Liquefied gas station safety risk assessment method and system

Through real-time data monitoring and Bayesian network analysis, combined with fuzzy set theory, the problems of accuracy and objectivity of traditional methods in the safety risk assessment of liquefied gas stations are solved, and risk assessment and accident prevention of liquefied gas stations are realized.

CN118982224APending Publication Date: 2024-11-19SHANGHAI INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410835205.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-26
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

Traditional accident trees, event trees and hierarchical analysis methods have problems such as difficulty in calculation and unified processing, lack of accuracy, reliability and objectivity in the safety risk assessment of liquefied gas stations.

Method used

Using real-time data monitoring, Bayesian network and fuzzy set theory, the risk factors of liquefied gas stations are obtained through sensors, and a safety risk index system is built, and Bayesian inference and reverse reasoning analysis is used to calculate the risk probability of accidents and identify the fundamentals of the accident. reason.

Benefits of technology

Real-time monitoring and rapid response to the safety risk assessment of liquefied gas stations is achieved, improving the accuracy and adaptability of the assessment, and providing in-depth insights to help effectively manage and mitigate safety risks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118982224A_ABST
    Figure CN118982224A_ABST
Patent Text Reader

Abstract

The invention discloses a liquefied gas station safety risk assessment method. The method comprises the following steps: acquiring real-time data of a sensor and risk factors of a liquefied gas station; presetting a first judgment criterion, performing index selection on the real-time data of the sensor and the risk factors of the liquefied gas station, and constructing a liquefied petroleum gas station safety risk index system according to the first index set; presetting a second judgment criterion, performing index selection on the constructed Bayesian network structure according to the liquefied petroleum gas station safety risk index system, and obtaining a network structure according to the second index set in combination with a constraint learning method; fuzzifying values of nodes in the network structure, calculating a parameter value of conditional probability distribution of each node in the network structure, calculating a risk probability of an accident of the liquefied gas station by using a Bayesian inference method and conditional probability distribution among the nodes in the network structure for the parameter values, and performing reverse reasoning analysis to obtain a risk probability of the accident of the liquefied gas station. And factors of accidents of the liquefied gas station can be obtained, so that safety risks can be better managed and reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of risk assessment of liquefied gas stations, and in particular to a method and system for safety risk assessment of liquefied gas stations. Background Art

[0003] In order to cope with the serious hazards that may be caused by liquefied petroleum gas stations, it is necessary to strengthen management systems, standardize operating procedures, improve monitoring systems and optimize risk assessment methods to improve the level of safety management and establish a scientific and reasonable monitoring and safety risk assessment system. Use artificial intelligence, data transmission technology and big data technology to achieve real-time monitoring of equipment operating status, and promptly detect and handle abnormal situations. Combine Bayesian networks with historical data and expert knowledge to analyze the association between multiple variables to help predict and evaluate safety risks. Comprehensively apply dynamic monitoring and Bayesian networks to improve the ability to identify safety hazards of liquefied petroleum gas stations, provide continuous data support and feedback during the implementation of preventive measures, thereby effectively strengthening the safety management system and ensuring the safety of personnel and facilities. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the present invention solves the problem that the traditional accident tree can clearly show the cause, process and result of the accident, but is not conducive to calculation and unified processing; the event tree can clearly analyze the possible consequences of the event, but cannot perform quantitative analysis; the hierarchical analysis method can perform trade-off evaluation, but it is highly subjective and lacks accuracy, reliability and objectivity.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for assessing safety risks of a liquefied gas station, comprising:

[0008] Get real-time data from sensors and risk factors of LPG stations;

[0009] A first judgment criterion is preset, and indicators are selected for the real-time data of the sensor and the risk factors of the liquefied gas station to obtain a first indicator set, and a liquefied petroleum gas station safety risk indicator system is constructed according to the first indicator set;

[0010] A second judgment criterion is preset, and indicators are selected for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system to obtain a second indicator set, and a network structure is obtained according to the second indicator set in combination with a constraint learning method;

[0011] The values ​​of the nodes in the network structure are fuzzified by using fuzzy set theory, the parameter value of the conditional probability distribution of each node in the network structure is calculated, and the conditional probability distribution between the nodes in the network structure is obtained by using the Bayesian inference method for the parameter value;

[0012] The risk probability of an accident at a liquefied gas station is calculated using the conditional probability distribution between nodes in the network structure, and the posterior probabilities of risk factors are obtained through reverse reasoning analysis. The posterior probabilities are sorted to obtain the factors causing the accident at the liquefied gas station.

[0013] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, wherein: presetting a first judgment criterion, selecting indicators for the real-time data of the sensor and the risk factors of the liquefied gas station includes:

[0014] The risk factors include personnel risk-related parameters, equipment risk-related parameters, performance risk-related parameters, and environmental risk-related parameters as primary indicators of liquefied gas stations;

[0015] Calculating an abnormality detection rate of the liquefied gas station according to real-time data of the sensor, and setting a first threshold and a second threshold according to the abnormality detection rate;

[0016] According to the first judgment criterion, if the abnormal detection rate of the first-level indicator is less than or equal to the first threshold, at least three personnel risk-related parameters are selected as second-level evaluation indicators; at least two equipment risk-related parameters are selected as second-level evaluation indicators; at least three performance risk-related parameters are selected as second-level evaluation indicators; at least one environment-related parameter is selected as a second-level evaluation indicator;

[0017] If the anomaly detection rate of the first-level indicator is greater than the first threshold and less than the second threshold, at least four personnel risk-related parameters are selected as second-level evaluation indicators; at least three equipment risk-related parameters are selected as second-level evaluation indicators; at least four performance risk-related parameters are selected as second-level evaluation indicators; at least two environment-related parameters are selected as second-level evaluation indicators;

[0018] If the anomaly detection rate of the first-level indicator is greater than or equal to the second threshold, at least five personnel risk-related parameters are selected as second-level evaluation indicators; at least four equipment risk-related parameters are selected as second-level evaluation indicators; at least five performance risk-related parameters are selected as second-level evaluation indicators; and at least three environment-related parameters are selected as second-level evaluation indicators.

[0019] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, the selection of indicators for constructing the Bayesian network structure includes:

[0020] Calculate the abnormal detection rate of each secondary indicator of the risk factor of the liquefied gas station, and compare the abnormal detection rate of each secondary indicator of the risk factor with the abnormal detection rate of the liquefied gas station through the second judgment criterion.

[0021] If the abnormal detection rate of the secondary indicator is less than or equal to the first threshold, the secondary indicator is not selected to construct the Bayesian network structure;

[0022] If the abnormality detection rate of the secondary indicator is greater than the first threshold and less than the second threshold, a third judgment is performed, wherein the third judgment includes calculating the abnormality degree of the secondary indicator. If it is greater than the set third threshold, the secondary indicator is selected to construct the Bayesian network structure; if it is less than or equal to the set third threshold, the secondary indicator is not selected to construct the Bayesian network structure;

[0023] If the anomaly detection rate of the secondary indicator is greater than the second threshold, the secondary indicator is selected to construct the Bayesian network structure.

[0024] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, wherein: according to the second indicator, the network structure obtained based on the constraint learning method includes:

[0025] Let X i , X j are two risk indicator nodes on the network structure X, x i , x j X i , X j Let C be the value of X that is not X. i , X j The set of risk indicator nodes, then X i , X j The conditional mutual information about C is expressed as:

[0026]

[0027] Among them, I(X i ,X j |C) is the condition under which C is given, X i , X j The mutual information between i ,x j ,c) is a given x i , x j The probability of C occurring at the same time, P(c) is the probability of a given C occurring, P(x i ,c) and P(x j ,c) is the given C occurs, x i and x j Probability of occurrence;

[0028] If I(X i ,X j |C) is larger, the risk indicator node X will be affected by other risk indicator node sets C. i , X j The mutual influence between them is large; if I(X i ,X j |C) is close to zero, then under the influence of other risk indicator node set C, X i , X j Independent of each other, we get the optimal network structure.

[0029] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, wherein: using fuzzy set theory to fuzzify the values ​​of the nodes in the network structure includes:

[0030] The values ​​of nodes in the network structure are divided into different levels, and the value of each node is scored by experts to obtain the scoring results. The membership of the scoring results is calculated by triangular fuzzy numbers.

[0031] Assume u and l represent the upper and lower limits of the fuzzy number value range of the node, respectively, and m is the mean value. Then the fuzzy number of the node is expressed as (l, m, u), and the membership degree φ(z) of the fuzzy number of the node is calculated as:

[0032]

[0033] The membership degree of each node is combined with the corresponding level to obtain the fuzzy set of nodes.

[0034] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, wherein: using the Bayesian inference method for the parameter value includes:

[0035] The expert judgment matrix is ​​established through expert level scoring, and the expert weights are obtained according to the judgment matrix. Using the expert scores and weights, the node probability of the network structure is expressed as:

[0036]

[0037] Among them, θ is the weight index of the node falling into the kth interval, k is the interval position of the expert score, and p l is the probability that a node falls into the lth interval, a k -a l is the interpolation between the k-th expert and the l-th expert.

[0038] As a preferred solution of the liquefied gas station safety risk assessment method of the present invention, the posterior probability of the risk factors obtained through reverse reasoning analysis includes:

[0039] Extract independent risk indicator nodes (x1, x2, ..., x n ), the joint probability density function of the nodes is P(x,α)=π(α)P(x|α), where π(α) is the prior probability, and the posterior probability of the nodes in the network structure is expressed as:

[0040]

[0041] The parameter α is expressed as:

[0042]

[0043] The posterior probabilities are sorted to obtain the factors causing accidents at liquefied gas stations.

[0044] In a second aspect, the present invention provides a system for safety risk assessment of a liquefied gas station, comprising:

[0045] A data acquisition module, used to obtain real-time data of sensors and risk factors of liquefied gas stations;

[0046] A first indicator selection module is used to preset a first judgment criterion, select indicators based on the real-time data of the sensor and the risk factors of the liquefied gas station, obtain a first indicator set, and construct a liquefied petroleum gas station safety risk indicator system according to the first indicator set;

[0047] A second indicator selection module is used to preset a second judgment criterion, select indicators for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system, obtain a second indicator set, and obtain a network structure according to the second indicator set in combination with a constraint learning method;

[0048] A calculation module is used to fuzzify the values ​​of the nodes in the network structure by using fuzzy set theory, calculate the parameter value of the conditional probability distribution of each node in the network structure, and use the Bayesian inference method to obtain the conditional probability distribution between the nodes in the network structure;

[0049] The sorting module is used to calculate the risk probability of an accident occurring in a liquefied gas station by using the conditional probability distribution between nodes in the network structure, obtain the posterior probability of the risk factors through reverse reasoning analysis, sort the posterior probabilities, and obtain the factors causing the accident occurring in the liquefied gas station.

[0050] In a third aspect, the present invention provides a computing device, comprising:

[0051] Memory and processor;

[0052] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the liquefied gas station safety risk assessment method are implemented.

[0053] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the liquefied gas station safety risk assessment method.

[0054] Beneficial effects of the present invention: The liquefied gas station safety risk assessment method proposed in the present invention can detect abnormal situations in time and make rapid responses through real-time data monitoring, and applies fuzzy set theory to effectively deal with uncertainty and ambiguity in risk assessment. The Bayesian network allows the risk assessment to be continuously updated as new data is obtained, thereby improving the accuracy of the assessment. Reverse reasoning analysis helps to identify the root causes of accidents and provide a basis for preventive measures. It can adapt to dynamically changing environments, integrate multi-source data, and provide in-depth insights to better manage and mitigate safety risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0056] Figure 1 An overall flow chart of a liquefied gas station safety risk assessment method provided by the present invention;

[0057] Figure 2 A liquefied gas station safety risk index system for a liquefied gas station safety risk assessment method provided by the present invention;

[0058] Figure 3 A liquefied gas station accident Bayesian network structure of a liquefied gas station safety risk assessment method provided by the present invention. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0062] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0063] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0064] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0065] Example 1

[0066] Reference Figure 1 , as an embodiment of the present invention, provides a liquefied gas station safety risk assessment method, comprising:

[0067] S100: Obtaining real-time data from sensors and risk factors of liquefied gas stations;

[0068] It should be noted that the construction of a remote monitoring system for liquefied gas stations and real-time data collection, and the construction of a remote monitoring system framework for liquefied gas stations based on industrial gateways, include the perception layer, network layer, and application layer. The perception layer is the part of the remote monitoring system for liquefied gas stations responsible for real-time data collection and transmission. Through field instruments, field buses, and communication modules, the perception layer realizes the collection and transmission of liquefied gas station operation data, and uses explosion-proof field instruments, mainly including temperature, pressure, liquid level, flow, gas concentration alarms, and video monitoring. These data reflect the real-time operation of the liquefied gas station and provide data support for subsequent risk assessments.

[0069] In addition to personnel risk-related parameters, equipment risk-related parameters, performance risk-related parameters, and environmental risk-related parameters, risk factors can also include laws and regulations, emergency response, social and economic risks, and other aspects.

[0070] S200: Preset a first judgment criterion, select indicators based on the real-time data of the sensor and the risk factors of the liquefied gas station, obtain a first indicator set, and construct a liquefied petroleum gas station safety risk indicator system according to the first indicator set;

[0071] Furthermore, the risk factors include personnel risk-related parameters, equipment risk-related parameters, performance risk-related parameters, and environmental risk-related parameters as primary indicators of liquefied gas stations;

[0072] Calculate the abnormality detection rate of the liquefied gas station according to the real-time data of the sensor, and set a first threshold and a second threshold according to the abnormality detection rate;

[0073] According to the first judgment criterion, if the abnormal detection rate of the first-level indicator is less than or equal to the first threshold, at least three personnel risk-related parameters are selected as second-level evaluation indicators; at least two equipment risk-related parameters are selected as second-level evaluation indicators; at least three performance risk-related parameters are selected as second-level evaluation indicators; at least one environment-related parameter is selected as a second-level evaluation indicator;

[0074] If the anomaly detection rate of the first-level indicator is greater than the first threshold and less than the second threshold, at least four personnel risk-related parameters are selected as second-level evaluation indicators; at least three equipment risk-related parameters are selected as second-level evaluation indicators; at least four performance risk-related parameters are selected as second-level evaluation indicators; at least two environment-related parameters are selected as second-level evaluation indicators;

[0075] If the anomaly detection rate of the first-level indicator is greater than or equal to the second threshold, at least five personnel risk-related parameters are selected as second-level evaluation indicators; at least four equipment risk-related parameters are selected as second-level evaluation indicators; at least five performance risk-related parameters are selected as second-level evaluation indicators; and at least three environment-related parameters are selected as second-level evaluation indicators.

[0076] It should be noted that historical normal operation data is obtained, the normal range and distribution model of each monitoring indicator is established, the real-time data is substituted into the normal model, the degree of deviation between the current data and the normal model is calculated, and the current anomaly detection rate is calculated according to the degree of deviation. The higher the anomaly detection rate, the more abnormal the current data is. The first threshold and the second threshold are set according to the risk tolerance;

[0077] It should also be noted that the personnel risk-related parameters include the number of safety production personnel, the quality of the main person in charge's supervisor, the quality of the maintenance personnel, safety operating procedures, safety production system, regular inspections, and emergency drills. The legal operation of a liquefied gas station involves multiple considerations; equipment risk-related parameters include liquefied gas storage tanks, loading and unloading systems, gasification devices, electrical systems, lightning protection and anti-static devices, etc. The safety considerations within the liquefied gas station include but are not limited to whether the overall volume of the liquefied gas tank and the distance from the station road meet the regulations and whether the fire protection distance is sufficient. Whether the loading and unloading system is equipped with anti-collision devices at fixed parking positions, whether measures are taken to prevent the movement of vehicles during loading and unloading operations, and whether interlocking protection devices are set up to prevent the pulling-off of loading and unloading pipes; performance risk-related parameters include humidity, temperature, pressure, liquid level, flow, gas concentration alarms and video monitoring, etc., to conduct all-round monitoring of the liquefied gas station, and to be able to promptly detect abnormal situations and take corresponding measures in a timely manner; environmental risk-related parameters include site layout, dangerous buildings in the station, site maintenance, etc. The rationality of the site layout, including but not limited to whether the production area is equipped with a solid fence of not less than 2m, and whether the fire protection distance between different areas meets the requirements, are all important considerations.

[0078] In one embodiment, the risk factors take into account the personnel risk-related parameters, equipment risk-related parameters, performance risk-related parameters and environmental risk-related parameters, the degree of deviation between the current data and the normal model, and according to the degree of deviation, the current abnormality detection rate is calculated to be 80%, then the first threshold is set to 70%, and the second threshold is set to 90%; the abnormality detection rate of the first-level indicator is calculated, the abnormality detection rate of personnel behavior is 95%, the abnormality detection rate of facilities and equipment is 75%, the abnormality detection rate of physical performance is 80%, and the abnormality detection rate of the site environment is 65%;

[0079] If the abnormal detection rate of personnel behavior is 95% and greater than the second threshold, then the personnel risk can select asset proof, number of production safety personnel, quality of main supervisors, quality of maintenance personnel, production safety specifications, etc. as secondary evaluation indicators; if the abnormal detection rate of facilities and equipment is 75% and greater than the first threshold but less than the second threshold, then liquefied gas storage tanks, loading and unloading systems, electrical systems, gas telephone devices, etc. can be selected as secondary evaluation indicators; if the abnormal detection rate of physical properties is 80% and greater than the first threshold, then the performance risk can select temperature, humidity, pressure, etc. as secondary evaluation indicators; if the abnormal detection rate of site environment is 65% and less than the first threshold, then the environmental risk can select site layout, etc. as secondary evaluation indicators;

[0080] The safety risk indicator system for liquefied gas stations includes primary indicators and secondary indicators.

[0081] S300: Preset a second judgment criterion, select indicators for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system, obtain a second indicator set, and obtain a network structure according to the second indicator set in combination with a constraint learning method;

[0082] Furthermore, the abnormal detection rate of each secondary indicator of the risk factor of the liquefied gas station is calculated, and the abnormal detection rate of each secondary indicator of the risk factor is compared with the abnormal detection rate of the liquefied gas station through the second judgment criterion.

[0083] If the abnormal detection rate of the secondary indicator is less than or equal to the first threshold, the secondary indicator is not selected to construct the Bayesian network structure;

[0084] If the abnormality detection rate of the secondary indicator is greater than the first threshold and less than the second threshold, a third judgment is performed, which includes calculating the abnormality degree of the secondary indicator. If it is greater than the set third threshold, the secondary indicator is selected to construct the Bayesian network structure. If it is less than or equal to the set third threshold, the secondary indicator is not selected to construct the Bayesian network structure.

[0085] If the anomaly detection rate of the secondary indicator is greater than the second threshold, the secondary indicator is selected to construct the Bayesian network structure.

[0086] It should be noted that the third threshold is set according to the degree of abnormality of the indicator;

[0087] In one possible embodiment, the current anomaly detection rate is 80%, so the first threshold is set to 70%, the second threshold is set to 90%, and the secondary indicators of personnel risk are asset proof, number of production safety personnel, quality of main guardians, quality of maintenance personnel, production safety specifications, etc. The anomaly detection rates are 30%, 70%, 80%, 85%, and 90% respectively. If the asset proof 30% and the number of production safety personnel 70% are both less than the first threshold, then the indicator is not selected to construct the Bayesian network structure; the quality of the main guardian is 80%, and the quality of the maintenance personnel is 85%, which is between the rate greater than the first threshold and less than the second threshold, then their abnormality levels are 1 and 2, and the third threshold is set to 1.5. If the abnormality level of the main guardian is less than the third threshold, then it is not selected; if the quality of the maintenance personnel is greater than the abnormality level, then the Bayesian network structure is selected to be constructed; if the production safety specifications are 90% greater than the second threshold, then the Bayesian network structure is selected to be constructed.

[0088] Furthermore, suppose X i , X j are two risk indicator nodes on the network structure X, x i , x j X i , X j Let C be the value of X that is not X. i , X j The set of risk indicator nodes, then X i , X j The conditional mutual information about C is expressed as:

[0089]

[0090] Among them, I(X i ,X j |C) is the condition under which C is given, X i , X j The mutual information between i ,x j ,c) is a given x i , x j The probability of C occurring at the same time, P(c) is the probability of a given C occurring, P(x i ,c) and P(x j ,c) is the given C occurs, x i and x j Probability of occurrence;

[0091] If I(X i ,X j |C) is larger, the risk indicator node X will be affected by other risk indicator node sets C. i , X j The mutual influence between them is large; if I(Xi ,X j |C) is close to zero, then under the influence of other risk indicator node set C, X i , X j Independent of each other, we get the optimal network structure.

[0092] It should be noted that the constraint-based learning method learns the Bayesian network structure by calculating the conditional independence and confidence between nodes, and obtains the optimal Bayesian network structure through the dependency relationship between nodes.

[0093] S400: Fuzzify the values ​​of nodes in the network structure using fuzzy set theory, calculate the parameter value of the conditional probability distribution of each node in the network structure, and use the Bayesian inference method for the parameter value to obtain the conditional probability distribution between the nodes in the network structure;

[0094] Furthermore, the values ​​of nodes in the network structure are divided into different levels, and the value of each node is scored by experts to obtain the scoring results. The membership of the scoring results is calculated by triangular fuzzy numbers.

[0095] Assume u and l represent the upper and lower limits of the fuzzy number value range of the node, respectively, and m is the mean value. Then the fuzzy number of the node is expressed as (l, m, u), and the membership degree φ(z) of the fuzzy number of the node is calculated as:

[0096]

[0097] The membership degree of each node is combined with the corresponding level to obtain the fuzzy set of nodes.

[0098] It should be noted that different methods of determining the probability of occurrence are used for different nodes. For nodes with sample data, the ratio of the number of accidents calculated from historical data to the total number of samples is used as the prior probability. For nodes that are difficult to quantify, expert experience combined with triangular fuzzy numbers are used to describe them.

[0099] Occurrence probability, let B1, B2, …, B n is a risk event in the sample space τ, if B i The probability of occurrence P(B i ) can be estimated based on the analysis of existing data or expert knowledge, i.e., prior knowledge, then P(B i ) is the prior probability. The conditional probability is that A and B represent two risk events, and P(A)≠0. Given the condition that risk event A occurs, the conditional probability of risk event B occurring is:

[0100] Furthermore, an expert judgment matrix is ​​established through expert level scoring, and expert weights are obtained according to the judgment matrix. Using the expert scores and weights, the node probability of the network structure is expressed as:

[0101]

[0102] Among them, θ is the weight index of the node falling into the kth interval, k is the interval position of the expert score, and p l is the probability that a node falls into the lth interval, a k -a l is the interpolation between the k-th expert and the l-th expert.

[0103] It should be noted that the occurrence level of each risk event is divided into four status levels, namely "high", "medium", "low" and "very low". Four experts are selected to grade the selected nodes, and the fuzzy set of the selected nodes is calculated according to the basic operation rules of fuzzy numbers;

[0104] It should also be noted that, according to the expert weight grading method, the evaluation experts will use the experts' education level, professional title and years of work experience as evaluation indicators to establish an expert indicator evaluation system. The relative importance of each indicator is obtained through the judgment matrix to obtain the expert weight; the expert confidence index method is used to score the experts, which is an evaluation of the reliability and recognition of the judgments made by the experts. The index range is clearly determined to be [0,1]. The expert confidence weight index of the probability level of the basic event is the product of the weight index and the information index.

[0105] The experts' scoring form is the conversion of probability values ​​into probability intervals. Experts score the selected probability intervals, indicating that the probability of the basic event occurring in this interval is the highest. At the same time, the triangular fuzzy numbers approximately satisfy the characteristics of the normal distribution, and the probability of being on both sides of the scoring interval decreases successively.

[0106] S500: Calculate the risk probability of an accident at the liquefied gas station using the conditional probability distribution between nodes in the network structure, obtain the posterior probability of the risk factors through reverse reasoning analysis, sort the posterior probabilities, and obtain the factors of the accident at the liquefied gas station;

[0107] Furthermore, independent risk indicator nodes (x1, x2, ..., x n ), the joint probability density function of the nodes is P(x,α)=π(α)P(x|α), where π(α) is the prior probability, and the posterior probability of the nodes in the network structure is expressed as:

[0108]

[0109] The parameter α is expressed as:

[0110]

[0111] The posterior probabilities are sorted to obtain the factors causing accidents at liquefied gas stations.

[0112] It should be noted that Python is used for Bayesian network modeling and simulation, and the Bayesian network is used to calculate the risk probability of accidents at liquefied gas stations. The posterior probability of risk factors under given conditions can also be inferred, and the key factors leading to accidents at liquefied gas stations can be found out by sorting by probability.

[0113] Through the reverse reasoning analysis of the Bayesian network, we can study the influence of different nodes on the probability of occurrence of the root node. By setting the state of the central node to occur, that is, the probability of occurrence of the liquefied gas station accident is 100%, the posterior probabilities of different nodes can be obtained. Through the probabilities of each node in the presented Bayesian network diagram, we can see which nodes have the most prominent influence on the central node.

[0114] The above is a schematic scheme of a liquefied gas station safety risk assessment method of this embodiment. It should be noted that the technical scheme of the liquefied gas station safety risk assessment device and the technical scheme of the above-mentioned liquefied gas station safety risk assessment method belong to the same concept, and the details of the technical scheme of the liquefied gas station safety risk assessment device not described in detail in this embodiment can be found in the description of the technical scheme of the above-mentioned liquefied gas station safety risk assessment method.

[0115] The device for safety risk assessment of a liquefied gas station in this embodiment includes:

[0116] A data acquisition module, used to obtain real-time data of sensors and risk factors of liquefied gas stations;

[0117] A first indicator selection module is used to preset a first judgment criterion, select indicators based on the real-time data of the sensor and the risk factors of the liquefied gas station, obtain a first indicator set, and construct a liquefied petroleum gas station safety risk indicator system according to the first indicator set;

[0118] A second indicator selection module is used to preset a second judgment criterion, select indicators for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system, obtain a second indicator set, and obtain a network structure according to the second indicator set in combination with a constraint learning method;

[0119] A calculation module is used to fuzzify the values ​​of the nodes in the network structure by using fuzzy set theory, calculate the parameter value of the conditional probability distribution of each node in the network structure, and use the Bayesian inference method to obtain the conditional probability distribution between the nodes in the network structure;

[0120] The sorting module is used to calculate the risk probability of an accident occurring in a liquefied gas station by using the conditional probability distribution between nodes in the network structure, obtain the posterior probability of the risk factors through reverse reasoning analysis, sort the posterior probabilities, and obtain the factors causing the accident occurring in the liquefied gas station.

[0121] This embodiment also provides a computing device suitable for evaluating the safety risk of a liquefied gas station, including:

[0122] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the liquefied gas station safety risk assessment method proposed in the above embodiment.

[0123] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for evaluating the safety risk of a liquefied gas station proposed in the above embodiment is implemented.

[0124] The storage medium proposed in this embodiment and the method for implementing the safety risk assessment of a liquefied gas station proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0125] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0126] Example 2

[0127] Reference Figure 2-3 As shown, this embodiment is different from the first embodiment and provides a verification test of a liquefied gas station safety risk assessment method to verify and illustrate the technical effects adopted in this method.

[0128] Take liquefied gas station as an example.

[0129] The following real-time monitoring data were obtained: Temperature sensor data: normal range is -20℃ to 50℃, current temperature is 35℃; Pressure sensor data: normal range is 0.8MPa to 1.2MPa, current pressure is 1.1MPa; Liquid level sensor data: normal range is 80% to 95%, current liquid level is 90%; Flow sensor data: normal range is 50m 3 / h to 100m 3 / h, the current flow rate is 80m 3 / h;

[0130] There are also potential risk factors, equipment failure risk: According to historical data, the probability of equipment failure is 8%;

[0131] Risk of human error: According to statistical analysis, the probability of human error is 12%;

[0132] External environmental factors: The current ambient temperature is 25°C and the humidity is 60%;

[0133] According to the first judgment criterion, the safety risk index system of liquefied petroleum gas stations is obtained, such as Figure 2 As shown, the first-level indicators include: personnel behavior, facilities and equipment, physical properties, and site environment; the second-level indicators include: asset certification for personnel behavior, number of safety production personnel, quality of main guardians, quality of maintenance personnel, safety production specifications, safety operation specifications, regular inspections, and emergency drills; liquefied gas storage tanks, loading and unloading systems, electrical systems, gasification devices, and lightning protection and static electricity devices for facilities and equipment; humidity, temperature, pressure, liquid level, flow, gas concentration alarms, and video surveillance for physical properties; site environment site layout, dangerous buildings in the station, and site maintenance;

[0134] like Figure 3 As shown in the figure, according to the second judgment criterion, the Bayesian network structure is obtained: root node: equipment failure probability (8%), operation error probability (12%), ambient temperature (25°C), ambient humidity (60%); intermediate nodes: temperature (35°C), pressure (1.1MPa), liquid level (90%), flow rate (80m 3 / h); Target node: probability of liquefied gas leakage accident;

[0135] Using fuzzy set theory and Bayesian inference, the following conditional probability distribution of network nodes is obtained:

[0136] P (temperature abnormality | equipment failure) = 60%; P (pressure abnormality | equipment failure) = 50%;

[0137] P(liquid level abnormality|equipment failure)=45%; P(flow abnormality|equipment failure)=55%;

[0138] P (abnormal flow | operation error) = 75%; P (liquefied gas leakage | abnormal temperature, abnormal pressure, abnormal liquid level, abnormal flow) = 85%;

[0139] Through reverse reasoning analysis, the posterior probability of each risk factor can be obtained:

[0140] P(equipment failure|liquefied gas leakage)=60%;

[0141] P(operation error|liquefied gas leakage)=45%;

[0142] P(environmental factors|liquefied gas leakage)=30%;

[0143] It is found that equipment failure and operating errors are key risk factors leading to liquefied gas leakage accidents.

[0144] Through real-time sensor data such as temperature, pressure, liquid level, and flow, abnormal equipment operation can be discovered in a timely manner and accidents can be prevented in a timely manner; the main risk factors such as equipment failure risk, human operation error risk, and external environmental factors can be accurately identified to lay the foundation for the next step of risk analysis. The Bayesian network model is used for probability inference, the probability of accidents is quantitatively calculated, and the importance of each risk factor is analyzed to provide a basis for risk control.

[0145] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for assessing safety risks of a liquefied gas station, characterized in that: include: Get real-time data from sensors and risk factors of LPG stations; A first judgment criterion is preset, and indicators are selected for the real-time data of the sensor and the risk factors of the liquefied gas station to obtain a first indicator set, and a liquefied petroleum gas station safety risk indicator system is constructed according to the first indicator set; A second judgment criterion is preset, and indicators are selected for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system to obtain a second indicator set, and a network structure is obtained according to the second indicator set in combination with a constraint learning method; The values ​​of the nodes in the network structure are fuzzified by using fuzzy set theory, the parameter value of the conditional probability distribution of each node in the network structure is calculated, and the conditional probability distribution between the nodes in the network structure is obtained by using the Bayesian inference method for the parameter value; The risk probability of an accident at a liquefied gas station is calculated using the conditional probability distribution between nodes in the network structure, and the posterior probabilities of risk factors are obtained through reverse reasoning analysis. The posterior probabilities are sorted to obtain the factors causing the accident at the liquefied gas station.

2. The liquefied gas station safety risk assessment method according to claim 1, characterized in that: The first judgment criterion is preset to select indicators for the real-time data of the sensor and the risk factors of the liquefied gas station, including: The risk factors include personnel risk-related parameters, equipment risk-related parameters, performance risk-related parameters, and environmental risk-related parameters as primary indicators of liquefied gas stations; Calculating an abnormality detection rate of the liquefied gas station according to real-time data of the sensor, and setting a first threshold and a second threshold according to the abnormality detection rate; According to the first judgment criterion, if the abnormal detection rate of the first-level indicator is less than or equal to the first threshold, at least three personnel risk-related parameters are selected as second-level evaluation indicators; at least two equipment risk-related parameters are selected as second-level evaluation indicators; at least three performance risk-related parameters are selected as second-level evaluation indicators; at least one environment-related parameter is selected as a second-level evaluation indicator; If the anomaly detection rate of the first-level indicator is greater than the first threshold and less than the second threshold, at least four personnel risk-related parameters are selected as second-level evaluation indicators; at least three equipment risk-related parameters are selected as second-level evaluation indicators; at least four performance risk-related parameters are selected as second-level evaluation indicators; at least two environment-related parameters are selected as second-level evaluation indicators; If the anomaly detection rate of the first-level indicator is greater than or equal to the second threshold, at least five personnel risk-related parameters are selected as second-level evaluation indicators; at least four equipment risk-related parameters are selected as second-level evaluation indicators; at least five performance risk-related parameters are selected as second-level evaluation indicators; and at least three environment-related parameters are selected as second-level evaluation indicators.

3. The liquefied gas station safety risk assessment method according to claim 2, characterized in that: The selection of indicators for constructing the Bayesian network structure includes: Calculate the abnormal detection rate of each secondary indicator of the risk factor of the liquefied gas station, and compare the abnormal detection rate of each secondary indicator of the risk factor with the abnormal detection rate of the liquefied gas station through the second judgment criterion. If the abnormal detection rate of the secondary indicator is less than or equal to the first threshold, the secondary indicator is not selected to construct the Bayesian network structure; If the abnormality detection rate of the secondary indicator is greater than the first threshold and less than the second threshold, a third judgment is performed, wherein the third judgment includes calculating the abnormality degree of the secondary indicator. If it is greater than the set third threshold, the secondary indicator is selected to construct the Bayesian network structure; if it is less than or equal to the set third threshold, the secondary indicator is not selected to construct the Bayesian network structure; If the anomaly detection rate of the secondary indicator is greater than the second threshold, the secondary indicator is selected to construct the Bayesian network structure.

4. The liquefied gas station safety risk assessment method according to claim 1 or 3, characterized in that: According to the second indicator, the network structure obtained based on the constraint learning method includes: Let X i , X j are two risk indicator nodes on the network structure X, x i , x j X i , X j Let C be the value of X that is not X. i , X j The set of risk indicator nodes, then X i , X j The conditional mutual information about C is expressed as: Among them, I(X i ,X j |C) is the condition under which C is given, X i , X j The mutual information between i ,x j ,c) is a given x i , x j The probability of C occurring at the same time, P(c) is the probability of a given C occurring, P(x i ,c) and P(x j ,c) is the given C occurs, x i and x j Probability of occurrence; If I(X i ,X j |C) is larger, the risk indicator node X will be affected by other risk indicator node sets C. i , X j The mutual influence between them is large; if I(X i ,X j |C) is close to zero, then under the influence of other risk indicator node set C, X i , X j Independent of each other, we get the optimal network structure.

5. The liquefied gas station safety risk assessment method according to claim 4, characterized in that: Using fuzzy set theory to fuzzify the values ​​of nodes in the network structure includes: The values ​​of nodes in the network structure are divided into different levels, and the value of each node is scored by experts to obtain the scoring results. The membership of the scoring results is calculated by triangular fuzzy numbers. Assume u and l represent the upper and lower limits of the fuzzy number value range of the node, and m is the mean value. Then the fuzzy number of the node is expressed as (l, m, u), and the membership degree φ(z) of the fuzzy number of the node is calculated as: The membership degree of each node is combined with the corresponding level to obtain the fuzzy set of nodes.

6. The liquefied gas station safety risk assessment method according to claim 5, characterized in that: Using the Bayesian inference method for the parameter value includes: The expert judgment matrix is ​​established through expert level scoring, and the expert weights are obtained according to the judgment matrix. Using the expert scores and weights, the node probability of the network structure is expressed as: Among them, θ is the weight index of the node falling into the kth interval, k is the interval position of the expert score, and p l is the probability that a node falls into the lth interval, a k -a l is the interpolation between the k-th expert and the l-th expert.

7. The liquefied gas station safety risk assessment method according to claim 6, characterized in that: Through reverse reasoning analysis, the posterior probabilities of risk factors include: Extract independent risk indicator nodes (x1, x2, ..., x n ), the joint probability density function of the nodes is P(x,α)=π(α)P(x|α), where π(α) is the prior probability, and the posterior probability of the nodes in the network structure is expressed as: The parameter α is expressed as: The posterior probabilities are sorted to obtain the factors causing accidents at liquefied gas stations.

8. A system for a liquefied gas station safety risk assessment method, characterized in that: include: A data acquisition module, used to obtain real-time data of sensors and risk factors of liquefied gas stations; A first indicator selection module is used to preset a first judgment criterion, select indicators based on the real-time data of the sensor and the risk factors of the liquefied gas station, obtain a first indicator set, and construct a liquefied petroleum gas station safety risk indicator system according to the first indicator set; A second indicator selection module is used to preset a second judgment criterion, select indicators for constructing a Bayesian network structure according to the liquefied petroleum gas station safety risk indicator system, obtain a second indicator set, and obtain a network structure according to the second indicator set in combination with a constraint learning method; A calculation module is used to fuzzify the values ​​of the nodes in the network structure by using fuzzy set theory, calculate the parameter value of the conditional probability distribution of each node in the network structure, and use the Bayesian inference method to obtain the conditional probability distribution between the nodes in the network structure; The sorting module is used to calculate the risk probability of an accident occurring in a liquefied gas station by using the conditional probability distribution between nodes in the network structure, obtain the posterior probability of the risk factors through reverse reasoning analysis, sort the posterior probabilities, and obtain the factors causing the accident occurring in the liquefied gas station.

9. An electronic device, characterized in that: The device comprises: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.