Substation safety operation situation assessment method and device

Abstract

The application discloses a substation safety operation situation assessment method and device. The method comprises the following steps: obtaining safety situation data of operation personnel and equipment in a substation; and based on the safety situation data, using a safety situation assessment model previously constructed based on a neural network to assess the current operation safety situation of the operation personnel and equipment; wherein the safety situation data is data related to operation safety of the operation personnel and equipment in the substation. The application solves the technical problem of inaccurate substation safety operation situation assessment.

Description

Technical Field

[0001] This application relates to the power industry, and more specifically, to a method and apparatus for assessing the safe operation status of a substation. Background Technology

[0002] Substation safety operation status assessment technology is a relatively effective proactive defense technology to deal with safety threats. Using this technology, staff can fully understand the safety risks and threats faced in the current environment, grasp the dynamic status of operational safety as a whole, and judge, evaluate and predict the safety status and development trend so as to take corresponding remedial and preventive measures in a timely manner.

[0003] When a task has multiple risk points, it's difficult to determine which hazard is greater. In complex work environments, dynamic changes in factors have a significant impact on safety risks, and effective early warning measures are lacking during the operation. Traditional safety situation assessment methods directly use raw safety data for evaluation. Due to the large dimensionality of the input data, the model becomes complex, increasing the computational burden of training and reducing timeliness, thus affecting the efficiency of the assessment method. Furthermore, it cannot effectively perform feature learning and representation. Current safety situation assessment methods suffer from low fitting between assessment results and actual values, poor fault tolerance, and high energy consumption.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a method and apparatus for assessing the safety status of substation operations, which at least solves the technical problem of inaccurate assessment of the safety status of substation operations.

[0006] According to one aspect of the embodiments of this application, a method for assessing the safety status of substation operations is provided, comprising: acquiring safety status data of substation personnel and equipment; and assessing the current operational safety status of the personnel and equipment based on the safety status data using a safety status assessment model pre-constructed based on a neural network; wherein the safety status data is data related to operational safety of the personnel and equipment in the substation.

[0007] In an exemplary embodiment, the safety situation assessment model is constructed through the following steps: acquiring historical safety data information related to the substation to construct the data layer of the safety situation assessment model; and constructing the assessment layer of the safety situation assessment model using the neural network based on the data layer.

[0008] In an exemplary embodiment, before constructing the security situation assessment model, the method further includes: extracting factors that influence the security situation assessment from the security data information as situation assessment elements; and constructing a situation assessment index system based on the situation assessment elements to assess the security situation of the substation.

[0009] In an exemplary embodiment, extracting factors that influence security situation assessment from the security data information includes: preprocessing the security data information, the preprocessing including label segmentation and normalization; performing layer-by-layer loss compensation on the preprocessed security data information; and extracting factors that influence security situation assessment from the layer-by-layer loss-compensated security data information.

[0010] In an exemplary embodiment, performing layer-by-layer loss compensation on the preprocessed security data information includes: encoding the preprocessed security data information; restoring the encoded security data information to obtain the loss of the encoded security data information during the encoding process; calculating the loss of the feature information of the security data information using a loss function based on the obtained loss; and compensating the loss of the feature information of the feature information to the situation assessment elements.

[0011] In an exemplary embodiment, constructing a situation assessment index system based on the situation assessment elements includes: quantifying the situation assessment elements to obtain attribute vectors of the situation assessment elements; clustering the situation assessment elements with similar functions based on the attribute vectors; calculating the importance of the situation assessment elements in each cluster using the analytic hierarchy process (AHP), and selecting representative situation assessment elements according to the importance to construct the situation assessment index system.

[0012] In an exemplary embodiment, constructing an evaluation layer of the security situation assessment model based on the neural network includes: using the neural network to perform information transmission and feature learning on the dataset in the data layer based on the situation assessment index system, so as to train the security situation assessment model; and using the forward propagation of the neural network to optimize the security situation assessment model to obtain the evaluation layer.

[0013] In one exemplary embodiment, an alarm is triggered when the current operational safety situation of the personnel and equipment is determined to be unsafe.

[0014] According to another aspect of the embodiments of this application, a substation safety operation status assessment device is also provided, comprising: an acquisition module configured to acquire safety status data of substation operators and equipment; and an assessment module configured to assess the current operational safety status of the operators and equipment based on the safety status data and using a safety status assessment model pre-constructed based on a neural network; wherein the safety status data is data related to operational safety of the operators and equipment in the substation.

[0015] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, on which a program is stored, which, when executed, causes a computer to perform the method described above.

[0016] In this embodiment, a safety situation assessment model pre-built based on a neural network is used to assess the current operational safety situation of the personnel and equipment, thus solving the technical problem of inaccurate assessment of the safety situation of substation operations. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a flowchart of a substation safety operation status assessment method according to an embodiment of this application;

[0019] Figure 2 This is a flowchart of another substation safety operation status assessment method according to an embodiment of this application;

[0020] Figure 3 This is a flowchart illustrating the design of a security situation assessment model based on embodiments of this application;

[0021] Figure 4 This is a flowchart of a security situation assessment element extraction method according to an embodiment of this application;

[0022] Figure 5 This is a flowchart illustrating the construction of a security situation indicator system based on embodiments of this application;

[0023] Figure 6 This is a flowchart of a situation assessment method based on the Subagging algorithm and a deep recurrent neural network according to an embodiment of this application;

[0024] Figure 7 This is a schematic diagram of the substation safety operation status assessment device according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] Example 1

[0028] According to an embodiment of this application, a method for assessing the safe operating status of a substation is provided, such as... Figure 1 As shown, the method includes:

[0029] Step S102: Obtain safety status data of substation personnel and equipment.

[0030] In this embodiment, the safety situation data refers to the data related to the safety of the workers and equipment in the substation.

[0031] Step S104: Based on the safety situation data, use a safety situation assessment model pre-built based on a neural network to assess the current operational safety situation of the personnel and equipment.

[0032] In one example, before constructing the security situation assessment model, it is necessary to extract situation assessment elements and, based on these elements, construct a situation assessment index system. For instance, factors that influence the security situation assessment are extracted from the security data information as situation assessment elements; and based on these elements, a situation assessment index system is constructed to assess the security situation of the substation.

[0033] In one example, extracting factors influencing security situation assessment from the security data information may include: preprocessing the security data information, which includes label segmentation and normalization; performing layer-by-layer loss compensation on the preprocessed security data information; and extracting factors influencing security situation assessment from the layer-by-layer loss-compensated security data information. Here, layer-by-layer loss compensation may include: encoding the preprocessed security data information; restoring the encoded security data information to obtain the loss incurred during the encoding process; calculating the loss of the feature information of the security data information using a loss function based on the obtained loss; and compensating the feature information loss of the security data information to the situation assessment elements.

[0034] In one example, constructing a situation assessment index system may include: quantifying the situation assessment elements to obtain attribute vectors for the situation assessment elements; clustering the situation assessment elements with similar functions based on the attribute vectors; using the analytic hierarchy process (AHP) to calculate the importance of the situation assessment elements in each cluster, and selecting representative situation assessment elements based on the importance to construct the situation assessment index system.

[0035] In one example, the safety situation assessment model can be constructed through the following steps: acquiring historical safety data information related to the substation to construct the data layer of the safety situation assessment model; and constructing the assessment layer of the safety situation assessment model using the neural network based on the data layer. For example, based on the situation assessment index system, the neural network is used to perform information transfer and feature learning on the dataset in the data layer to train the safety situation assessment model; and the forward propagation of the neural network is used to optimize the safety situation assessment model to obtain the assessment layer.

[0036] In one example, an alarm is triggered if the current operational safety situation of the personnel and equipment is deemed unsafe.

[0037] This application addresses the problem of inaccurate safety situation assessment in the current substation safety operation environment. Starting from the behavior of operators and the characteristics of equipment operation, a safety situation assessment model is established, and deep learning methods are introduced and applied to various important aspects of substation safety operation situation assessment, such as situation element extraction and situation assessment. This achieves accurate, efficient and comprehensive safety operation situation assessment in the current operating environment, providing strong support for substation operation safety management.

[0038] Example 2

[0039] According to an embodiment of this application, another method for assessing the safe operating status of a substation is provided, such as... Figure 2 As shown, the method includes:

[0040] Step S202: Obtain information on operators, vehicles, and equipment.

[0041] By utilizing multi-source heterogeneous sensors to acquire raw safety status data of workers, vehicles, and equipment, information affecting operational safety status is extracted and converted into a unified data format to provide necessary data support for safety status assessment, model building, and trend prediction.

[0042] Step S204, design of security situation assessment model.

[0043] Using the information extracted above that affects the operational safety situation, a safety situation assessment model is constructed. By analyzing the parameters in the model, a comprehensive and quantitative description of the operational safety situation is obtained.

[0044] Figure 3 This is a design flowchart of the security situation assessment model according to an embodiment of this application, such as... Figure 3 As shown, the method includes:

[0045] Step S302: Construct the data layer.

[0046] The data layer is the foundational layer of the safety situation assessment model. It obtains data sources for safety situation assessment from safety data information from systems and equipment such as data acquisition devices, detection systems, and monitoring systems, providing a data foundation for assessing the safety situation of personnel and vehicles.

[0047] Step S304: Construct the evaluation layer.

[0048] The assessment layer is the core layer of the security situation assessment model, comprising three sub-layers: the situation assessment element extraction sub-layer, the situation assessment indicator system construction sub-layer, and the situation assessment sub-layer. The situation assessment element extraction sub-layer is used to extract situation elements, the situation assessment indicator system construction sub-layer is used to establish assessment indicators, and the situation assessment sub-layer is used to assess the security situation.

[0049] Step S306: Construct the knowledge layer.

[0050] The knowledge layer is the top layer of the model. It combines the security posture assessment results from the network domain and the behavior domain for integrated analysis to derive the overall security posture of the entire system.

[0051] Step S206: Extraction of security situation elements.

[0052] Security situation element extraction is the prerequisite and foundation for security situation assessment. Extracting and identifying factors that profoundly impact security situation assessment from a large amount of security data, and then forming situation assessment elements through statistical analysis, the core issue is how to achieve efficient dimensionality reduction and feature extraction of security data.

[0053] Figure 4 This is a flowchart of a situation feature extraction method based on a loss module compensation deep autoencoder according to an embodiment of this application. The method includes:

[0054] Step S402, input raw data.

[0055] Collect raw data for security situation awareness.

[0056] Step S404, data preprocessing.

[0057] Data preprocessing is performed on the collected raw data, mainly including label segmentation and normalization.

[0058] Step S406: Layer-by-layer loss compensation.

[0059] The processed data is then input into a layer-by-layer loss compensation encoder for dimensionality reduction and feature extraction. The layer-by-layer loss compensation encoder adds a loss compensation module to each encoding layer to compensate for the loss of feature information of security situation assessment elements during encoding.

[0060] A deep autoencoder is an unsupervised learning model that uses the backpropagation algorithm and optimization methods to guide the neural network to learn a mapping relationship in order to obtain a reconstructed output data.

[0061] The specific process of loss compensation using a depth autoencoder is as follows:

[0062] 1) Encode the i-th layer of the depth autoencoder. When the output data of the (i-1)-th layer enters the i-th encoding layer, the resulting dimension is M. i Output data Its expression is:

[0063] 2) To restore the decoded data, in order to obtain the loss of the encoded data during the encoding process, it is necessary to analyze the output data of the i-th encoding layer. By using the corresponding decoding layer to reconstruct the data, we obtain output data with the same dimension as the (i-1)th encoding layer. The expression is: Among them, decX i M represents the data restored by the i-th coding layer. i-1 This represents the data dimension of the (i-1)th encoding layer.

[0064] 3) Obtain the loss value, calculate the loss of feature information using the loss function, and use the input data of the i-th coding layer. With decX i The difference is calculated, and the i-th encoding layer is used to encode the differenced data to obtain the loss value lossX. i The expression is:

[0065] 4) Compensate for the loss value by replacing the corresponding loss value lossX of the i-th coding layer during the coding process. i The compensation is applied to the output data of the i-th coding layer by summation, as shown in the expression:

[0066] Step S408: Determine whether the training period is met.

[0067] The encoded situation assessment feature information is passed to the decoder for decoding and reconstruction. Then, the difference between the reconstructed data and the original data is evaluated by minimizing the MSE loss function. The situation assessment feature extraction model is then trained, where the MSE loss function can be expressed as:

[0068]

[0069] Where X represents the input variable, X′ represents the output variable, L(X,X′) represents the loss function, and k represents the number of samples. j Let X′ represent the input variable for the j-th sample. j This represents the output variable for the j-th sample.

[0070] Determine whether the situation assessment element extraction model meets the training period requirements. If not, proceed to step S406; otherwise, proceed to step S410.

[0071] Step S410: Extract situational elements.

[0072] The trained situation assessment element extraction model is used to extract situation assessment elements from the overall data to obtain the final set of assessment factors.

[0073] Step S412: Output the evaluation factors.

[0074] Step S208: Construct a situation assessment indicator system.

[0075] The situation assessment indicator system is a unified whole composed of several interconnected and complementary indicators, used to assess the current security situation and predict the trend of the situation.

[0076] First, a hierarchical structure model of the indicator system is established; second, the evaluation factors are quantified; third, the evaluation indicators are clustered; and finally, the evaluation indicators are optimized. The construction process is as follows: Figure 5 As shown, the specific steps include:

[0077] Step S502: Establish a hierarchical structure model for the indicator system.

[0078] First, it is necessary to determine the comprehensive indicators in the target layer and the criterion layer. The target layer is the overall safety situation, and the criterion layer is the behavioral information of the operators.

[0079] Step S504: Quantify the evaluation factors.

[0080] Quantification is a prerequisite for indicator clustering. Clustering requires the similarity of the attributes of the evaluation factors. The source information of the indicators is used as the evaluation attribute to represent the membership relationship between the evaluation factors and the criteria layer indicators, expressed as: source={s1,s2,s3,…,s…} n}, where n is the number of types of source information, s1..s n These represent different sources of indicator information, where "source" represents the set of indicator source information.

[0081] This leads to the attribute vector f = (f1, f2, f3, ... f) of the evaluation factors. n ), which represents the membership relationship between the evaluation factor and each element in source. The larger the value of an element in f, the more likely the evaluation factor is to belong to the corresponding source information.

[0082] Step S506: Cluster the evaluation indicators.

[0083] Evaluation factors with similar functions are automatically clustered into one category, and the clustering results form a hierarchical relationship with the comprehensive indicators, replacing the direct selection of evaluation factors by humans, thereby reducing the subjectivity in constructing the indicator system.

[0084] Step S508: Optimize the evaluation indicators.

[0085] The indicator system is optimized by screening evaluation factors. The importance of evaluation factors in each category is calculated using the analytic hierarchy process (AHP), and representative evaluation factors are selected to construct the final indicator system.

[0086] Step S210, security situation assessment.

[0087] Safety situation assessment evaluates the current operational safety situation by analyzing the safety status of the assessment model and operating conditions. Based on the safety situation assessment index system constructed above, it analyzes factors such as equipment information and event information, and uses situation assessment methods for feature learning and representation to improve the performance of safety situation assessment.

[0088] This application proposes a situation assessment method based on the Subagging algorithm and a deep recurrent neural network to address the problem of long-term dependency of assessment data. The method flow is as follows: Figure 6 As shown, the specific implementation steps are as follows:

[0089] Step S602: Use the Subagging algorithm to sample the evaluation dataset.

[0090] The Subagging algorithm, also known as the bagging algorithm, is a group learning algorithm in the field of machine learning. This algorithm uses a random subsampling scheme instead of the bootstrap resampling scheme in the Bagging algorithm. It allows control over the ratio of the original data and its feature values ​​during the sampling process, and also controls the ratio of positive samples to negative samples in the sampled dataset.

[0091] For a given training sample S, assuming its feature set is F, each time, training subsamples are randomly sampled from the training sample S and the feature set F according to a set ratio. The subsample data size is m = αS (0 < α < 1), where α represents the sampling ratio of the dataset, and the feature size is f = βF (0 < β < 1), where β represents the sampling ratio of the feature set. This process is repeated n times to obtain n training sample sets.

[0092] Step S604: Obtain the evaluation model for the sample set.

[0093] After obtaining the sample set, an evaluation model is obtained by using one sample set at a time. For n sample sets, n evaluation models can be obtained.

[0094] Step S606: Initialize the deep recurrent neural network.

[0095] Deep recurrent neural networks (RNNs) are a type of neural network with short-term memory capabilities. In RNNs, neurons can receive information not only from other neurons but also from themselves, forming a network structure with loops.

[0096] Initialize the deep recurrent neural network and use model averaging to obtain the final evaluation model.

[0097] Step S608: Train the evaluation model.

[0098] The deep recurrent neural network algorithm is used to extract the dataset from the evaluation model in step S606 for information transfer and feature learning. Assume the input x... t For the security situation assessment dataset, the assessment state h transmitted from the previous moment is used. t-1 and the input x of the current node t To get the status of the reset door and update the door:rt =σ(w r ·[h t-1 ,x t ]), where r t This indicates a reset gate neuron, where σ refers to the sigmoid activation function, and w r h represents the weight parameters of the gate neuron being reset. t-1 This indicates the evaluation state at the previous moment. t =σ(w z ·[h t-1 ,x t ]), where z t This indicates updating the gate neuron, w z This indicates that the weight parameters of the gate neuron are being updated.

[0099] Step S610: Output the optimal parameter combination.

[0100] Get the current candidate set The state at the current moment as remembered above: Where tanh represents the activation function used by the memory gate neuron, w h This represents the weight parameters of the memory gate neuron.

[0101] Step S612: Optimize model training.

[0102] Update the deep recurrent neural network units in the memory phase. Where, h t This indicates the current evaluation status.

[0103] The forward propagation output y is obtained using the update equation. t =σ(w o ·h t ), where w o y represents the output gate weight parameters. t This indicates the output result of the forward propagation.

[0104] During the forward propagation of a deep recurrent neural network, w is learned. r w z w h w o Parameters: w r =w rx +w rh , where w rx w represents the weight parameter under the reset gate data input state. rh This indicates resetting the gate evaluation weight parameters; w z =w zx +w zh , where w zx This indicates the weight parameters w under the updated gate data input state.zh This indicates updating the gate evaluation weight parameters; w h =w hx +w hh , where w hx w represents the weight parameters under the data input state of the memory gate. hh This represents the evaluation weight parameters of the memory gate. The input to the output layer is obtained using the trained weight parameters: Where w o Let h represent the output weight parameters, and h represent the output evaluation state. The output of the output layer is: At this time This is the security situation score at the current moment, and the security situation assessment level can be obtained based on this value.

[0105] Step S614 yields the situation assessment model.

[0106] Step S212, Behavior detection and assessment.

[0107] Real-time acquisition of operator and equipment operational behaviors. Utilizing a situational safety assessment model to detect and evaluate these behaviors. For example, processed operational behavior data is input into a trained situational safety assessment model, and threatening and potentially threatening behaviors are detected by calculating behavioral errors. The severity of the threat is then further assessed based on the detected threatening behaviors.

[0108] Step S214, intelligent early warning.

[0109] Based on the above behavioral detection and assessment results, intelligent early warning is provided for dangerous operations within the substation.

[0110] For example, to determine safe distances, real-time laser scanning data from the substation is collected using 3D LiDAR. Image processing algorithms are then used to obtain 3D point cloud data of the physical space. A constructed 3D model of the substation is then matched and calculated with the original 3D model point cloud data to determine the distances between various external equipment, vehicles, and personnel and the energized parts of existing equipment. This distance is then combined with operational safety measures to determine whether the energized parts of the equipment are currently energized. When the safe distance is less than the specified safe distance, an early warning is issued to the work site and remote monitoring terminal, prompting personnel to take safety measures.

[0111] The embodiments of this application solve the following technical problems and have the following technical effects:

[0112] Because operational safety situation awareness data is characterized by multi-source heterogeneity, large volume, numerous features, and nonlinearity, assessment factors cannot accurately reflect the risks present in the behavioral state. The processing capabilities for multi-feature, high-dimensional nonlinear data have limitations, leading to inaccurate and inefficient extraction of safety situation assessment elements. This application proposes a situation assessment element extraction method based on a layer-by-layer loss-compensated deep autoencoder, which can effectively reduce the loss of feature information in assessment elements.

[0113] This application proposes a situation assessment method based on the Subagging algorithm and deep recurrent neural networks, which is suitable for assessing large datasets with long-term data sample dependencies. This method can effectively perform feature learning and representation, thereby improving the performance of security situation assessment.

[0114] The embodiments of this application achieve a comprehensive, accurate, and effective safety situation assessment, ensuring the safety of substation workers and vehicles.

[0115] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0116] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0117] Example 3

[0118] According to an embodiment of this application, a substation safety operation status assessment device is also provided, such as... Figure 7 As shown, the device includes an acquisition module 72 and an evaluation module 74.

[0119] The acquisition module 72 is configured to acquire safety status data of personnel and equipment operating in the substation. This safety status data refers to data related to operational safety for the personnel and equipment within the substation.

[0120] The assessment module 74 is configured to assess the current operational safety status of the personnel and equipment based on the safety status data and using a pre-built safety status assessment model based on a neural network.

[0121] The substation safety operation status assessment device in this embodiment can realize the substation safety operation status assessment method in the above embodiments, therefore, it will not be described again here.

[0122] Example 4

[0123] Embodiments of this application also provide a storage medium configured to store program code for performing the methods in embodiments 1 and 2 above.

[0124] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0125] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0126] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0127] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0128] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0129] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0130] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0131] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for assessing the safe operating status of a substation, characterized in that, include: Obtain safety status data of substation personnel and equipment; Based on the aforementioned safety situation data, a safety situation assessment model pre-built based on a neural network is used to assess the current operational safety situation of the personnel and equipment. The safety situation data refers to the data related to the safety of the personnel and equipment in the substation. The security situation assessment model is obtained as follows: The assessment dataset is sampled using a subagging algorithm. This subagging algorithm employs a random subsampling scheme instead of the bootstrap resampling scheme in the Bagging algorithm. It also controls the ratio of the original data and its feature values ​​during sampling, as well as the ratio of positive to negative samples in the sampled dataset. For a given training sample S, assuming its feature set is F, training subsamples are randomly extracted from the training sample S and feature set F according to a set ratio each time. The subsample data size is m = αS (0 < α < 1), where α represents the sampling ratio of the dataset, and the feature size is f = βF (0 < β < 1), where β represents the sampling ratio of the feature set. This process is repeated n times to obtain n training sample sets. Based on a set of n training samples, n evaluation models are obtained. A deep recurrent neural network is initialized, and the final evaluation model is obtained by averaging the models.

2. The method according to claim 1, characterized in that, Before constructing the security situation assessment model, the method further includes: Extract factors that influence security situation assessment from security data and information, and use them as situation assessment elements; Based on the aforementioned situation assessment elements, a situation assessment index system is constructed, wherein the situation assessment index system is used to assess the safety status of the substation.

3. The method according to claim 2, characterized in that, Factors influencing security situation assessment are extracted from the security data information, including: The security data information is preprocessed, including label segmentation and normalization. The preprocessed security data information is then compensated for loss layer by layer. Factors that influence security situation assessment are extracted from the security data information after layer-by-layer loss compensation.

4. The method according to claim 3, characterized in that, The preprocessed security data information is subjected to layer-by-layer loss compensation, including: The preprocessed security data information is encoded; The encoded security data information is restored to obtain the loss of the encoded security data information during the encoding process; Based on the obtained loss, the loss of the feature information of the security data information is calculated using the loss function; The loss of the feature information of the elements is compensated to the situation assessment elements.

5. The method according to claim 1, characterized in that, Based on the aforementioned situation assessment elements, a situation assessment index system is constructed, including: The situation assessment elements are quantified to obtain the attribute vectors of the situation assessment elements; Based on the attribute vector, the situation assessment elements with similar functions are clustered; The importance of the situation assessment elements in each cluster is calculated using the analytic hierarchy process (AHP), and representative situation assessment elements are selected based on the importance to construct the situation assessment index system.

6. The method according to any one of claims 1 to 5, characterized in that, If the current operational safety situation of the personnel and equipment is deemed unsafe, an alarm will be triggered.

7. A substation safety operation status assessment device, characterized in that, The acquisition module is configured to acquire safety status data of personnel and equipment in the substation. The assessment module is configured to assess the current operational safety status of the personnel and equipment based on the safety status data and using a pre-built safety status assessment model based on a neural network. The safety situation data refers to the data related to the safety of the personnel and equipment in the substation. The security situation assessment model is obtained as follows: The assessment dataset is sampled using a subagging algorithm. This subagging algorithm employs a random subsampling scheme instead of the bootstrap resampling scheme in the Bagging algorithm. It also controls the ratio of the original data and its feature values ​​during sampling, as well as the ratio of positive to negative samples in the sampled dataset. For a given training sample S, assuming its feature set is F, training subsamples are randomly extracted from the training sample S and feature set F according to a set ratio each time. The subsample data size is m = αS (0 < α < 1), where α represents the sampling ratio of the dataset, and the feature size is f = βF (0 < β < 1), where β represents the sampling ratio of the feature set. This process is repeated n times to obtain n training sample sets. Based on a set of n training samples, n evaluation models are obtained. A deep recurrent neural network is initialized, and the final evaluation model is obtained by averaging the models.

8. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed, the computer performs the method as claimed in any one of claims 1 to 5.

Images