. Power grid production operations involve a large number of planning tasks, equipment maintenance and operation management, which need to be organized efficiently while ensuring the safety of operators and equipment.

[0031] However, the existing power grid security management system has the following limitations when dealing with complex operation scenarios:

[0032] 1. Untimely risk identification: Traditional methods rely on static rules or historical experience and cannot capture potential risk factors in real time in the dynamic interaction of work tasks, changes in personnel status and environmental conditions, and equipment operation and maintenance data. This lag can easily lead to untimely responses to emergencies.

[0033] 2. Multi-dimensional data is not fully utilized: Power grid operations involve a large amount of multi-dimensional data, including operation plans, personnel capabilities, environmental conditions, and equipment operation and maintenance records. However, existing systems usually analyze these data in isolation, failing to achieve the integration and comprehensive evaluation of multi-source data, and thus failing to fully evaluate risk status.

[0034] In response to the above-mentioned problems, the present application provides a method and device for perceiving the risk status of power grid production safety operations and equipment, which can timely identify potential risks in power grid production scenarios, and comprehensively assess the risk status of power grid production operations and equipment, thereby achieving accurate perception of the global risk status of power grid production scenarios.

[0035] Figure 1This is one of the flow charts of a method for sensing the risk status of power grid safety production operations and equipment provided in an embodiment of the present application. The method can be performed by a device or electronic device for sensing the risk status of power grid safety production operations and equipment. The device or electronic device can be implemented in software and / or hardware, and the device or electronic device can be integrated in any intelligent device with network communication function. Figure 1 As shown, the method for sensing the risk status of power grid safety production operations and equipment may include the following steps:

[0036] S101. Acquire power grid production operation data in a power grid operation management system in real time; the power grid production operation data includes: operation plan data, operation personnel task allocation data, operation environment condition data and equipment operation and maintenance data.

[0037] In an embodiment of the present application, multi-dimensional power grid production operation data is obtained from a power grid operation management system, including operation plan data, operator task allocation data, operation environment condition data, and equipment operation and maintenance data.

[0038] Among them, the work plan data includes annual plan, monthly plan, weekly plan and daily plan tasks.

[0039] Specifically, the annual plan data is used to describe the overall layout of the year's operations, covering the time nodes and priorities of key tasks; the monthly plan refines the annual goals, including specific task decomposition and resource allocation; the weekly plan dynamically adjusts weekly tasks and optimizes resource allocation based on short-term forecast results; the daily plan uses hours as units to refine task execution time and scenario conditions. By analyzing plans at different levels, a task dependency model is established to ensure the logical consistency and feasibility of multi-level plans.

[0040] The task allocation data of operators includes position information, work experience, task type and task time.

[0041] Specifically, the task allocation data of operators focuses on the detailed representation of job information, work experience, task type and task time. Job information is used to clarify the division of responsibilities and role positioning of each operator. Work experience generates a comprehensive score by analyzing the completion of historical tasks to measure the professional ability and adaptability of personnel. Task type records the specific category of the job, such as inspection, maintenance or inspection, and combines the task time to ensure the rationality and timeliness of task allocation. Through data cleaning and standardization, redundant information is eliminated to ensure the integrity and accuracy of the allocation data.

[0042] The operating environment condition data includes the climate and geographical location of the operating environment, etc. The equipment operation and maintenance data includes the equipment operating status, historical maintenance records and fault information.

[0043] S102. Process the power grid production operation data based on the feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data.

[0044] In the present application example, the feature extraction model can be an embedding layer model.

[0045] Specifically, the embedding layer model is first used to encode the priority features of the job plan data, and the time attributes, dependencies and types of tasks are converted into low-dimensional feature vectors.

[0046] The ability characteristics of operators are obtained based on their task allocation data. The task completion time, execution accuracy and complexity in historical task records are used as indicators. A unified ability characteristic representation is generated through an embedded model, which is recorded as the operator's ability characteristic vector.

[0047] The working environment condition characteristics are obtained based on the working environment condition data, and the working environment condition feature vector is formed by layered extraction of climate, geographical location and historical impact data.

[0048] These features are fused after specific standardization and normalization operations to generate a comprehensive feature vector.

[0049] S103. Process the comprehensive feature vector based on the bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of power grid production operation data at different time points.

[0050] In an embodiment of the present application, the comprehensive feature vector is input into BiLSTM, and the forward and backward time series characteristics are combined to generate a dynamic risk feature sequence to characterize the dynamic changes of tasks, personnel and environment at different time points.

[0051] S104. Calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, weight the corresponding risk feature based on the attention score at each time point, and generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data.

[0052] In the embodiment of the present application, based on the dynamic risk feature sequence, the attention score corresponding to the risk feature at each time point is calculated. Specifically, the correlation between the risk feature at each time point and the global context is calculated through the trained weight matrix and activation function, and the correlation score is normalized into the attention score. The risk features are weighted using the attention score, and a comprehensive risk perception feature vector is generated by weighted summation to strengthen the contribution of key time points to the global risk assessment.

[0053] S105. Determine the real-time risk status of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector.

[0054] In an embodiment of the present application, based on a comprehensive risk perception feature vector, multi-dimensional task risks, personnel capability deviations, and environmental anomaly indicators are analyzed, and real-time data and historical risk records are combined to generate real-time risk status values ​​for operation plans and equipment through calculation.

[0055] The real-time risk status value is compared with the preset risk level threshold and divided into low risk, medium risk and high risk. A risk assessment result is generated including risk description, impact scope and optimization suggestions to provide support for subsequent operation plan optimization and risk management.

[0056] The risk state perception method for power grid safety production operations and equipment provided in the embodiment of the present application constructs a real-time dynamic risk state identification framework through a feature extraction model and a BiLSTM model. The feature extraction model uniformly characterizes the operation plan data, the task allocation data of the operator, the operation environment condition data and the equipment operation and maintenance data; the BiLSTM model captures the time series characteristics of these multi-dimensional data, and realizes the accurate perception of the task execution status and the dynamic changes of the environment, so that the potential risk factors can be identified in real time, and the potential risks in the power grid production scene can be identified in time, which significantly improves the timeliness of emergency response. The attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence is calculated, and the corresponding risk feature is weighted based on the attention score at each time point, the contribution of the key features is strengthened, and a comprehensive risk perception feature vector is generated to comprehensively characterize the relationship between multi-dimensional data. Through the comprehensive risk perception vector, the risk factors such as the operation plan data, the task allocation data of the operator, the operation environment condition data and the equipment operation and maintenance data are uniformly quantified and integrated, which overcomes the problem of data isolation in traditional technologies and realizes the accurate perception and evaluation of the global risk state of the power grid production scene.

[0057] In some examples, the feature extraction model is trained by:

[0058] Step 1) obtaining a sample of power grid production operation data and corresponding label information, the label information including: a plan execution result label and a risk status label;

[0059] Step 2) Train the feature extraction model based on the power grid production operation data samples and the corresponding label information until convergence to obtain a trained feature extraction model.

[0060] In an embodiment of the present application, a preprocessing method is adopted for the acquired power grid production operation data samples, including data cleaning, formatting processing and missing value filling. At the same time, the power grid production operation data samples are marked by comparing with historical records to generate label information.

[0061] Among them, the labeling of risk status is combined with task execution deviation, equipment abnormality and environmental fluctuation. Through preset rules and classification algorithms, the risk status labels are divided into normal, low risk, medium risk and high risk to form a high-quality training data set.

[0062] Label information is obtained by marking in the following ways:

[0063] 1) Obtain the task planning time, task actual execution time and equipment operation data in the power grid production operation data sample;

[0064] 2) Compare the task plan time and actual execution time, and label the plan execution result based on the comparison result.

[0065] Specifically, compare the task plan time with the actual execution time, and mark the plan execution result label, including: whether there is a delayed, advanced or unfinished task status.

[0066] 3) Determine abnormal information based on the plan execution result label and equipment operation data; abnormal information includes historical fault records and operation plan execution deviations.

[0067] Specifically, based on the task completion status and equipment operation data, mark whether any abnormal information occurs during the task completion process, including equipment failure, personnel scheduling conflicts, and changes in environmental conditions.

[0068] 4) Based on historical failure records and deviations from work plan execution, risk status labels are marked.

[0069] Specifically, risk status labels are marked based on historical fault records and deviations from the operation plan execution, including: grid risk status is normal, low risk, medium risk and high risk.

[0070] In some embodiments, first, the planned time of the task is compared with the actual execution time, and the task is marked as delayed, advanced or unfinished by matching and analyzing the planned timeline and the actual completion timeline. The judgment of delayed tasks is based on whether the actual start or completion time exceeds the planned time range, and the delay duration is recorded; the judgment of early tasks is based on the actual completion time being earlier than the planned time range; unfinished tasks are marked by comparing the task end mark and the planned completion deadline. This step not only identifies the deviation of task time, but also provides an important basis for subsequent task adjustments.

[0071] Secondly, based on the task completion status and equipment operation data, detailed annotations are made to determine whether any abnormal situations occur during the task completion process. In the equipment dimension, by comparing the real-time collected equipment operation status parameters (such as temperature, vibration, operating efficiency, etc.) with the equipment's historical operation baseline, detect whether there are signs of equipment failure; in the personnel scheduling dimension, analyze the task assignment records and actual execution records, and mark whether there are personnel scheduling conflicts, such as multiple tasks assigned to the same person or the task execution personnel not arriving as planned; in the environmental condition dimension, through the actual weather, geographical conditions and other environmental data within the planned task time, mark whether there are abnormal changes, such as the impact of bad weather or unexpected geological events on the task. The abnormal annotations in each dimension are associated with the task identifier to form a multi-dimensional risk description of the task completion process.

[0072] Finally, based on historical fault records and deviations from the execution of the job plan, the execution results of the task are classified and labeled into different risk states, including normal, low risk, medium risk, and high risk. The normal state means that the task is completed as planned without exception; low risk means that the task has slight time deviations or environmental changes, but it does not affect the completion of the task; medium risk indicates that a large deviation or a single abnormality occurs in the task (such as equipment failure or scheduling conflict); high risk indicates that multi-dimensional abnormalities occur in the task or that there may be a chain reaction on the overall job plan. The classification of risk states adopts a method that combines rule bases with historical statistical analysis, and is determined by calculating the comprehensive deviation index value and comparing it with the risk threshold.

[0073] In some examples, the comprehensive feature vector includes: 1) a job priority feature vector corresponding to the job plan data; 2) a worker capability feature vector corresponding to the worker task allocation data; 3) a job environment condition feature vector corresponding to the job environment condition data; 4) an equipment operation and maintenance feature vector corresponding to the equipment operation and maintenance data. The comprehensive feature vector is obtained by fusing the job priority feature vector, the worker capability feature vector, the job environment condition feature vector, and the equipment operation and maintenance feature vector;

[0074] Among them: 1): The job priority feature vector is obtained by the following method:

[0075] Based on the time attributes, task types and the dependencies between the time attributes and task types in the job planning data, a task priority sequence is constructed; the task priority sequence is input into the feature extraction model so that the feature extraction model embeds and encodes the task priority sequence to generate a job priority feature vector.

[0076] Specifically, a task priority sequence is constructed based on the time attributes, task types and dependencies of the job planning data; the priority sequence is embedded and encoded through the embedding layer model to generate a low-dimensional job priority feature vector.

[0077] 2): The operator capability feature vector is obtained by:

[0078] From the historical task completion records of operators in the operator task allocation data, the task completion time, execution accuracy and operation complexity are extracted; the task completion time, execution accuracy and operation complexity are weighted to establish an operation capability scoring matrix; the operation capability scoring matrix is ​​input into the feature extraction model so that the feature extraction model embeds the operation capability scoring matrix to generate the operator capability feature vector.

[0079] Based on the historical task completion records of operators in the operator task allocation data, the task completion time, execution accuracy and operation complexity are extracted and weighted to establish an operation capability scoring matrix; the operation capability features in the scoring matrix are mapped to a low-dimensional vector space through an embedding layer model to generate an operator capability feature vector.

[0080] 3): The characteristic vector of the working environment conditions is obtained by:

[0081] Acquire weather, geographic location, temperature and humidity data from the working environment condition data; input the weather, geographic location, temperature and humidity data into the feature extraction model so that the feature extraction model embeds and encodes the weather, geographic location, temperature and humidity data to generate a working environment condition feature vector.

[0082] Specifically, the working environment condition data is stratified, including weather, geographic location, and temperature and humidity data; embedding processing is performed through the embedding layer model to generate a working environment condition feature vector.

[0083] Finally, the job priority feature vector, operator capability feature vector, job environment condition feature vector and equipment operation and maintenance feature vector are concatenated, and a comprehensive feature vector is generated through standardization and normalization operations.

[0084] In practical applications, first, extract the job plan data, including the time attributes of the task (such as the planned start time, duration), the task type (such as maintenance, inspection or emergency handling), and the dependencies between tasks (such as predecessor tasks and follow-up tasks). Based on these data, construct a task priority sequence, which is generated by analyzing the dependencies and timing characteristics between tasks. The embedding layer model maps the task priority sequence to a low-dimensional feature representation, and each task is represented as a feature vector of a fixed dimension. The embedding process optimizes the objective function to ensure that the mapped vector can retain the relative relationship between tasks in the vector space. For example, tasks with high priority are closer to the center point in the vector space, and tasks with low priority are farther away from the center.

[0085] Secondly, for the operator's capability feature vector, core indicators such as task completion time, execution accuracy and operation complexity are extracted from the historical task completion records. These indicators are weighted to generate an operation capability scoring matrix, where the weight setting is adjusted according to the contribution of the indicator to the task completion efficiency. For example, execution accuracy is given a higher weight, while completion time has a relatively low weight in complex tasks. The scoring matrix is ​​feature mapped through the embedding layer model, compressing the complex multi-dimensional scoring data into a low-dimensional feature representation to form the capability feature vector of each operator. During the training process, the embedding layer model uses the objective function to minimize the error between the personnel capability feature vector and the historical task completion performance to ensure that the generated vector can accurately reflect the ability differences of personnel.

[0086] At the same time, the operating environment condition data is decomposed at multiple levels to extract multi-dimensional information including weather (such as temperature and rainfall), geographic location (such as region type and altitude), and environmental parameters of equipment operation (such as humidity and vibration intensity). Each layer of environmental data is feature encoded through the embedding layer model. The embedding process not only retains the physical meaning of each dimension, but also captures the potential correlation between dimensions through feature learning. For example, high temperature and high humidity environments may have a common impact on certain task types. Finally, each environmental feature is mapped to a low-dimensional vector of a fixed length.

[0087] The job priority feature vector, operator capability feature vector, operating environment condition feature vector, and equipment operation and maintenance feature vector are spliced ​​into a joint vector (i.e., the comprehensive feature vector mentioned above). The spliced ​​vectors are standardized and normalized to ensure that each feature has the same scale in subsequent calculations. The standardization process calculates the mean and variance to make the numerical distribution of each feature dimension consistent; the normalization operation ensures that the spliced ​​vector is within a unified numerical range, so as to meet the input requirements of the subsequent deep learning model.

[0088] In some examples, the comprehensive feature vector is processed based on the BiLSTM model to obtain a dynamic risk feature sequence, which can be specifically achieved through the following steps:

[0089] Step 1) Convert the comprehensive feature vector into a comprehensive feature vector sequence according to the time sequence of the operation plan data.

[0090] Step 2) Input the comprehensive feature vector sequence into the BiLSTM model to obtain the forward hidden state and backward hidden state corresponding to each time point output by the BiLSTM model; the forward hidden state is used to record the historical information of the current time point, and the backward hidden state is used to record the future information of the current time point.

[0091] Specifically, the comprehensive feature vector sequence is input into the BiLSTM model, the time-dependent characteristics of task execution are captured through forward propagation, and the potential impact of task history on the current job is captured through backward propagation, generating forward and backward hidden states at each time point.

[0092] Step 3) For any time point, calculate the weight coefficients of the forward hidden state and the backward hidden state; weight the weight coefficients to generate a fused hidden state; concatenate the fused hidden state with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point.

[0093] Specifically, feature fusion is performed based on the forward and backward latent states, and the comprehensive feature vector at the current time point is combined to generate time-series dynamic risk features.

[0094] Specifically, the weight coefficients of the forward hidden state and the backward hidden state are calculated, the time-related fused hidden state is generated using the weighted average method, the fused hidden state is concatenated with the current comprehensive feature vector, and a dynamic risk feature sequence is generated through linear transformation.

[0095] Step 4) Connect the dynamic risk features corresponding to each time point to generate a dynamic risk feature sequence.

[0096] In practical applications, first, the comprehensive feature vectors generated in the early stage are organized into feature sequences according to the time sequence of the job tasks. Each comprehensive feature vector contains a multi-dimensional feature representation of task priority, personnel capabilities, and environmental conditions. In the organization process, ensure that the feature vector sequence is strictly sorted according to the task timeline so that the input sequence can reflect the actual time dependency of task execution. This sorting provides accurate time sequence input for the subsequent learning of the time series model, ensuring that the model can capture the temporal dynamic association between tasks.

[0097] Subsequently, the comprehensive feature vector sequence is input into the Bi-LSTM model. In the forward propagation, the Bi-LSTM model takes the past data of the time series as input, generates the forward hidden state at each time point, and captures the time-dependent characteristics of the task during execution, such as the possible impact of the previous task on the completion of the current task. In the backward propagation, the Bi-LSTM takes the future data as input, generates the backward hidden state at each time point, and reflects the potential impact of the future task on the current task, such as the risk pressure of the urgency of the subsequent task on the current task.

[0098] For each time point, the forward hidden state and the backward hidden state record the historical information and future information of the task respectively. On this basis, the forward and backward hidden states are feature fused by the weighted fusion method. The weight coefficient is calculated by an adaptive function, which dynamically adjusts the weight according to the relative importance of the forward and backward hidden states to ensure that the key time-dependent information is fully reflected. After the fused hidden state is generated, it is concatenated with the comprehensive feature vector of the current time point to form a complete vector representation containing historical, future and current feature information.

[0099] Finally, the concatenated feature vectors are mapped through linear transformation to generate a dynamic risk feature sequence. The weight matrix of the linear transformation is continuously updated during the model training process, with the goal of minimizing the prediction error so that the generated dynamic risk feature sequence can accurately characterize the risk status changes of tasks, personnel, and environment. This method can efficiently capture the dynamic risk characteristics in the time series and provide high-quality input for subsequent risk perception and prediction modules.

[0100] In some examples, the fused latent state is concatenated with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point, which is calculated by the following formula (1):

[0101]

[0102] Among them, h t Characterize the dynamic risk characteristics corresponding to time point t; and Represent the forward hidden state and the backward hidden state respectively; x t Characterizes the comprehensive feature vector corresponding to time point t; W f The weight matrix representing the forward hidden state; W b The weight matrix representing the backward hidden state; W c Represents the weight matrix of the comprehensive feature vector corresponding to time point t; b represents the bias; σ represents the activation function.

[0103] In practical applications, the Bi-LSTM model starts from the starting point of the time series and gradually calculates the forward hidden state of each time point. The forward hidden state is used to capture historical information and reflect the extent to which the current task is affected by the previous task, such as whether the delay of the previous task or the lack of personnel capabilities will cause risk accumulation for the current task. At the same time, it backpropagates from the end point of the time series to the starting point to calculate the backward hidden state of each time point. The backward hidden state reflects the potential impact of future tasks on the current task, such as the urgency of subsequent tasks or the risk pressure that competition for resources may pose to the current task. During the calculation process, the comprehensive feature vector at the current time point is used to supplement the static information of the current task, including the priority of the task, environmental conditions, and the capabilities of the operators.

[0104] By combining the forward hidden state and the backward hidden state, the Bi-LSTM model achieves unified modeling of historical, current, and future information. In the specific implementation, this combination is completed through weighted summation, in which the weight matrix is ​​used to adjust the contribution ratio of each part of the information to the final risk feature. The addition of the bias term is used to correct the model output to ensure that the numerical range of the feature vector is reasonable. In order to introduce nonlinear characteristics and improve the expressiveness of the model, activation functions are used in the calculation to transform the results. For example, the commonly used Sigmoid function can limit the output range to a certain interval, making the model more sensitive to features, thereby more accurately capturing the complex relationship between tasks.

[0105] In some examples, the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence is calculated, which is specifically implemented by the following steps:

[0106] Step 1) For the risk features at each time point, the correlation score between the risk features and the global context is calculated based on the pre-trained weight matrix and activation function.

[0107] Step 2) Normalize the relevance score to obtain the attention score.

[0108] Specifically, first, for the risk feature at each time point in the dynamic risk feature sequence, its correlation is calculated through the pre-trained weight matrix and the global context vector.

[0109] Specifically, the global context vector represents the overall attention direction of the risk feature sequence, while the risk feature at each time point reflects the information of the current task or state. The calculation of the correlation is achieved through the inner product or a specific matching function, and the result value reflects the degree of match between the risk feature at the current time point and the global risk assessment target. Subsequently, the correlation result is normalized into an attention score through an activation function (such as Softmax). The purpose of normalization is to ensure that the sum of the attention scores of all time points is 1, so that the attention score can directly reflect the relative importance of each time point in the overall risk assessment.

[0110] Next, the calculated attention score is used to perform a weighted operation on the risk features at each time point in the dynamic risk feature sequence.

[0111] Specifically, the risk features of each time point will be weighted by its corresponding attention score, and the features of time points with high attention scores have a greater weight in the overall calculation. The weighted risk features are fused into a comprehensive risk-aware feature vector through a summation operation. This feature vector integrates the information of all key points in the time series, but the attention mechanism ensures that the features of important time points are strengthened, and the influence of minor time points is ignored or weakened. For example, in a power grid inspection task, some time points may correspond to abnormal equipment status or personnel scheduling conflicts. The features of these time points will be given higher attention scores due to their high correlation with the global context, thereby occupying a larger weight in the comprehensive feature vector. In contrast, the risk features of other time points may be given lower weights due to the lack of key information, and their impact is limited.

[0112] In some examples, based on the comprehensive risk perception feature vector, determining the real-time risk status of the power grid production operation data at the current moment specifically includes the following steps:

[0113] Step 1), based on the cosine similarity algorithm, calculate the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector in the preset abnormal risk vector library;

[0114] Step 2), when the similarity score between the comprehensive risk perception feature vector and the target abnormal risk vector in each abnormal risk vector reaches a preset threshold, the risk state corresponding to the target abnormal risk vector is determined as the real-time risk state;

[0115] When the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector does not reach a preset threshold, the real-time risk state is determined to be a risk-free state.

[0116] In the embodiment of the present application, the cosine similarity algorithm is used to match the risk feature vector with the preset abnormal type library, automatically identify the abnormal type and generate warning information. At the same time, combined with the preset solution strategy library, accurate optimization suggestions are provided for different risk levels to ensure efficient response to complex problems in the operation scenario.

[0117] First, the comprehensive risk perception feature vector is used, which integrates multi-dimensional information of task priority, personnel capabilities and environmental conditions to comprehensively characterize the risk status of the current operation plan and equipment. In order to identify the abnormal type, the system constructs a preset abnormal risk vector library, in which each abnormal risk vector represents a typical abnormal type, such as equipment failure, high-risk weather conditions, personnel scheduling conflicts, etc. The preset abnormal risk vector library is constructed through cluster analysis and manual annotation of historical data to ensure that each abnormal type vector can accurately cover common risk scenarios.

[0118] During the matching process, the cosine similarity algorithm is used to compare the comprehensive risk perception feature vector with each abnormal type vector in the preset abnormal risk vector library one by one. Cosine similarity measures the similarity between the feature vector and the abnormal type vector by calculating the cosine value of the angle between the two vectors. When the traversal is completed, the risk state corresponding to the target abnormal risk vector with the highest similarity is selected as the matching result of the current risk state, such as incomplete coverage of the pre-trial annual plan.

[0119] In order to improve the robustness of the match, the system can also set a preset threshold. Only when the maximum similarity value exceeds the threshold, the matching result is considered valid, thereby avoiding misjudgment. After completing the abnormal type matching, the system retrieves the corresponding solution from the preset solution strategy library according to the matching result. The solution strategy library contains the optimization suggestions and specific measures for each abnormal type that are manually set in advance. The risk feature vector is matched with the preset abnormal type library through the cosine similarity algorithm, and the abnormal type is automatically identified and warning information is generated. At the same time, combined with the preset solution strategy library, accurate optimization suggestions are provided for different risk levels to ensure efficient response to complex problems in the operation scenario.

[0120] Figure 2 The second flow chart of the method for sensing the risk status of power grid safety production operations and equipment provided in an embodiment of the present application specifically includes the following steps:

[0121] S201. Acquire power grid production operation data in the power grid operation management system in real time; the power grid production operation data includes: operation plan data, operation personnel task allocation data, operation environment condition data and equipment operation and maintenance data.

[0122] S202. Process the power grid production operation data based on the feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data.

[0123] S203, converting the comprehensive feature vector into a comprehensive feature vector sequence according to the time sequence of the operation plan data.

[0124] S204. Input the comprehensive feature vector sequence into the BiLSTM model to obtain the forward hidden state and backward hidden state corresponding to each time point output by the BiLSTM model; the forward hidden state is used to record the historical information of the current time point, and the backward hidden state is used to record the future information of the current time point.

[0125] S205. For any time point, calculate the weight coefficients of the forward hidden state and the backward hidden state; weight the weight coefficients to generate a fused hidden state; concatenate the fused hidden state with the current comprehensive feature vector to generate a dynamic risk feature corresponding to the current time point.

[0126] S206. Connect the dynamic risk features corresponding to each time point to generate a dynamic risk feature sequence.

[0127] S207. For the risk features at each time point, the correlation score between the risk features and the global context is calculated based on the pre-trained weight matrix and activation function, and the correlation score is normalized to obtain the attention score corresponding to the risk features at each time point in the dynamic risk feature sequence.

[0128] S208. Weight the corresponding risk features based on the attention score at each time point to generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data.

[0129] S209. Based on the cosine similarity algorithm, calculate the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector in the preset abnormal risk vector library.

[0130] S210. When the similarity score between the comprehensive risk perception feature vector and the target abnormal risk vector in each abnormal risk vector reaches a preset threshold, the risk state corresponding to the target abnormal risk vector is determined as the real-time risk state; when the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector does not reach the preset threshold, the real-time risk state is determined to be a risk-free state.

[0131] Figure 3A schematic diagram of the structure of a risk status perception device for power grid safety production operations and equipment provided in an embodiment of the present application. Figure 3 As shown, the risk status perception device for power grid safety production operations and equipment includes:

[0132] The acquisition module 301 is used to acquire the power grid production operation data in the power grid operation management system in real time; the power grid production operation data includes: operation plan data, operation personnel task allocation data, operation environment condition data and equipment operation and maintenance data;

[0133] The first processing module 302 is used to process the power grid production operation data based on the feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0134] The second processing module 303 is used to process the comprehensive feature vector based on the bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of power grid production operation data at different time points;

[0135] The calculation module 304 is used to calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, and weight the corresponding risk feature based on the attention score at each time point to generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operation personnel task allocation data, the operation environment condition data and the equipment operation and maintenance data;

[0136] The determination module 305 is used to determine the real-time risk status of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector.

[0137] The risk state perception device for power grid safety production operations and equipment provided in the embodiment of the present application constructs a real-time dynamic risk state identification framework through a feature extraction model and a BiLSTM model. The feature extraction model uniformly characterizes the operation plan data, the task allocation data of the operator, the operation environment condition data and the equipment operation and maintenance data; the BiLSTM model captures the time series characteristics of these multi-dimensional data, and realizes the accurate perception of the task execution status and the dynamic changes of the environment, so that the potential risk factors can be identified in real time, and the potential risks in the power grid production scene can be identified in time, which significantly improves the timeliness of emergency response. The attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence is calculated, and the corresponding risk feature is weighted based on the attention score at each time point, the contribution of the key features is strengthened, and a comprehensive risk perception feature vector is generated to comprehensively characterize the relationship between multi-dimensional data. Through the comprehensive risk perception vector, the risk factors such as the operation plan data, the task allocation data of the operator, the operation environment condition data and the equipment operation and maintenance data are uniformly quantified and integrated, which overcomes the problem of data isolation in traditional technologies and realizes the accurate perception and evaluation of the global risk state of the power grid production scene.

[0138] The comprehensive feature vector includes: a job priority feature vector corresponding to the job plan data, an operator capability feature vector corresponding to the operator task allocation data, an operation environment condition feature vector corresponding to the operation environment condition data, and an equipment operation and maintenance feature vector corresponding to the equipment operation and maintenance data; the comprehensive feature vector is obtained by fusing the job priority feature vector, the operator capability feature vector, the operation environment condition feature vector, and the equipment operation and maintenance feature vector;

[0139] Among them: The job priority feature vector is obtained by the following method:

[0140] Based on the time attributes, task types and the dependency relationship between the time attributes and the task types in the job plan data, a task priority sequence is constructed; the task priority sequence is input into the feature extraction model so that the feature extraction model embeds and encodes the task priority sequence to generate a job priority feature vector;

[0141] The operator capability feature vector is obtained by:

[0142] Extracting task completion time, execution accuracy and operation complexity from the historical task completion records of operators in the operator task allocation data; weighting the task completion time, execution accuracy and operation complexity to establish an operation capability scoring matrix; inputting the operation capability scoring matrix into the feature extraction model so that the feature extraction model embeds and encodes the operation capability scoring matrix to generate an operator capability feature vector;

[0143] The operating environment condition feature vector is obtained by:

[0144] Acquire weather, geographic location, temperature and humidity data from the working environment condition data; input the weather, geographic location, temperature and humidity data into the feature extraction model so that the feature extraction model embeds and encodes the weather, geographic location, temperature and humidity data to generate a working environment condition feature vector.

[0145] Optionally, the second processing module 303 is further configured to:

[0146] Convert the comprehensive feature vector into a comprehensive feature vector sequence according to the time sequence of the operation plan data;

[0147] The comprehensive feature vector sequence is input into the BiLSTM model to obtain the forward hidden state and backward hidden state corresponding to each time point output by the BiLSTM model; the forward hidden state is used to record the historical information of the current time point, and the backward hidden state is used to record the future information of the current time point;

[0148] For any time point, calculate the weight coefficients of the forward hidden state and the backward hidden state; weight the weight coefficients to generate a fused hidden state; concatenate the fused hidden state with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point;

[0149] The dynamic risk features corresponding to each time point are spliced ​​together to generate a dynamic risk feature sequence.

[0150] Optionally, the fused latent state is concatenated with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point, which is calculated by the following formula (1):

[0151]

[0152] Among them, h t Characterize the dynamic risk characteristics corresponding to time point t; and Represent the forward hidden state and the backward hidden state respectively; x t Characterizes the comprehensive feature vector corresponding to time point t; W f The weight matrix representing the forward hidden state; W b The weight matrix representing the backward hidden state; W c Represents the weight matrix of the comprehensive feature vector corresponding to time point t; b represents the bias; σ represents the activation function.

[0153] Optionally, the calculation module 304 is further configured to:

[0154] For the risk features at each time point, the correlation score between the risk features and the global context is calculated based on the pre-trained weight matrix and activation function;

[0155] The relevance scores are normalized to obtain the attention scores.

[0156] Optionally, the determination module 305 is further configured to:

[0157] Based on the cosine similarity algorithm, the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector in the preset abnormal risk vector library is calculated;

[0158] When the similarity score between the comprehensive risk perception feature vector and the target abnormal risk vector in each abnormal risk vector reaches a preset threshold, the risk state corresponding to the target abnormal risk vector is determined as the real-time risk state;

[0159] When the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector does not reach a preset threshold, the real-time risk state is determined to be a risk-free state.

[0160] Optionally, the feature extraction model is trained in the following way:

[0161] Obtaining power grid production operation data samples and corresponding label information, the label information includes: plan execution result label and risk status label;

[0162] The feature extraction model is trained based on the power grid production operation data samples and the corresponding label information until convergence, thereby obtaining a trained feature extraction model;

[0163] Label information is obtained by marking in the following ways:

[0164] Obtain the task planning time, task actual execution time and equipment operation data in the power grid production operation data sample;

[0165] Compare the task plan time and actual execution time, and label the plan execution result based on the comparison results;

[0166] Determine abnormal information based on the plan execution result label and equipment operation data; abnormal information includes historical fault records and operation plan execution deviations;

[0167] Based on historical failure records and deviations from work plan execution, risk status labels are marked.

[0168] The above-mentioned risk state perception device for power grid safety production operations and equipment can execute the method provided by any embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in this embodiment, please refer to the risk state perception method for power grid safety production operations and equipment provided by any embodiment of the present application.

[0169] Figure 4A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. Figure 4 A block diagram of an exemplary electronic device suitable for implementing the embodiments of the present application is shown. Figure 4 The electronic device 12 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0170] like Figure 4 As shown, the electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0171] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0172] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0173] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 4 not shown, usually called a "hard drive"). Although Figure 4 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical medium) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present application.

[0174] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in the memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0175] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 12, and / or may communicate with any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the electronic device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although Figure 4 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0176] The processing unit 16 executes various functional applications and data processing by running the programs stored in the system memory 28, such as implementing the risk status perception method for power grid safety production operations and equipment provided in the embodiment of the present application.

[0177] The embodiment of the present application also provides a computer storage medium.

[0178] The computer-readable storage medium of the embodiment of the present application can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, - but not limited to - a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive lists) of computer-readable storage media include: an electrical connection with one or more wires, 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 above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0179] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0180] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0181] Computer program code for performing the operation of the present application can be written in one or more programming languages ​​or a combination thereof, 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 (e.g., using an Internet service provider to connect through the Internet).

[0182] The embodiment of the present application also provides a computer program product.

[0183] Various embodiments 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 embodiments can include: being implemented in one or more computer program products, which can include one or more computer programs, which 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, which 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.

[0184] Note that the above are only preferred embodiments of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A method for perceiving the risk status of power grid safety production operations and equipment, characterized in that: The method comprises: Real-time acquisition of power grid production operation data in the power grid operation management system; the power grid production operation data includes: operation plan data, operator task allocation data, operation environment condition data and equipment operation and maintenance data; Processing the power grid production operation data based on a feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operator task allocation data, the operation environment condition data, and the equipment operation and maintenance data; The comprehensive feature vector is processed based on a bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of the power grid production operation data at different time points; Calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, weight the corresponding risk feature based on the attention score at each time point, and generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; Based on the comprehensive risk perception feature vector, the real-time risk status of the power grid production operation data at the current moment is determined.

2. The method according to claim 1, characterized in that The comprehensive feature vector includes: a job priority feature vector corresponding to the job plan data, an operator capability feature vector corresponding to the operator task allocation data, an operation environment condition feature vector corresponding to the operation environment condition data, and an equipment operation and maintenance feature vector corresponding to the equipment operation and maintenance data; the comprehensive feature vector is obtained by fusing the job priority feature vector, the operator capability feature vector, the operation environment condition feature vector, and the equipment operation and maintenance feature vector; Wherein: the job priority feature vector is obtained by the following method: Based on the time attributes, task types and the dependency relationship between the time attributes and the task types in the job plan data, a task priority sequence is constructed; the task priority sequence is input into the feature extraction model so that the feature extraction model embeds and encodes the task priority sequence to generate the job priority feature vector; The operator capability feature vector is obtained in the following way: Extracting task completion time, execution accuracy and operation complexity from the operator's historical task completion records in the operator's task assignment data; weighting the task completion time, the execution accuracy and the operation complexity to establish an operation capability scoring matrix; inputting the operation capability scoring matrix into the feature extraction model so that the feature extraction model embeds the operation capability scoring matrix to generate the operator capability feature vector; The operating environment condition feature vector is obtained by: Acquire weather, geographic location, and temperature and humidity data from the operating environment condition data; input the weather, geographic location, and temperature and humidity data into the feature extraction model so that the feature extraction model embeds and encodes the weather, geographic location, and temperature and humidity data to generate the operating environment condition feature vector.

3. The method according to claim 1, characterized in that: The comprehensive feature vector is processed based on the BiLSTM model to obtain a dynamic risk feature sequence, including: Converting the comprehensive feature vector into a comprehensive feature vector sequence according to the time sequence of the operation plan data; Input the comprehensive feature vector sequence into the BiLSTM model to obtain a forward hidden state and a backward hidden state corresponding to each time point output by the BiLSTM model; the forward hidden state is used to record historical information at the current time point, and the backward hidden state is used to record future information at the current time point; For any time point, calculate the weight coefficients of the forward hidden state and the backward hidden state; weight the weight coefficients to generate a fused hidden state; concatenate the fused hidden state with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point; The dynamic risk features corresponding to each of the time points are concatenated to generate the dynamic risk feature sequence.

4. The method according to claim 3, characterized in that The fusion latent state is concatenated with the current comprehensive feature vector to generate the dynamic risk feature corresponding to the current time point, which is calculated by the following formula (1): Among them, h t Characterize the dynamic risk characteristics corresponding to time point t; and Respectively represent the forward hidden state and the backward hidden state; x t Characterizes the comprehensive feature vector corresponding to time point t; W f The weight matrix representing the forward hidden state; W b The weight matrix representing the backward hidden state; W c represents the weight matrix of the comprehensive feature vector corresponding to time point t; b represents the bias; σ represents the activation function.

5. The method according to claim 1, characterized in that The calculating the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence includes: For the risk feature at each time point, calculate the correlation score between the risk feature and the global context based on the pre-trained weight matrix and activation function; The relevance score is normalized to obtain the attention score.

6. The method according to claim 1, characterized in that Determining the real-time risk status of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector includes: Based on the cosine similarity algorithm, the similarity score between the comprehensive risk perception feature vector and each abnormal risk vector in the preset abnormal risk vector library is calculated; When the similarity score between the comprehensive risk perception feature vector and the target abnormal risk vector in each of the abnormal risk vectors reaches a preset threshold, determining the risk state corresponding to the target abnormal risk vector as the real-time risk state; When the similarity score between the comprehensive risk perception feature vector and each of the abnormal risk vectors does not reach the preset threshold, the real-time risk state is determined to be a risk-free state.

7. The method according to any one of claims 1 to 6, characterized in that The feature extraction model is trained in the following way: Obtaining a power grid production operation data sample and corresponding label information, wherein the label information includes: a plan execution result label and a risk status label; Training the feature extraction model based on the power grid production operation data sample and the corresponding label information until convergence to obtain a trained feature extraction model; The label information is obtained by marking in the following manner: Obtaining the task planning time, task actual execution time and equipment operation data in the power grid production operation data sample; Comparing the task plan time with the actual execution time, and marking the plan execution result label based on the comparison result; Based on the plan execution result label and the equipment operation data, determining abnormal information; the abnormal information includes historical fault records and operation plan execution deviations; Based on the historical fault records and the operation plan execution deviation, the risk status label is marked.

8. A risk status perception device for power grid safety production operations and equipment, characterized in that: The device comprises: An acquisition module is used to acquire power grid production operation data in the power grid operation management system in real time; the power grid production operation data includes: operation plan data, operation personnel task allocation data, operation environment condition data and equipment operation and maintenance data; A first processing module is used to process the power grid production operation data based on a feature extraction model to obtain a comprehensive feature vector corresponding to the power grid production operation data; the comprehensive feature vector is a unified vector representation of the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; The second processing module is used to process the comprehensive feature vector based on a bidirectional long short-term memory network BiLSTM model to obtain a dynamic risk feature sequence; the dynamic risk feature sequence is used to reflect the dynamic changes of the power grid production operation data at different time points; A calculation module, used to calculate the attention score corresponding to the risk feature at each time point in the dynamic risk feature sequence, weight the corresponding risk feature based on the attention score at each time point, and generate a comprehensive risk perception feature vector; the comprehensive risk perception feature vector is used to characterize the correlation between the operation plan data, the operator task allocation data, the operation environment condition data and the equipment operation and maintenance data; A determination module is used to determine the real-time risk status of the power grid production operation data at the current moment based on the comprehensive risk perception feature vector.

9. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the risk status perception method for power grid safety production operations and equipment as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for perceiving the risk status of power grid safety production operations and equipment as described in any one of claims 1 to 7 is implemented.

Images