Method and device for evaluating fault hazard of key equipment in gas field station

Abstract

The invention provides a method and a device for evaluating fault hazards of key equipment in a gas field station. The method comprises the following steps: determining key equipment of a gas station, and establishing a Bayesian network model based on factors causing faults of the key equipment and hazards generated by the faults of the key equipment; calculating the fault probability of the key equipment and the fault probability of the gas station by using a Bayesian network model; using a Bayesian network model to calculate the fault probability of each key device when the gas station has a fault; and carrying out quantitative evaluation on the harm generated by the key equipment fault, and proposing decision suggestions based on a quantitative evaluation result. Through the evaluation model established on the Bayesian network, probability distribution can be dynamically updated according to real-time data, quantitative real-time evaluation of gas field station harm is achieved, and the influence on upstream and downstream equipment when key equipment breaks down can be effectively evaluated.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault assessment, and in particular relates to a method and a device for assessing the hazard of key equipment failure in a gas station. Background Art

[0002] With the acceleration of urbanization, the demand for gas continues to grow, and it has become an indispensable energy cornerstone to support the daily operation of cities and the lives of residents. As the hub of the gas supply chain, gas stations play a vital role in ensuring the safety and stability of urban energy. However, in recent years, extreme weather events have frequently occurred, such as abnormally heavy rainfall, sudden cold waves and extreme low temperatures, which have brought severe challenges to the city's energy infrastructure and increased the risk of safety hazards and economic losses. The impact of heavy rainfall on gas stations is both far-reaching and complex, and its consequences will gradually emerge, mainly including equipment immersion caused by water accumulation on the site, damage to infrastructure, and possible system failures, which will interfere with the normal operation of core links such as gas purification, pressure regulation, power distribution and automation control. At the same time, low temperatures and cold waves may cause the gas pipelines and equipment in the station to face the dilemma of "freezing", seriously hindering the smooth transmission and processing efficiency of gas.

[0003] However, the current technical system seems to be unable to cope with these challenges. Due to the lack of a comprehensive gas station risk assessment system, it is impossible to fully consider the impact of multiple sources (such as operating environment, equipment failure rate and human factors) on the normal operation of the system, resulting in an incomplete risk assessment. In addition, traditional methods mostly use simple statistical or empirical models, which cannot dynamically reflect the real-time changing operating environment and equipment status, and the accuracy and timeliness of the prediction results are poor. When quantifying risks, they often only consider the probability of failure, ignoring the harmfulness of the failure, making the risk assessment results not comprehensive enough and unable to effectively assist operation and maintenance decisions. More importantly, existing technologies mainly rely on historical data for static evaluation, which cannot update and predict possible risks in the future in real time, and it is difficult to take preventive measures in a timely manner. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a method and device for evaluating the hazard of failure of key equipment in a gas station.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions.

[0006] In a first aspect, the present invention provides a method for evaluating the hazard of failure of key equipment in a gas station, comprising the following steps:

[0007] Determine the key equipment of the gas station, and establish a Bayesian network model based on the factors that cause key equipment failures and the hazards caused by key equipment failures;

[0008] Using the Bayesian network model, the probability of failure of key equipment and the probability of failure of gas stations are calculated;

[0009] Using the Bayesian network model, the probability of failure of each key device when a gas station fails is calculated;

[0010] Conduct quantitative assessment of the hazards caused by failures of key equipment and make decision recommendations based on the quantitative assessment results.

[0011] Furthermore, key equipment within a gas station includes pipelines, valves, filters, flow meters, pressure regulators and odorizing devices.

[0012] Furthermore, the Bayesian network model includes a network structure consisting of nodes and directed edges and a conditional probability table, each node corresponds to a conditional probability table, which represents the influence of the node's parent node on the node, and the conditional probability table of the node without a parent node includes a priori probability calculated based on historical data; the nodes of the Bayesian network model include nodes related to each key equipment and gas station nodes with each key equipment node as a parent node; the nodes related to each key equipment include: key equipment; the parent node of the key equipment, including weather factors, equipment factors and management factors; the child node gas parameters of the key equipment, the gas parameters include pressure and flow rate; the child node of the gas parameter affects the equipment, and the affecting equipment is the key equipment that may fail due to changes in the gas parameters.

[0013] Furthermore, the calculation method of the conditional probability table of the Bayesian network model includes:

[0014] Calculate the unconditional probability of nodes without parent nodes based on historical data or expert evaluation, including the probability of occurrence and the probability of non-occurrence;

[0015] Calculate the conditional probability of node A based on historical data: If node A has n parent nodes, namely B1, B2, ..., B n , the conditional probability of node A occurring is P(A|B1,B2,...,B n ), B1, B2, ..., B n The values ​​of include true and false, a total of 2 n There are different combinations, and the conditional probability that node A does not occur is 1-P(A|B1,B2,...,B n ).

[0016] Furthermore, methods for calculating the probability of failure of key equipment include:

[0017] Calculate the probability of failure of the i-th key equipment, the formula is:

[0018] P(X i )=1-(1-P1(X i ))(1-P2(X i ))(1-P3(X i )) (1)

[0019] In the formula, P(X i ) is the probability of failure of the i-th key equipment, P1(X i )、P2(X i )、P3(X i ) are the probabilities of failure of the i-th key equipment due to weather factors, equipment factors and management factors, where i = 1, 2, ..., N, and N is the number of key equipment;

[0020] P1(X i ) include:

[0021] Divide weather factors into M weather levels C j , j = 1, 2, ..., M;

[0022] Calculate the probability of each weather level occurring based on historical data P(C j );

[0023] Based on historical data, the probability P(X) of each weather level causing the failure of the i-th key equipment is calculated. i |C j );

[0024] Calculate P1(X i ), the formula is:

[0025]

[0026] P2(X i ) is calculated as:

[0027]

[0028] In the formula, t i is the service time of the i-th key equipment, λ i , β i is the fitting parameter;

[0029] P3(X i ) include:

[0030] An expert group is formed, and the members of the expert group score the i-th key equipment from the three aspects of safety awareness, operating specifications and enterprise management. The weighted sum of the average scores of the expert group members in each aspect is used to obtain P3(X i ).

[0031] Furthermore, the formula for calculating the probability of failure of a gas station is:

[0032]

[0033] Where P(Y) is the probability of failure of the gas station, P(X i ) is the probability of failure of the i-th key device, where i = 1, 2, ..., N, and N is the number of key devices.

[0034] Furthermore, using the Bayesian network model, the formula for calculating the probability of failure of key equipment when a known gas station fails is as follows according to the posterior probability formula:

[0035]

[0036] In the formula, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, P(X i ) is the probability of failure of the i-th key equipment, P(Y|X i )、P(Y|X k ) are the probability of failure of the gas station when the i-th and k-th key equipment fail, respectively.

[0037] Furthermore, methods for quantitatively assessing the hazards caused by failure of key equipment include:

[0038] The formula for calculating the possibility of each key equipment causing harm when a gas station fails is:

[0039] Q(X i |Y)=5P(X i |Y) (6)

[0040] In the formula, Q(X i |Y) is the possibility of the i-th key equipment causing harm when a gas station fails, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, where i = 1, 2, ..., N, and N is the number of key equipment;

[0041] The evaluation score of the hazard caused by each key equipment is calculated using the following formula:

[0042] Z(X i |Y)=Q(X i |Y)×W(X i |Y) (7)

[0043] In the formula, Z(X i |Y) is the evaluation score of the hazard caused by the i-th key equipment, W(Xi |Y) is the criticality of the i-th key equipment failure, which is obtained by expert evaluation and ranges from 0 to 5 points.

[0044] Furthermore, methods for making decision recommendations based on quantitative evaluation results include:

[0045] The hazard level is divided according to the evaluation score, and decision-making suggestions are given based on the hazard level:

[0046] If 0 <Z(X i |Y)≤3, the hazard level is level 1, no monitoring is required;

[0047] If 3 <Z(X i |Y)≤12, hazard level is 2, irregular monitoring;

[0048] If 12 <Z(X i |Y)≤18, the hazard level is level 3, and appropriate protective measures should be taken;

[0049] If 18 <Z(X i |Y)≤22, the hazard level is level 4, and maintenance should be carried out in time;

[0050] If 22 <Z(X i |Y)≤25, the hazard level is level 5, pay high attention, red alert.

[0051] In a second aspect, the present invention provides a device for evaluating the hazard of failure of key equipment in a gas station, comprising:

[0052] Model building module, used to identify key equipment in gas stations and build a Bayesian network model based on the factors causing key equipment failures and the hazards caused by key equipment failures;

[0053] The first calculation module is used to calculate the probability of failure of key equipment and the probability of failure of gas stations by using a Bayesian network model;

[0054] The second calculation module is used to calculate the probability of failure of each key equipment when a gas station fails by using a Bayesian network model;

[0055] The hazard assessment module is used to quantitatively assess the hazards caused by key equipment failures and make decision recommendations based on the quantitative assessment results.

[0056] Compared with the prior art, the present invention has the following beneficial effects.

[0057] The present invention determines the key equipment of the gas station, establishes a Bayesian network model based on the factors causing the failure of the key equipment and the hazards caused by the failure of the key equipment, uses the Bayesian network model to calculate the probability of failure of the key equipment and the probability of failure of the gas station, calculates the probability of failure of each key equipment when the gas station fails, quantitatively evaluates the hazards caused by the failure of the gas station, and makes decision-making suggestions based on the quantitative evaluation results, thereby realizing the evaluation of the hazards of failure of key equipment in the gas station. The present invention can dynamically update the probability distribution according to real-time data through the evaluation model established on the Bayesian network, realize the quantitative real-time evaluation of the hazards of the gas station, and can effectively evaluate the impact of the failure of the key equipment on its upstream and downstream equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 The present invention is a flowchart of a method for evaluating the hazard of failure of key equipment in a gas station according to an embodiment of the present invention.

[0059] Figure 2 Schematic diagram of the network structure of the Bayesian network model.

[0060] Figure 3 This is a schematic diagram of the distribution of key equipment and influencing factors in a gas station.

[0061] Figure 4 The present invention is a block diagram of a device for evaluating the hazard of failure of key equipment in a gas station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention clearer and more understandable, the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0063] Figure 1 This is a flow chart of a method for evaluating the hazard of failure of key equipment in a gas station according to an embodiment of the present invention, comprising the following steps:

[0064] Step 101, determining the key equipment of the gas station, and establishing a Bayesian network model based on the factors causing the failure of the key equipment and the harm caused by the failure of the key equipment;

[0065] Step 102, using the Bayesian network model, calculating the probability of failure of key equipment and the probability of failure of the gas station;

[0066] Step 103, using the Bayesian network model, calculating the probability of each key device failing when a gas station fails;

[0067] Step 104, quantitatively evaluate the hazards caused by the gas station failure, and make decision recommendations based on the quantitative evaluation results.

[0068] In this embodiment, step 101 is mainly used to establish a Bayesian network model based on the key equipment of the gas station. First, according to the process flow of the gas station, the key equipment is determined according to the actual needs in the evaluation process. The process flow of the gas station is as follows: the natural gas is supplied through pipelines upstream, enters the gas station through the station entry area, and is transported out of the station after gas quality analysis, heat exchange, filtration, metering, pressure regulation, and odorization processes, and supplied to downstream users. Therefore, key equipment should include pipelines, valves, filters, and pressure regulators. Then analyze the factors that can cause key equipment failures and the possible hazards caused by key equipment failures, and on this basis, establish a Bayesian network model for gas station damage assessment.

[0069] The Bayesian model is a method based on probability theory to represent the relationship between prior knowledge (signs, omens) and posterior knowledge (phenomena, conclusions). It is based on the Bayesian formula, Bayesian statistical inference and Bayesian network. The random variables in the system can be represented by nodes in the Bayesian network. The nodes are connected in the form of lines to show the sequential directional relationship. The posterior probability is obtained using the prior probability and sample information. It is mainly used to process random information in uncertain information. The Bayesian network is a directed acyclic graph model, such as Figure 2 As shown in Figure 1, it consists of nodes that describe random variables and directed edges that describe the dependencies between variables. A Bayesian network model usually consists of nodes, directed edges (network structure), and conditional probability tables (CPTs).

[0070] In this embodiment, step 102 is mainly used to calculate the probability of failure of key equipment and the probability of failure of gas stations using the Bayesian network model. The Bayesian network model has two very important applications: one is prediction; the other is diagnosis. Prediction means that a certain cause is known, and the Bayesian network model is used for calculation to solve the probability of the cause leading to the result; diagnosis means that a certain result is known to have occurred, and the Bayesian network model is used for reasoning to obtain the probability of the cause of the result. Prediction and diagnosis are two opposite processes, prediction is top-down, and diagnosis is bottom-up. Calculating the probability of failure of key equipment and the probability of failure of gas stations are both predictions using the Bayesian network model. According to the information of the parent node (probability of occurrence and non-occurrence), the probability of the current node is calculated using the conditional probability table. When the number of parent nodes is multiple, the probability of the current node is calculated according to the total probability formula.

[0071] In this embodiment, step 103 is mainly used to use the Bayesian network model to calculate the probability of each key device failing when a gas station fails. Step 103 belongs to diagnosis using the Bayesian network model, which is to find the cause of the failure when a gas station failure is known: by calculating the probability that each key device may fail, the faulty device is determined according to the size of the probability, and the key device with the highest probability generally has the highest probability of failure. The probability of predicting (diagnosing) the cause based on the known result can be calculated according to the posterior probability formula, that is, first calculate the probability that each parent node causes the result, and sum all the probabilities, and then divide the probability of each parent node causing the result by the sum, and the probability of each parent node occurring when the result fails can be obtained.

[0072] In this embodiment, step 104 is mainly used to quantitatively evaluate the hazards caused by the failure of key equipment. The existing technology generally only predicts the probability of failure of key equipment, and rarely evaluates the hazards caused by the failure of key equipment. For example, the failure of a key equipment will cause changes in gas parameters (pressure, flow, etc.), and the change in gas parameters will further affect the operation of other key equipment (the impact of upstream equipment on downstream equipment is particularly obvious), and in severe cases, it will cause other key equipment failures. Therefore, it is necessary to quantitatively evaluate the hazards caused by gas station failures, and make decision-making recommendations based on the quantitative evaluation results to prevent them before they happen.

[0073] This embodiment can conveniently perform fault prediction, diagnosis and fault hazard assessment by establishing an assessment model based on a Bayesian network; and can also use the Bayesian formula to update the probability distribution of nodes in real time. When new data is observed, the model converts the prior probability into the posterior probability according to the Bayesian formula, thereby achieving real-time and dynamic assessment of equipment damage risk.

[0074] As an optional embodiment, key equipment in the gas station includes pipelines, valves, filters, flow meters, pressure regulators and odorizing devices.

[0075] This embodiment provides a specific configuration of key equipment in a gas station. The key equipment in this embodiment is, in order of the upstream and downstream relationship of the installation position, a pipeline, a valve, a filter, a flow meter, a pressure regulator, and an odorizing device. The following describes the function of each key equipment in a gas station and the possible causes of failure. Figure 3 shown.

[0076] Pipelines are used to transport gas and are also the link between the upstream of gas stations, gas stations and downstream of gas stations. The layout of pipelines is limited by terrain factors. Some pipelines are exposed to the soil during construction. Heavy rains may also wash away some soil, exposing the originally buried pipelines. Exposed pipelines are exposed to the sun and rain for a long time. Therefore, pipelines are easily corroded and damaged, causing gas leakage during transportation, which in turn affects the pressure and flow rate of the gas in the pipeline.

[0077] Valves are devices that control the flow and pressure of gas in gas stations. Common valves include ball valves, butterfly valves, gate valves, etc. The frequency of use, pressure, temperature, etc. of the valve are all factors that cause valve failure. When the valve is in an open or closed state for a long time, it may cause valve failure, reduced sealing performance, and other damages; when the valve is impacted by high-pressure gas, it may cause the valve to rupture, deform, and other damages.

[0078] Filters are devices used to filter impurities in gas stations. Common filters include mesh filters, membrane filters, etc. Factors that affect the performance of filters mainly include the filter's inlet pressure, flow rate, etc. When the filter is impacted by high-pressure gas, it may cause filter failure, filter screen wear and other damage; when too many impurities accumulate in the filter, it may cause the filter to become clogged and lose its filtering function.

[0079] Flowmeter is a device used to measure gas flow. Factors that affect flowmeter mainly include flow rate, pressure, etc. When too many impurities accumulate in the flowmeter, it may cause measurement errors; when the flowmeter is disturbed by flow fluctuations, it may cause inaccurate measurement.

[0080] The pressure regulator is a device used to regulate the gas pressure in a gas station. Common pressure regulators include diaphragm pressure regulators and piston pressure regulators. Factors affecting the pressure regulator mainly include the inlet pressure and outlet flow of the pressure regulator. When the pressure regulator is impacted by the inlet pressure, it may cause failure of the pressure regulator, decrease in regulation accuracy, and other damages; when too much impurities accumulate inside the pressure regulator, it may cause the pressure regulator to become blocked and lose its pressure regulation function.

[0081] Odorizer is a device used to add odor to gas in gas stations. Common odorizers include liquid odorizers, solid odorizers, etc. Factors affecting odorizers mainly include odorant concentration, supply, etc. When the odorant concentration is insufficient, the gas may not meet the required odor standards; when the odorant supply is too large, it may cause excessive accumulation of odorant, affecting the gas quality.

[0082] As an optional embodiment, the Bayesian network model includes a network structure consisting of nodes and directed edges and a conditional probability table, each node corresponds to a conditional probability table, which represents the influence of the parent node of the node on the node, and the conditional probability table of the node without a parent node includes a priori probability calculated based on historical data; the nodes of the Bayesian network model include nodes related to each key equipment and gas station nodes with each key equipment node as a parent node; the nodes related to each key equipment include: key equipment; the parent node of the key equipment, including weather factors, equipment factors and management factors; the child node gas parameters of the key equipment, the gas parameters include pressure and flow rate; the child node of the gas parameter affects the equipment, and the affecting equipment is the key equipment that may fail due to changes in the gas parameters.

[0083] This embodiment provides a technical solution for the Bayesian network model. The Bayesian network model includes a network structure and a conditional probability table. The network structure is composed of nodes and directed edges. The network structure of this embodiment takes the key equipment as the main line (main node). By analyzing the factors causing the failure of the key equipment and the impact caused by the failure of the key equipment, the parent nodes and child nodes related to each key equipment node and the directed edges connecting the nodes are determined. The network structure of the Bayesian network model of this embodiment is as follows: Figure 2 As shown, N (the number of key equipment) approximately identical network areas can be divided, and each network area corresponds to a key equipment. The parent node of each key equipment node is the same, including weather factors, equipment factors and management factors; the child (grandchild) nodes of the key equipment nodes are slightly different. For example, the gas parameters that change due to failures of different key equipment may be different; the devices affected by different gas parameters are also different. For example, excessive gas pressure will cause the pressure regulator and filter to face the risk of damage, and excessive gas flow rate will cause the flow meter to face the risk of damage. In addition, it also includes gas station nodes with all key equipment nodes as parent nodes. Each node corresponds to a conditional probability table, which represents the influence of the parent node of the node on the node.

[0084] It is worth noting that the name of each node should be the name of a random variable (or event) in principle, but in order to simplify the node names, most of the node names in this embodiment are not real random variable names. For example, the random variable corresponding to the key equipment node should be the failure of the key equipment, and the random variable corresponding to the gas parameter node should be the change of the gas parameter.

[0085] As an optional embodiment, a method for calculating the conditional probability table of the Bayesian network model includes:

[0086] Calculate the unconditional probability of nodes without parent nodes based on historical data or expert evaluation, including the probability of occurrence and the probability of non-occurrence;

[0087] Calculate the conditional probability of node A based on historical data: If node A has n parent nodes, namely B1, B2, ..., B n , the conditional probability of node A occurring is P(A|B1,B2,...,B n ), B1, B2, ..., B n The values ​​of include true and false, a total of 2 n There are different combinations, and the conditional probability that node A does not occur is 1-P(A|B1,B2,...,B n ).

[0088] This embodiment provides a technical solution for calculating the conditional probability table of the Bayesian network model. The node at the top of the Bayesian network has no parent node (such as weather factors, equipment factors, and management factors). The probability of a node without a parent node is an unconditional probability, which can only be calculated based on historical data or expert evaluation, including the probability of occurrence and the probability of non-occurrence, and the sum of the two is 1.

[0089] The following example illustrates the method of calculating unconditional probability. When calculating the probability of a weather factor node, it is necessary to calculate the probability of each weather level occurring P(C j ), for example, the probability of the general level appearing is P(C1). P(C1) can be calculated as follows: In the 365 days of the year, count the number of days D of the general level appearing in each of the past K years. k , take the average value to get the probability of a general level appearing each year, and then divide it by 365 to get the probability of a general level appearing every day P(C1). The calculation formula is:

[0090]

[0091] The conditional probability of having a parent node can be calculated based on historical data; when historical data is not available, it can be determined based on subjective experience or expert evaluation. For a node A with multiple parent nodes (such as a key equipment node), if node A has n parent nodes B1, B2, ..., B n , the conditional probability of node A occurring is P(A|B1,B2,...,B n ), total 2 n different combinations, corresponding to B1, B2, ..., B n respectively true (1) and false (0); the conditional probability that node A does not occur is 1-P(A|B1,B2,...,B n ).

[0092] Below is the conditional probability table for the gas parameter (pressure) node.

[0093] Table 1 Conditional probability table of gas parameter (pressure) nodes

[0094]

[0095] Table 1 gives the probability of the impact of key equipment (valve) failure on gas parameters (pressure), including: the probability of normal gas parameters when the valve fails, the probability of gas parameters increasing by 0-20%, 20-50%, and the probability greater than 50%, and the probability of normal gas parameters when the valve is normal, the probability of gas parameters increasing by 0-20%, 20-50%, and the probability greater than 50%.

[0096] As an optional embodiment, the method for calculating the probability of failure of key equipment includes:

[0097] Calculate the probability of failure of the i-th key equipment, the formula is:

[0098] P(X i )=1-(1-P1(X i ))(1-P2(X i ))(1-P3(X i )) (1)

[0099] In the formula, P(X i ) is the probability of failure of the i-th key equipment, P1(X i )、P2(X i )、P3(X i ) are the probabilities of failure of the i-th key equipment due to weather factors, equipment factors and management factors, where i = 1, 2, ..., N, and N is the number of key equipment;

[0100] P1(X i ) include:

[0101] Divide weather factors into M weather levels C j , j = 1, 2, ..., M;

[0102] Calculate the probability of each weather level occurring based on historical data P(C j );

[0103] Based on historical data, the probability P(X) of each weather level causing the failure of the i-th key equipment is calculated. i |C j );

[0104] Calculate P1(X i ), the formula is:

[0105]

[0106] P2(X i ) is calculated as:

[0107]

[0108] Where, t i is the service time of the i-th key equipment, λ i , β i is the fitting parameter;

[0109] P3(X i ) include:

[0110] An expert group is formed, and the members of the expert group score the i-th key equipment from the three aspects of safety awareness, operating specifications and enterprise management. The weighted sum of the average scores of the expert group members in each aspect is used to obtain P3(X i ).

[0111] This embodiment provides a technical solution for calculating the failure probability of key equipment. Figure 2 , the key equipment has three parent nodes, namely weather factors, equipment factors and management factors. The probability of failure of the i-th key equipment caused by each factor is recorded as P1(X i )、P2(X i )、P3(X i ), the probability that the i-th key equipment is normal is P 正常 =(1-P1(X i ))(1-P2(X i ))(1-P3(X i )), the probability of failure is 1-P 正常 , as shown in formula (1).

[0112] The following introduces the solution methods of these three probabilities respectively.

[0113] P1(X i ) calculation method: first divide the weather into M weather levels, such as general, severe, serious and extremely severe according to the degree of impact on key equipment. Then calculate the probability of each weather level P(C j ), and the probability P(X i |C j ), and finally calculate P1(X i ), as shown in formula (2). Dividing weather factors into M types is actually equivalent to adding M parent nodes to weather factors.

[0114] P2(X i ) calculation method: Equipment factors mainly consider the failure caused by the service time of key equipment. Obviously, the longer the service time, the greater the probability of failure of key equipment. i) is calculated as shown in formula (3). Formula (3) is an empirical formula expressed by an exponential function, where the parameter λ i , β i To fit the parameters, we can construct a likelihood function based on the sample data and solve the λ that maximizes the likelihood function. i , β i .

[0115] P3(X i ) calculation method: Management factors are subjective factors, so there is no objective data available. This embodiment uses the expert scoring method to calculate the probability P3 (X) of key equipment failure caused by management factors. i ). First, the management factors are divided into three aspects: safety awareness, operating specifications, and enterprise management; then an expert group is formed, and the expert group members score the key equipment for these three aspects respectively (these three aspects can also be further divided into three levels: low, medium, and high, and then the level of each aspect is evaluated), and the average of the scores of the expert group members in each aspect is calculated, and finally P3(X) is obtained through weighted summation. i ).

[0116] As an optional embodiment, the formula for calculating the probability of a gas station failure is:

[0117]

[0118] Where P(Y) is the probability of failure of the gas station, P(X i ) is the probability of failure of the i-th key device, where i = 1, 2, ..., N, and N is the number of key devices.

[0119] This embodiment provides a technical solution for calculating the probability P(Y) of a gas station failure. The calculation method of the probability P(Y) of a gas station failure is similar to the probability P(X) of each key device failure. i ) are calculated in a similar way, such as in formula (1) and formula (4), where the probability of normal operation is first calculated, and then the probability of normal operation is subtracted from 1 to obtain the probability of failure.

[0120] As an optional embodiment, the formula for calculating the probability of failure of key equipment when a known gas station fails according to the posterior probability formula using the Bayesian network model is:

[0121]

[0122] In the formula, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, P(X i ) is the probability of failure of the i-th key equipment, P(Y|X i)、P(Y|X k ) are the probability of failure of the gas station when the i-th and k-th key equipment fail, respectively.

[0123] This embodiment provides a method for calculating the probability P(X) of failure of key equipment when a known gas station fails. i |Y). As mentioned above, P(X i The calculation of |Y) belongs to the diagnostic application of the Bayesian network model and can be calculated according to the posterior probability formula shown in formula (5). First, the probability that each key device may cause a gas station failure when the gas station is working normally is calculated, and then the sum is obtained to obtain the probability of a failure when the gas station is working normally. The probability that each key device may cause a gas station failure is divided by the probability of a failure when the gas station is working normally, and the probability of each key device failing can be obtained when a gas station failure is known.

[0124] As an optional embodiment, a method for quantitatively evaluating the hazards caused by failure of key equipment includes:

[0125] The formula for calculating the possibility of each key equipment causing harm when a gas station fails is:

[0126] Q(X i |Y)=5P(X i |Y) (6)

[0127] In the formula, Q(X i |Y) is the possibility of the i-th key equipment causing harm when a gas station fails, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, where i = 1, 2, ..., N, and N is the number of key equipment;

[0128] The evaluation score of the hazard caused by each key equipment is calculated using the following formula:

[0129] Z(X i |Y)=Q(X i |Y)×W(X i |Y) (7)

[0130] In the formula, Z(X i |Y) is the evaluation score of the hazard caused by the i-th key equipment, W(X i |Y) is the criticality of the i-th key equipment failure, which is obtained by expert evaluation and ranges from 0 to 5 points.

[0131] This embodiment is a technical solution for quantitatively evaluating the hazards caused by failure of key equipment. This embodiment calculates the evaluation score Z(X) of the hazard caused by each key equipment. i|Y), quantitatively evaluate the hazards caused by failure of key equipment. Z(X i The calculation formula of Z(X i |Y) is equal to two factors Q(X i |Y)、W(X i |Y). Q(X i |Y) is the probability of each key device causing harm when a gas station fails. In fact, it is the posterior probability value P(X i |Y), since P(X i |Y) is too small to be used directly for hazard assessment, so P(X i |Y) is magnified to 5 times as Q(X i |Y), as shown in formula (6), so that Q(X i |Y) is between 0 and 5. i |Y) is the criticality of each key equipment failure, which is obtained by expert evaluation and ranges from 0 to 5. i |Y) value is between 0 and 25.

[0132] As an optional embodiment, the method for proposing decision suggestions based on the quantitative evaluation results includes:

[0133] The hazard level is divided according to the evaluation score, and decision-making suggestions are given based on the hazard level:

[0134] If 0 <Z(X i |Y)≤3, the hazard level is level 1, no monitoring is required;

[0135] If 3 <Z(X i |Y)≤12, hazard level is 2, irregular monitoring;

[0136] If 12 <Z(X i |Y)≤18, the hazard level is level 3, and appropriate protective measures should be taken;

[0137] If 18 <Z(X i |Y)≤22, the hazard level is level 4, and maintenance should be carried out in time;

[0138] If 22 <Z(X i |Y)≤25, the hazard level is level 5, pay high attention, red alert.

[0139] This embodiment provides a technical solution for making decision suggestions based on quantitative evaluation results. i |Y) is compared with the thresholds of each hazard level to determine the hazard level. The hazard levels of this embodiment include 1 to 5 levels, and the corresponding Z(X iThe value of |Y) increases gradually from small to large. Then, according to the obtained hazard level, the decision to be taken is determined. The five levels and the corresponding decision suggestions for each level are shown in the following table.

[0140] Table 2 Hazard level classification and decision-making recommendations

[0141]

[0142]

[0143] Figure 4 The present invention is a schematic diagram of a device for evaluating the hazard of failure of key equipment in a gas station according to an embodiment of the present invention. The device comprises:

[0144] The model building module 11 is used to determine the key equipment of the gas station and to establish a Bayesian network model based on the factors causing the failure of the key equipment and the hazards caused by the failure of the key equipment;

[0145] The first calculation module 12 is used to calculate the probability of failure of key equipment and the probability of failure of gas stations by using a Bayesian network model;

[0146] The second calculation module 13 is used to calculate the probability of failure of each key equipment when a failure occurs in the gas station by using a Bayesian network model;

[0147] The hazard assessment module 14 is used to quantitatively assess the hazards caused by failures of key equipment and to make decision recommendations based on the quantitative assessment results.

[0148] The device of this embodiment can be used to perform Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects, which will not be repeated here.

[0149] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for evaluating the hazard of failure of key equipment in a gas station, characterized in that: The following steps are involved: Determine the key equipment of the gas station, and establish a Bayesian network model based on the factors that cause key equipment failures and the hazards caused by key equipment failures; Using the Bayesian network model, the probability of failure of key equipment and the probability of failure of gas stations are calculated; Using the Bayesian network model, the probability of failure of each key device when a gas station fails is calculated; Conduct quantitative assessment of the hazards caused by failures of key equipment and make decision recommendations based on the quantitative assessment results.

2. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 1 is characterized in that: The key equipment in the gas station includes pipelines, valves, filters, flow meters, pressure regulators and odorizing devices.

3. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 1 is characterized in that: The Bayesian network model includes a network structure consisting of nodes and directed edges and a conditional probability table, each node corresponds to a conditional probability table, which represents the influence of the parent node of the node on the node, and the conditional probability table of the node without a parent node includes the prior probability calculated based on historical data; the nodes of the Bayesian network model include nodes related to each key equipment and gas station nodes with each key equipment node as a parent node; the nodes related to each key equipment include: key equipment; the parent node of the key equipment, including weather factors, equipment factors and management factors; the child node gas parameters of the key equipment, the gas parameters include pressure and flow rate; the child node of the gas parameter affects the equipment, and the affecting equipment is the key equipment that may fail due to the change of the gas parameter.

4. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 3 is characterized in that: The calculation method of the conditional probability table of the Bayesian network model includes: Calculate the unconditional probability of nodes without parent nodes based on historical data or expert evaluation, including the probability of occurrence and the probability of non-occurrence; Calculate the conditional probability of node A based on historical data: If node A has n parent nodes, namely B1, B2, ..., B n , the conditional probability of node A occurring is P(A|B1,B2,...,B n ), B1, B2, ..., B n The values ​​of include true and false, a total of 2 n There are different combinations, and the conditional probability that node A does not occur is 1-P(A|B1,B2,...,B n ).

5. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 3 is characterized in that: Methods for calculating the probability of failure of critical equipment include: Calculate the probability of failure of the i-th key equipment, the formula is: P(X i )=1-(1-P1(X i ))(1-P2(X i ))(1-P3(X i )) (1) In the formula, P(X i ) is the probability of failure of the i-th key equipment, P1(X i )、P2(X i )、P3(X i ) are the probabilities of failure of the i-th key equipment due to weather factors, equipment factors and management factors, where i = 1, 2, ..., N, and N is the number of key equipment; P1(X i ) include: Divide weather factors into M weather levels C j , j = 1, 2, ..., M; Calculate the probability of each weather level occurring based on historical data P(C j ); Based on historical data, the probability P(X) of each weather level causing the failure of the i-th key equipment is calculated. i |C j ); Calculate P1(X i ), the formula is: P2(X i ) is calculated as: Where, t i is the service time of the i-th key equipment, λ i , β i is the fitting parameter; P3(X i ) include: An expert group is formed, and the members of the expert group score the i-th key equipment from the three aspects of safety awareness, operating specifications and enterprise management. The weighted sum of the average scores of the expert group members in each aspect is used to obtain P3(X i ).

6. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 3 is characterized in that: The formula for calculating the probability of failure of a gas station is: Where P(Y) is the probability of failure of the gas station, P(X i ) is the probability of failure of the i-th key device, where i = 1, 2, ..., N, and N is the number of key devices.

7. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 3 is characterized in that: Using the Bayesian network model, the formula for calculating the probability of failure of key equipment when a known gas station fails is as follows according to the posterior probability formula: In the formula, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, P(X i ) is the probability of failure of the i-th key equipment, P(Y|X i )、P(Y|X k ) are the probability of failure of the gas station when the i-th and k-th key equipment fail, respectively.

8. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 3 is characterized in that: Methods for quantitatively assessing the hazards caused by failure of key equipment include: The formula for calculating the possibility of each key equipment causing harm when a gas station fails is: Q(X i |Y)=5P(X i |Y) (6) In the formula, Q(X i |Y) is the possibility of the i-th key equipment causing harm when a gas station fails, P(X i |Y) is the probability of failure of the i-th key equipment when a gas station fails, where i = 1, 2, ..., N, and N is the number of key equipment; The evaluation score of the hazard caused by each key equipment is calculated using the formula: Z(X i |Y)=Q(X i |Y)×W(X i |Y) (7) In the formula, Z(X i |Y) is the evaluation score of the hazard caused by the i-th key equipment, W(X i |Y) is the criticality of the i-th key equipment failure, which is obtained by expert evaluation and ranges from 0 to 5 points.

9. The method for evaluating the hazard of failure of key equipment in a gas station according to claim 8 is characterized in that: Methods for making decision recommendations based on quantitative evaluation results include: The hazard level is divided according to the evaluation score, and decision-making suggestions are given based on the hazard level: If 0 <Z(X i |Y)≤3, the hazard level is level 1, no monitoring is required; If 3 <Z(X i |Y)≤12, hazard level is 2, irregular monitoring; If 12 <Z(X i |Y)≤18, the hazard level is level 3, and appropriate protective measures should be taken; If 18 <Z(X i |Y)≤22, the hazard level is level 4, and maintenance should be carried out in time; If 22 <Z(X i |Y)≤25, the hazard level is level 5, pay close attention, red alert.

10. A device for evaluating the hazard of failure of key equipment in a gas station, characterized in that: include: Model building module, used to identify key equipment in gas stations and build a Bayesian network model based on the factors causing key equipment failures and the hazards caused by key equipment failures; The first calculation module is used to calculate the probability of failure of key equipment and the probability of failure of gas stations by using a Bayesian network model; The second calculation module is used to calculate the probability of failure of each key equipment when a gas station fails by using a Bayesian network model; The hazard assessment module is used to quantitatively assess the hazards caused by key equipment failures and make decision recommendations based on the quantitative assessment results.

Images