Risk prediction model training method, risk prediction method and device

Abstract

The embodiment of the invention provides a risk prediction model training method and device and a risk prediction method and device, and belongs to the field of production safety assessment. The risk prediction model training method is used for a refining production process, and comprises the steps of obtaining a plurality of risk events of a plurality of refining production processes; performing preliminary risk quantification on each risk event to obtain a quantitative index of each risk event; the risk occurrence probability and the risk consequence degree of each refining production process are determined according to all the quantitative indexes of each refining production process; labeling the corresponding quantitative indexes by using the risk occurrence probability and the risk consequence degree, and taking all the labeled quantitative indexes as a training set; and performing model training on the to-be-trained model by using the training set to obtain a risk prediction model. The problems that in the prior art, refining production safety accident early warning depends on artificial experience, fault discovering is not conducted in time, and the alarm time is not in time are solved.

Description

Technical Field

[0001] The present invention relates to the field of production safety assessment, and in particular to a risk prediction model training method and device, and a risk prediction method and device. Background Art

[0002] The production process of refining and chemical enterprises is complex and the process conditions are harsh. They are flammable and explosive, severely corrosive, highly toxic, high temperature and high pressure, and have strong production continuity. During continuous operation, they are affected by many factors such as process equipment, personnel operating level, and reliability of instrument equipment, and production safety accidents are prone to occur.

[0003] At present, the method of quantitative safety risk assessment of the refining and chemical production process, although it meets the scientific nature of the results, is complex in calculation and difficult to be conveniently applied. Moreover, the chemical safety assessment system is complex and cumbersome, heavily dependent on manual experience and low informatization, which leads to deviations in the assessment results and poor timeliness. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to overcome the problems in the prior art that risk assessment of refining and chemical production processes is cumbersome and complicated, heavily relies on manual experience and has a low degree of informatization, and to provide a training method and device for a risk prediction model, as well as a risk prediction method and device.

[0005] The first aspect of the present application provides a risk prediction model training method for a refining production process, the method comprising:

[0006] Obtain multiple risk events of multiple refining and production processes;

[0007] Conduct preliminary risk quantification for each risk event to obtain quantitative indicators for each risk event; determine the risk probability and risk consequence level of each refining and production process based on all quantitative indicators of each refining and production process;

[0008] Use the risk probability and risk consequence to label the corresponding quantitative indicators, and use all the labeled quantitative indicators as training sets;

[0009] The training set is used to train the model to be trained to obtain a risk prediction model.

[0010] In one embodiment of the present application, the risk occurrence probability and risk consequence degree of each refining production process are determined according to all quantitative indicators of each refining production process, including:

[0011] Determine the risk factor category corresponding to each risk event included in the refining and chemical production process;

[0012] According to all risk factor categories, the analytic hierarchy process is used to modify the quantitative indicators, and the probability of risk occurrence is obtained based on all modified quantitative indicators;

[0013] According to all risk factor categories, the analytic hierarchy process is used to modify the quantitative indicators, and the degree of risk consequences is obtained based on all modified quantitative indicators.

[0014] In one embodiment of the present application, the risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, and according to all risk factor categories, the analytic hierarchy process is used to modify the quantitative indicators, and the risk occurrence probability is obtained according to all modified quantitative indicators, including:

[0015] The analytic hierarchy process is used to determine the occurrence probability weights of all first-level categories and the occurrence probability weights of all second-level categories included in each first-level category;

[0016] Multiply the occurrence probability weight of each secondary category by the corresponding occurrence probability weight of the primary category to obtain the occurrence probability correction weight of each quantitative indicator;

[0017] Multiply each quantitative indicator by the corresponding occurrence probability correction weight and add them up to obtain the probability of risk occurrence.

[0018] In one embodiment of the present application, the risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, and according to all risk factor categories, the quantitative indicators are corrected using the hierarchical analysis method, and the risk consequence degree is obtained according to all the corrected quantitative indicators, including:

[0019] The analytic hierarchy process is used to determine the consequence weights of all first-level categories and the consequence weights of all second-level categories included in each first-level category;

[0020] Multiply the consequence degree weight of each secondary category by the consequence degree weight of the corresponding primary category to obtain the consequence degree correction weight of each quantitative indicator;

[0021] Multiply each quantitative indicator by the corresponding consequence degree correction weight and add them up to obtain the risk consequence degree.

[0022] In one embodiment of the present application, each risk event is subjected to preliminary risk quantification to obtain a quantitative index of each risk event, including:

[0023] Mathematical fuzzy method is used to conduct preliminary risk quantification of each risk event to obtain quantitative indicators of each risk event.

[0024] In one embodiment of the present application, the model to be trained is a BP neural network model.

[0025] The second aspect of the present application provides a risk prediction method for a refining production process, the method comprising:

[0026] Obtain risk events to be predicted in the refining and chemical production process;

[0027] Conduct preliminary risk quantification for each risk event to obtain quantitative indicators for each risk event;

[0028] All quantitative indicators are combined into a data set to be predicted;

[0029] Inputting the data set to be predicted into the risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining and petrochemical production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model provided in the first aspect of the present application;

[0030] Determine the risk level of the refining and chemical production process to be predicted based on the probability of risk occurrence and the degree of risk consequences.

[0031] In one embodiment of the present application, the method further includes:

[0032] According to the risk level, determine the emergency measures corresponding to the risk level and output the emergency measures.

[0033] The third aspect of the present application provides a risk prediction model training device for a refining production process, the device comprising:

[0034] A risk event acquisition module is used to acquire multiple risk events of multiple refining and petrochemical production processes;

[0035] The preliminary quantification module is used to perform preliminary risk quantification on each risk event to obtain quantitative indicators for each risk event;

[0036] The probability and degree determination module is used to determine the probability of risk occurrence and the degree of risk consequence of each refining and production process according to all quantitative indicators of each refining and production process;

[0037] The labeling module is used to label the corresponding quantitative indicators using the risk probability and risk consequence degree, and all the labeled quantitative indicators are used as training sets;

[0038] The model training module is used to train the model to be trained using the training set to obtain a risk prediction model.

[0039] In a fourth aspect, the present application provides a risk prediction device for use in a refining production process, the device comprising:

[0040] The risk event acquisition module is used to obtain the risk events of the refining and chemical production process to be predicted;

[0041] The risk event quantification module is used to perform preliminary risk quantification on each risk event to obtain quantitative indicators for each risk event;

[0042] A module for determining a data set to be predicted, used to group all quantitative indicators into a data set to be predicted;

[0043] A prediction result acquisition module, used for inputting the data set to be predicted into the risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model provided in the first aspect of the present application;

[0044] The risk level determination module is used to determine the risk level of the refining and chemical production process to be predicted based on the probability of risk occurrence and the degree of risk consequences.

[0045] The fifth aspect of the present application provides an electronic device, including a processor and a memory, the memory storing machine executable instructions that can be executed by the processor, and the processor executing the machine executable instructions to implement the risk prediction model training method provided in the first aspect of the present application, or the risk prediction method provided in the second aspect of the present application.

[0046] The sixth aspect of the present application provides a machine-readable storage medium, on which instructions are stored. When the instructions are executed by a processor, the processor implements the risk prediction model training method provided in the first aspect of the present application, or the risk prediction method provided in the second aspect of the present application.

[0047] Through the above technical scheme, a training set for training a risk prediction model can be established based on the quantitative indicators of risk events in multiple refining and chemical production processes, as well as the risk probability and risk consequence degree of the refining and chemical production processes determined according to the quantitative indicators. The risk prediction model for the refining and chemical production process obtained after the training can accurately predict the risk probability and risk consequence degree of the quantitative production process to be predicted based on the quantitative indicators of the refining and chemical production process to be predicted, thereby realizing the quantification, standardization and rapidity of risk prediction of the refining and chemical production process, and solving the problems in the prior art that the early warning of refining and chemical production safety accidents relies on manual experience, and the fault detection and alarm time are not timely.

[0048] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0050] Figure 1 A flowchart of a risk prediction model training method according to an embodiment of the present application is schematically shown;

[0051] Figure 2 A flow chart of a risk prediction method according to an embodiment of the present application is schematically shown;

[0052] Figure 3 A structural block diagram of a risk prediction model training device according to an embodiment of the present application is schematically shown;

[0053] Figure 4 The structural block diagram of a risk prediction device according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0054] The specific implementation of the present application is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present application, and is not used to limit the present application.

[0055] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, reversal status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0056] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0057] There are many factors that affect the safety of the refining and chemical production process, but the assessment of the safety of the refining and chemical production process still relies on manual experience and has a low level of informatization. At present, the production risk detection and early warning system for chemical safety has accumulated a large amount of field data. Therefore, how to mine and analyze the existing monitoring data of the refining and chemical production process is the key to providing support for the safe operation of refining and chemical enterprises.

[0058] Figure 1 A flowchart of a risk prediction model training method according to an embodiment of the present application is schematically shown. Figure 1 As shown, in one embodiment of the present application, a risk prediction model training method is provided for use in a refining production process, the method comprising steps S100 to S500.

[0059] Step S100: obtaining multiple risk events of multiple refining production processes;

[0060] First, the processor can retrieve multiple refining and chemical production processes from the relevant production risk detection and early warning system. The technical personnel then identify the risks of each refining and chemical production process and determine multiple risk events for each refining and chemical production process, such as the high altitude of the production location, aging of production equipment, and failure to meet the relevant professional qualification standards of the staff, and other risk events related to refining and chemical production safety.

[0061] Step S200: Perform preliminary risk quantification on each risk event to obtain a quantitative indicator for each risk event.

[0062] After determining the multiple risk events in the above-mentioned refining and production processes, it is necessary to quantify the degree of danger of each risk event, that is, preliminary risk quantification, and determine the quantitative indicators of each risk event, that is, to convert abstract risk events into numerical values ​​that can be recognized and understood by deep learning models.

[0063] In one embodiment of the present application, each risk event is subjected to preliminary risk quantification to obtain a quantitative index of each risk event (step S200), including:

[0064] Mathematical fuzzy method is used to conduct preliminary risk quantification of each risk event to obtain quantitative indicators of each risk event.

[0065] Mathematical fuzzy method can be used for preliminary risk quantification of risk events. The basic idea of ​​mathematical fuzzy method is to first determine the set of influencing factors, the classification type of evaluation results, the evaluation matrix, and the evaluation model, and then use the evaluation model to determine which specific evaluation result the object belongs to. Among them, the evaluation model is a fuzzy mathematical model, and its evaluation object and evaluation background both have a certain probability or randomness.

[0066] Step S300: Determine the risk occurrence probability and risk consequence degree of each refining production process according to all quantitative indicators of each refining production process.

[0067] The quantitative indicators of all risk events can form an indicator data set, which must also be labeled before it can be used as a training set for model training. The annotation of the indicator data set is for the probability of risk occurrence and the degree of risk consequences. Specifically, each quantitative indicator is labeled with the probability of risk occurrence and the degree of risk consequences of the refining and chemical production process to which the quantitative indicator belongs. Therefore, to complete the labeling process of the indicator data set, it is necessary to determine the probability of risk occurrence and the degree of risk consequences of each refining and chemical production process based on all the quantitative indicators included in each refining and chemical production process.

[0068] In one embodiment of the present application, the risk occurrence probability and risk consequence degree of each refining production process are determined according to all quantitative indicators of each refining production process (step S300), including:

[0069] Step S310: determining the risk factor category corresponding to each risk event included in the refining production process;

[0070] Step S320: According to all risk factor categories, the quantitative indicators are modified using the hierarchical analysis method, and the risk occurrence probability is obtained according to all the modified quantitative indicators;

[0071] Step S330: According to all risk factor categories, the quantitative indicators are modified using the hierarchical analysis method, and the risk consequence degree is obtained according to all the modified quantitative indicators.

[0072] For each refining and chemical production process, first determine the risk factor category of each risk event included in the refining and chemical production process. The risk factor category is the specific impact type of the risk event in the safety impact. The determined risk factor category is used in the subsequent quantitative index correction process. The correction process of the quantitative index adopts the Analytic Hierarchy Process (AHP). The Analytic Hierarchy Process is a simple method for making decisions on some more complex and more ambiguous problems. It is particularly suitable for those problems that are difficult to fully quantitatively analyze. The embodiment of the present application adopts the Analytic Hierarchy Process to eliminate the influence of subjective factors on the quantitative results in the preliminary risk quantification process. Specifically, the mathematical method of the judgment matrix is ​​used to correct the quantitative indicators. The probability of risk occurrence and the degree of risk consequences of each refining and chemical production process are calculated by the quantitative indicators after the correction is completed.

[0073] Specifically, in one embodiment of the present application, the risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, and according to all risk factor categories, the analytic hierarchy process is used to modify the quantitative indicators, and the risk occurrence probability is obtained according to all modified quantitative indicators (step S320), including:

[0074] Step S321: using the analytic hierarchy process to determine the occurrence probability weights of all first-level categories and the occurrence probability weights of all second-level categories included in each first-level category;

[0075] Step S322: multiplying the occurrence probability weight of each secondary category by the occurrence probability weight of the corresponding primary category to obtain the occurrence probability correction weight of each quantitative indicator;

[0076] Step S323: multiply each quantitative indicator by the corresponding occurrence probability correction weight, and add them up to obtain the risk occurrence probability.

[0077] The method provided in the embodiment of the present application determines the risk factor categories for risk events, including primary categories and secondary categories. The secondary categories are subordinate to the primary categories, that is, the secondary categories are subordinate categories of the primary categories. For example, the secondary category of aging of production equipment is equipment operation problems, and its primary category is material factors. The primary category of material factors can also include secondary categories such as imperfect system operation and imperfect safety-related facilities.

[0078] According to the hierarchical analysis method, a judgment matrix is ​​first constructed. The importance of each first-level category and each second-level category can be obtained through discussions among technicians and experts. The first-level categories are compared two by two, or the second-level categories are compared two by two. Then, the occurrence probability weight of each first-level category or second-level category is determined according to the comparison results of the importance between two first-level categories or between two second-level categories.

[0079] For example, in the embodiment of the present application, the first-level categories include four, namely, human factors, material factors, management factors and environmental factors. Each first-level category includes multiple second-level categories. The judgment matrix of the first-level category is recorded as A 0 , the judgment matrix of the secondary categories of human factors is recorded as A 1 The judgment matrix of the secondary categories of the factors of the object is recorded as A 2 The judgment matrix of the secondary categories of management factors is recorded as A 3 The judgment matrix of the secondary categories of environmental factors is recorded as A 4 , A 0 , A 1 , A 2 , A 3 , A 4 They can be expressed by the following matrix A:

[0080]

[0081] Among them, all a in the matrix expression jjis the comparison result of the importance of the ith first-level category and the jth first-level category, or the ith second-level category and the jth second-level category, and n is the number of first-level categories or second-level categories in the matrix.

[0082] After getting an A 0 , A 1 , A 2 , A 3 , A 4 After that, the average geometric calculation and normalization processing are performed on each judgment matrix to obtain the occurrence probability weights of all first-level categories and the occurrence probability weights of all second-level categories included in each first-level category. The occurrence probability weights of all first-level categories and the occurrence probability weights of all second-level categories included in each first-level category can be expressed by a matrix or vector set. Specifically, A 0 What is obtained is the probability weight of all first-level categories, which will be represented by A 0 All the occurrence probability weights obtained are collected into a matrix, denoted as W 0 ; by A 1 What is obtained is the probability weight of the secondary category of human factors, which will be obtained by A 1 All the occurrence probability weights obtained are collected into a matrix, denoted as W 1 ; by A 2 The obtained factors include the probability weights of the secondary categories, which will be 2 All the occurrence probability weights obtained are collected into a matrix, denoted as W 2 ; by A 3 What is obtained is the probability weight of the secondary category included in the management factor, which will be 3 All the occurrence probability weights obtained are collected into a matrix, denoted as W 3 ; by A 4 What is obtained is the probability weight of the secondary category included in the environmental factor, which will be 4 The set of all occurrence probability weights obtained is denoted as W 4 .W 0 , W 1 , W 2 , W 3 , W 4 It can be expressed by the following matrix W:

[0083]

[0084] After obtaining all the above-mentioned occurrence probability weights, the occurrence probability weight of each secondary category is multiplied by the occurrence probability weight of the corresponding primary category to obtain the occurrence probability correction weight of each quantitative indicator. 0 =(b 1 , b2 , b 3 , b 4 ), classified as quantitative indicators of human factors, the matrix composed of their occurrence probability weight correction weights can be recorded as b 1 W 1 , that is, the quantitative indicators of all factors can be classified according to their secondary categories in b 1 W 1 Find the corresponding occurrence probability correction weight.

[0085] Finally, all quantitative indicators of the refining and chemical production process are multiplied by the corresponding occurrence probability correction weights and then added up to obtain the risk occurrence probability of the refining and chemical production process.

[0086] The specific process of determining the degree of risk consequences of the refining and chemical production process is similar to the process of determining the probability of risk occurrence mentioned above.

[0087] In one embodiment of the present application, the risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, and according to all risk factor categories, the analytic hierarchy process is used to modify the quantitative indicators, and the risk consequence degree is obtained according to all modified quantitative indicators (step S330), including:

[0088] Step S331: using the analytic hierarchy process to determine the consequence degree weights of all first-level categories and the consequence degree weights of all second-level categories included in each first-level category;

[0089] Step S332: multiplying the consequence degree weight of each secondary category by the consequence degree weight of the corresponding primary category to obtain the consequence degree correction weight of each quantitative indicator;

[0090] Step S333: multiply each quantitative indicator by the corresponding consequence degree correction weight, and add them up to obtain the risk consequence degree.

[0091] Similar to the above-mentioned process of determining the probability of risk occurrence, the risk consequence degree of the refining and chemical production process is also determined by first using the hierarchical analysis method to determine the consequence degree weights of the primary and secondary categories, and then multiplying the consequence degree weight of each secondary category with the corresponding consequence degree weight of the primary category to obtain the consequence degree correction weight. After multiplying and accumulating all the quantitative indicators included in the refining and chemical production process with the corresponding consequence degree correction weight, the risk consequence degree of the refining and chemical production process can be obtained.

[0092] Step S400: using the risk occurrence probability and the risk consequence degree, the corresponding quantitative indicators are labeled, and all the labeled quantitative indicators are used as training sets.

[0093] After determining the risk occurrence probability and risk consequence degree of each refining production process through the above step S300, all quantitative indicators included in each refining production process are labeled accordingly. The indicator data set after labeling is the training set input into the model to be trained for training.

[0094] Step S500: Use the training set to perform model training on the model to be trained to obtain a risk prediction model.

[0095] The training set is used to train the model to be trained. After the model convergence conditions are reached, the model training is completed, and a risk prediction model that can accurately predict the risks of the refining production process is obtained.

[0096] In one embodiment of the present application, the model to be trained is a BP neural network model.

[0097] The embodiment of the present application utilizes the BP neural network model as the model to be trained, and the established BP neural network model includes only one hidden layer, which is a single hidden layer network. It can be understood by those skilled in the art that the basic learning process of the BP neural network model is composed of two processes: the forward propagation of the signal and the back propagation of the error. Forward propagation is to input the sample into the input layer and transmit the data to the output layer after the hidden layer is processed. If the actual output data of the output layer does not match the expected data, the back propagation stage of the error will be transferred to. The process of adjusting the weights of each layer of the forward propagation of the signal and the back propagation of the error is repeated, and this process is called the self-learning stage of the BP neural network model. This process is carried out until the error output by the network model is reduced to an acceptable level (i.e., the model convergence condition is met) or until the number of learning times set in advance.

[0098] Through the above steps S100-S500, a training set for training a risk prediction model can be established based on the quantitative indicators of risk events in multiple refining and chemical production processes, and the risk probability and risk consequence degree of the refining and chemical production processes determined based on the quantitative indicators. Among them, the risk probability and risk consequence degree are obtained by correcting the quantitative indicators through the hierarchical analysis method, eliminating the influence of subjective factors on the quantitative results, making the obtained risk probability and risk consequence degree more reasonable and accurate. The risk prediction model for the refining and chemical production process obtained after training can accurately predict the risk probability and risk consequence degree of the quantitative production process to be predicted based on the quantitative indicators of the refining and chemical production process to be predicted, realizing the quantification, standardization and rapidity of the risk prediction of the refining and chemical production process, and solving the problems in the prior art that the early warning of safety accidents in refining and chemical production depends on manual experience, and the fault detection and alarm time are not timely.

[0099] Figure 2A flow chart of a risk prediction method according to an embodiment of the present application is schematically shown. Figure 2 As shown, in one embodiment of the present application, a risk prediction method is provided for a refining production process, the method comprising:

[0100] Step S600: Obtain risk events of the refining production process to be predicted;

[0101] Step S700: Perform preliminary risk quantification on each risk event to obtain a quantitative indicator for each risk event;

[0102] Step S800: All quantitative indicators are grouped as a data set to be predicted;

[0103] Step S900: inputting the data set to be predicted into the risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model in the above embodiment;

[0104] Step S1000: Determine the risk level of the refining and chemical production process to be predicted based on the probability of risk occurrence and the degree of risk consequence.

[0105] In the process of applying the risk prediction model obtained through steps S100 to S500, the technician first identifies the risk events of the refining and petrochemical production process to be predicted. After the risk events are initially quantified, all the quantitative indicators obtained can form a data set to be predicted that is input into the risk prediction model. After the processor inputs the data set to be predicted into the risk prediction model, the risk probability and risk consequence degree of the refining and petrochemical production process to be predicted can be obtained, and the risk level of the refining and petrochemical production process to be predicted can be determined based on the risk probability and risk consequence degree, providing a reference for the selection of subsequent emergency measures.

[0106] For example, the risk level of the refining production process to be predicted can be determined according to the probability of risk occurrence and the degree of risk consequence by searching and matching the following Tables 1, 2, and 3.

[0107] Table 1

[0108] Probability Qualitative Frequent occurrence May occur Occurrence Rarely occurs Very unlikely Probability level 5 4 3 2 1 Probability of risk occurring [1-0.8) [0.8-0.6) [0.6-0.4) [0.4-0.2) [0.2-0]

[0109] Table 2

[0110] Consequences Catastrophic Very serious Severe Larger Mild Consequence level E D C B A The magnitude of the risk consequences [1-0.8) [0.8-0.6) [0.6-0.4) [0.4-0.2) [0.2-0]

[0111] Table 3

[0112]

[0113] Specifically, after determining the risk probability and risk consequence degree of the refining and chemical production process to be predicted, first determine the probability level corresponding to the risk probability and the consequence level corresponding to the risk consequence degree through Table 1 and Table 2 respectively, and then match the risk level of the refining and chemical production process to be predicted in Table 3 according to the determined probability level and consequence level - very high, high, medium, and low.

[0114] In one embodiment of the present application, the method further includes:

[0115] According to the risk level, determine the emergency measures corresponding to the risk level and output the emergency measures.

[0116] The risk level of the refining and production process to be predicted can be used as a reference for the selection of emergency measures. For example, the processor can find successful emergency measures for refining and production processes with the same risk level in the historical database, and output the emergency measures to achieve timely warning and timely processing, thereby eliminating risk factors in the refining and production process to be predicted.

[0117] Figure 3 The structure block diagram of a risk prediction model training device according to an embodiment of the present application is schematically shown. Figure 3 As shown, in one embodiment of the present application, a risk prediction model training device 10 is provided, which is used in a refining production process. The device 10 includes:

[0118] A risk event acquisition module 11 is used to acquire multiple risk events of multiple refining and chemical production processes;

[0119] A preliminary quantification module 12 is used to perform preliminary risk quantification on each risk event to obtain a quantitative indicator for each risk event;

[0120] The probability and degree determination module 13 is used to determine the probability of risk occurrence and the degree of risk consequence of each refining and production process according to all quantitative indicators of each refining and production process;

[0121] The labeling module 14 is used to label the corresponding quantitative indicators using the risk occurrence probability and the risk consequence degree, and use all the labeled quantitative indicators as a training set;

[0122] The model training module 15 is used to perform model training on the model to be trained using the training set to obtain a risk prediction model.

[0123] The risk prediction model training device 10 provided in the embodiment of the present application can implement each process of step S100 to step S500 in the method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0124] Figure 4A structural block diagram of a risk prediction device according to an embodiment of the present application is schematically shown. Figure 4 As shown, in one embodiment of the present application, a risk prediction device 20 is provided, which is used in a refining production process. The device 20 includes:

[0125] The risk event acquisition module 21 is used to acquire the risk events of the refining and chemical production process to be predicted;

[0126] The risk event quantification module 22 is used to perform preliminary risk quantification on each risk event to obtain a quantitative indicator for each risk event;

[0127] A module 23 for determining a data set to be predicted, used to group all quantitative indicators into a data set to be predicted;

[0128] A prediction result acquisition module 24 is used to input the data set to be predicted into the risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model in the above-mentioned embodiment;

[0129] The risk level determination module 25 is used to determine the risk level of the refining production process to be predicted based on the probability of risk occurrence and the degree of risk consequence.

[0130] The risk prediction device 20 provided in the embodiment of the present application can implement each process of step S600 to step S1000 in the method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0131] In one embodiment of the present application, an electronic device is provided, including a processor and a memory, the memory storing machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the risk prediction model training method in the above embodiment, or the risk prediction method in the above embodiment.

[0132] In one embodiment of the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor implements the training method of the risk prediction model in the above embodiment, or the risk prediction method in the above embodiment.

[0133] In one embodiment of the present application, a computer program product is also provided, including a computer program, which, when executed by a processor, implements the training method of the risk prediction model in the above embodiment, or the risk prediction method in the above embodiment.

[0134] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD ROM, optical storage, etc.) that include computer-usable program code.

[0135] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0136] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0137] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0138] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0139] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0140] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A training method for a risk prediction model, characterized in that: For use in a refining production process, the method comprises: Obtain multiple risk events of multiple refining and production processes; Conducting preliminary risk quantification on each of the risk events to obtain quantitative indicators for each of the risk events; Determine the risk occurrence probability and risk consequence degree of each refining and production process according to all the quantitative indicators of each refining and production process; Using the risk occurrence probability and the risk consequence degree, the corresponding quantitative indicators are labeled, and all the labeled quantitative indicators are used as a training set; The training set is used to perform model training on the model to be trained to obtain the risk prediction model.

2. The method according to claim 1, characterized in that: Determining the risk occurrence probability and risk consequence degree of each refining and production process according to all the quantitative indicators of each refining and production process includes: Determine the risk factor category corresponding to each of the risk events included in the refining production process; According to all the risk factor categories, the quantitative indicators are modified using the hierarchical analysis method, and the probability of the risk occurrence is obtained according to all the modified quantitative indicators; According to all the risk factor categories, the quantitative indicators are corrected using the hierarchical analysis method, and the risk consequence degree is obtained according to all the corrected quantitative indicators.

3. The method according to claim 2, characterized in that The risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, the analytic hierarchy process is used to modify the quantitative indicators based on all the risk factor categories, and the risk occurrence probability is obtained based on all the modified quantitative indicators, including: Determine the occurrence probability weights of all the first-level categories and the occurrence probability weights of all the second-level categories included in each first-level category by using the hierarchical analysis method; Multiply the occurrence probability weight of each of the secondary categories by the occurrence probability weight of the corresponding primary category to obtain the occurrence probability correction weight of each of the quantitative indicators; Each of the quantitative indicators is multiplied by the corresponding occurrence probability correction weight, and the results are accumulated to obtain the risk occurrence probability.

4. The method according to claim 2, characterized in that: The risk factor categories include primary categories and secondary categories, the secondary categories are subordinate to the primary categories, the analytic hierarchy process is used to modify the quantitative indicators based on all the risk factor categories, and the risk consequence degree is obtained based on all the modified quantitative indicators, including: Determine the consequence degree weights of all the first-level categories and the consequence degree weights of all the second-level categories included in each first-level category by using the analytic hierarchy process; Multiplying the consequence degree weight of each of the secondary categories by the consequence degree weight of the corresponding primary category to obtain the consequence degree correction weight of each of the quantitative indicators; Each of the quantitative indicators is multiplied by the corresponding consequence degree correction weight, and the results are added up to obtain the risk consequence degree.

5. The method according to claim 1, characterized in that The preliminary risk quantification of each risk event to obtain a quantitative indicator of each risk event includes: A mathematical fuzzy method is used to perform preliminary risk quantification on each of the risk events to obtain a quantitative indicator for each of the risk events.

6. The method according to claim 1, characterized in that The model to be trained is a BP neural network model.

7. A risk prediction method, characterized in that: For use in a refining production process, the method comprises: Obtain risk events to be predicted in the refining and chemical production process; Conducting preliminary risk quantification on each of the risk events to obtain quantitative indicators for each of the risk events; All the quantitative indicators are combined into a data set to be predicted; Inputting the data set to be predicted into a risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model as described in claims 1-6; The risk level of the refining and petrochemical production process to be predicted is determined based on the probability of risk occurrence and the degree of risk consequence.

8. The method according to claim 7, characterized in that The method further comprises: According to the risk level, emergency measures corresponding to the risk level are determined, and the emergency measures are output.

9. A training device for a risk prediction model, characterized in that: Used in a refining production process, the device comprises: A risk event acquisition module is used to acquire multiple risk events of multiple refining and petrochemical production processes; A preliminary quantification module, used to perform preliminary risk quantification on each of the risk events to obtain a quantitative indicator for each of the risk events; A probability and degree determination module, used to determine the probability of risk occurrence and the degree of risk consequence of each refining and production process according to all the quantitative indicators of each refining and production process; A labeling module, used to label the corresponding quantitative indicators using the risk occurrence probability and the risk consequence degree, and use all the labeled quantitative indicators as a training set; The model training module is used to perform model training on the model to be trained using the training set to obtain the risk prediction model.

10. A risk prediction device, characterized in that: Used in a refining production process, the device comprises: The risk event acquisition module is used to obtain the risk events of the refining and chemical production process to be predicted; A risk event quantification module, used to perform preliminary risk quantification on each of the risk events to obtain a quantitative indicator for each of the risk events; A module for determining a data set to be predicted, used for grouping all the quantitative indicators into a data set to be predicted; A prediction result acquisition module, used for inputting the data set to be predicted into a risk prediction model to obtain the risk occurrence probability and risk consequence degree of the refining production process to be predicted output by the risk prediction model, wherein the risk prediction model is obtained according to the training method of the risk prediction model as described in claims 1-6; The risk level determination module is used to determine the risk level of the refining production process to be predicted based on the probability of risk occurrence and the degree of risk consequence.

11. An electronic device, characterized in that: It includes a processor and a memory, the memory storing machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the risk prediction model training method described in any one of claims 1 to 6, or the risk prediction method described in claims 7 to 8.

12. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions, which, when executed by a processor, enable the processor to implement the risk prediction model training method according to any one of claims 1 to 6, or the risk prediction method according to claims 7 to 8.

Images