Risk management and control method and device based on safety eye system and readable storage medium

Abstract

The invention relates to the technical field of management informatization, in particular to a risk management and control method and device based on a safety eye system and a readable storage medium. The method comprises the following steps: acquiring historical hidden danger data, and performing data clustering on the historical hidden danger data to determine cause elements; the historical hidden danger data is data collected by a safety system when a historical accident occurs; the cause factors represent cause factors causing historical accidents in the historical hidden danger data; establishing a Bayesian network model of a to-be-analyzed system, and determining hidden danger Bayesian network nodes; determining an accident probability and a hidden danger grading list of the hidden danger Bayesian network nodes according to the cause elements; and performing risk management and control on the to-be-analyzed system based on the accident probability and the hidden danger grading list. Through the method, certain analysis capability can be adapted to informatization risk hidden danger supervision and control, and intelligent analysis and risk grading are performed based on the risk hidden danger data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information management technology, in particular to a risk management and control method and device based on an eye safety system and a readable storage medium. BACKGROUND

[0002] The process flow of a natural gas purification plant is complex, involving various high-risk equipment such as pressure vessels and pressure pipelines. At the same time, unsafe behaviors of people on site and unsafe states of objects can easily cause major safety accidents, resulting in huge property losses and casualties.

[0003] In recent years, some oil and gas field natural gas purification plants have installed eye safety systems to extract on-site illegal behaviors and abnormal conditions of process parameters for hidden danger data collection. However, risk hidden danger analysis is still mainly completed by personnel, which requires high on-site work experience of personnel. At the same time, due to the lag of personnel analysis, the real-time data collected by the eye safety system is difficult to analyze the real-time dynamic changes of hidden dangers, which causes huge data waste and low hidden danger identification efficiency. The existing technology has developed an intelligent identification system for risk hidden dangers to solve the above problems, but still lacks an intelligent analysis method for hidden danger identification data, which cannot accurately distinguish the risks of industrial equipment and facilities based on objective hidden danger data. Therefore, it is urgent to develop an intelligent identification and analysis method for hidden dangers based on an eye safety system. SUMMARY

[0004] The purpose of the present application is to provide a risk management and control method and device based on an eye safety system, which solves the problem that the information technology risk hidden danger supervision and control system of the natural gas purification plant in the prior art still lacks adaptive analysis capability and cannot perform intelligent analysis and risk classification based on risk hidden danger data.

[0005] To achieve the above purpose, the present application relates to a risk management and control method based on an eye safety system, comprising:

[0006] obtaining historical hidden danger data, performing data clustering on the historical hidden danger data to determine causation elements; the historical hidden danger data is data collected by an eye safety system when a historical accident occurs; the causation elements represent reason elements that cause the historical accident in the historical hidden danger data;

[0007] establishing a Bayesian network model of a system to be analyzed to determine hidden danger Bayesian network nodes;

[0008] determining an accident probability of the hidden danger Bayesian network nodes and a hidden danger classification list according to the causation elements;

[0009] performing risk management and control on the system to be analyzed based on the accident probability and the hidden danger classification list.

[0010] Optionally, after determining the accident probability of the hidden danger Bayesian network node and the hidden danger classification list according to the cause element, the method comprises:

[0011] Based on the eye protection system, real-time data of the system to be analyzed is acquired to establish a field risk hidden danger data set;

[0012] Based on the accident probability and the hidden danger classification list, the field risk hidden danger data set is analyzed to perform field risk control on the system to be analyzed.

[0013] Optionally, establishing the Bayesian network model of the system to be analyzed comprises:

[0014] The system to be analyzed is divided into multiple process units;

[0015] Operation steps of each process unit are acquired, and deviation data of each operation step is analyzed; the deviation data represents possible deviations of the operation step;

[0016] According to the deviation data, process failure factors corresponding to the operation step are acquired; the process failure factors are causes of accidents of the system to be analyzed caused by the deviation data;

[0017] Based on the operation steps and the process failure factors of the process unit, the hidden danger Bayesian network node is set.

[0018] Optionally, after determining the hidden danger Bayesian network node, the method comprises: determining the hierarchy and the dependency relationship of each hidden danger Bayesian network node according to the correlation between the hidden danger Bayesian network nodes, and then obtaining a topology graph of the Bayesian network.

[0019] Optionally, the hidden danger Bayesian network node comprises a child node, an intermediate node and a parent node; the method comprises: setting the child node as the process unit, setting the intermediate node as the operation step corresponding to the process unit, and setting the parent node as the process failure factor corresponding to the operation step.

[0020] Optionally, determining the accident probability of the hidden danger Bayesian network node according to the cause element comprises:

[0021] A parameter matrix of the system to be analyzed is constructed, and the parameter matrix represents a learning matrix of the cause element of the system to be analyzed;

[0022] The parameter matrix is learned by using the Bayesian network model to obtain the probability of failure of each hidden danger Bayesian network node.

[0023] Optionally, after determining the accident probability of the hidden danger Bayesian network node, the method further comprises: correcting the accident probability of a child node in the hidden danger Bayesian network node.

[0024] Optionally, the method further comprises: performing sensitivity analysis on the process failure factor.

[0025] The present application relates to a risk management and control device based on an eye safety system, the device comprising:

[0026] a data acquisition unit configured to acquire historical hidden danger data, and perform data clustering on the historical hidden danger data to determine a cause element; the historical hidden danger data is data collected by the eye safety system when a historical accident occurred; the cause element represents a reason element that caused the historical accident in the historical hidden danger data;

[0027] a model establishment unit configured to establish a Bayesian network model of a system to be analyzed, and determine a hidden danger Bayesian network node;

[0028] a probability determination unit configured to determine an accident probability of the hidden danger Bayesian network node and a hidden danger classification list according to the cause element;

[0029] a risk management and control unit configured to perform risk management and control on the system to be analyzed based on the accident probability and the hidden danger classification list.

[0030] The present application also relates to a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method according to any one of the above.

[0031] The present application also relates to a computer readable storage medium storing one or more programs, wherein the one or more programs are executable by one or more processors to implement the steps of the method according to any one of the above.

[0032] The risk management and control method based on the eye safety system can sort the risk factors in the system to be analyzed by screening and analyzing the historical hidden danger data, extracting the cause element and the accident probability having a strong correlation with the nodes of the Bayesian network model from the historical hidden danger data, and providing a solution to the problems of low efficiency and inability to intelligently classify risks in the risk hidden danger analysis of the system to be analyzed. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a flowchart of the risk management and control method based on the eye safety system according to an embodiment of the present application;

[0034] Figure 2It is a Bayesian network model topological graph of the risk management and control method based on the eye safety system according to an embodiment of the application;

[0035] Figure 3 It is a structural schematic diagram of the analysis management and control device based on the eye safety system according to an embodiment of the application;

[0036] Figure 4 It is a structural diagram of a computer device according to an embodiment of the application.

[0037] Explanation of the drawing symbols:

[0038] 301, data acquisition unit;

[0039] 302, model establishment unit;

[0040] 303, probability determination unit;

[0041] 304, risk management and control unit;

[0042] 402, computer device;

[0043] 404, processor;

[0044] 406, memory;

[0045] 408, driving mechanism;

[0046] 410, input / output module;

[0047] 412, input device;

[0048] 414, output device;

[0049] 416, presentation device;

[0050] 418, graphical user interface;

[0051] 420, network interface;

[0052] 422, communication link;

[0053] 424, communication bus. DETAILED DESCRIPTION

[0054] The specific embodiments of the application will be described in detail below with reference to the accompanying drawings, but it should be understood that the scope of protection of the application is not limited by the specific embodiments.

[0055] Unless otherwise explicitly stated, throughout the specification and claims, the term "comprise" or variations such as "comprises" or "comprising" will be understood to imply the inclusion of a stated element or group of elements but not the exclusion of any other element or group of elements.

[0056] Based on the existing safety management system of a natural gas purification plant, the application proposes a risk management and control method, which can intelligently analyze the identified accident data of the natural gas purification plant, and control the risk through objective analysis conclusions. Figure 1 As shown in the risk management and control method,

[0057] Step 101, obtaining historical hidden danger data, and performing data clustering on the historical hidden danger data to determine the cause element; the historical hidden danger data is the data collected by the safety management system when the historical accident occurred; the cause element represents the reason element that causes the historical accident in the historical hidden danger data. Specifically, the risk management and control method in the embodiment of the application obtains historical hidden danger data through a case analysis system. The historical hidden danger data is the accident data related to the accident when the natural gas purification plant has occurred in history, or the risk data at the time of the accident and before the accident. The historical hidden danger data can be displayed through a single accident. In the embodiment of the application, the historical hidden danger data is filtered and then processed by data clustering, which can clearly determine the cause element involved in the historical accident in the natural gas purification plant. Specifically, the data clustering processing is to determine the reason elements involved in the same or similar accidents from multiple accidents and aggregate them by category, and the cause element and its set are obtained after the data clustering processing. Optionally, the data clustering can use K-means clustering method, hierarchical clustering method and other clustering methods.

[0058] Step 102, establishing a Bayesian network model of the system to be analyzed, and determining the hidden danger Bayesian network node. The Bayesian network model is a probabilistic graphical model, which can simulate the uncertainty processing model of the causal relationship in the human reasoning process, and its network topology structure is a directed acyclic graph. In order to understand the related hidden dangers and accident probabilities in the historical hidden danger data, a Bayesian network model is first constructed in the application, the hidden danger Bayesian network node is determined in the model, and then the accident probability of each node is analyzed through the constructed Bayesian network model. In a specific embodiment, the system to be analyzed can be a natural gas purification plant, and the operation process and possible risk factors of the natural gas purification plant are taken as the nodes of the hidden danger Bayesian network and the correlation of the hidden danger Bayesian network nodes is constructed. The risk factor refers to the cause of the process failure that may exist in each operation process of the natural gas purification plant, that is, the cause that may directly or indirectly lead to process failure when the operation process is performed. The correlation can include the hierarchy and / or progressive relationship between nodes, and the topology graph of the Bayesian network model is drawn according to the determination of the hidden danger Bayesian network node. Specifically, the causes of process failure mainly focus on personnel violation, abnormal process parameters, abnormal environmental parameters, and management conditions, such as: personnel misoperation, not wearing a respirator, valve failure, heavy rain, and inadequate daily maintenance, etc.

[0059] Step 103, determining the accident probability of the hidden danger Bayesian network node and the hidden danger classification list according to the cause element. Specifically, the cause element in the historical hidden danger data is compared with the risk factors that may exist in the hidden danger Bayesian network node, if the cause element in the historical hidden danger data coincides with the risk factors that may exist in the hidden danger Bayesian network node, the hidden danger Bayesian network node is set as the first parameter; if the cause element in the historical hidden danger data is different from the risk factors that may exist in the hidden danger Bayesian network node, the hidden danger Bayesian network node is set as the second parameter. The hidden danger Bayesian network node and the cause element are compared one by one according to the above determination method, and a fault reason learning parameter matrix is constructed. Parameter learning is performed on the fault reason learning parameter matrix, and the accident probability of each hidden danger Bayesian network node can be obtained, that is, the probability that the risk factors corresponding to the node may cause a fault. The hidden danger Bayesian network node is classified according to the accident probability, the importance of the operation process corresponding to the node, the link and the like, and a hidden danger classification list is established according to the classified results. For example, the risk hidden danger with an accident probability greater than a preset threshold and high importance is a first-level risk, and the risk hidden danger with an accident probability less than a preset threshold and low importance is an N-level risk. The above classification is not limited to the embodiments of the present application.

[0060] Step 104, risk control of the system to be analyzed based on the accident probability and the hidden danger classification list. Specifically, according to the accident probability and the hidden danger classification list obtained in step 103, the importance of the operation process of the natural gas purification plant and the possible risk factors is determined, the risk control is performed according to the importance, and a guidance basis is provided for subsequent production, so that the natural gas purification plant can take hidden danger rectification measures.

[0061] The risk control method based on the safety eye system provided by the present application can sort the risk factors in the system to be analyzed by screening the accident cases in the accident investigation report as the historical hidden danger data, analyzing the historical hidden danger data, establishing the Bayesian network model of the system to be analyzed, extracting the cause elements and the accident probability having strong correlation with the hidden danger Bayesian network model node from the historical hidden danger data, and can provide a solution to the problems of low efficiency of risk hidden danger analysis and inability of intelligent risk classification of the system to be analyzed.

[0062] The risk control method based on the safety eye system provided by the embodiment of the present application, preferably, after determining the accident probability of the hidden danger Bayesian network node and the hidden danger classification list according to the cause element, the method comprises:

[0063] Based on the eye safety system, real-time data of the system to be analyzed is acquired to establish a field risk hidden danger data set. In the embodiment, the eye safety system mainly acquires reasons for process failure in three aspects of personnel, process and environment in the natural gas purification plant, such as abnormal process parameters and illegal operation behaviors. Real-time data in the field is collected by various detectors, and the basic situation of the risk hidden danger in the plant is determined by the hidden danger intelligent recognition algorithm to establish the field risk hidden danger data set.

[0064] Based on the accident probability and the hidden danger classification list, the field risk hidden danger data set is analyzed to control the field risk of the system to be analyzed. In the embodiment, the risk hidden danger data existing in the field risk hidden danger data set is compared with the hidden danger classification list, repeated risk hidden danger contents are screened out, and the risk hidden danger data in the plant is sorted according to the order of the hidden danger classification list, so as to control the field risk of the system to be analyzed. Preferably, the hidden danger classification list contains all possible risk hidden danger contents, forms a sorted risk hidden danger library of the natural gas purification plant, and can provide sorting and grading basis for the field risk hidden danger classification.

[0065] In the preferred embodiment, the eye safety system arranged in the natural gas purification plant includes the following modules:

[0066] Process parameter abnormality monitoring module: Sensing devices are arranged inside and around the devices in the natural gas purification plant region to acquire state data of temperature, pressure, liquid level, pH and gas concentration in the device;

[0067] Illegal operation monitoring module: High-definition cameras are arranged in the personnel operation area of the natural gas purification plant to monitor illegal operation behaviors in real time;

[0068] Environment abnormality monitoring module: Sensing devices are arranged in the natural gas purification plant region to acquire state data of temperature, humidity, visibility, noise vibration and gas concentration in the environment; and high-definition cameras are arranged around the plant to monitor severe weather conditions such as lightning, hail, heavy rain and snow;

[0069] The hidden danger intelligent identification module: through a hidden danger intelligent identification algorithm, risk hidden danger data collected by the process parameter abnormality monitoring module and the environment abnormality monitoring module are integrated, state data is compared with a set parameter threshold, the state data exceeding the threshold is identified as an abnormal parameter, and is identified as a hidden danger in the process parameter and the environment; through the hidden danger intelligent identification algorithm, personnel operation recorded by the personnel violation monitoring module is compared with the entered violation operation behavior, if similar behavior occurs, it is identified as personnel unsafe operation hidden danger, and finally historical hidden danger data is formed.

[0070] The risk management and control method based on the eye protection system according to the embodiment of the application, preferably, the establishment of the Bayesian network model of the system to be analyzed comprises:

[0071] The system to be analyzed is divided into a plurality of process units. Specifically, in order to establish the Bayesian network model of the system to be analyzed, the system to be analyzed is first divided into process units, and then the specific elements involved in each process unit and the correlation between them are analyzed to determine possible deviations. Preferably, the process units are divided according to the actual production process characteristics of the system to be analyzed.

[0072] The operation steps of each process unit are obtained, and the biasable data of each operation step are analyzed; the biasable data represent possible deviations of the operation steps. Specifically, according to the division of the process units, the operation steps of each process unit are analyzed to determine possible deviations of the operation steps, i.e. possible risk factors.

[0073] The process failure factors corresponding to the operation steps are obtained according to the biasable data; the process failure factors are the causes of accidents of the system to be analyzed caused by the biasable data.

[0074] The hidden danger Bayesian network nodes are set based on the operation steps and process failure factors of the process units. Specifically, the operation steps and process failure factors are taken as hidden danger Bayesian network nodes, the hierarchy and dependency relationship of each hidden danger Bayesian network node are determined according to the correlation between the hidden danger Bayesian network nodes, and then the topological graph of the Bayesian network model is obtained, such as Figure 2 The topological graph of the Bayesian network model established by the embodiment of the application is shown in the figure.

[0075] The risk management and control method based on the eye safety system according to the embodiment of the present application, preferably, the hidden danger Bayesian network node comprises a child node, an intermediate node and a parent node; the method comprises: setting the child node as the process unit, setting the intermediate node as the operation step corresponding to the process unit, and setting the parent node as the process failure factor corresponding to the operation step. The process unit division on the system to be analyzed helps to orderly split the overall risk to each process unit, preventing the omission of nodes due to the complexity of the system, facility and operation process when establishing the Bayesian network model.

[0076] The risk management and control method based on the eye safety system according to the embodiment of the present application, preferably, determining the accident probability of the hidden danger Bayesian network node according to the cause element comprises:

[0077] Constructing a parameter matrix of the system to be analyzed, wherein the parameter matrix represents a learning matrix of the cause element of the system to be analyzed. In a specific embodiment, the construction of the parameter matrix is as follows:

[0078] Parent node

[0079]

[0080] Wherein, Q represents an n-row and n-column parameter matrix composed of q ab

[0081] Learning the parameter matrix by using the Bayesian network model to obtain the failure probability of each hidden danger Bayesian network node.

[0082] Preferably, the accident probability of the child node in the Bayesian network model is corrected to obtain the posterior probability of the Bayesian network model, and the difference between the posterior probability and the prior probability before correction is compared to find the node with a larger probability value change. The posterior probability of the Bayesian network model is calculated according to the following formula:

[0083]

[0084] Wherein, P(M|N) represents the accident probability of the parent node under the condition that the child node fails, i.e. the posterior probability, P(M) and P(N) respectively represent the prior probability of the failure of the parent node and the failure of the child node, and P(N|M) represents the accident probability of the child node under the condition that the parent node fails.

[0085] ​In the preferred embodiment, a sensitivity analysis is performed on the process failure factors, the sensitivity analysis is performed on the process units of the child nodes, a sensitivity value of the process failure factors of the parent node is calculated, the process failure factors with high sensitivity in the process units are determined, and the process failure factors are sorted according to the sensitivity values, and a hidden danger classification list can be listed.

[0086] In a specific embodiment of the present application, the accident probability value of the child node is adjusted to 100%, the posterior probability is determined through the posterior probability calculation formula of the Bayesian network model, and the difference between the posterior probability and the prior probability is obtained. By determining the parent node corresponding to the process failure factor with high sensitivity, combined with the sensitivity analysis result, a risk hidden danger classification list is listed.

[0087] Through the sensitivity and posterior probability analysis, the key problems of each process unit can be found out, the key problems have a high probability of occurrence and will cause serious consequences once they occur, and the key problems can provide guidance for subsequent production, so that the natural gas purification plant can take targeted hidden danger rectification measures.

[0088] The present application also provides a specific embodiment of a natural gas purification plant as a system to be analyzed. Raw natural gas is usually obtained from the exploitation of natural gas fields or oil fields, and contains certain impurities, hydrogen sulfide, organic sulfur and other harmful substances, and needs to be purified to meet the use requirements. The natural gas purification process involves complex process links, including a large number of special equipment such as pressure pipelines and pressure vessels, and natural gas is a flammable and explosive gas, which has the risk of fire and explosion; the raw material gas contains hydrogen sulfide gas, and if it leaks, it will also cause serious consequences of poisoning personnel. Therefore, a safety risk control method is needed for the natural gas purification plant, which can realize intelligent identification of risk hidden dangers, and classify and grade the risk hidden dangers in the natural gas purification plant based on hidden danger data.

[0089] Specifically, the step of constructing a Bayesian network model comprises:

[0090] (1) Process flow division: according to the actual production process characteristics of the natural gas purification plant, the process units of the natural gas purification plant are divided, specifically, the natural gas purification plant process flow is divided into 11 process units, including raw material gas filtration separation unit, desulfurization and decarburization unit, dehydration unit, sulfur recovery unit, tail gas treatment unit, acid water stripping unit, sulfur forming unit, sewage treatment unit, venting unit, fire fighting unit and circulating water unit;

[0091] (2) Process failure factor: According to the process unit setting of the natural gas purification plant, the operation steps of each process unit are analyzed, possible deviations of the operation steps are determined, and process failure factors possibly caused by the operation steps are obtained. Specifically, the process failure factors mainly include four aspects of personnel violation, process parameter abnormality, environmental parameter abnormality and management condition;

[0092] (3) Hidden danger Bayesian network node setting: Hidden danger Bayesian network nodes are set according to the operation steps of the process unit and the process failure factors. The nodes include child nodes, intermediate nodes and parent nodes. The child nodes are the process units of the natural gas purification plant, the intermediate nodes are the operation steps corresponding to the process units, and the parent nodes are the process failure factors corresponding to the operation steps. Specifically, as shown in FIG. 3, taking the "sulfur recovery unit" as an example, the parent nodes are the process failure factors related to personnel violation, process parameter abnormality, environmental parameter abnormality and management condition in each operation step, the intermediate nodes are operation steps such as blowdown, parameter adjustment, equipment start-stop, equipment switching, main combustion furnace ignition, acid water pressure sending, waste heat boiler blowdown, acid gas separator blowdown, sulfur condenser blowdown, protection gas flow adjustment, air distribution adjustment, main combustion furnace temperature adjustment, pressure adjustment, liquid sulfur pump start-stop, air cooling fan start-stop, main fan switching and condensate pump switching, and the child nodes are the sulfur recovery unit. Figure 2

[0093] (4) Bayesian network model construction: According to the association between the hidden danger Bayesian network nodes, the node hierarchy and the dependency relationship are determined, and then the topology diagram of the Bayesian network is drawn. Taking the "sulfur recovery unit" as an example, Figure 2 The Bayesian network model topology diagram provided by the present application is shown in FIG. 4. The hidden danger Bayesian network node content is: the sulfur recovery unit (child node), the operation steps (intermediate nodes) include: acid water pressure sending, main combustion furnace ignition and acid gas separator blowdown, and the process failure factors (parent nodes) corresponding to the operation steps include: personnel misoperation, not wearing a respirator, valve failure, heavy rain and daily maintenance not in place.

[0094] The process units of the natural gas purification plant are divided, which helps to logically split the overall risk of the natural gas purification plant to each process unit, and prevents missing nodes due to too complex system, facility and operation process when establishing the Bayesian network learning model.

[0095] In specific examples, the historical hidden danger data is obtained by the TF-IDF method, and the characteristic words of the cause element having a strong association with the hidden danger Bayesian network node are extracted therefrom. The weight of the characteristic words of the cause element is proportional to the frequency of occurrence, and the weight calculation formula of the characteristic words of the cause element is as follows:

[0096] S = TF ij x IDF ij ​

[0097] wherein S represents the weight of the jthcausal element feature word in the ithhistorical hidden danger data, TF ij represents the frequency of the jthcausal element feature word in the ithhistorical hidden danger data, IDF ij represents the inverse text frequency of the jthcausal element feature word in the ithhistorical hidden danger data.

[0098] The specific calculation formula of the inverse text frequency of the causal element feature word is as follows:

[0099]

[0100] wherein x represents the number of historical hidden danger data containing the jthcausal element feature word in the ithhistorical hidden danger data, X represents the total number of historical hidden danger data containing the jthcausal element feature word in the ithhistorical hidden danger data, Y represents the total number of accident cases, and z is a constant and z>0.

[0101] The causal element feature words with higher weights are represented as corresponding TF-IDF vectors, a vector set of the causal element feature words is established, and the causal elements of the natural gas purification plant accidents are determined through the data clustering method.

[0102] In specific embodiments, a vector set of multiple causal element feature words is established according to the multiple causal element feature words, and the causal elements of the safety production accidents in the natural gas purification plant are determined, including: clustering the multiple causal element feature words through the data clustering method; and dividing the causal element feature words of different risk hidden dangers into different causal element sets according to the clustering relationship, wherein the causal element set and the causal element causing the safety production accident have a determined corresponding relationship.

[0103] The causal elements provided by the case analysis system are matrix processed, and the parameter learning function of the Bayesian network model is used to determine the accident probability of each node in the Bayesian network model. Specifically, taking the “acid water pressure sending” operation step as an example, among the causes of the corresponding process failure, the accident probability of “acid gas leakage” is 52%, and the accident probability of “not wearing a respirator” is 44%, so it can be considered that the possibility of failure of these two nodes is larger in daily production.

[0104] The embodiment of the safety risk management and control method based on the eye safety system provided by the application adjusts the accident probability of the "sulfur recovery unit" to 100%, and the rest of the nodes remain unchanged. By comparing the difference between the prior probability and the posterior probability, it is clear that in the case of failure of the "sulfur recovery unit" node, the key nodes with greater importance are analyzed. In this embodiment, the process failure with the largest difference between the prior probability and the posterior probability is "acid water pressure sending", and the difference between the prior probability and the posterior probability is 10%. It can be considered that this node is the key process failure factor leading to the failure of the "sulfur recovery unit".

[0105] The Bayesian network model of the sulfur recovery unit in the safety risk management and control method based on the eye safety system provided by the application performs sensitivity analysis on the hidden danger Bayesian network nodes of the "sulfur recovery unit", including nodes with higher sensitivity. According to the analysis result of the accident probability, the risk hidden danger is sorted, so as to realize the hierarchical management and control of the risk. According to the sensitivity analysis and the posterior probability analysis result, the node with higher sensitivity value and the difference between the posterior probability and the prior probability greater than 5% is defined as a high-risk node, the node with medium sensitivity value and the difference between the posterior probability and the prior probability greater than 2.5% and less than 5% is defined as a medium-risk node, and the node with lower sensitivity value and the difference between the posterior probability and the prior probability less than 2.5% is defined as a low-risk node.

[0106] In specific embodiments, the "acid water pressure sending" node has higher sensitivity and the difference between the posterior probability and the prior probability is greater than 5%, which is defined as a high-risk node. Therefore, when classifying risk hidden dangers, the causes of the process failure leading to the failure of the "sulfur condenser blowdown" operation step, i.e., "acid gas leakage", "acid water leakage" and "not wearing a respirator", should be sorted in the front; and the "sulfur condenser blowdown" node has lower sensitivity and the difference between the posterior probability and the prior probability is less than 2.5%, which is defined as a low-risk node. Therefore, when classifying hidden dangers, the process failure factors leading to the failure of the "sulfur condenser blowdown" operation step, i.e., "condenser failure", "personnel misoperation" and "inadequate daily maintenance", should be sorted in the back.

[0107] In specific embodiments, the intelligent identification algorithm of the eye protection system is equipped with a personnel violation behavior library, and the personnel violation behaviors recorded in the library include: not wearing a safety belt, not wearing a safety helmet, not wearing protective glasses, personnel intrusion, and violation of dangerous machinery operation, etc. When the personnel in the natural gas purification plant operate, the hidden danger intelligent identification module compares the behaviors of the on-site operating personnel with the violation operation behaviors in the violation operation behavior database. If there is a behavior with high similarity, it is determined as a violation behavior. The process and environmental parameters collected by the sensor and detector are compared with the parameter threshold values in the parameter threshold database, and abnormal process parameters and abnormal environmental parameters are intelligently identified, recorded, classified, and formed into a set of on-site risk hidden danger data. The set of on-site risk hidden danger data is compared with the risk hidden danger classification list, and the on-site risk hidden dangers identified by the eye protection system in the natural gas purification plant are sorted according to the cause factor sorting in the risk hidden danger classification list.

[0108] The embodiment of the present application also provides a risk management and control device based on the eye protection system, as shown in the figure, the device comprises: Figure 3

[0109] The data acquisition unit 301 is used for acquiring historical hidden danger data, and performing data clustering on the historical hidden danger data to determine a cause element. The historical hidden danger data is data collected by the eye protection system when a historical accident occurs. The cause element represents a reason element causing the historical accident in the historical hidden danger data.

[0110] The model establishment unit 302 is used for establishing a Bayesian network model of a system to be analyzed, and determining a hidden danger Bayesian network node.

[0111] The probability determination unit 303 is used for determining an accident probability of the hidden danger Bayesian network node and a hidden danger classification list according to the cause element.

[0112] The risk management and control unit 304 is used for performing risk management and control on the system to be analyzed based on the accident probability and the hidden danger classification list.

[0113] The embodiment of the present application also provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the method according to any one of the above embodiments are implemented.

[0114] The embodiment of the present application also provides a computer readable storage medium, which stores one or more programs executable by one or more processors to implement the steps of the method according to any one of the above embodiments.

[0115] As shown in the figure, Figure 4 ​A computer device is shown for providing an embodiment herein. The risk management and control method based on an eye safety system in the embodiment herein can be executed by the computer device in the embodiment. The computer device 402 can include one or more processors 404, such as one or more central processing units (CPUs), each of which can implement one or more threads. The computer device 402 can also include any memory 406 for storing any kind of information, such as code, settings, data, etc. Without limitation, for example, the memory 406 can include any one or combination of: any type of RAM, any type of ROM, a flash memory device, a hard disk, an optical disk, etc. More generally, any memory can use any technology for storing information. Further, any memory can provide volatile or non-volatile retention of information. Further, any memory can represent a fixed or removable component of the computer device 402. In one case, the computer device 402 can perform any operation of the associated instructions when executed by the processor 404, which are stored in any memory or combination of memories. The computer device 402 also includes one or more drive mechanisms 408, such as a hard disk drive mechanism, an optical disk drive mechanism, etc., for interacting with any memory.

[0116] The computer device 402 can also include an input / output module 410 (I / O) for receiving various inputs (via input devices 412) and for providing various outputs (via output devices 414). One particular output mechanism can include a presentation device 416 and an associated graphical user interface (GUI) 418. In other embodiments, the input / output module 410 (I / O), the input devices 412, and the output devices 414 can also not be included, just as a computer device in a network. The computer device 402 can also include one or more network interfaces 420 for exchanging data with other devices via one or more communication links 422. One or more communication buses 424 couple the above-described components together.

[0117] The communication links 422 can be implemented in any manner, such as through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication links 322 can include any combination of hardwired links, wireless links, router functionality, gateway functionality, name server functionality, etc., governed by any protocol or combination of protocols.

[0118] A computer readable storage medium storing one or more programs, the one or more programs executable by one or more processors to implement the steps of the method according to any one of the above described specific embodiments.

[0119] It should be understood that the size of the serial number of each process described above in various embodiments herein does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments herein.

[0120] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete embodiment, a complete software embodiment, or an embodiment combined with software and aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0121] The present application is described with reference to flowcharts and / or block diagrams according to the method, device (system), and computer program product of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the functions specified in the flowchart and / or block diagram.

[0122] These computer program instructions can also be stored in a computer-readable memory that can guide the computer or other programmable data processing apparatus to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or means for performing the functions specified in the flowchart and / or block diagram.

[0123] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or means for performing the functions specified in the flowchart and / or block diagram.

[0124] The foregoing description of specific exemplary embodiments of the application has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the application to the precise forms disclosed, and obviously many modifications and variations are possible in light of the above teaching. It is intended that the scope of the application be limited not with this detailed description, but rather by the claims appended hereto.

Claims

1. A risk management method based on an eye safety system, characterized in that, The method comprises: acquiring historical hazard data, performing data clustering on the historical hazard data to determine causation elements; the historical hazard data is data collected by an eye safety system when an accident occurred in the past; the causation elements represent reason elements that caused the accident in the past; establishing a Bayesian network model of a system to be analyzed to determine hazard Bayesian network nodes; determining accident probabilities of the hazard Bayesian network nodes and a hazard classification list according to the causation elements; controlling risks of the system to be analyzed based on the accident probabilities and the hazard classification list.

2. The risk management method based on the eye-care system according to claim 1, wherein, After determining the accident probabilities of the hazard Bayesian network nodes and the hazard classification list according to the causation elements, the method comprises: acquiring real-time data of the system to be analyzed based on the eye safety system to establish a field risk hazard data set; analyzing the field risk hazard data set based on the accident probabilities and the hazard classification list to control field risks of the system to be analyzed.

3. The risk management method based on the eye-care system according to claim 1, wherein, Establishing a Bayesian network model of a system to be analyzed comprises: dividing the system to be analyzed into multiple process units; acquiring operation steps of each process unit and analyzing deviable data of each operation step; the deviable data represents possible deviations of the operation step; acquiring process failure factors corresponding to the operation step according to the deviable data; the process failure factors are reasons for accidents of the system to be analyzed caused by the deviable data; setting the hazard Bayesian network nodes based on the operation steps of the process units and the process failure factors.

4. The risk management method based on the eye-care system according to claim 3, wherein, After determining the hazard Bayesian network nodes, the method comprises: determining levels and dependency relationships of the hazard Bayesian network nodes according to correlations between the hazard Bayesian network nodes to obtain a topology graph of the Bayesian network.

5. The risk management method based on the eye-care system according to claim 4, wherein, The hazard Bayesian network nodes comprise child nodes, intermediate nodes and parent nodes; the method comprises: setting the child nodes as the process units, setting the intermediate nodes as operation steps corresponding to the process units, and setting the parent nodes as process failure factors corresponding to the operation steps.

6. The risk management method based on the eye-care system according to claim 5, wherein, Determining accident probabilities of the hazard Bayesian network nodes according to the causation elements comprises: constructing a parameter matrix of the system to be analyzed, the parameter matrix representing a learning matrix of causation elements of the system to be analyzed; learning the parameter matrix by using the Bayesian network model to obtain the accident probabilities of the hazard Bayesian network nodes.

7. The risk management method based on the eye-care system according to claim 6, wherein, After determining the accident probabilities of the hazard Bayesian network nodes, the method further comprises: correcting the accident probabilities of child nodes in the hazard Bayesian network nodes.

8. The risk management method based on the eye-care system according to claim 5, wherein, The method further comprises: performing sensitivity analysis on the process failure factors.

9. A risk management and control device based on an eye safety system, characterized in that, The device comprises: a data acquisition unit configured to acquire historical hazard data, perform data clustering on the historical hazard data to determine causation elements; the historical hazard data is data collected by an eye safety system when an accident occurred in the past; the causation elements represent reason elements that caused the accident in the past; A model establishing unit is configured to establish a Bayesian network model of a system to be analyzed and determine a hidden danger Bayesian network node; A probability determining unit is configured to determine an accident probability of the hidden danger Bayesian network node and a hidden danger classification list according to the cause elements. A risk management and control unit is configured to perform risk management and control on the system to be analyzed based on the accident probability and the hidden danger classification list.

10. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the steps of the method in any one of claims 1 to 8 when executing the program.

11. A computer readable storage medium, characterized in that, The computer readable storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the method in any one of claims 1 to 8.

12. A computer program product, characterised in that, The computer program product comprises at least one instruction or at least one program, which is loaded and executed by the processor to implement the method in any one of claims 1 to 8.

Images