Production risk early warning method and device, electronic equipment and storage medium

Abstract

The invention discloses a production risk early warning method and device, electronic equipment and a storage medium, and relates to the technical field of safety production. The scheme comprises the following steps: acquiring production data in a production process; dividing the production data into different labeled sample data sets according to a preset risk level; determining feature data of each marked sample data set according to an association degree between the production data and a preset risk level; determining a risk early warning model and an early warning monitoring data dimension based on the data dimension of the feature data; monitoring and collecting early warning key data in real time according to the early warning monitoring data dimension, and determining a prediction risk level based on the risk early warning model and the early warning key data; and determining a risk reason and a disposal suggestion according to the predicted risk level, and generating risk early warning according to the risk reason and the disposal suggestion. According to the embodiment of the invention, accurate early warning of the production risk is realized, and the management efficiency of the production risk is improved.

Description

Technical Field

[0001] The present invention relates to the field of production safety technology, and in particular to a production risk early warning method, device, electronic equipment and storage medium. Background Art

[0002] Dangerous chemicals are flammable, explosive, toxic, and highly corrosive. Safety accidents often occur during the production of hazardous chemicals, especially the casualties and property losses caused by major accidents, which are unacceptable. In order to effectively prevent and control safety accidents in the production of hazardous chemicals, enterprises need to strengthen the management of safety production. At present, the traditional risk measurement and early warning methods for hazardous chemical production safety mainly rely on manual inspections and experience judgment. However, the production process of hazardous chemicals is complex, and the traditional risk measurement and early warning methods for hazardous chemical production safety are difficult to deal with sudden hazardous chemical production safety accidents. Therefore, how to achieve accurate early warning of hazardous chemical production safety risks and improve the management efficiency of hazardous chemical production safety risks are urgent problems in the field of hazardous chemical production safety. Summary of the invention

[0003] The present invention provides a production risk early warning method, device, electronic equipment and storage medium, which solve the problem of low accuracy of production risk early warning and improve the management efficiency of production risks.

[0004] In one aspect of an embodiment of the present invention, a production risk early warning method is provided, comprising:

[0005] Obtain production data during the production process;

[0006] Divide production data into different labeled sample data sets according to preset risk levels;

[0007] Determine the characteristic data of each labeled sample data set according to the degree of association between the production data and the preset risk level;

[0008] Determine the risk warning model and warning monitoring data dimension based on the data dimension of the characteristic data;

[0009] According to the early warning monitoring data dimension, the key early warning data is collected in real time, and the predicted risk level is determined based on the risk early warning model and the key early warning data;

[0010] Determine the risk causes and disposal suggestions based on the predicted risk level, and generate risk warnings based on the risk causes and disposal suggestions.

[0011] Another aspect of an embodiment of the present invention provides a production risk early warning device, including:

[0012] A production data acquisition module is used to acquire production data during the production process;

[0013] A labeled sample data set acquisition module, used to divide the production data into different labeled sample data sets according to preset risk levels;

[0014] A feature data determination module, used to determine the feature data of each labeled sample data set according to the degree of association between the production data and the preset risk level;

[0015] A model determination module is used to determine the risk warning model and the warning monitoring data dimension based on the data dimension of the characteristic data;

[0016] The risk level determination module is used to monitor and collect early warning key data in real time according to the early warning monitoring data dimension, and determine the predicted risk level based on the risk early warning model and the early warning key data;

[0017] The risk warning generation module is used to determine the risk causes and disposal suggestions based on the predicted risk level, and generate risk warnings according to the risk causes and disposal suggestions.

[0018] Another aspect of an embodiment of the present invention provides an electronic device, including:

[0019] at least one processor;

[0020] and a memory communicatively coupled to the at least one processor;

[0021] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the production risk warning method described in any one of the embodiments of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions, and the computer instructions are used to enable a processor to implement the production risk warning method of any embodiment of the present invention when executed.

[0023] In an embodiment of the present invention, production data in a production process is obtained, a preset risk level is obtained, and based on the preset risk level, the obtained production data is classified and labeled, and the production data is divided into labeled sample data sets of different risk levels. In each labeled sample data set, the degree of correlation between the production data and the preset risk level is analyzed, and corresponding production data is selected in each labeled sample data set according to the degree of correlation between the production data and the preset risk level. The production data selected according to the degree of correlation is used as feature data, and a preset algorithm model is trained using the data dimension of the feature data. The trained algorithm model is used as a risk warning model, and the warning monitoring data dimension is determined according to the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data collected by real-time monitoring is input into the risk warning model. The risk warning model outputs a predicted risk level, and according to the predicted risk level, the risk cause and disposal suggestion are further determined. Factors of both the risk cause and disposal suggestion are comprehensively considered to generate a risk warning, and subsequent operations are performed based on the risk warning. In an embodiment of the present invention, production data obtained during the production process provides data support for the subsequent determination of predicted risk levels and generation of risk warnings. The data is divided into sample data sets of different risk levels, and feature data is selected from each labeled sample data set based on the degree of correlation between the production data and the preset risk level. A risk warning model is constructed based on the data dimension of the feature data. By constructing a risk warning model using the feature data screened out from the production data, the prediction accuracy of the risk warning model is increased, and the false alarm rate and missed alarm rate are reduced. The warning monitoring data dimension is determined based on the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data are input into the risk warning model. The preset risk level is obtained by real-time monitoring and collection of warning key data corresponding to the warning monitoring data dimension, and the warning key data are input into the risk warning model, thereby ensuring the accurate warning of the risk warning model. By predicting the risk level, the risk cause and disposal suggestions are determined, so that the enterprise can quickly formulate effective response strategies and improve the management efficiency of production risks.

[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 This is a flow chart of a production risk early warning method provided according to Embodiment 1 of the present invention;

[0027] Figure 2 is a flow chart of another production risk early warning method provided according to the second embodiment of the present invention;

[0028] Figure 3 is a bar chart of the number of production accidents from 2016 to 2023 provided in the third embodiment of the present invention;

[0029] Figure 4 This is a risk level classification standard diagram provided by the third embodiment of the present invention;

[0030] Figure 5 This is a flow chart of another production risk early warning method provided by Embodiment 3 of the present invention;

[0031] Figure 6 is a feature correlation degree histogram provided in the third embodiment of the present invention;

[0032] Figure 7 is a corresponding relationship diagram between decision tree depth and accuracy provided in the third embodiment of the present invention;

[0033] Figure 8 This is a left branch structure diagram of a decision tree provided by Embodiment 3 of the present invention;

[0034] Fig. 9 It is a right branch structure diagram of a decision tree provided by Embodiment 3 of the present invention;

[0035] Fig.10 It is a schematic diagram of the structure of a device for implementing a production risk early warning method provided by an embodiment of the present invention;

[0036] Fig.11 It is a structural schematic diagram of an electronic device for implementing the production risk early warning method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0039] Embodiment 1

[0040] Figure 1 A flowchart of a production risk warning method is provided for an embodiment of the present invention. This embodiment is applicable to the case of predicting production risks. The method can be executed by a production risk warning device. The production risk warning device can be implemented in the form of hardware and / or software. The production risk warning device can be configured in an electronic device. Figure 1 As shown, the method includes:

[0041] S101. Acquire production data during the production process.

[0042] In an embodiment of the present invention, production data can be understood as information generated by an enterprise during the production process of hazardous chemicals. By way of example, production data includes but is not limited to the enterprise's online rate, hidden danger inspection completion rate, and average risk value, etc. The methods for obtaining production data may include: obtaining through manual data recording or obtaining through real-time collection and analysis using data acquisition technology.

[0043] Specifically, the production data generated during production is collected and recorded, including but not limited to information such as enterprise online rate, hidden danger inspection completion rate and average risk value. The production data provides data support for risk warning.

[0044] S102. Divide the production data into different labeled sample data sets according to preset risk levels.

[0045] In an embodiment of the present invention, the preset risk level can be understood as a pre-set risk classification standard. For example, the preset risk level can be pre-set according to four classification standards: major risk, greater risk, general risk and low risk, or pre-set according to three classification standards: high risk, medium risk and low risk.

[0046] The labeled sample data set can be understood as a data set formed by classifying the production data according to the preset risk level and assigning the corresponding risk level label. For example, the risk level label may include: a major risk level label or a general risk level label. Specifically, the preset risk level is obtained, wherein the preset risk level can be set according to historical data or expert experience. Based on the preset risk level, the production data obtained in the production process is classified and labeled, and divided into labeled sample data sets of different risk levels. Each labeled sample data set is clearly marked as the corresponding preset risk level. The labeled sample data of different preset risk levels reflect different degrees of risk. The production data can be accurately analyzed and processed through classification and labeling.

[0047] S103: Determine characteristic data of each labeled sample data set according to the degree of association between the production data and the preset risk level.

[0048] In the embodiment of the present invention, the degree of association may be understood as the correlation coefficient between the production data and the preset risk level. For example, the method for calculating the degree of association may include: a recursive feature elimination method or a Selectkbest algorithm, etc.

[0049] Feature data can be understood as production data extracted from the labeled sample data set that is highly correlated with the preset risk level and is used to determine the risk warning model.

[0050] Specifically, in each labeled sample data set, the degree of correlation between the production data and the preset risk level is analyzed, and the corresponding production data is selected in each labeled sample data set based on the degree of correlation between the production data and the preset risk level, and the production data selected based on the degree of correlation is used as the feature data of each labeled sample data set.

[0051] For example, the correlation between production data and preset risk levels is analyzed, and production data with a high correlation with the preset risk level is selected as feature data in each labeled sample data set. Risk warning based on the determined feature data can improve the accuracy and timeliness of the warning.

[0052] S104. Determine the risk warning model and warning monitoring data dimension based on the data dimension of the characteristic data.

[0053] In the embodiment of the present invention, the data dimension can be understood as an indicator for measuring risks at different levels of an enterprise. For example, the data dimension can include information such as the completion rate of hidden danger inspection or the completion rate of hidden danger rectification.

[0054] The risk warning model can be understood as an algorithm model built based on characteristic data, which is used to predict the production risk level.

[0055] The early warning monitoring data dimension can be understood as the data dimension of the production data that needs to be paid attention to when predicting the risk level. It can be understood that the early warning monitoring data dimension is equal to the data dimension of the feature data.

[0056] Specifically, feature data is extracted from each labeled sample data set, and data dimensions contained in the feature data for measuring risks at different levels of the enterprise are determined based on the extracted feature data. The early warning monitoring data dimensions are determined based on the data dimensions of the feature data, wherein the early warning monitoring data dimensions refer to production data that only need to be paid attention to when predicting the risk level. The preset algorithm model is trained based on the data dimensions of the feature data, and the trained algorithm model is used as a risk early warning model.

[0057] S105. According to the early warning monitoring data dimension, the early warning key data is collected in real time, and the predicted risk level is determined based on the risk early warning model and the early warning key data.

[0058] In an embodiment of the present invention, the early warning key data refers to data with the same dimension as the early warning monitoring data in the production data collected in real time. For example, if the early warning monitoring data dimension is the hidden danger investigation completion rate and the hidden danger rectification completion rate, then the early warning key data collected and monitored in real time according to the early warning monitoring data dimension is the hidden danger investigation completion rate value and the hidden danger rectification completion rate value.

[0059] Predicted risk level refers to the risk level predicted by the risk warning model. For example, the predicted risk level output by the risk warning model may include: major risk, greater risk, general risk and low risk or high risk, medium risk and low risk.

[0060] Specifically, the early warning monitoring data dimension is determined based on the data dimension of the feature data, and the early warning key data in the production process is monitored and collected in real time according to the early warning monitoring data dimension. The early warning key data monitored and collected in real time is input into the trained risk early warning model. The risk early warning model processes and analyzes the early warning key data and outputs the predicted risk level.

[0061] S106. Determine the risk causes and disposal suggestions based on the predicted risk level, and generate risk warnings based on the risk causes and disposal suggestions.

[0062] In the embodiment of the present invention, the risk cause may be understood as the reason that causes the enterprise to be at a predicted risk level. For example, the risk cause may include: a low completion rate of hidden danger investigation or a low completion rate of hidden danger rectification.

[0063] Treatment recommendations can be understood as countermeasures proposed for the causes of risks, aimed at reducing or eliminating the risks.

[0064] Risk warning can be understood as a warning message. For example, risk warning may include: risk impact scope or operation process.

[0065] Specifically, the risk warning model is used to obtain the predicted risk level. According to the predicted risk level, the risk warning model can further output the risk reasons that cause the enterprise to be at the predicted risk level and the disposal suggestions for the risk reasons. The risk reasons and disposal suggestions are comprehensively considered to generate a risk warning, and subsequent operations are performed based on the risk warning. For example, the generated risk warning is sent to relevant personnel or related equipment so that the relevant personnel or related equipment can take corresponding measures in time according to the risk warning to reduce or eliminate the risk.

[0066] In an embodiment of the present invention, production data in a production process is obtained, a preset risk level is obtained, and based on the preset risk level, the obtained production data is classified and labeled, and the production data is divided into labeled sample data sets of different risk levels. In each labeled sample data set, the degree of correlation between the production data and the preset risk level is analyzed, and corresponding production data is selected in each labeled sample data set according to the degree of correlation between the production data and the preset risk level. The production data selected according to the degree of correlation is used as feature data, and a preset algorithm model is trained using the data dimension of the feature data. The trained algorithm model is used as a risk warning model, and the warning monitoring data dimension is determined according to the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data collected by real-time monitoring is input into the risk warning model. The risk warning model outputs a predicted risk level, and according to the predicted risk level, the risk cause and disposal suggestion are further determined. Factors of both the risk cause and disposal suggestion are comprehensively considered to generate a risk warning, and subsequent operations are performed based on the risk warning. In an embodiment of the present invention, production data obtained during the production process provides data support for the subsequent determination of predicted risk levels and generation of risk warnings. The data is divided into sample data sets of different risk levels, and feature data is selected from each labeled sample data set based on the degree of correlation between the production data and the preset risk level. A risk warning model is constructed based on the data dimension of the feature data. By constructing a risk warning model using the feature data screened out from the production data, the prediction accuracy of the risk warning model is increased, and the false alarm rate and missed alarm rate are reduced. The warning monitoring data dimension is determined based on the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data are input into the risk warning model. The preset risk level is obtained by real-time monitoring and collection of warning key data corresponding to the warning monitoring data dimension, and the warning key data are input into the risk warning model, thereby ensuring the accurate warning of the risk warning model. By predicting the risk level, the risk cause and disposal suggestions are determined, so that the enterprise can quickly formulate effective response strategies and improve the management efficiency of production risks.

[0067] Embodiment 2

[0068] Figure 2 This is a flow chart of another production risk early warning method provided by an embodiment of the present invention. Based on the above embodiment, the embodiment of the present invention provides another production risk early warning method. Figure 2 As shown, the method includes:

[0069] S201. Obtain work logs during the production process.

[0070] In the embodiment of the present invention, the work log can be understood as a document used to record various data in the production process. For example, the work log can include: operation records, equipment status or production volume data.

[0071] Specifically, the work log of the production process is obtained, which records various data in the production process, including but not limited to operation records, equipment status or production volume, etc., wherein the method of obtaining the work log may include: obtaining it by accessing the enterprise cloud platform or obtaining it through a third-party organization.

[0072] S202. Extracting work data of the production process from the work log.

[0073] In an embodiment of the present invention, work data can be understood as data extracted from the work log and directly related to the generation of risk warnings. By way of example, work data may include: commitment completion rate, video online rate, risk analysis completion rate, average number of alarms, or number of special operations.

[0074] Specifically, obtain the production work log, filter out the data directly related to the generation of risk warnings in the work log, and use the data directly related to the generation of risk warnings as the work data in the production process.

[0075] S203. Eliminate abnormal data from the working data, and use the working data as production data.

[0076] In an embodiment of the present invention, abnormal data can be understood as data that deviates from a normal range. For example, the abnormal data may include: data in which the video online rate in the working data is greater than 1 or data in which the average number of alarms in the working data is greater than the total number of alarms. The reasons for generating the abnormal data may include: equipment failure or environmental changes.

[0077] Specifically, the work data in the production process is obtained, and the work data is processed using a data processing method to eliminate abnormal data in the work data, and the work data that does not contain abnormal data is used as the production data in the production process. By way of example, the data processing method may include: the Laida criterion or the percentile method.

[0078] S204: normalize the production data to obtain normalized data of the production data.

[0079] In the embodiment of the present invention, normalization processing refers to a method of converting data from different magnitudes to the same magnitude. For example, the normalization processing method may include: a minimum-maximum normalization method or a Z-score normalization method.

[0080] Normalized data can be understood as production data that has been normalized. Normalized data has a unified dimension, which facilitates comparison between data.

[0081] Specifically, the production data in the production process is obtained, and according to the data characteristics and subsequent analysis requirements, a suitable normalization method is selected to normalize the production data, and the production data of different magnitudes are converted to the same magnitude to obtain normalized data with a unified dimension. For example, the number of special operations in the obtained production data is set to be 100, 90, and 80, respectively. The minimum-maximum normalization method is used to normalize the obtained production data 100, 90, and 80, and the production data 100, 90, and 80 are scaled to the range of [0, 1], and the normalized data of the production data is obtained to be 1, 0.5, and 0.

[0082] S205: Input the normalized data into the clustering model to obtain data clusters corresponding to preset risk levels output by the clustering model, wherein the number of data clusters set by the clustering model is equal to the number of preset risk levels.

[0083] In the embodiment of the present invention, the clustering model can be understood as a supervised learning algorithm for grouping normalized data according to similarity, wherein normalized data within the same group has high similarity and normalized data between different groups has low similarity.

[0084] A data cluster can be understood as a data group formed by normalized data with high similarity. These data groups are divided based on a clustering model. Within a data cluster, each normalized data has a high degree of similarity.

[0085] Specifically, the production data is normalized to obtain normalized data of the production data, the number of data clusters output by the clustering model is set to the number of preset risk levels, the normalized data is input into the trained clustering model, the clustering model performs clustering operations on the normalized data, and outputs the same number of data clusters as the preset risk levels.

[0086] S206: Using the data clusters corresponding to the preset risk levels output by the clustering model as the labeled sample data set, wherein the clustering model at least includes: a K-MEANS clustering model.

[0087] Specifically, the data clusters output by the clustering model, which have the same number as the preset risk level, are aggregated to form a labeled sample data set, wherein the clustering model is obtained by at least: training using a K-MEANS algorithm.

[0088] S207: Obtain normalized data of the production data, and determine a correlation value between the normalized data and a level value of a preset risk level.

[0089] In the embodiment of the present invention, the level value may be understood as a quantitative value, and the level value is a quantitative representation of a preset risk level. For example, the level value may include: 0, 1, 2 or 3.

[0090] The correlation value can be understood as a quantitative description of the closeness of the relationship between the normalized data and the preset risk level. It can be understood that the degree of correlation between the production data and the preset risk level can be determined based on the correlation value.

[0091] Specifically, the preset risk level is quantified to obtain the level value of the preset risk level, the normalized data of the production data is obtained, the normalized data of the production data is compared with the level value of the preset risk level, and the correlation value between the normalized data and the level value is determined, wherein the correlation value is used to represent the degree of correlation between the production data and the preset risk level.

[0092] S208. When the correlation value is greater than a preset threshold, determine that the correlation degree between the production data and the preset risk level is a high correlation degree.

[0093] In the embodiment of the present invention, the preset threshold value may be understood as a preset value used to determine the degree of association between the production data and the preset risk level.

[0094] Specifically, the correlation value between the normalized data and the level value of the preset risk level and the preset threshold are obtained, and the correlation value is compared with the preset threshold. When the correlation value is greater than the preset threshold, the correlation degree between the production data and the preset risk level is determined to be a high correlation degree. When the correlation value is less than the preset threshold, the correlation degree between the production data and the preset risk level is determined to be a low correlation degree.

[0095] S209: extracting production data corresponding to a high correlation degree from each labeled sample data set as feature data.

[0096] Specifically, according to the comparison result of the correlation value and the preset threshold value, the correlation degree between the production data and the preset risk level is determined, and based on the determined correlation degree, the production data with a high degree of correlation with the preset risk level is searched in each labeled sample data set, and the production data with a high degree of correlation with the preset risk level is extracted from each labeled sample data set, and the production data with a high degree of correlation with the preset risk level extracted from each labeled sample data set is used as the characteristic data of the production data.

[0097] S210: Use the feature data to train a preset decision tree model, and use the trained preset decision tree model as a risk warning model.

[0098] In an embodiment of the present invention, the preset decision tree model can be understood as an untrained and tree-structured machine learning algorithm, which is used to classify data into different categories through a series of judgment conditions. Each node of the decision tree represents a judgment condition, each branch represents a judgment result, and the leaf node represents the final risk level.

[0099] Specifically, based on the feature data extracted from each labeled sample data set, an untrained preset decision tree model based on a tree structure is trained, and the trained preset decision tree model is used as a risk warning model for generating risk warnings.

[0100] S211. Extract data dimensions of feature data, and use the data dimensions of feature data as early warning monitoring data dimensions.

[0101] Specifically, based on the feature data extracted from each labeled sample data set, the data dimension of the feature data is determined, and the determined data dimension of the feature data is used as the early warning monitoring data dimension.

[0102] S212. Monitor and collect key warning data in real time according to the warning monitoring data dimensions.

[0103] Specifically, the production process is monitored in real time, and normalized data with the same dimension as the early warning monitoring data is collected, and the normalized data with the same dimension as the early warning monitoring data collected in real time is used as the key early warning data.

[0104] S213. Input the key warning data into the risk warning model to obtain the predicted risk level output by the risk warning model.

[0105] Specifically, the key warning data collected in real time is input into the risk warning model that has been trained. The risk warning model processes and analyzes the key warning data, predicts the risk level in the production process, and outputs the predicted risk level.

[0106] S214. Determine the risk causes and disposal suggestions based on the predicted risk level, and generate risk warnings based on the risk causes and disposal suggestions.

[0107] Specifically, based on the predicted risk level, the risk warning model further analyzes the causes of risks in the production process and suggestions for dealing with risks, outputs risk causes and disposal suggestions corresponding to the predicted risk level, and generates risk warnings corresponding to the predicted risk level according to the specific information in the risk causes and disposal suggestions.

[0108] In an embodiment of the present invention, a work log in a production process is obtained, work data directly related to generating risk warnings is screened in the work log, abnormal data in the work data is eliminated, the work data not containing abnormal data is used as production data in the production process, the production data is normalized, production data of different magnitudes are converted to the same magnitude to obtain normalized data with a unified dimension, the number of data clusters output by a clustering model is set to be the number of preset risk levels, the normalized data is input into the clustering model, the clustering model performs a clustering operation on the normalized data, and outputs data clusters with the same number as the preset risk level, the data clusters are aggregated to form a labeled sample data set, the preset risk level is quantified to obtain a level value of the preset risk level, the normalized data of the production data is compared with the level value of the preset risk level, and the correlation between the normalized data and the level value is determined. value, compare the correlation value with the preset threshold value, when the correlation value is greater than the preset threshold value, determine that the correlation degree between the production data and the preset risk level is high, use the production data with a high correlation degree with the preset risk level as the characteristic data of the production data, train the preset decision tree model based on the characteristic data to obtain the risk warning model, determine the data dimension of the characteristic data, use the data dimension of the characteristic data as the warning monitoring data dimension, monitor and collect the warning key data in real time according to the warning monitoring data dimension, input the warning key data collected in real time into the risk warning model, the risk warning model processes and analyzes the warning key data and outputs the predicted risk level, based on the predicted risk level, the risk warning model further outputs the risk causes and disposal suggestions corresponding to the predicted risk level, and generates the risk warning corresponding to the predicted risk level according to the specific information in the risk causes and disposal suggestions. In an embodiment of the present invention, by removing abnormal data from working data and using working data that does not contain abnormal data as production data, the influence of data bias on analysis results can be avoided, thereby improving the reliability and accuracy of risk warnings; calculating the correlation value between normalized data and a preset risk level value, comparing the correlation value with a preset threshold, and screening out feature data through the comparison result of the correlation value with the preset threshold, thereby ensuring the accuracy of feature data screening; training a preset decision tree model based on the data dimension of the feature data to construct a risk warning model, thereby making the risk warning model more robust, thereby reducing the false alarm rate and missed alarm rate of the risk level predicted by the risk warning model output, thereby achieving accurate warning of production risks, generating risk warnings according to risk causes and disposal suggestions, and enabling enterprises to effectively respond to risks based on risk warnings, thereby improving the management efficiency of production risks.

[0109] On the basis of the above embodiment, the work data of the production process is extracted from the work log, including: extracting work data of different data dimensions in the production process from the work log, wherein the data dimensions of the work data include at least the following data dimensions: enterprise online rate, hidden danger investigation completion rate, average risk value, hidden danger rectification completion rate, safety commitment completion rate, video online rate, risk analysis completion rate, average number of alarms and number of special operations.

[0110] In the embodiment of the present invention, the enterprise online rate can be understood as the proportion of time that all key production equipment and systems of the enterprise operate normally within a specified time range. The enterprise online rate reflects the stability and reliability of the enterprise's ability to maintain normal operation of equipment and systems during production; the hidden danger investigation completion rate can be understood as the degree to which the enterprise identifies and records potential hidden dangers in production facilities, equipment or operations within a specified time; the average risk value can be understood as the average value generated by evaluating and summarizing all relevant risk factors, which is used to evaluate the overall risk level of the enterprise; the hidden danger rectification completion rate can be understood as the degree to which the enterprise rectify potential hidden dangers in production facilities, equipment and operations within a specified time; the safety commitment completion rate can be understood as the degree to which the enterprise has achieved its commitments and goals set in production. This indicator The indicators illustrate the importance that an enterprise attaches to its safety commitment and its ability to execute it; the video online rate can be understood as the proportion of time that an enterprise continuously monitors the production process through online technology. A high video online rate indicates the ability of an enterprise to supervise production activities in real time through technical means; the risk analysis completion rate can be understood as the proportion of an enterprise's analysis of potential and existing risks and generation of reports over a period of time, revealing the company's execution in systematically assessing and managing risks; the average number of alarms can be understood as the average number of alarms triggered in the production process within a specified period of time. Monitoring the alarm frequency can help enterprises identify production links that need attention and improvement; the number of special operations can be understood as the number of operations involving specific risks or complex operations in the production process. This indicator helps identify high-risk activities in production and take additional measures.

[0111] Specifically, work data of multiple data dimensions in the production process are collected from work logs, wherein the collected work data at least covers the following data dimensions: enterprise online rate, hidden danger investigation completion rate, average risk value, hidden danger rectification completion rate, safety commitment completion rate, video online rate, risk analysis completion rate, average number of alarms and number of special operations.

[0112] Embodiment 3

[0113] Based on the above-mentioned embodiments, the embodiments of the present invention provide a bar chart of the number of production accidents from 2016 to 2023, a classification standard chart of risk levels and another flow chart of a production risk early warning method.

[0114] like Figure 3 As shown in the figure, the change in the number of production accidents from 2016 to 2023 is intuitively displayed. The X-axis in the figure is from 2016 to 2023, representing the timeline of each year. The Y-axis in the figure is the number of accidents, ranging from 0 to 160. The bar chart in the figure represents the total number of accidents in the corresponding year. The line chart and data points in the figure intuitively show the trend of the change in the number of accidents, and the specific number of accidents is marked above the data points. Figure 1 It can be seen that from 2016 to 2019, the number of accidents remained at a high level, 158, 162, and 158 respectively. In 2020, there was a significant decline, with the number being 148. 2021 was the lowest point, falling to 121, and the number rose slightly in 2022 and 2023, reaching 123 and 144 respectively. Overall, the number of accidents has risen again after reaching its lowest point in 2021, reflecting that reducing production risks and curbing the occurrence of major or even extremely serious accidents are still the essential needs of production.

[0115] like Figure 4 As shown in the figure, a classification standard of risk level is described in detail, and the risk level is divided into four risk levels: major risk, greater risk, general risk and low risk. Major risk is described as unacceptable risk, and high-level pre-control needs to be initiated, and immediate and comprehensive action is taken until the risk is eliminated or reduced to a safe range. Greater risk is described as undesirable risk, and medium-level pre-control is initiated, and local action is taken until the risk is eliminated or reduced to a safe range. General risk is described as limited acceptance risk, and some restricted operations are implemented, low-level pre-control is implemented, and production is carried out under control measures at the same time to eliminate risks at a certain cost. Low risk is described as acceptable risk, and conventional operations and conventional pre-control can be carried out through on-site response, combined with the conditions of the police field and attention. The warning color corresponding to each risk level can be represented by red, orange, yellow and blue respectively. When studying the method for the classification of production risk warning levels, the implementation principle of risk warning is matching theory. The so-called matching theory refers to the matching between risk level and risk warning level, and the matching between risk level and risk prevention measures, that is, seeking the optimal matching combination with resources.

[0116] like Figure 5As shown in the figure, another production risk early warning method is described in detail. The management, production and model conclusion data of Kenli District, Dongying City, Shandong Province for nearly 700 days are selected as the research object. The data collection frequency is processed to prepare for subsequent analysis with a frequency of one sample per day. The number of special operations, risk assessment indicators and corresponding indicators of dual prevention conditions that occur every day in 60 major hazardous source enterprises in Kenli District, Dongying City are aggregated. Among them, the number of special operations includes the number of hot work operations (including special hot work, first-level hot work and second-level hot work), the number of confined space operations, the number of blind plate extraction and plugging operations, the number of high-altitude operations, the number of hoisting operations, the number of temporary power operations, the number of earthwork operations and the number of circuit breaking operations. The risk assessment indicators include the average risk value, the video online rate, the enterprise online rate and the total number of alarms. The dual prevention indicators include the risk analysis completion rate, the hidden danger investigation completion rate, the hidden danger rectification completion rate and the safety commitment completion rate. After processing, there are a total of 9 dimensions. After removing 24 days of data with obvious abnormal values ​​through data cleaning and other methods, data samples are obtained. The data samples of 9 characteristic dimensions of the remaining 675 days in Kenli District, Shandong Province are connected to the clustering algorithm based on Euclidean distance, and it is detected whether the data samples are successfully connected. The clustering algorithm based on Euclidean distance is used to cluster and mark the data of 9 characteristic dimensions of the remaining 675 days in Kenli District, Shandong Province, and the data samples are divided into major risk, greater risk, general risk, and low risk. Further detection is made as to whether the data samples are marked successfully. Among them, the data samples with levels of major risk, greater risk, general risk, and low risk are 511, 51, 57, and 56, respectively, accounting for 75.7%, 7.5%, 8.4%, and 8.3% of the total samples, respectively. The silhouette coefficient is close to 0.75, and the classification effect is good.

[0117] The data of 9 dimensions are screened, and the correlation value between the features of each dimension and the target variable is calculated according to the Selectkbest algorithm. The correlation value represents the degree of correlation between the features and the target variable. The features with low correlation with the target variable are screened out, and only the features with high correlation are retained. The features with high correlation have stronger predictive ability. Figure 6 As shown in the figure, the feature scores between the features and the target variables are intuitively displayed. Figure 6 It can be seen that 4 low-correlation features can be screened out from the 9-dimensional data, namely, enterprise online rate, hidden danger investigation completion rate, average risk value and hidden danger rectification completion rate, and the remaining 5 high-correlation features are retained, namely, safety commitment completion rate, video online rate, risk analysis completion rate, average number of alarms and number of special operations. To further test whether the data is successfully screened, the successfully screened and marked data samples are trained using classification algorithms to build a risk model.

[0118] After supervised training based on various classification technologies such as statistics, rules, distances, and neural networks, the model is evaluated from multiple angles, including accuracy, timeliness, stability, and interpretability. Finally, the decision tree algorithm based on the C4.5 model is selected as the final model for classification training from among several classic algorithms such as Bayesian networks, decision trees, support vector machines, and neural networks. The successfully screened and labeled data samples are divided into training data and test data. The decision tree model is trained with the training data, and the decision tree model is tested with the test data. New samples are input into the trained decision tree model, mapped and graded, and the preset risk level is output. The predicted risk is tested to see if it is output normally. Finally, the risk cause and disposal suggestions are output. Figure 7 As shown in Figure 1, when the depth of the decision tree model is 5, the accuracy of the test data reaches the maximum, nearly 99.5%. Figure 7 It can be seen that when the depth of the tree gradually increases from 1, the accuracy of the test data is 84%, 87%, 99%, 99% and 99.5% respectively. That is, when the depth of the tree reaches 5, the accuracy of the decision tree on the test set no longer continues to increase, indicating that the decision tree model training is complete. The corresponding decision tree is as follows Figure 8 and Fig. 9 As shown, Figure 8 is the left branch of the decision tree, Fig. 9 It is the right branch of the decision tree. In this decision tree, each node shows the conditions for splitting, and also contains the entropy value. The entropy value indicates the purity of the data. The lower the entropy, the higher the purity. It also contains the number of samples. Samples indicates the number of data samples contained in the current node, the value indicates the number of samples of different categories, and the class indicates the predicted risk category of the current node. Starting from the root node, the decision tree splits the data according to the conditions. The left branch indicates that the condition is true, and the right branch indicates that the condition is false. The leaf node shows the final classification result, and no further splitting is required. Among them, the four categories of excellent, good, medium, and poor correspond to major risks, large risks, general risks, and low risks. A sample explainability analysis was performed on the data sample whose risk category is excellent in the decision tree. The video online rate is >= 0.769, the enterprise's commitment rate is >= 0.546, and the enterprise's average risk analysis completion rate is >= 0.902. According to the tree structure shaped by the decision tree, the risk reasons why the risk category of this data sample is excellent are the high commitment rate, high video online rate, and high enterprise average risk analysis completion rate. Based on this reason, the business results reflected by the data model can be fed back to the corresponding management unit, and the corresponding management personnel can take corresponding measures for the enterprise or region according to the reason.

[0119] Embodiment 4

[0120] Fig.10 A schematic structural diagram of a production risk early warning device provided in an embodiment of the present invention.

[0121] like Fig.10 As shown, the device comprises:

[0122] The production data acquisition module 301 is used to acquire production data during the production process;

[0123] A marked sample data set acquisition module 302 is used to divide the production data into different marked sample data sets according to the preset risk level;

[0124] The characteristic data determination module 303 is used to determine the characteristic data of each labeled sample data set according to the degree of association between the production data and the preset risk level;

[0125] A model determination module 304 is used to determine a risk warning model and a warning monitoring data dimension based on the data dimension of the feature data;

[0126] The risk level determination module 305 is used to monitor and collect the early warning key data in real time according to the early warning monitoring data dimension, and determine the predicted risk level based on the risk early warning model and the early warning key data;

[0127] The risk warning generation module 306 is used to determine the risk cause and treatment suggestion according to the predicted risk level, and generate a risk warning according to the risk cause and treatment suggestion.

[0128] In an embodiment of the present invention, production data in a production process is obtained, a preset risk level is obtained, and based on the preset risk level, the obtained production data is classified and labeled, and the production data is divided into labeled sample data sets of different risk levels. In each labeled sample data set, the degree of correlation between the production data and the preset risk level is analyzed, and corresponding production data is selected in each labeled sample data set according to the degree of correlation between the production data and the preset risk level. The production data selected according to the degree of correlation is used as feature data, and a preset algorithm model is trained using the data dimension of the feature data. The trained algorithm model is used as a risk warning model, and the warning monitoring data dimension is determined according to the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data collected by real-time monitoring is input into the risk warning model. The risk warning model outputs a predicted risk level, and according to the predicted risk level, the risk cause and disposal suggestion are further determined. Factors of both the risk cause and disposal suggestion are comprehensively considered to generate a risk warning, and subsequent operations are performed based on the risk warning. In an embodiment of the present invention, production data obtained during the production process provides data support for the subsequent determination of predicted risk levels and generation of risk warnings. The data is divided into sample data sets of different risk levels, and feature data is selected from each labeled sample data set based on the degree of correlation between the production data and the preset risk level. A risk warning model is constructed based on the data dimension of the feature data. By constructing a risk warning model using the feature data screened out from the production data, the prediction accuracy of the risk warning model is increased, and the false alarm rate and missed alarm rate are reduced. The warning monitoring data dimension is determined based on the data dimension of the feature data. According to the warning monitoring data dimension, real-time monitoring and collection of warning key data are performed, and the warning key data are input into the risk warning model. The preset risk level is obtained by real-time monitoring and collection of warning key data corresponding to the warning monitoring data dimension, and the warning key data are input into the risk warning model, thereby ensuring the accurate warning of the risk warning model. By predicting the risk level, the risk cause and disposal suggestions are determined, so that the enterprise can quickly formulate effective response strategies and improve the management efficiency of production risks.

[0129] Furthermore, the production data acquisition module 301 also includes: a work log acquisition unit, which is used to obtain the production work log; a work data extraction unit, which is used to extract the work data of the production process from the work log; and an abnormal data elimination unit, which is used to eliminate abnormal data in the work data and use the work data as production data.

[0130] Optionally, the work data extraction unit is specifically used to extract work data of different data dimensions in the production process from the work log, wherein the data dimensions of the work data include at least the following data dimensions: enterprise online rate, hidden danger investigation completion rate, average risk value, hidden danger rectification completion rate, safety commitment completion rate, video online rate, risk analysis completion rate, average number of alarms and number of special operations.

[0131] Optionally, the labeled sample data set acquisition module 302 is specifically used for:

[0132] Normalizing the production data to obtain normalized data of the production data;

[0133] Inputting the normalized data into the clustering model to obtain data clusters corresponding to the preset risk levels output by the clustering model, wherein the number of data clusters set by the clustering model is equal to the number of preset risk levels;

[0134] The data clusters corresponding to the preset risk levels output by the clustering model are used as the labeled sample data set, wherein the clustering model at least includes: a K-MEANS clustering model.

[0135] Optionally, the feature data determination module 303 is specifically used for:

[0136] Obtaining normalized data of the generated data, and determining a correlation value between the normalized data and a level value of a preset risk level;

[0137] When the correlation value is greater than a preset threshold, it is determined that the correlation degree between the production data and the preset risk level is a high correlation degree;

[0138] The production data corresponding to the high correlation degree is extracted from each labeled sample data set as feature data.

[0139] Optionally, the model determination module 304 is specifically used for:

[0140] The preset decision tree model is trained using the characteristic data, and the trained preset decision tree model is used as a risk warning model;

[0141] The data dimensions of the characteristic data are extracted, and the data dimensions of the characteristic data are used as the early warning monitoring data dimensions.

[0142] Optionally, the risk level determination module 305 is specifically used for:

[0143] Detect and collect key warning data in real time based on the warning monitoring data dimensions;

[0144] Input the key warning data into the risk warning model to obtain the predicted risk level output by the risk warning model.

[0145] Embodiment 5

[0146] Fig.11 A schematic diagram of an electronic device that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown in the embodiments of the present invention, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present invention described and / or required herein.

[0147] like Fig.11 As shown, the electronic device includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the ROM 12 or the computer program loaded from the storage unit 18 to the RAM 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0148] A number of components in the electronic device are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0149] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit, a graphics processing unit, various special artificial intelligence computing chips, various processors running machine learning model algorithms, digital signal processors, and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, for example, the production risk early warning method.

[0150] In some embodiments, the production risk early warning method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on an electronic device via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the production risk early warning method may be performed. Alternatively, in other embodiments, the processor 11 may be configured as a production risk early warning method in any other appropriate manner (e.g., by means of firmware).

[0151] Various implementations of the systems and techniques described above in the embodiments of the present invention may be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays, application specific integrated circuits, application specific standard products, systems on a chip, load programmable logic devices, computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: being implemented in one or more computer programs, which may be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general programmable processor, which may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0152] The computer programs for implementing the methods of the embodiments of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer programs are executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer programs may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0153] In the context of an embodiment of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0154] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a cathode ray tube or a liquid crystal display monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including: acoustic input, voice input, or tactile input).

[0155] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area networks, wide area networks, blockchain networks, and the Internet.

[0156] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and virtual private server services.

[0157] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0158] The above specific implementations do not constitute a limitation on the protection scope of the embodiments of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A production risk early warning method, characterized in that: The method comprises: Obtain production data during the production process; Dividing the production data into different labeled sample data sets according to preset risk levels; Determining characteristic data of each of the labeled sample data sets according to the degree of association between the production data and the preset risk level; Determine a risk warning model and a warning monitoring data dimension based on the data dimension of the characteristic data; According to the early warning monitoring data dimension, real-time monitoring and collection of early warning key data are performed, and the predicted risk level is determined based on the risk early warning model and the early warning key data; The risk cause and treatment suggestion are determined according to the predicted risk level, and a risk warning is generated according to the risk cause and the treatment suggestion.

2. The method according to claim 1, characterized in that: The obtaining of production data during the production process includes: Get the work log of the production process; Extracting work data of the production process from the work log; Abnormal data of the working data are eliminated, and the working data are used as the production data.

3. The method according to claim 2, characterized in that: Extracting the work data in the production process from the work log includes: The work data of different data dimensions in the production process are extracted from the work log, wherein the data dimensions of the work data include at least the following data dimensions: enterprise online rate, hidden danger investigation completion rate, average risk value, hidden danger rectification completion rate, safety commitment completion rate, video online rate, risk analysis completion rate, average number of alarms and number of special operations.

4. The method according to claim 1, characterized in that: The dividing the production data into different labeled sample data sets according to the preset risk level includes: Normalizing the production data to obtain normalized data of the production data; Inputting the normalized data into a clustering model to obtain data clusters corresponding to the preset risk levels output by the clustering model, wherein the number of data clusters set by the clustering model is equal to the number of the preset risk levels; The data cluster corresponding to the preset risk level output by the clustering model is used as the labeled sample data set, wherein the clustering model at least includes: a K-MEANS clustering model.

5. The method according to claim 1, characterized in that: The determining of the characteristic data of each of the labeled sample data sets according to the degree of association between the production data and the preset risk level includes: Acquire normalized data of the production data, and determine a correlation value between the normalized data and a level value of the preset risk level; When the correlation value is greater than a preset threshold, determining that the correlation degree between the production data and the preset risk level is a high correlation degree; The production data corresponding to the high correlation degree is extracted from each of the labeled sample data sets as the feature data.

6. The method according to claim 1, characterized in that: The determining of the risk warning model and the warning monitoring data dimension based on the data dimension of the feature data includes: The characteristic data is used to train a preset decision tree model, and the trained preset decision tree model is used as the risk warning model; The data dimension of the characteristic data is extracted, and the data dimension of the characteristic data is used as the early warning monitoring data dimension.

7. The method according to claim 1, characterized in that: The real-time monitoring and collection of early warning key data according to the early warning monitoring data dimension, and determining the predicted risk level based on the risk early warning model and the early warning key data, include: Monitor and collect the early warning key data in real time according to the early warning monitoring data dimension; The key warning data is input into the risk warning model to obtain the predicted risk level output by the risk warning model.

8. A production risk early warning device, characterized in that: include: A production data acquisition module is used to acquire production data during the production process; A marked sample data set acquisition module, used to divide the production data into different marked sample data sets according to preset risk levels; A feature data determination module, used to determine the feature data of each of the labeled sample data sets according to the degree of association between the production data and the preset risk level; A model determination module, used to determine a risk warning model and a warning monitoring data dimension based on the data dimension of the characteristic data; A risk level determination module, used to monitor and collect early warning key data in real time according to the early warning monitoring data dimension, and determine the predicted risk level based on the risk early warning model and the early warning key data; The risk warning generation module is used to determine the risk cause and disposal suggestion according to the predicted risk level, and generate the risk warning according to the risk cause and the disposal suggestion.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the production risk warning method described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the production risk early warning method according to any one of claims 1 to 7 when executed.

Images