A process technology operation risk management method and system

Abstract

The present application relates to the technical field of petrochemical safety, and in particular to a process technology operation risk management method and system, wherein the method includes the steps of establishing an SDG model, a node classification step, a deviation determination step, a risk determination step, and an alert management step; wherein the system includes a model building module, a node classification module, a deviation determination module, a risk determination module, and an alert management module. The present application has the effect of reducing the inaccuracy of human judgment and improving the safety of petrochemical process technology operations.

Description

Technical Field

[0001] The present application relates to the technical field of petrochemical safety, and in particular to a process technology operation risk management method and system. Background Art

[0002] At present, HAZOP analysis is a qualitative evaluation technology based on guide words, which aims to identify and evaluate the risks that may cause personal injury or property loss. Through a systematic approach, HAZOP analysis helps identify potential hazards and operability problems in designs or existing devices, and thus puts forward improvement opinions and suggestions to improve the safety and operability of petrochemical plant processes.

[0003] In the existing technology, the analysis team divides the continuous process flow into many parts (node ​​division), combines the existing process design requirements, and under the guidance of guide words, looks for situations where the process or operation deviates from the design standard parameters, identifies potential hazards and finds the cause of the deviation. After finding the cause of the deviation, it is necessary to manually judge whether there will be serious consequences or affect production based on the importance of the node and the propagation path, and whether intervention is needed. Human judgment is prone to errors, resulting in reduced safety. Summary of the invention

[0004] In order to reduce the inaccuracy of human judgment and improve the safety of petrochemical process operation, the present application provides a process operation risk management method and system.

[0005] In the first aspect, the present application provides a process technology operation risk management method, which adopts the following technical solution:

[0006] A process technology operation risk management method comprises the following steps:

[0007] Establish SDG model: Establish SDG model of each equipment unit, and merge and simplify the SDG model of each equipment unit according to the process flow;

[0008] Node classification: Classify nodes according to their criticality and assign weights based on the classification;

[0009] Deviation determination: Determine the node where the parameter that generates the deviation is located, and manually pull the deviation according to the SDG model to determine the deviation propagation path and form an operational directed acyclic graph;

[0010] Risk determination: including relevance judgment step, relevance intervention step, risk value calculation step and risk judgment step;

[0011] Correlation judgment: judge whether there is correlation among multiple node deviations. If so, execute the correlation intervention step. Otherwise, execute the risk value calculation step.

[0012] Risk value calculation: Calculate the node deviation risk value , Value at Risk The calculation model is as follows:

[0013] ;

[0014] In the formula, is the probability of deviation of the ith node at level n, Indicates taking the larger The calculation model is as follows:

[0015] ; In the formula, is a constant, and , is the basis for the value-at-risk calculation, Assign a value to the weight of the i-th node at level n;

[0016] Association intervention: Intervene in the risk value calculation step and update the risk value The calculation model is as follows:

[0017] , where to rank the nodes, and , where the higher the criticality, The bigger;

[0018] Risk judgment: judge the node deviation risk value Is it greater than the risk threshold? , if yes, then execute the warning control steps;

[0019] Warning control: remind relevant personnel to intervene.

[0020] By adopting the above technical solution, according to the production process and related data of the petrochemical plant, the SDG model is established, and then the SDG model of each equipment unit is integrated and simplified. According to the production process and accident risk level, the criticality of the node is determined and graded, and then the weight is assigned according to the grade. Then, according to the HAZOP analysis report, the parameter where the deviation occurs is determined, and then according to the SDG model, manual deviation is performed to determine the deviation propagation path, and then the node deviation risk value is calculated, and the maximum is taken after comparing all the node risk values, and then it is determined whether the node deviation risk value is greater than the preset risk critical value. If it is greater than, the relevant personnel are reminded to intervene and control to reduce the occurrence of accidents, or correct them in time to reduce the occurrence of inferior products, control production quality, and reduce safety risks; and through the calculation of the risk value, the consequences of node deviation are more intuitive, reducing the error of manual judgment, and being able to control the safety of the process in time, reduce the occurrence of accidents, and improve risk management capabilities;

[0021] When multiple node deviations occur, a directed acyclic graph is used to determine whether the nodes of the multiple deviations have a causal relationship or are on a propagation path. If so, the risk value calculation model is used to calculate the risk value. If they are not on a propagation path, the associated intervention step is performed to update the risk value calculation model so that the risk value coefficient is increased, and then the risk value is judged to determine whether intervention is needed. By recalculating multi-node deviations without correlation, the importance of multi-node deviations is increased, the generation of risks is reduced, and equipment problems can be discovered in a timely manner, reducing continuous damage to equipment and improving safety.

[0022] Optionally, a report push step is also provided after the warning control step;

[0023] Report push: An operational directed acyclic graph is formed based on the SDG model and the deviation propagation path, and an operational report is formed and pushed to relevant personnel.

[0024] By adopting the above technical solution, when relevant personnel are notified to perform maintenance or correction, an operational directed acyclic graph is formed into a report and sent to the relevant personnel, so that the relevant personnel can determine the intervention position according to the deviation node, and intervene quickly and in a targeted manner, thereby improving efficiency, and being able to handle it in a timely manner, reduce the accumulation of hidden dangers, and improve safety.

[0025] Optionally, a deviation qualitative step is further provided after the deviation determination step;

[0026] Deviation characterization: Determine whether the deviation is an unusual deviation. If so, execute the shutdown inspection step. Otherwise, execute the risk value calculation step.

[0027] Shutdown and inspection: According to the deviation location, check the location of the parameter variable one by one to determine whether it is an abnormal deviation. If so, shut down and repair it.

[0028] By adopting the above technical solution, when there are unconventional deviations, such as pipeline leakage, equipment failure and other abnormal parameter deviations, the machine is shut down for inspection, and the locations where the parameter variables occur are checked one by one for repair; through the above settings, frequent shutdowns or interventions in production are reduced, so that production activities can be carried out with quality and efficiency, and production can be carried out under the premise of ensuring safety; and repairs can be carried out in a timely manner to reduce excessive damage to equipment and reduce the risk of accidents.

[0029] Optionally, between the relevance judgment step and the relevance intervention step, a relevance parameter acquisition step, a relevance judgment step and a risk update step are further provided;

[0030] Correlation parameter acquisition: According to the directed acyclic graph, determine the location of the cause node and obtain the correlation interval between the cause nodes , get the associated interval Greater than the associated interval limit The number of and in the associated interval The proportion of the total number ;

[0031] Relevance judgment: judgment ratio Is it greater than the percentage limit? , if yes, then execute the risk update step, otherwise, execute the associated intervention step;

[0032] Risk update: Intervene in the risk value calculation step and update the risk value The calculation model is as follows:

[0033] , where the value model of S is as follows: , ,and , The value model of is as follows:

[0034] .

[0035] By adopting the above technical solution, when multiple deviation nodes are related, but there is a long correlation interval, the degree of correlation is low. Through the statistics of the correlation interval exceeding the set correlation interval limit value and the calculation of the proportion, it is determined whether the risk update step needs to be executed, the risk value calculation model is optimized, and the calculation is performed, and then the risk judgment is performed. Through further determination of the above correlation degree, recalculation is performed. When deviations occur in multiple nodes, further subdivision is performed to further improve the accuracy of risk value calculation, facilitate timely control of risks, reduce accidents caused by safety hazards not being discovered in time, and thus improve safety.

[0036] Optionally, before the risk value calculation step, a deviation position determination step and a risk value adjustment step are further provided;

[0037] Determine the deviation position: obtain the nodes that are prone to deviation. If there is a node that is prone to deviation and deviation occurs again, perform the risk value adjustment step;

[0038] Risk value adjustment: Intervene in the risk value calculation step and update the risk value The calculation model is as follows:

[0039] , where is the upward adjustment coefficient, .

[0040] By adopting the above technical solution, through multiple or regular analysis, the nodes where deviations often occur are determined. When deviations occur at this node, the risk value is increased to pay attention to this node, pay more attention to it, and promptly stop the expansion of the node deviation, reduce damage to equipment, improve the accuracy of risk identification and control, and improve production safety and the accuracy of the method.

[0041] In the second aspect, the present application provides a risk management system, which adopts the following technical solutions:

[0042] A risk management and control system includes the following modules:

[0043] Model building module: establish the SDG model of each equipment unit and simplify and merge them according to the process;

[0044] Node grading module: The input end is connected to the output end of the model building module and is used to grade the nodes and assign weights;

[0045] Deviation determination module: The input end is connected to the output end of the model building module, and is used to determine the node where the deviation is located and determine the propagation path;

[0046] Risk determination module: The input end is connected to the output end of the node classification module and the output end of the deviation determination module, and is used to calculate the risk value of the node deviation and determine whether the risk value exceeds the risk critical value;

[0047] Warning control module: The input end is connected to the output end of the risk determination module to remind relevant personnel to deal with deviations.

[0048] By adopting the above technical solution, according to the petrochemical production process and related data, the SDG model is built through the model building module, and then after merging and simplifying, an overall SDG model is formed, and then the node classification module is executed to classify the nodes according to their criticality, and then the weights are assigned; then according to the HAZOP analysis results, the deviation determination module is started to determine the nodes where the deviation occurs, and the propagation path is determined on the SDG model by manual deviation, and the nodes where the deviation occurs and their propagation paths are formed into an operational directed acyclic graph, and then the warning control module is executed to send the operational directed acyclic graph and the accident or risk that is likely to be caused to the relevant personnel, and remind the relevant personnel to repair; through the above system, it is easy to find the risk nodes that need to be corrected in time, and through the operational directed acyclic graph, the nodes can be corrected intuitively, the deviation correction efficiency is improved, the further propagation of node deviation is reduced, the occurrence of accidents is curbed in time, and safety is improved.

[0049] Optionally, it further comprises: a deviation qualitative module: the input end of which is connected to the output end of the deviation determination module, and the output end of which is connected to the input end of the risk determination module, for determining whether the deviation is an unconventional deviation;

[0050] Shutdown and inspection module: The input end is connected to the output end of the deviation qualitative module, and is used to execute the shutdown operation after determining that the deviation is an unconventional deviation, and check the nodes where the deviation occurs one by one.

[0051] By adopting the above technical solution and setting up the deviation qualitative module, the types of deviations can be distinguished, equipment damage caused by unconventional failures can be reduced, and they can be discovered and stopped in time, thereby reducing losses and the risk of accidents.

[0052] In summary, the present application includes at least one of the following beneficial technical effects:

[0053] 1. According to the production process and relevant information of the petrochemical plant, the SDG model is established, and then the SDG models of each equipment unit are integrated and simplified. According to the production process and accident risk level, the criticality of the node is determined and graded, and then weights are assigned according to the grade. Then, according to the HAZOP analysis report, the parameters of the deviation are determined, and then according to the SDG model, manual deviation is performed to determine the deviation propagation path, and then the node deviation risk value is calculated, and the maximum value is taken after comparing all node risk values. Then, it is determined whether the node deviation risk value is greater than the preset risk threshold value. If it is greater than, the relevant personnel are reminded to intervene and control to reduce the occurrence of accidents, or correct them in time to reduce the occurrence of inferior products, control production quality, and reduce safety risks. And through the calculation of risk values, the consequences of node deviation are more intuitive, reducing the error of manual judgment, and being able to control the safety of the process in time, reduce the occurrence of accidents, and improve risk management capabilities;

[0054] 2. When multiple node deviations occur, determine whether the nodes of the multiple deviations have a causal relationship or are on a propagation path through a directed acyclic graph. If so, execute the risk value calculation model to calculate the risk value. If not on a propagation path, execute the associated intervention step, update the risk value calculation model, increase the risk value coefficient, and then judge the risk value to determine whether intervention is needed; by recalculating the multi-node deviations without correlation, increase the attention to the multi-node deviations, reduce the risk, and be able to find equipment problems in time, reduce the continuous damage to the equipment, and improve safety;

[0055] 3. Through multiple or regular analysis, determine the nodes where deviations often occur. When deviations occur at this node, pay attention to this node by increasing the risk value, pay more attention to it, and promptly stop the expansion of the node deviation, reduce damage to equipment, improve the accuracy of risk identification and control, improve production safety, and the accuracy of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of the risk management method in Example 1 of the present application;

[0057] Figure 2 is a schematic diagram of the deviation propagation path in the SDG model in Example 1 of the present application;

[0058] Figure 3 This is a flowchart from the associated parameter acquisition step to the risk update step in Example 2 of the present application;

[0059] Figure 4 It is a flow chart from the deviation position determination step to the risk value adjustment step in Example 2 of the present application;

[0060] Figure 5 It is a system diagram of the risk management system in the embodiment of the present application. DETAILED DESCRIPTION

[0061] The following combination Figures 1 to 5 This application is described in further detail.

[0062] This embodiment discloses a process technology operation risk management method.

[0063] Example 1: Reference Figure 1 , the process technology operation risk control method includes the following steps:

[0064] S1: Establish SDG model: Establish the SDG model of each equipment unit, and merge and simplify the SDG model of each equipment unit according to the process flow;

[0065] Specifically, the production processes of petrochemical plants are different, and the corresponding process flows are different. The SDG model is built according to the products and processes of the actual plants. After the construction is completed, it is integrated according to the degree of process correlation, and then simplified to form an overall SDG model;

[0066] S2: Node classification: Classify nodes according to their criticality and assign weights based on the classification;

[0067] Specifically, the criticality of the node is determined according to the impact of the node on subsequent parameter deviations, or the length of the propagation path, or the impact on production quality and output, or the scope and degree of impact of the accident. The nodes are then graded according to the criticality, and the grades can be divided into 1 to n levels, where n is an integer. Weights are then assigned according to the grades, where the higher the criticality, the higher the grade and the corresponding weight assignment.

[0068] S3: Deviation determination: Determine the node where the parameter that generates the deviation is located, and manually pull the deviation according to the SDG model to determine the deviation propagation path and form an operational directed acyclic graph;

[0069] Specifically, after the HAZOP analysis report, the node where the deviation parameter is located is determined, and then according to the overall SDG model, the node is found, and according to the associated route, the propagation path is determined, and the path is recorded and extracted to form an operational directed acyclic graph.

[0070] S4: Risk determination: including relevance judgment step, relevance intervention step, risk value calculation step and risk judgment step;

[0071] S4-1: Correlation judgment: judge whether there is correlation among multiple node deviations. If so, execute the correlation intervention step. Otherwise, execute the risk value calculation step.

[0072] Specifically, when there are parameter deviations in multiple cause nodes, multiple cause nodes may cause changes in the parameters of the same intermediate node or consequence node. By judging whether there is a correlation between multiple cause nodes, that is, whether two or more nodes cause the same node parameter to change or cause the same node parameter to change through conduction, the mutual influence between nodes may increase the probability of risk occurrence. By executing the association intervention step, the risk calculation parameters are adjusted upward for the situation where multiple nodes influence each other, thereby improving the controllability of risks and reducing the occurrence of accidents.

[0073] S4-2: Risk value calculation: Calculate the node deviation risk value , Value at Risk The calculation model is as follows:

[0074] ;

[0075] In the formula, is the probability of deviation of the ith node at level n, Indicates taking the larger The calculation model is as follows:

[0076] ; In the formula, is a constant, and , is the basis for the value-at-risk calculation, Assign a value to the weight of the i-th node at level n;

[0077] Specifically, the calculation method of the probability of node deviation is the existing technology, and the specific steps are as follows: according to the deviation quantitatively described in the HAZOP analysis report, the fuzzy processing method is used to calculate the cause node probability; wherein, the determination steps of the cause node probability are as follows:

[0078] Calculate the average fuzzy number: According to the operation rules of interval numbers, calculate the relationship function of the average fuzzy number W;

[0079] The node probability occurrence level can be divided into seven levels: "very small, small, small, medium, large, large, very large", and triangular or trapezoidal fuzzy numbers can be used to replace the above natural language.

[0080] The left-right fuzzy number sorting method is adopted, wherein the left-right fuzzy number sorting method proposed by Lin CT and Wang MJ is adopted to convert the fuzzy number into a fuzzy possibility value; the maximum fuzzy set and the minimum fuzzy set defined by this method are:

[0081]

[0082]

[0083] Then the fuzzy probability of the fuzzy number W is:

[0084]

[0085]

[0086] in," "and" " respectively means taking the larger and taking the smaller, is the membership function corresponding to the fuzzy number W, W is the fuzzy number corresponding to the failure probability level;

[0087] The fuzzy probability formula for calculating W is:

[0088]

[0089] in," " means to take the larger one, is the left fuzzy number score, is the right fuzzy number score;

[0090] The fuzzy possibility value is converted into the fuzzy failure rate FFR (i.e., the cause node probability of the SDG model). The conversion formula for converting the fuzzy possibility value into the fuzzy failure rate FFR is as follows:

[0091] ;

[0092] in, ;

[0093] Reference Figure 2 In the SDG model, there are several basic influence relationships between nodes, and the propagation paths are different. There are the following situations:

[0094] a: The occurrence of the deviation or consequence of the latter variable depends on the occurrence of the deviation or bias of the former variable;

[0095] b: Multiple causes or variables lead to a certain consequence or variable deviation;

[0096] c: Different consequences caused by the same cause or variable deviation;

[0097] When the above three situations occur, the transmission path can be decomposed into Figure 2 The relationship between:

[0098] Figure 2 In , V2, F11, N1, p1, C1, L1, N2 and N3 can all be variable nodes or cause nodes. When they are cause nodes, their probabilities are cause node probabilities and can be directly used for calculation.

[0099] Then for the attachment Figure 2 The probability of each node in is calculated as follows:

[0100] When situation a occurs, the probability of deviation is calculated as follows:

[0101] ;

[0102] Where C1 is the rear deviation node, p1 is the middle deviation node, V2 is the front cause node or variable deviation node, P1 and P2 are propagation paths, V2 generates deviation p1 through propagation path P1, and p1 generates deviation C1 through propagation path P2;

[0103] It can be seen that there is a similar "series" relationship between nodes on compatible branches in the SDG model, and when the propagation path is turned on, the deviation or consequence of the latter variable depends on the deviation or deviation of the previous variable. There is a mutual influence relationship between the nodes, which needs to be considered comprehensively when calculating the probability.

[0104] When situation b occurs, the probability of deviation is calculated as follows:

[0105] ;

[0106] Where, F11 is the cause or variable, V2 is the cause or variable, L1 is the consequence or variable deviation caused by F11 and / or V2, P3 is the propagation path between F11 and L1, and P4 is the propagation path between V2 and L1;

[0107] When several causes or variables cause a certain consequence or variable deviation, the graph forms a "parallel" relationship, that is, the occurrence of any cause or forward variable will lead to the consequence or variable deviation. In this case, it is necessary to consider the mutual influence relationship between each cause or forward variable.

[0108] When situation c occurs, the probability of deviation is calculated as follows:

[0109]

[0110]

[0111] Where N1 is the cause or variable, N2 and N3 are the consequences or variable deviations caused by N1, P5 is the propagation path between N1 and N2, and P6 is the propagation path between N1 and N3;

[0112] Through the calculation of the probability of deviation, it can be seen that different consequences caused by the same cause or variable deviation have no influence on each other when calculated, and they are calculated independently of each other.

[0113] Specifically, when the lopa protective layer is used normally, the values ​​of P1, P2, P3, P4 and P5 in the above formula are usually 1. In other cases, they can be obtained according to the data set given by the lopa database;

[0114] The probability of node deviation and the node deviation risk value calculation model are obtained according to the calculation: , calculate the risk value, and obtain the maximum node deviation where an accident may occur by taking the largest value in the calculation model. Assign a value to the weight of the i-th node at level n, is a constant, and ,and The value of is determined according to the criticality of the node or the node classification. The larger the value of is, and the maximum value does not exceed 2, where The value of is determined based on historical data or historical accident data.

[0115] S4-3: Related intervention: Intervene in the risk value calculation step and update the risk value calculation model as follows:

[0116] , where to rank the nodes, and , where the higher the criticality, The bigger.

[0117] S4-4: Risk judgment: judge the risk value of node deviation Is it greater than the risk threshold? , if yes, then execute the warning control steps;

[0118] S5: Warning control: remind relevant personnel to intervene.

[0119] Specifically, by calculating the risk value that may be caused by deviations in each node, and then calculating the increase coefficient through weight intervention, and then taking the maximum value among the increase coefficients, and then adjusting the risk value, special attention is paid to the situation where multiple nodes influence each other, and the importance of the above situation is strengthened, so as to facilitate timely discovery of the situation where multiple nodes interfere with each other, and when the risk value is greater than the risk limit value, timely intervention is carried out to improve the reliability and accuracy of risk management and control, and thus improve safety.

[0120] Specifically, after the risk value is calculated, the risk value is compared with the risk threshold. When it is greater than the risk threshold, the warning control step is executed to remind relevant staff to intervene, correct the deviation parameters, or adjust the cause node; The value is calculated based on historical accident data, and the risk critical value is calculated through the parameter changes of historical accidents.

[0121] The implementation principle of Example 1 of the present application is as follows: according to the actual production products, production process flow and related information of the petrochemical plant, an SDG model is constructed, and then according to the relationship between each sub-process or sub-process, integration and optimization are performed to form an overall SDG model, and then the criticality of the node is determined according to the number of nodes in the overall SDG model, the propagation path of each node and the degree of danger of possible accidents or accidents, and then weights are assigned according to the criticality, and then according to the HAZOP analysis report, the position of the parameter deviation is determined, and then the deviation is manually pulled according to the propagation path, and then it is extracted separately to form an operational directed acyclic graph, and then the qualitative description of the HAZOP analysis report is converted into a quantitative description to obtain the cause node probability, and then the deviation probability is obtained through the directed acyclic graph and the propagation path calculation, and then the risk value of the node deviation is calculated according to the risk value calculation model, and then the risk value is compared to determine whether warning control is needed.

[0122] When multiple node parameter deviations occur, a preliminary distinction is made by judging whether there is a correlation between the multiple nodes to distinguish whether the nodes with deviations are related. If there is a correlation, the risk value calculation model is updated, the upward adjustment coefficient is increased, and then the calculation is performed. After the calculation is completed, the risk value is compared to determine whether warning control is required.

[0123] Example 2: Reference Figure 3 , the difference between this embodiment and embodiment 1 is that a report push step is also provided after the warning control step;

[0124] S6: Report push: An operational directed acyclic graph is formed based on the SDG model and the deviation propagation path, and an operational report is formed and pushed to relevant personnel.

[0125] Specifically, by forming reports and pushing them to relevant personnel, the transmission path can be displayed to relevant personnel in a more intuitive manner, making it convenient for relevant personnel to make targeted corrections based on operational reports, thereby improving work efficiency, quickly eliminating safety hazards, and improving safety.

[0126] Reference Figure 3 In other embodiments, a deviation qualitative step is further provided after the deviation determination step;

[0127] S7: Deviation characterization: Determine whether the deviation is an unusual deviation. If so, execute the shutdown inspection step. Otherwise, execute the correlation judgment step.

[0128] Specifically, non-routine deviations include equipment failures: control system failures: failures in control systems such as temperature, pressure, and flow may cause related parameters to deviate from set values. For example, a temperature control system failure may cause the temperature to exceed the set range, and a pressure control system failure may cause the pressure to be too high or too low.

[0129] Sensor failure: Failure or damage of various sensors (such as temperature sensors, pressure sensors, flow sensors, etc.) will directly affect the accurate measurement of parameters, resulting in deviations.

[0130] Failure of equipment such as valves and pumps: Valve leakage, change in pump capacity, pump damage, etc. will affect the stability of parameters such as flow and pressure.

[0131] Changes in medium state: Changes in medium properties: Changes in the physical properties of the medium, such as density, viscosity, and temperature, will affect its flow state in the pipeline, and thus affect the measurement of parameters such as flow rate and pressure.

[0132] Chemical reaction of media: In some process, chemical reactions may occur between media, resulting in the formation or consumption of products, thus affecting the stability of related parameters.

[0133] Environmental factors: Changes in ambient temperature and humidity: Changes in ambient temperature and humidity may affect the operating efficiency and measurement accuracy of the equipment, resulting in parameter deviations.

[0134] External vibration and electromagnetic interference: External vibration and electromagnetic interference may interfere with the normal operation of the equipment and the accurate transmission of signals, causing parameter deviations.

[0135] Process design issues: Unreasonable process parameter design: The actual situation and process requirements are not fully considered when designing process parameters, resulting in parameter setting values ​​not meeting actual requirements, thus causing deviations.

[0136] Improper equipment selection: When selecting equipment, a reasonable match is not made between process requirements and equipment performance, resulting in the equipment being unable to meet process requirements during operation, causing parameter deviations.

[0137] When deviations due to the above-mentioned unconventional reasons occur, the shutdown and inspection steps are executed. Conversely, if parameter changes may be caused by operating errors or other reasons, for example, the flow rate decreases due to a smaller opening of the pipeline valve, the correlation judgment step is executed.

[0138] S8: Shutdown and check: According to the deviation position, check the location of the parameter variable one by one to determine whether it is an abnormal deviation. If so, shut down and repair it.

[0139] Specifically, deviations caused by unconventional reasons are difficult to correct and require individual investigation and analysis of causes to reduce the likelihood of problems being left unsolved and leading to large-scale equipment failures or accidents.

[0140] In other embodiments, a correlation parameter acquisition step is further provided between the correlation determination step and the correlation intervention step;

[0141] That is, after determining that multiple node deviations are associated, the associated parameter acquisition step is performed;

[0142] S4-5: Obtaining associated parameters: According to the directed acyclic graph, determine the location of the cause node and obtain the associated interval between the cause nodes , get the associated interval Greater than the associated interval limit The number of and in the associated interval The proportion of the total number ;

[0143] S4-6: Determination of correlation: Determination of proportion Is it greater than the percentage limit? , if yes, then execute the risk update step, otherwise, execute the associated intervention step;

[0144] Specifically, when multiple nodes deviate, it is determined whether there is a correlation. When multiple cause nodes have problems, the result may be the same. When the node deviations are on the same directed acyclic graph, or the result is the same or the intermediate variables have an intersection, it means that there is a correlation. Then the correlation interval is obtained to determine whether the correlation interval is greater than the correlation interval limit value, that is, whether the correlation is low. When the correlation is low, the risk value calculation model is updated again by executing the risk update step; the so-called correlation interval is the number of interval nodes transmitted from two nodes to the same node. For example, two cause nodes cause parameter changes in the same intermediate node, and the interval between the two cause nodes and the node is four different intermediate nodes, then the correlation interval is four; where The value of is obtained based on historical data, and after determining how many nodes are separated, it is determined whether there is a slight impact or no impact on a certain event. In this embodiment, it is determined to be 4 to 6 node intervals based on historical data. Therefore, in this embodiment The value is between 4 and 6, and it is an integer; The value of is three-fifths, that is, when more than half of the associated intervals are large, the risk update step is executed. The value of is obtained by analyzing historical data.

[0145] S4-7: Risk update: Intervene in the risk value calculation step and update the risk value calculation model as follows:

[0146] , where the value model of S is as follows: ,and , The value model of is as follows:

[0147] .

[0148] Specifically, by distinguishing and calculating the risk values ​​for the three types of strong correlation, weak correlation and no correlation respectively, the risk adjustment coefficient is strengthened, so that the risk value calculated by multiple uncorrelated deviation nodes is maximized, the coefficient of multiple deviation nodes with strong correlation is increased, and the weight of the risk caused by multiple deviation nodes with strong correlation is enhanced, so as to facilitate timely discovery of the possibility of risks caused by problems with deviation nodes, and distinguish the weights between multiple deviation nodes with weak correlation and uncorrelated deviation nodes, so as to facilitate timely containment of accidents.

[0149] Among them, regarding the value of S, by comparing the correlation interval with the correlation interval boundary value, and when the value is greater than 2, the value is taken as 2, and the proportion of weaker correlation strength is intervened in the value of S, so that the calculation of the risk value is more reasonable and more convenient for timely risk control. In the formula, The value of is greater than 1, and The value of is determined according to the criticality of the node or the node classification. The larger the value, the greater the value, and the maximum value does not exceed 2.

[0150] Reference Figure 4 In other embodiments, between the risk value calculation step and the risk judgment step, a deviation position determination step and a risk value adjustment step are also provided;

[0151] S4-8: Determine the deviation position: obtain the nodes prone to deviation, and if there is a node prone to deviation and deviation occurs again, perform the risk value adjustment step;

[0152] S4-9: Risk value adjustment: Intervene in the risk value calculation steps and update the risk value calculation model as follows:

[0153] , where is the upward adjustment coefficient, .

[0154] Specifically, after multiple detection cycles have been analyzed and detected, if there is a node that often deviates and when it occurs again, the risk value is calculated upward by using the upward adjustment coefficient m through the risk value adjustment step to reduce the impact of excessive deviation or persistent deviation on the device; Regarding the value of m, m can be determined as a fixed value according to the weight of the node, and the fixed value is greater than 1 and less than 2. In other implementations, it can be calculated by designing a calculation model, and the calculation model of m can be:

[0155] ; In the formula, is the number of deviations that occur at the current node, z is the maximum number of times the current node is allowed to occur, .

[0156] The implementation principle of Example 2 of the present application is: when multiple node parameter deviations occur, a preliminary distinction is made by judging whether there is a correlation between the multiple nodes, distinguishing between the two situations of correlation and non-correlation, and then calculating the risk value when the non-correlated node deviation occurs, and then performing risk judgment to determine whether management and control are needed to reduce the expansion of risks. When there is a correlation, the degree of correlation is analyzed, the correlated node deviation is analyzed, and boundaries are drawn to distinguish the strength of the correlation. Then, the risk value is calculated separately to reduce the probability of weakly correlated node deviations, and the risk value is greater than the actual risk possibility, thereby optimizing the accuracy of risk management and control.

[0157] Mark nodes that are prone to deviation or often deviate. By strengthening management and control, timely curb excessive deviations at the nodes, which may cause damage to equipment, reduce risks, and improve safety.

[0158] The embodiment of the present application also discloses a risk management system.

[0159] Reference Figure 5 , risk management and control system, including the following modules:

[0160] Model building module: establish the SDG model of each equipment unit and simplify and merge them according to the process;

[0161] Node grading module: The input end is connected to the output end of the model building module and is used to grade the nodes and assign weights;

[0162] Deviation determination module: The input end is connected to the output end of the model building module, and is used to determine the node where the deviation is located and determine the propagation path;

[0163] Deviation qualitative module: The input end is connected to the output end of the deviation determination module, and is used to detect the deviation node to determine whether the deviation is an unconventional deviation;

[0164] Shutdown and inspection module: The input end is connected to the output end of the deviation qualitative module, and is used to execute the shutdown operation when it is judged to be an abnormal deviation, and to inspect the nodes one by one and make corrections in time;

[0165] Risk determination module: The input end is connected to the output end of the node classification module and the output end of the deviation qualitative module. It is used to determine whether there is a correlation between multiple nodes when deviations occur in multiple nodes, and to intervene and update the risk value calculation model respectively, and calculate the risk value of the node deviation, and then determine whether the risk value exceeds the risk critical value;

[0166] Warning control module: The input end is connected to the output end of the risk determination module to remind relevant personnel to deal with deviations.

[0167] Report push module: The input end is connected to the output end of the warning control module, and is used to push the directed acyclic graph and operability report to relevant personnel when the risk value is greater than the risk threshold.

[0168] The implementation principle of a risk management system in an embodiment of the present application is as follows: a model building module builds an SDG model according to the production process technology and performs merge optimization, then executes a node grading module to grade the nodes according to their criticality and assign weights, then executes a deviation determination module to convert qualitative deviations into quantitative deviations according to a HAZOP analysis report, then determines the propagation path according to a directed acyclic graph, then calculates the probability of deviation occurrence, then starts a deviation qualitative module to determine whether the deviation is a routine deviation, and when it is an unconventional deviation, executes a shutdown inspection module to shut down the equipment and inspect the nodes one by one, and when it is a routine deviation, starts a risk determination module to analyze whether there is a correlation between the nodes with deviations, updates the risk value calculation model respectively, calculates the risk value, and then determines whether the risk value exceeds the set risk limit value, and if so, executes a warning management module to remind relevant personnel to conduct management intervention to reduce the occurrence of accidents and curb the expansion of risks.

[0169] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A process technology operation risk management method, characterized by: The following steps are involved: Establish SDG model: Establish SDG model of each equipment unit, and merge and simplify the SDG model of each equipment unit according to the process flow; Node classification: Classify nodes according to their criticality and assign weights based on the classification; Deviation determination: Determine the node where the parameter that generates the deviation is located, and manually pull the deviation according to the SDG model to determine the deviation propagation path and form an operational directed acyclic graph; Risk determination: including relevance judgment step, relevance intervention step, risk value calculation step and risk judgment step; Correlation judgment: judge whether there is correlation among multiple node deviations. If so, execute the correlation intervention step. Otherwise, execute the risk value calculation step. Risk value calculation: Calculate the node deviation risk value , Value at Risk The calculation model is as follows: ； In the formula, is the probability of deviation of the ith node at level n, Indicates taking the larger The calculation model is as follows: ; In the formula, is a constant, and , is the basis for the value-at-risk calculation, Assign a value to the weight of the i-th node at level n; Association intervention: Intervene in the risk value calculation step and update the risk value The calculation model is as follows: , where to rank the nodes, and , where the higher the criticality, The bigger; Risk judgment: judge the node deviation risk value Is it greater than the risk threshold? , if yes, then execute the warning control steps; Warning control: remind relevant personnel to intervene.

2. The process technology operation risk management method according to claim 1, characterized in that: After the warning control step, there is also a report push step; Report push: An operational directed acyclic graph is formed based on the SDG model and the deviation propagation path, and an operational report is formed and pushed to relevant personnel.

3. The process technology operation risk management method according to claim 1, characterized in that: A deviation qualitative step is also provided after the deviation determination step; Deviation characterization: Determine whether the deviation is an unusual deviation. If so, perform the shutdown inspection step. Otherwise, perform the risk value calculation step. Shutdown and inspection: According to the deviation location, check the location of the parameter variable one by one to determine whether it is an abnormal deviation. If so, shut down and repair it.

4. The process technology operation risk management method according to claim 1, characterized in that: Between the relevance judgment step and the relevance intervention step, a relevance parameter acquisition step, a relevance judgment step and a risk update step are also set; Correlation parameter acquisition: According to the directed acyclic graph, determine the location of the cause node and obtain the correlation interval between the cause nodes , get the associated interval Greater than the associated interval limit The number of and in the associated interval The proportion of the total number ; Relevance judgment: judgment ratio Is it greater than the percentage limit? , if yes, then execute the risk update step, otherwise, execute the associated intervention step; Risk update: Intervene in the risk value calculation step and update the risk value The calculation model is as follows: , where the value model of S is as follows: , ,and , The value model of is as follows: 。 5. The process technology operation risk management method according to any one of claims 1 to 4, characterized in that: Before the risk value calculation step, a deviation position determination step and a risk value adjustment step are also provided; Determine the deviation position: obtain the nodes that are prone to deviation. If there is a node that is prone to deviation and deviation occurs again, perform the risk value adjustment step; Risk value adjustment: Intervene in the risk value calculation step and update the risk value The calculation model is as follows: , where is the upward adjustment coefficient, .

6. A risk management system for implementing the process technology operation risk management method according to any one of claims 1 to 5, characterized in that: Includes the following modules: Model building module: establish the SDG model of each equipment unit and simplify and merge them according to the process; Node grading module: The input end is connected to the output end of the model building module and is used to grade the nodes and assign weights; Deviation determination module: The input end is connected to the output end of the model building module, and is used to determine the node where the deviation is located and determine the propagation path; Risk determination module: The input end is connected to the output end of the node classification module and the output end of the deviation determination module, and is used to calculate the risk value of the node deviation and determine whether the risk value exceeds the risk critical value; Warning control module: The input end is connected to the output end of the risk determination module to remind relevant personnel to deal with deviations.

7. The risk management and control system according to claim 6, characterized in that: Also includes: Deviation qualitative module: the input end is connected to the output end of the deviation determination module, and the output end is connected to the input end of the risk determination module, which is used to determine whether the deviation is an unconventional deviation; Shutdown and inspection module: The input end is connected to the output end of the deviation qualitative module, and is used to execute the shutdown operation after determining that the deviation is an unconventional deviation, and check the nodes where the deviation occurs one by one.

Images