A risk warning method and related device based on operation risk points

Abstract

A risk warning method and related device based on operation risk points, including obtaining the content of all key risk points from the text content of the operation plan; normalizing the content of the key risk points; screening out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions; matching the key risk point list obtained by the normalization process with the risk point list of each screened historical accident event to obtain the number of matched risk points; calculating the matching degree of the operation plan risk point and the matching degree of the historical accident event risk point according to the number of risk points; combining the matching degrees of the two risk points to obtain a comprehensive matching degree, and selecting the historical accident event with the highest comprehensive matching degree to conduct risk warning for the current operation plan. Starting from the operation plan, the present invention provides efficient help for the operation safety management of the current operation plan by identifying the risk points in the operation plan and matching and warning with the historical accident events.

Description

Technical Field

[0001] The present invention belongs to the technical field of risk warning for electric power operations, and in particular relates to a risk warning method based on operation risk points and related devices. Background Art

[0002] In the power industry, operational safety has always been a crucial consideration, especially given the complex power operation environment and the many potential dangers. Any slight negligence may lead to serious safety accidents, affecting personnel safety, power grid stability and social and economic activities. Therefore, it is particularly urgent to adopt advanced risk management strategies and technical means to achieve effective early warning and control of potential risks in the operation process.

[0003] Traditionally, the safety management of power operations relies on manual experience and post-accident analysis, which is not only inefficient but also difficult to systematically prevent accidents. With the development of technologies such as big data and artificial intelligence, power companies have begun to explore the use of technical means to improve the level of intelligent safety management and achieve a shift from passive response to active prevention. Summary of the invention

[0004] In view of this, the present invention takes the operation plan as the starting point and provides a risk warning method and related devices based on operation risk points to solve the above-mentioned problems existing in the existing safety management methods of power operations.

[0005] In order to solve the above technical problems, the technical solutions provided by the present invention are as follows:

[0006] In a first aspect, the present invention provides a risk warning method based on operation risk points, comprising the following steps:

[0007] Get the content of all key risk points from the text content of the work plan;

[0008] Standardize the content of key risk points to obtain a list of key risk points described in standardized terms;

[0009] Match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions;

[0010] Matching the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, wherein the risk point list of the historical accident event includes the risk points corresponding to the historical accident event;

[0011] The matching degree of the operation plan risk points and the matching degree of the historical accident event risk points are calculated based on the number of risk points. The matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all risk points in the current operation plan. The matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all corresponding risk points in different historical accident events.

[0012] The comprehensive matching degree is obtained by combining the matching degree of the risk points of the operation plan and the matching degree of the risk points of historical accident events, and the historical accident event with the highest comprehensive matching degree is selected to provide risk warning for the current operation plan.

[0013] Furthermore, the content of key risk points is standardized, including:

[0014] Use similarity matching method to match the text content of all key risk points with the pre-built key risk point standard library one by one. The key risk point standard library includes all key risk points described in standard terms.

[0015] The key risk points whose semantic similarity meets the set conditions are selected from the key risk point specification library as matching results, and the matching results of all key risk points are formed into a key risk point list.

[0016] Furthermore, the key risk point specification library is associated with the risk problem library, which contains several different types of risk problems. Each key risk point in the key risk point specification library is associated with a risk problem in the risk problem library.

[0017] Furthermore, for the work plan without key risk points filled in, the work plan is predicted through the pre-trained multi-label classification model to obtain the content of all key risk points. The prediction process of the multi-label classification model is as follows:

[0018] Convert the text of the job plan into several sentence vectors;

[0019] Input several sentence vectors into several pre-trained binary classification models to obtain the prediction results of all binary classification models. Several binary classification models are sub-models of the multi-label classification model. Each binary classification model corresponds to a risk point type and is used to determine whether the current input involves the risk point type corresponding to each model.

[0020] The prediction results of the binary classification models involving the corresponding risk points are summarized to obtain the prediction results of the multi-label classification model on the key risk points.

[0021] Furthermore, before using the pre-trained multi-label classification model for prediction, the multi-label classification model is also evaluated, and prediction is performed only after the evaluation criteria are met. The evaluation methods include: macro-average and / or accuracy scoring based on model prediction probability;

[0022] The calculation expression of macro-average is as follows:

[0023]

[0024] In the formula, acc represents the classification accuracy of the binary classification model on a single label; TP represents the number of positive samples predicted by the binary classification model as positive samples; TN represents the number of negative samples predicted by the binary classification model as negative samples; FP represents the number of negative samples predicted by the binary classification model as positive samples; FN represents the number of positive samples predicted by the binary classification model as negative samples; Represents the number of all samples; Micro_acc represents the macro average of the accuracy of multiple label classifications of n binary classification models;

[0025] The accuracy score based on the model's predicted probability is calculated as follows:

[0026]

[0027] In the formula, score represents the accuracy score, p(x) represents the true probability distribution under a category, q(x) represents the probability distribution predicted by the model under a category; m represents m labels, n represents n test data, i and j are the label and test data indexes respectively.

[0028] Furthermore, the comprehensive matching degree is obtained by combining the matching degree of the operation plan risk points with the matching degree of the historical accident event risk points, including:

[0029] Calculate the ratio of the length of the key risk point list of the operation plan and the length of the risk point list of historical accident events to the sum of the total length of the two lists, and use the two ratio values ​​as two scoring values;

[0030] Multiply the two scoring values ​​corresponding to the matching degree of the operation plan risk point and the matching degree of the historical accident event risk point respectively to obtain the real scores of the two matching degrees;

[0031] The actual scores of the two matches are added together to get the score of the comprehensive match.

[0032] Furthermore, the current operation plan is matched with all historical accident events, including:

[0033] Match the current operation plan with all historical accident events in sequence according to the professional attributes, level attributes and type attributes of the operation;

[0034] Filter out historical accident events belonging to the same profession, level and type.

[0035] In a second aspect, the present invention provides a risk warning device based on operation risk points, comprising:

[0036] The risk point extraction module is used to obtain the contents of all key risk points from the text content of the operation plan;

[0037] The normalization processing module is used to normalize the content of key risk points and obtain a list of key risk points described in standardized terms;

[0038] The first matching module is used to match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions;

[0039] The second matching module is used to match the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, and the risk point list of the historical accident event includes the risk points corresponding to the historical accident event;

[0040] The matching degree calculation module is used to calculate the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points according to the number of risk points. The matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all the risk points in the current operation plan, and the matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all the corresponding risk points in different historical accident events.

[0041] The early warning module is used to combine the risk point matching degree of the operation plan and the risk point matching degree of historical accident events to obtain a comprehensive matching degree, and select the historical accident event with the highest comprehensive matching degree to issue a risk warning for the current operation plan.

[0042] Accordingly, the present invention also provides a computer device, the device comprising a processor and a memory:

[0043] The memory is used to store the computer program and send the instructions of the computer program to the processor;

[0044] The processor executes a risk warning method based on operation risk points as described in the first aspect according to the instructions of the computer program.

[0045] Correspondingly, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, a risk warning method based on operational risk points as in the first aspect is implemented.

[0046] In summary, the present invention provides a risk warning method and related devices based on operation risk points, including obtaining the content of all key risk points from the text content of the operation plan; normalizing the content of the key risk points to obtain a list of key risk points described in standardized terms; matching the current operation plan with all historical accident events, screening out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions; matching the list of key risk points with the risk point list of each screened historical accident event to obtain the number of matched risk points; calculating the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points according to the number of risk points; combining the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points to obtain a comprehensive matching degree, and selecting the historical accident event with the highest comprehensive matching degree to conduct risk warning for the current operation plan. Starting from the operation plan, the present invention provides efficient help for the operation safety management of the current operation plan by identifying the risk points in the operation plan and matching and warning with the historical accident events. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A flow chart of a risk warning method based on operation risk points provided by an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of extracting key risk points provided by an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of a multi-label classification model for predicting risk points provided by an embodiment of the present invention;

[0051] Figure 4 A schematic diagram of the normalization effect of key risk points provided by an embodiment of the present invention;

[0052] Figure 5 A schematic block diagram of the early warning process provided by an embodiment of the present invention;

[0053] Figure 6 A block diagram of a risk warning device based on operation risk points provided by an embodiment of the present invention;

[0054] Figure 7 A block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] See also Figure 1 This embodiment provides a risk warning method based on operation risk points, including the following steps:

[0057] S11: Obtain the content of all key risk points from the text content of the work plan.

[0058] It should be noted that this step automatically extracts all key risk point information mentioned in the work plan text to be executed through natural language processing (NLP) and other technologies. These risk points may include but are not limited to the working environment, equipment conditions, potential safety hazards in the operation steps, etc., which are the basic data for risk assessment.

[0059] S12: Standardize the contents of key risk points to obtain a list of key risk points described in standardized terms.

[0060] It should be noted that when power supply company personnel enter work plans, they usually fill in the key risk points before or at the end of the plan content. However, due to differences in thinking and language organization of different personnel, the description of the same key risk point may be different in different plan sheets. It is necessary to standardize the risk point description terminology. As one of the parameters of the risk warning model, the standardization of the risk point has a great impact on the accuracy of the model prediction.

[0061] The extracted risk points are standardized to eliminate differences in expression and unify them into standard terms recognized by the industry or enterprise. This ensures that the description of risk points is consistent and highly comparable, which facilitates subsequent matching and analysis.

[0062] S13: Match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions.

[0063] It should be noted that the current operation plan is matched with the accident events in the historical database to find historical accident cases under the same or similar operation conditions and operation types. This process can use big data analysis to narrow the matching range and improve the pertinence of early warning by comparing factors such as operation type and operation environment.

[0064] S14: Match the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, wherein the risk point list of each historical accident event includes the risk points corresponding to the historical accident event.

[0065] It should be noted that the standardized risk point list of the operation plan is matched with the risk point list of each historical accident event that has been screened. By comparing the risk points in the two lists, it is counted how many risk points are common, providing data support for the subsequent matching degree calculation.

[0066] S15: Calculate the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points according to the number of risk points. The matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all risk points in the current operation plan. The matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all corresponding risk points in different historical accident events.

[0067] It should be noted that based on the number of matched risk points, the risk point matching degree of the operation plan and the risk point matching degree of historical accident events are calculated separately. The risk point matching degree of the operation plan reflects the proportion of risk points identified in the current plan to all risk points in the plan, while the risk point matching degree of historical accident events refers to the proportion of these matched risk points in the corresponding historical accidents. These two indicators help quantify the similarity and severity of risks.

[0068] The intersection of the operation plan risk point list and the historical accident event risk point list is called intersection, and the number of matching tags is calculated as The calculation of the matching degree of the operation plan risk point is as follows:

[0069]

[0070] The calculation of the matching degree of historical accident event risk points is as follows:

[0071]

[0072] S16: Combine the risk point matching degree of the operation plan and the risk point matching degree of historical accident events to obtain a comprehensive matching degree, and select the historical accident event with the highest comprehensive matching degree to issue a risk warning for the current operation plan.

[0073] It should be noted that a comprehensive matching degree is calculated by combining the matching degree of the risk points of the operation plan and the matching degree of the risk points of historical accident events. This comprehensive matching degree can more comprehensively reflect the similarity and potential risk size between the current operation plan and historical accident events. The historical accident event with the highest comprehensive matching degree is selected as the warning case, and a risk warning is issued for the operation plan to remind operators and managers to pay attention to preventive measures to avoid the recurrence of similar accidents.

[0074] This embodiment provides a risk warning method based on operational risk points. Through an intelligent, data-driven risk warning mechanism, it achieves a shift from passive response to active prevention, greatly improving the safety management level of power operations. This method can not only quickly and accurately identify potential risks in operations, but also provide targeted historical cases as references to help formulate more effective preventive measures, thereby reducing the occurrence of accidents and ensuring personnel safety and the stability of power grid operation. At the same time, this method improves the efficiency and accuracy of risk assessment and reduces the burden of manual analysis through standardized processing and automated matching.

[0075] At present, some power supply bureaus have clearly required that the key risk points in the operation process be accurately filled in according to the standards of the "Transparent Control Measures for Personal Safety Risks at Operation Sites" in the filling of operation plans. For operation plans that have already filled in the key risk points in accordance with this specification, relevant information can be directly extracted without re-analysis or evaluation.

[0076] In the process of extracting risk points, a variety of technical means can be used, such as writing algorithms, regular expressions, and NLP natural language processing. These technologies can help us accurately identify keywords, intercept sentences according to punctuation marks, and remove non-critical information such as serial numbers and symbols. Taking a typical work plan as an example, the work task may include specific work content such as making and installing cable heads on the load side of a switch in a specific substation. In the description of the work plan, the key risk nodes of the work will be clearly listed, such as working near live equipment and near waters. By applying the above technical means, these key risk points can be quickly and accurately extracted from the work plan, such as "working near live equipment" and "near waters". These extracted risk points will provide important reference information for the power supply bureau, helping them to better understand and control safety risks in the operation process, thereby ensuring the safety of operators and the stable operation of power facilities. Extract key risk points directly from the work plan such as Figure 2 shown.

[0077] In some embodiments, for a work plan that does not have key risk points filled in, this embodiment proposes a method for obtaining key risk points. This method uses historical work plans that have key risk points filled in as sample data, summarizes the semantic rules in the sample data, trains a multi-label classification model, and predicts the possible risk points of a new work plan through the classification model. The prediction process of the multi-label classification model is as follows:

[0078] Specifically, firstly, based on the pre-trained model, the plan text is converted into a fixed number of sentence vectors, and the sentence vectors are used as input data for the subsequent classification model.

[0079] The method of converting the plan text into a numerical sentence vector is called word vector embedding, which is very common in the field of natural language processing. Word vector embedding can capture the semantic and grammatical relationship of words, map the semantics of words into a continuous vector space, and after obtaining the word vector of each word, the word vectors of all words in the sentence can be summed and averaged to obtain the sentence vector of the sentence.

[0080] Next, train the multi-label classification model. Specifically, for each risk category, create and train a binary classification model to infer whether the work plan involves the risk category. The binary classification model is implemented using a fully connected neural network, and each binary classification model takes several sentence vectors obtained in the previous step as input.

[0081] The reason for creating multiple binary classification models is that a work plan may have several risk points, each of which corresponds to a certain risk category; and the multi-classification model only supports classifying samples into multiple mutually exclusive categories. Therefore, the classification task is decomposed into a combination of several binary classification tasks, thereby indirectly realizing multi-label classification.

[0082] Finally, by summarizing the prediction results output by multiple classification models, we can predict the possible risk points in the new work plan.

[0083] The process of predicting risk points using a multi-label classification model is as follows: Figure 3 As shown, including:

[0084] S21: Convert the text of the homework plan into several sentence vectors;

[0085] S22: Inputting the sentence vectors into the pre-trained binary classification models to obtain prediction results of all the binary classification models. The binary classification models are all sub-models of the multi-label classification model. Each binary classification model corresponds to a risk point type, and is used to determine whether the current input involves the risk point type corresponding to each model.

[0086] S23: Summarize the prediction results of the binary classification models involving corresponding risk points to obtain the prediction results of the multi-label classification model on the key risk points.

[0087] In a further embodiment, before using the pre-trained multi-label classification model for prediction, the multi-label classification model is also evaluated, and prediction is performed only after the evaluation criteria are met, and the evaluation method includes: macro-average and / or accuracy scoring based on model prediction probability;

[0088] The calculation expression of macro-average is as follows:

[0089]

[0090] In the formula, acc represents the classification accuracy of the binary classification model on a single label; TP represents the number of positive samples predicted by the binary classification model as positive samples; TN represents the number of negative samples predicted by the binary classification model as negative samples; FP represents the number of negative samples predicted by the binary classification model as positive samples; FN represents the number of positive samples predicted by the binary classification model as negative samples; Represents the number of all samples; Micro_acc represents the macro average of the accuracy of multiple label classifications of n binary classification models;

[0091] The accuracy score based on the model's predicted probability is calculated as follows:

[0092]

[0093] In the formula, score represents the accuracy score, p(x) represents the true probability distribution under a category, q(x) represents the probability distribution predicted by the model under a category; m represents m labels, n represents n test data, i and j are the label and test data indexes respectively.

[0094] In some embodiments, a planned processing method is proposed. The method realizes standardized processing of key risk points by constructing a key risk point standard library. Specifically, the constructed key risk point database uses a unified term to represent the description of each risk point. The standardized processing process includes:

[0095] S31: A similarity matching method is used to match the text content of all key risk points with a pre-built key risk point specification library one by one. The key risk point specification library includes all key risk points described in standard terms.

[0096] It should be noted that this step can be implemented by converting each key risk point of the work plan into a word vector, and then matching the key risk points in the dictionary library based on the cosine similarity of the semantics.

[0097]

[0098] Represents the word vector of text 1, Word vector representing text 2.

[0099] The specific steps of matching the key risk points in the dictionary library based on cosine similarity are as follows: First, calculate the cosine similarity between the key risk points of a work plan and all the key risk points in the dictionary library.

[0100] S32: Select key risk points whose semantic similarity meets the set conditions from the key risk point specification library as matching results, and form a key risk point list with the matching results of all key risk points.

[0101] It should be noted that in this step, the item with the largest similarity greater than a certain threshold is selected as the matching item.

[0102] S33: The normalization process can use a similarity matching algorithm to match the extracted key risk nodes with the standard library, with the aim of unifying the descriptions of various risk nodes into one description term.

[0103] For example, the model is asked to infer which risk point in the standard library matches the "live work" extracted from the work content of the plan. The matching results are: 87% similarity with "live work"; 64% similarity with "work near live equipment"; 61% similarity with "height work (including climbing and pole tower)"... Finally, it is judged that the risk point described in the text is "live work". The effect of normalization of key risk points is as follows: Figure 4 shown.

[0104] In a further embodiment, the key risk point specification library is associated with the risk problem library, and the risk problem library contains several different types of risk problems, and each key risk point in the key risk point specification library is associated with the risk problem in the risk problem library. The risk problem library extracts the causes of casualties from historical accident reports, and the summary includes the following: mechanical injuries, burns, electric shocks, vehicle injuries, pole tower collapses, gunpowder explosions, fires, falls from heights, explosions of power equipment, lifting injuries, and object strikes.

[0105] The key risk point specification library can be based on the key risk point description in the "Transparent Control Measures for Personal Safety Risks at Work Sites" as the standard. Key risk nodes include personnel qualifications, inspection of tools and personal protective equipment, high-altitude operations (including climbing and pole towers), electrical grounding, operations near live equipment, low-voltage live disconnection and connection, live operations, 10kV special transformer metering device transformation without power outages, scaffolding erection, dismantling, entering switch cabinet operations, lifting and hoisting and installation and disassembly of lifting machinery, clearing tree obstacles without power outages, temporary power supply, erecting and removing pole towers, placing, removing and tightening wires, electrical measurements, crossing railways, highways, navigable rivers and other important and special crossing operations, pipe jacking operations in areas near or crossing gas, oil pipelines, electricity, communications, water pipes, etc., operations near waters, open fire, acceptance of hidden projects in important parts, shield engineering, operations near flammable and explosive places, high formwork, deep foundation pits, open-cut tunnel operations, limited space operations, and cable cutting operations. The following table is a comparison of some key risk points and risk issues.

[0106]

[0107] In some embodiments, the comprehensive matching degree is obtained by combining the matching degree of the operation plan risk point and the matching degree of the historical accident event risk point, including:

[0108] S41: Calculate the ratio of the length of the key risk point list of the operation plan and the length of the risk point list of historical accident events to the sum of the total length of the two lists respectively, and use the two ratio values ​​as two scoring values.

[0109] This step determines the total score of the two indicators based on the length of the risk point list of the operation plan and the length of the risk point list of historical accident events:

[0110]

[0111] S42: Multiply the two scoring values ​​corresponding to the matching degree of the operation plan risk point and the matching degree of the historical accident event risk point respectively to obtain the real scores of the two matching degrees.

[0112] This step uses the matching scores of the work plan and historical accident events as the score ratio to calculate the actual scores of the current matching task in the two matching degrees:

[0113]

[0114] S43: Add the two real matching scores together to obtain a comprehensive matching score.

[0115] This step adds up the actual scores to get the final matching score:

[0116]

[0117] In some embodiments, matching the current operation plan with all historical incident events includes:

[0118] S51: Match the current operation plan with all historical accident events in sequence according to the professional attributes, level attributes and type attributes of the operation;

[0119] S61: Filter out historical accident events of the same profession, level and type.

[0120] Specifically, first, professional matching is performed to extract the "professional" attribute of the operation plan, where the profession includes substation, distribution, transmission, infrastructure, etc. The profession involved in the accident event is directly matched through the attribute name. Secondly, voltage level matching is performed, extracted from the work location or work content, and the voltage level with the word kV is extracted through regular expressions, and the extracted voltage level is compared with the voltage level of the historical accident event. Then the operation type is matched, the "operation type" attribute of the operation plan is extracted, and the extracted operation type is compared with the operation type of the historical accident event. The operation types include pre-test, regular inspection, maintenance, maintenance, troubleshooting, operation, installation\replacement operation, repair, technical transformation, emergency repair, substation civil engineering, substation electrical installation, transmission line, distribution network engineering, routine test, special test, relay protection, automation, communication engineering, marketing, network security, and others. Data preprocessing is required before matching, and the operation type of the operation plan is classified according to the operation type label of the historical accident event.

[0121] Secondly, extract the key risk points of the operation plan and match them with the risk points of historical accident events. Based on the key risk points extracted from the operation plan, match the risk points of each accident event. Then, when the current operation plan and the historical accident event meet the requirements of profession, voltage level, and operation type, risk point matching is performed. By calling the risk point matching algorithm, the risk points of each operation plan are matched with all risks in the accident event risk library. When the operation plan and the information of the historical accident event meet the requirements of profession, voltage level, operation type, and risk points, an early warning message is generated to remind the operators to pay attention to the risks and remind the safety supervision department to strengthen safety measures, prevention, control, and supervision.

[0122] In combination with the above embodiment, in a possible implementation, the early warning process is as follows: Figure 5As shown. The early warning process uses the operation plan as the source, extracts or identifies the key risk points of the operation, and matches them with the operation risk points associated with historical accident events. When the similarity of the operation risk reaches a certain threshold, a warning result is generated and the risk of the operation is prompted. Specifically, in order to effectively prevent safety accidents in power operations, based on rich historical accident data, common risk issues are deeply analyzed and summarized. On this basis, for each operation plan, potential risk points are carefully extracted and accurately identified. The identification of these risk points not only covers considerations in professional fields, but also involves multiple dimensions such as operation type and operation voltage level to ensure the comprehensiveness and accuracy of risk assessment.

[0123] Subsequently, with risk points as the core, combined with key factors such as professional characteristics, operation type and voltage level, an efficient risk assessment model was constructed. This model aims to predict the possible risks in the operation plan through scientific methods, thereby providing strong decision-making support for operation safety.

[0124] When the risk assessment model is run, it will automatically output a detailed risk report. The report contains basic information such as the plan number, management unit, and work unit. More importantly, it will clearly point out the risk issues and their basis, so that relevant personnel can quickly understand where the risks lie. In addition, the report will also list similar historical accident cases, including violation codes, violation descriptions, accident processes, response measures taken, and problems exposed, etc., to provide operators with valuable lessons and help them better avoid risks and ensure operational safety.

[0125] Based on the same inventive concept, the embodiment of the present application also provides a risk warning device based on operation risk points for implementing the risk warning method based on operation risk points involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiment of the risk warning device based on operation risk points provided below can refer to the limitations of the risk warning method based on operation risk points above, and will not be repeated here.

[0126] See also Figure 6 , this embodiment provides a risk warning device based on operation risk points, including:

[0127] The risk point extraction module is used to obtain the contents of all key risk points from the text content of the operation plan;

[0128] The normalization processing module is used to normalize the content of key risk points and obtain a list of key risk points described in standardized terms;

[0129] The first matching module is used to match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions;

[0130] The second matching module is used to match the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, and the risk point list of the historical accident event includes the risk points corresponding to the historical accident event;

[0131] The matching degree calculation module is used to calculate the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points according to the number of risk points. The matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all the risk points in the current operation plan, and the matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all the corresponding risk points in different historical accident events.

[0132] The early warning module is used to combine the risk point matching degree of the operation plan and the risk point matching degree of historical accident events to obtain a comprehensive matching degree, and select the historical accident event with the highest comprehensive matching degree to issue a risk warning for the current operation plan.

[0133] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0134] Reference Figure 7 An embodiment of the present invention further provides a computer device, comprising: a memory and a processor and a computer program stored in the memory. When the computer program is executed on the processor, a risk warning method based on operation risk points as described in any one of the above methods is implemented.

[0135] The computer device may be a desktop computer, a notebook, a PDA, a cloud server or other computing device. The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that Figure 7 It is only an example of a computer device and does not constitute a limitation of the computer device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0136] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0137] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device. Further, the memory may include both an internal storage unit and an external storage device of the computer device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program. The memory may also be used to temporarily store data that has been output or is to be output.

[0138] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the risk warning method based on operation risk points as described in any one of the above methods is implemented.

[0139] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the camera / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, USB flash drive, mobile hard disk, disk or optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0140] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0141] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0142] In the embodiments disclosed in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0143] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A risk warning method based on operational risk points, characterized in that: The steps include: Get the content of all key risk points from the text content of the work plan; Standardize the contents of the key risk points to obtain a list of key risk points described in standardized terms; Match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions; Matching the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, the risk point list of the historical accident event including the risk points corresponding to the historical accident event; The matching degree of the operation plan risk points and the matching degree of the historical accident event risk points are calculated respectively according to the number of risk points. The matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all risk points in the current operation plan. The matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all corresponding risk points in different historical accident events. The comprehensive matching degree is obtained by combining the matching degree of the risk point of the operation plan and the matching degree of the risk point of the historical accident event, and the historical accident event with the highest comprehensive matching degree is selected to perform risk warning for the current operation plan; For the operation plan without key risk points filled in, the operation plan is predicted by the pre-trained multi-label classification model to obtain the content of all key risk points. The prediction process of the multi-label classification model is as follows: Convert the text of the job plan into a number of sentence vectors; Input the plurality of sentence vectors into a plurality of pre-trained binary classification models to obtain prediction results of all the binary classification models, wherein the plurality of binary classification models are sub-models of the multi-label classification model, and each of the binary classification models corresponds to a risk point type, and is used to determine whether the current input involves a risk point type corresponding to each model; The prediction results of the binary classification model involving corresponding risk points are summarized to obtain the prediction results of the multi-label classification model on key risk points.

2. The risk warning method based on operation risk points according to claim 1 is characterized in that: Standardize the content of the key risk points, including: Using a similarity matching method, the text contents of all the key risk points are matched one by one with a pre-built key risk point standard library, wherein the key risk point standard library includes all the key risk points described in standard terms; The key risk points whose semantic similarity meets the set conditions are selected from the key risk point specification library as matching results, and the matching results of all the key risk points are formed into the key risk point list.

3. The risk warning method based on operation risk points according to claim 2 is characterized in that: The key risk point specification library is associated with a risk problem library, the risk problem library contains several different types of risk problems, and each key risk point in the key risk point specification library is associated with a risk problem in the risk problem library.

4. The risk warning method based on operation risk points according to claim 1 is characterized in that: Before using the pre-trained multi-label classification model for prediction, the multi-label classification model is also evaluated, and prediction is performed only after the evaluation criteria are met. The evaluation methods include: macro-average and / or accuracy scoring based on model prediction probability; The calculation expression of the macro average is as follows: ； ； In the formula, acc represents the classification accuracy of the binary classification model on a single label; TP represents the number of positive samples predicted by the binary classification model as positive samples; TN represents the number of negative samples predicted by the binary classification model as negative samples; FP represents the number of negative samples predicted by the binary classification model as positive samples; FN represents the number of positive samples predicted by the binary classification model as negative samples; Represents the number of all samples; Micro_acc represents the macro average of the accuracy of multiple label classifications of n binary classification models; The accuracy score based on the model's predicted probability is calculated as follows: ； In the formula, score represents the accuracy score, p(x) represents the true probability distribution under a category, q(x) represents the probability distribution predicted by the model under a category; m represents m labels, n represents n test data, i and j are the label and test data indexes respectively.

5. The risk warning method based on operation risk points according to claim 1 is characterized in that: The comprehensive matching degree is obtained by combining the matching degree of the risk points of the operation plan and the matching degree of the risk points of the historical accident events, including: Calculate the ratio of the length of the key risk point list of the operation plan and the length of the risk point list of the historical accident events to the sum of the total length of the two lists respectively, and use the two ratio values ​​as two scoring values; Multiply the two scoring values ​​corresponding to the matching degree of the operation plan risk point and the matching degree of the historical accident event risk point respectively to obtain the real scores of the two matching degrees; The actual scores of the two matching degrees are added together to obtain the score of the comprehensive matching degree.

6. The risk warning method based on operation risk points according to claim 1 is characterized in that: Match current operational plans with all historical incident events, including: Match the current operation plan with all historical accident events in sequence according to the professional attributes, level attributes and type attributes of the operation; Filter out historical accident events belonging to the same profession, level and type.

7. A risk warning device based on operational risk points, characterized in that: include: The risk point extraction module is used to obtain the contents of all key risk points from the text content of the operation plan; A normalization processing module, used to normalize the contents of the key risk points to obtain a list of key risk points described in standardized terms; The first matching module is used to match the current operation plan with all historical accident events, and filter out several historical accident events that belong to the same operation type as the current operation plan under the same operation conditions; A second matching module is used to match the key risk point list with the risk point list of each screened historical accident event to obtain the number of matched risk points, and the risk point list of historical accident events includes the risk points corresponding to the historical accident events; A matching degree calculation module, used to calculate the matching degree of the operation plan risk points and the matching degree of the historical accident event risk points according to the number of risk points, wherein the matching degree of the operation plan risk points is used to indicate the proportion of the matched risk points in all risk points in the current operation plan, and the matching degree of the historical accident event risk points is used to indicate the proportion of the matched risk points in all corresponding risk points in different historical accident events; An early warning module is used to combine the matching degree of the risk points of the operation plan and the matching degree of the risk points of the historical accident events to obtain a comprehensive matching degree, and select the historical accident event with the highest comprehensive matching degree to perform risk early warning on the current operation plan; For the operation plan without key risk points filled in, the operation plan is predicted by the pre-trained multi-label classification model to obtain the content of all key risk points. The prediction process of the multi-label classification model is as follows: Convert the text of the job plan into a number of sentence vectors; Input the plurality of sentence vectors into a plurality of pre-trained binary classification models to obtain prediction results of all the binary classification models, wherein the plurality of binary classification models are sub-models of the multi-label classification model, and each of the binary classification models corresponds to a risk point type, and is used to determine whether the current input involves a risk point type corresponding to each model; The prediction results of the binary classification model involving corresponding risk points are summarized to obtain the prediction results of the multi-label classification model on key risk points.

8. A computer device, characterized in that: The device comprises a processor and a memory: The memory is used to store a computer program and send instructions of the computer program to the processor; The processor executes a risk warning method based on operation risk points as described in any one of claims 1 to 6 according to the instructions of the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, a risk warning method based on operation risk points as described in any one of claims 1 to 6 is implemented.

Images