Quantitative risk assessment method for petroleum refining device protection layer

Abstract

The invention belongs to but not limited to the technical field of safety control, and particularly relates to a quantitative risk assessment method and system for a petroleum refining device protection layer, and the method comprises the steps: S1, carrying out the harmfulness analysis of a system, and forming a system HAZOP analysis table; s2, according to a system HAZOP analysis table, fusing system process knowledge such as operation regulations, and constructing a safety knowledge graph of the system; s3, comprehensively considering the influence of the individual background of the expert and the evaluation process on the weight of the expert, determining a weight coefficient, and assisting in making a quantitative risk evaluation rule; and S4, according to the structural characteristics of the security knowledge graph, considering the path risk propagation influence and the reduction effect of the protection layer on the influence, performing rule supplementation on the RPN calculation method to obtain a quantitative risk assessment rule, and completing system risk assessment based on the rule.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of safety control technology, and in particular relates to a quantitative risk assessment method and system for a protective layer of a petroleum refining device. Background Art

[0002] In petroleum refining units, catalytic cracking is a very important secondary processing process and a key means of converting heavy oil into light products. However, there are many dangerous factors involved in the catalytic cracking production process. Most of the raw materials, intermediate products and finished products in production are flammable, explosive and combustible substances, which pose a great risk of fire and explosion. If an abnormal situation occurs, it will cause casualties. Therefore, it is particularly important to carry out process hazard analysis and risk assessment.

[0003] Hazard and Operability Study (HAZOP) is one of the most widely used process hazard analysis methods. It can comprehensively identify potential hazards in the system, is applicable to various types of process flow units, and is not limited by the type and scale of process units. It can be flexibly applied in process research and development, design, operation and shutdown stages. However, the expert knowledge base generated by the traditional HAZOP method usually only stores information in a database, and there is a lack of correlation between the information, which makes it difficult for the HAZOP analysis method to guide dynamic risk assessment.

[0004] At present, the risk assessment method based on HAZOP is usually combined with other semi-quantitative or quantitative methods, including risk matrix, fault tree, layer of protection analysis (LOPA) and other methods. However, these methods generally rely on the probability of occurrence of events for risk assessment calculations, lack consideration of risk propagation, are highly subjective and have low accuracy. For example, the invention patent CN202310902199.6 discloses a risk identification method for key equipment in the shale oil and gas gathering and transportation process. This method is based on the qualitative analysis results of HAZOP, and then combines the LOPA analysis method and the F&EI analysis method to determine the accident probability and risk level of the system, respectively, and quantitatively analyzes the risk level of fire and explosion accidents in the system by combining the risk matrix. However, the accuracy of this traditional method is too low. Even if multiple methods are integrated for analysis, the accuracy and objectivity still need to be improved. Invention patent application CN202310016238.2 discloses a risk assessment method for hydrocracking units based on fuzzy fault trees. This method builds a fault tree model for explosion or fire accidents in the unit based on the results of HAZOP qualitative analysis, and calculates the probability of occurrence of basic events by combining expert review methods with fuzzy theory. This method uses fuzzy theory to reduce the subjectivity of the assessment results, but risk assessment based solely on the probability of occurrence is not comprehensive enough, and the calculation of the impact of risk propagation is not objective enough.

[0005] As a structured knowledge representation method, knowledge graph displays entities and their relationships through nodes and edges, making information easier to understand, manage and apply. It can effectively improve the accuracy and ease of use of HAZOP analysis in quantitative risk assessment of catalytic cracking units.

[0006] In view of the above analysis, the technical problems that urgently need to be solved in the existing technology are: the existing methods generally rely on the probability of occurrence of events for risk assessment calculations, lack consideration of risk propagation, are highly subjective, have low accuracy, and increase the complexity of calculations. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a quantitative risk assessment method and system for the protective layer of a petroleum refining device.

[0008] The present invention is implemented as follows: a quantitative risk assessment method for a protective layer of a petroleum refining device, comprising:

[0009] S1, conduct hazard analysis on the system and form a system HAZOP analysis table;

[0010] S2, based on the system HAZOP analysis table, integrates system process knowledge such as operating procedures to build a system safety knowledge graph;

[0011] S3, comprehensively consider the impact of the expert's personal background and the evaluation process on the expert's weight, determine the weight coefficient, and assist in the formulation of quantitative risk assessment rules;

[0012] S4, according to the structural characteristics of the security knowledge graph, taking into account the impact of path risk propagation and the reduction effect of the protection layer on the impact, the RPN calculation method is supplemented with rules to obtain quantitative risk assessment rules, and the system risk assessment is completed based on the rules.

[0013] Further, the specific process of S1 is:

[0014] (1.1) The analysis target is clearly defined as the petroleum refining unit. The analysis scope mainly includes the reaction regeneration unit, the fractionation unit and the absorption stabilization unit. An expert group is established to carry out HAZOP (hazard and operability study) analysis based on the safety requirements of the petroleum refining unit.

[0015] (1.2) Be familiar with the object and purpose of analysis, collect and organize relevant information, including process flow diagrams (PFDs), process and equipment diagrams (P&IDs), etc., agree on the record form format, and complete the schedule planning.

[0016] (1.3) Use P&IDs to divide nodes, analyze the deviations generated by each node, the causes of the deviations and the impacts caused, and use risk matrix analysis to obtain the semi-quantitative risk level of accident consequences. First, determine the guide words, and determine the deviations based on the guide words. The deviations are in the form of "parameters + guide words" as the starting point for dangerous event analysis. There are two main causes: direct causes and root causes. Direct analysis of root causes takes a lot of time and requires the use of other analysis tools. Therefore, only direct causes are considered when conducting HAZOP analysis, including equipment failure, instrument failure, and operational errors.

[0017] (1.4) Propose corresponding protective measures for the causes of deviations that lead to excessively high risk levels, so as to promptly detect the occurrence of deviations or mitigate the consequences of failures.

[0018] Further, the specific process of S2 is:

[0019] (2.1) Analyze the elements in the HAZOP analysis table and design the safety knowledge graph ontology based on it. The knowledge graph ontology is mainly composed of classes, attributes, relationships, and rules. Determine the node categories and the relationship categories between nodes, add node attributes according to the evaluation requirements, and build a network topology diagram suitable for the system.

[0020] (2.2) Filling the knowledge graph data. Since the analysis text string for the petroleum refining unit is relatively long, after trying multiple entity recognition algorithms, it was found that the automatic extraction accuracy is very low. In order to ensure the accuracy of the subsequent risk assessment, the present invention adopts manual extraction to convert the HAZOP information into triple data of the "entity-relationship-entity" structure, and uses the Neo4j graph database to store the structured data in the form of a graph. Specifically, the entities are stored as node types in Neo4j, and the relationships are stored as edges, so that the connection between the HAZOP information can be more intuitive and accurate. At the same time, the graph query language can be used to quickly retrieve the required data for subsequent failure path extraction.

[0021] Furthermore, the specific process of S3 is as follows:

[0022] (3.1) Comprehensively consider the three factors of expert position, years of work experience, and education level to formulate a scoring standard for expert personal background.

[0023] (3.2) Calculate the expert background weight, assuming that the total background score of the i-th expert is g i , then the expert background weight

[0024] (3.3) Calculate the average distance of the evaluation interval. Suppose there are two intervals [x 1 ,x 2 ],[y 1 ,y2 ], using the Euclidean distance function Calculate the distance between two interval numbers. Assume that the distance between the scoring interval of the i-th expert and the scoring interval of the j-th expert is d ij , then the average distance between the evaluation intervals of the i-th expert and other experts is

[0025] (3.4) Calculate the consistency of the evaluation intervals, use normalization to calculate the consistency of the expert evaluation intervals, and the relative consistency of the evaluation interval of the i-th expert

[0026] (3.5) Comprehensively consider the consistency between the expert background score and the expert evaluation results, and determine the final weight K of the i-th expert evaluation interval i , where α and β reflect the relative importance of the two weights, satisfying 0≤α≤1, 0≤β≤1, α+β=1. The influence of expert consensus on the weight is determined by adjusting the values ​​of the two. K i =αk i +βFA ij .

[0027] Further, the specific process of S4 is:

[0028] (4.1) Formulate an evaluation comparison table. For the i-th path, the j-th expert gives the score interval number [S] for the three risk indicators according to the evaluation comparison table according to the evaluation rules. Ij1 ,S Ij2 ],[O Ij1 ,O Ij2 ],[D Ij1 ,D Ij2 ].

[0029] (4.2) The specific rules for the three indicators in the formula RPN = S*O*D are as follows, where S is the severity index, O is the probability of occurrence index, and D is the detectability index. S refers to the final variable that affects the severity of the consequences of the current analysis deviation. The value of this index can be fuzzily evaluated in combination with the risk level part of the HAZOP analysis table and the evaluation comparison table. D represents the nature of no abnormality detected after the deviation occurs. This index requires experts to judge the importance of each variable involved in the deviation in the system in combination with process knowledge. The higher the importance of the variable, the more vulnerable it is to attack, and the higher the value of D. If the variable has alarm protection, the general standard value of the alarm failure probability is 10. -1According to the evaluation comparison table, the occurrence of each alarm will cause the evaluation value of D to drop by one level. It should be noted that only the impact of the alarm failure of the variable involved in the current deviation is considered. Whether the alarm of other variables fails has no effect on whether the abnormality of the current analysis variable can be detected. The value of O is based on the probability evaluation value of the cause (fault), taking into account the reduction of the propagation impact caused by the protection layer factor and other reasons. Assuming that there is only one cause node R in the tth path t , [O ti1 ,O ti2 ] represents R t The i-th evaluation interval of the probability of occurrence, a represents the reduction coefficient of the propagation impact caused by other factors, x t represents the number of alarms and basic process control systems (BPCS) on the tth path, then the value of O on the tth path is

[0030] To solve the problem of multiple cause nodes in a single path, we need to perform logical accumulation calculation on O. S and D remain unchanged. Assuming that there are m cause nodes in the tth path, O ti Represents the i-th reason R that causes the variable abnormality in the t-th path ti The probability of occurrence is evaluated, then the value of the tth path O is

[0031] When performing risk ranking later, since S is much more important among the three indicators, for paths with larger S values, the risk ranking should be higher even if the RPN value is relatively small.

[0032] (4.3) According to the above rules, the node path RPN value and node RPN value are calculated based on the expert evaluation interval, and the values ​​are sorted. Based on the risk threshold given by the experts, it is determined which paths need to add protection layers, and specific suggestions for adding protection layers are made for specific paths.

[0033] Another object of the present invention is to provide a quantitative risk assessment system for a protective layer of a petroleum refining device for realizing the quantitative risk assessment method for a protective layer of a petroleum refining device, comprising:

[0034] Hazard analysis module: performs hazard analysis on the system and forms a system HAZOP analysis table;

[0035] The knowledge graph construction module builds the system's safety knowledge graph based on the system HAZOP analysis table and the system process knowledge such as operating procedures;

[0036] The weight coefficient determination module comprehensively considers the impact of the expert's personal background and the evaluation process on the expert's weight, determines the weight coefficient, and assists in the formulation of quantitative risk assessment rules;

[0037] The risk assessment module, based on the structural characteristics of the security knowledge graph, takes into account the impact of path risk propagation and the reduction effect of the protection layer on the impact, supplements the RPN calculation method with rules, obtains quantitative risk assessment rules, and completes system risk assessment based on the rules.

[0038] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the quantitative risk assessment method for the protective layer of a petroleum refining device.

[0039] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the quantitative risk assessment method for the protective layer of a petroleum refining device.

[0040] Another object of the present invention is to provide an information data processing terminal, which includes the quantitative risk assessment system for the protective layer of a petroleum refining device.

[0041] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0042] First, the present invention overcomes the problem of poor reusability of traditional HAZOP analysis results, increases the reliability of subsequent research, and has higher accuracy and more comprehensive considerations in evaluation. In general, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0043] (1) The method for constructing a safety knowledge graph based on HAZOP analysis results provided by the present invention parses the HAZOP analysis results stored in a table form and stores them in a knowledge graph to achieve visual display, thereby solving the problem of difficulty in sharing and reusing HAZOP analysis results. At the same time, due to the structured characteristics of the knowledge graph, quantitative risk assessment rules can be designed in combination with the impact of path risk propagation and the reduction effect of the protection layer on the impact;

[0044] (2) The rule-based quantitative risk assessment method provided by the present invention improves the basic RPN calculation method, making the calculation of the three indicators of severity, occurrence probability, and detectability more suitable for failure path risk assessment analysis, while reducing the subjectivity of the analysis. Compared with the traditional RPN calculation method, the weight distribution of this method is more reasonable, and the hazard propagation factors and protection layer reduction factors are considered more comprehensively, and the results are more practical.

[0045] Second, the expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0046] First, it helps enterprises reduce the deviations and omissions in the traditional manual risk assessment process, optimizes the safety management process, and thus significantly reduces the probability of safety accidents and reduces economic losses. This invention introduces systematic HAZOP analysis, safety knowledge graphs, and quantitative risk assessment rules to make the risk assessment process more accurate, reduce subjectivity, and improve the comprehensiveness of risk identification. At the same time, the risk reduction effect of the protection layer is accurately considered during the assessment process, making the risk calculation more realistic and reducing unnecessary safety redundancy.

[0047] Second, the invention is versatile and scalable, with a huge potential market size, covering industries such as petrochemicals, chemicals, and energy. The modular design of the method and system enables it to adapt to different devices and production processes, and expand to other fields involving the production, storage, and transportation of hazardous chemicals.

[0048] The third is to enhance the technical capabilities and industry competitiveness of enterprises and seize market opportunities in the increasingly stringent security regulatory environment. By building a security knowledge graph, enterprises can achieve the transition from traditional manual management to intelligent management and enhance digital transformation capabilities. The evaluation results based on the knowledge graph and risk propagation model can serve as decision support to help enterprise management formulate more scientific security strategies.

[0049] The technical solution of the present invention fills the technical gap in the industry at home and abroad: At present, research ideas combining HAZOP analysis with knowledge graphs have begun to emerge at home and abroad, but there are few follow-up studies on structured HAZOP data, especially in the field of risk assessment. The current research focus is on fusion assessment with other semi-quantitative or quantitative methods, which generally rely on the probability of occurrence of events, lack consideration of risk propagation, and cannot maximize the use of the structured characteristics of knowledge graphs. Based on graph nodes and node relationships, the present invention obtains risk assessment results under the influence of risk propagation factors and protection layer reduction factors, filling the technical gap in this technical field in the industry at home and abroad, and has broad market prospects and social value. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flow chart of a quantitative risk assessment method for a petroleum refining unit provided by an embodiment of the present invention;

[0051] Figure 2 It is a system security knowledge graph ontology model diagram provided by an embodiment of the present invention;

[0052] Figure 3 is a failure path diagram of each deviation type node provided by an embodiment of the present invention;

[0053] Figure 4 is a single deviation node relationship subgraph provided by an embodiment of the present invention;

[0054] Figure 5 It is a quantitative risk assessment result diagram of a single deviation node provided by an embodiment of the present invention;

[0055] Figure 6 It is a schematic diagram of a quantitative risk assessment system for a petroleum refining unit provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] The quantitative risk assessment method for the protective layer of a petroleum refining device provided by the present invention has the following process: Figure 1 As shown; the regeneration system is crucial in the catalytic cracking unit. It not only ensures the regeneration of the catalyst and the stable operation of the device, but also improves the heat utilization efficiency and environmental protection level, thus having a profound impact on the economic benefits and environmental protection performance of the entire production process. Therefore, based on a semi-physical simulation experimental platform of a certain petroleum refining unit, the reaction regenerator was selected as the experimental object, and the method proposed in the present invention was experimentally verified. The system contains a total of 46 process variables related to temperature, pressure, flow and liquid level, and has 27 controllers. The implementation of the method is as follows:

[0058] Step 1: Establish an expert team to conduct HAZOP analysis.

[0059] (1) The analysis team consists of a team leader, a recorder, and three professional members, a total of 5 people, including a member who is proficient in HAZOP analysis method, a member who is familiar with process production, and a member who is familiar with equipment maintenance and repair.

[0060] (2) For the above experimental scenario, combined with supporting data such as P&IDs and PFDs, the experimental object is divided into four parts: reactor, regenerator, settler and external heat exchanger, and the deviations generated in each part, the causes of the deviations, the impacts caused and the existing protection measures are analyzed. The parameters involved in the deviations mainly include temperature, liquid level, pressure, flow rate, etc. IEC 61882 stipulates common guide words such as too large, too small, reverse, etc. The protection measures of the example objects selected in the present invention mainly include basic process control system (BPCS), safety instrument system (SIS) and alarm, etc.

[0061] (3) An example of system HAZOP analysis is shown in Table 1.

[0062] Table 1 HAZOP analysis results of the recirculation system of a catalytic cracking unit

[0063]

[0064] Step 2: Based on the system HAZOP analysis table, integrate system process knowledge such as operating procedures to build a system safety knowledge graph.

[0065] (1) Perform HAZOP element analysis. HAZOP elements mainly include deviation, cause, consequence, protective measures and risk level. The consequence and risk level are used as references for the expert system evaluation later. The construction of the system safety knowledge graph is only for the generation of node failure paths. Therefore, the elements required to construct the graph include deviation, cause and protective measures.

[0066] (2) Design the safety knowledge graph ontology. Based on the above element analysis, the HAZOP ontology mainly includes deviation class, cause class and protective measure class. In order to meet the special needs of subsequent risk assessment application scenarios, the protective measure class can be divided into three protection layer categories: BPCS class, SIS class and alarm class. Among them, SIS interlocking further derives interlocking action class, which is convenient for subsequent calculation and analysis. Each of the above categories can be abstracted as "entity + state" for description. For example, the deviation class: the temperature at the bottom of the regenerator is too high, which can be divided into two parts: regenerator (entity) and temperature too high (state). The attributes of the deviation class include name and bit number, the attributes of the cause class, SIS class, interlocking action class and alarm class include name, and the attributes of the BPCS class include controller bit number, actuator bit number and name. There are three types of relationships within and between classes: cause and effect, protection and interlocking.

[0067] (3) Extract specific data from the system HAZOP analysis table to obtain the relationship between specific nodes. Perform text extraction on the HAZOP analysis results, combine the content of the process manual such as the catalytic cracking operating procedures, perform data processing, remove duplicate content, and standardize the text format. Finally, 81 nodes, 118 nodes, 10 nodes, 25 nodes, 6 nodes, and 19 nodes were extracted for the selected experimental objects. The SIS interlocking process was disassembled and the interlocking relationship was converted into a causal relationship. Finally, there were 408 causal relationships and 90 protection relationships. The nodes and relationships were imported into the Neo4j graph database to store the structured data in the form of a graph.

[0068] Step 3: Integrate similarity aggregation method to calculate expert weights.

[0069] (1) Score the background of the expert panel based on the scoring criteria in Table 2 to obtain the background score of each person.

[0070] Table 2 Expert background scoring criteria

[0071]

[0072] Based on each person's background score, based on the formula Get the expert background weight k i , where g i is the total background score of the i-th expert.

[0073] (2) Comprehensively consider the consistency between the expert background score and the expert evaluation results, and combine the similarity aggregation method to determine the final weight K of the i-th expert evaluation interval i =αk i +βFA ij ; Among them, α and β use the most commonly used values ​​α=β=0.5, FA ij is the relative consistency of the evaluation interval of the i-th expert, Ad ij is the average distance between the evaluation intervals of the i-th expert and other experts, where d ij is the distance between the scoring interval of the i-th expert and the scoring interval of the j-th expert, which can be calculated using the Euclidean distance function.

[0074] Step 3: Extract the failure path of each deviation node based on the security knowledge graph and perform expert evaluation according to the rules formulated in this paper. Figure 3 The following are specific embodiments of the present invention.

[0075] (1) First, develop an evaluation comparison table, as shown in Table 3.

[0076] Table 3 Evaluation comparison table

[0077]

[0078] (2) Lead to Figure 3 There are 8 failure paths for the deviation of "the pressure difference between the two devices is too small", such as Figure 4 As shown, the ID numbers are 1-8. For the i-th path, the j-th expert gives the score interval number ([S Ij1 ,S Ij2 ],[O Ij1 ,O Ij2 ],[D Ij1 ,D Ij2 ]), and determine the number of protection layers for each path and the number of propagation nodes to the deviation node. The obtained evaluation table is used as the initial data for subsequent rule calculation. Table 4 is an example of the evaluation table.

[0079] Table 4 Example of expert interval evaluation table

[0080]

[0081] Based on the expert interval evaluation, the final evaluation index is calculated based on the rules, and the evaluation results are as follows: Figure 5 (1) The absolute size of the evaluation value is meaningless, only the relative size is meaningful. The risk threshold of different systems is different. After evaluation by the expert group, the risk threshold of the experimental object selected by the present invention is 124.5. Figure 4 It can be seen that the risk value of the seventh path that caused the deviation is higher. After calculation, it is found that adding an independent protection layer in this path can meet the security requirements. The system risk assessment results after the addition are as follows: Figure 5 (2) According to the evaluation of the expert group, the risk ranking obtained by this method is basically consistent with the actual situation of the system and has a high degree of credibility.

[0082] like Figure 6 As shown in the figure, the HAZOP analysis method is used to generate the system's HAZOP analysis table. By analyzing the device operating conditions, parameter deviations and the resulting dangerous consequences node by node, a structured hazard list is formed. This module ensures that the system's potential risks are systematically identified and provides data support for subsequent risk assessments.

[0083] The knowledge graph construction module is based on the HAZOP analysis table and system process operating procedures to graphically express the risk nodes, process parameters, protection layers and their relationships in the system. By integrating system process knowledge and equipment operation experience, a safety knowledge graph for petroleum refining units is constructed. The graph can intuitively display the relationship between risk propagation paths and protection layers, which helps to quickly locate high-risk nodes and influencing factors.

[0084] The weight coefficient determination module conducts quantitative analysis on the personal background (such as experience, professional field) of the experts involved in the assessment and the factors affecting the assessment process to determine the expert weight coefficient. Through comprehensive weight allocation, an objective and fair risk assessment weight system is formed to assist in the formulation of subsequent quantitative risk assessment rules. The rationality of the weight coefficient directly affects the accuracy and scientificity of the assessment results.

[0085] The risk assessment module improves the traditional risk priority number (RPN) calculation method based on the structural characteristics of the safety knowledge graph, taking into full account the propagation characteristics of risks in different paths and the reduction effect of the protection layer. Through rule supplementation and correction, a set of quantitative risk assessment rules suitable for the protection layer of petroleum refining units is formed. This rule quantifies the risk propagation path, protection layer effectiveness and risk severity, and realizes an accurate assessment of system risks.

[0086] The system conducts quantitative risk analysis on oil refining units based on the constructed quantitative risk assessment rules. It automatically calculates the risk priority number (RPN), evaluates the risk level of each node, and determines the final risk level in combination with the reduction factor of the protection layer. The system can intuitively display the effects of high-risk nodes, transmission paths, and protection layers, helping managers to formulate precise safety prevention and control measures.

[0087] The computer device provided by the present invention includes a memory and a processor, and the memory stores a computer program for executing a quantitative risk assessment method for a protective layer of a petroleum refining device. After the program is executed by the processor, steps such as hazard analysis, knowledge graph construction, weight coefficient determination, and risk assessment are completed. In addition, the information data processing terminal integrates this system, which can be quickly deployed and applied in actual industrial scenarios, thereby improving the safety management and risk control level of petroleum refining devices.

[0088] 1. Specific application fields or related products of the present invention.

[0089] The application fields of the present invention cover petrochemical, chemical, energy and other industries. By constructing a safety knowledge graph, enterprises can achieve the transformation from traditional manual management to intelligent management; the evaluation results based on the knowledge graph and risk propagation model can be used as decision support to help enterprise management formulate more scientific safety strategies. The modular design of the technical solution of the present invention enables it to adapt to different devices and production processes, and expand to other fields involving the production, storage and transportation of hazardous chemicals.

[0090] 2. Relevant evidence of the technical effects obtained by the embodiments of the present invention.

[0091] (1) In practical applications, the quantitative risk assessment method of the present invention is compared with the traditional assessment method, which can show the difference between the two in terms of risk identification and assessment accuracy. For example, by introducing the security knowledge graph, some potential risks that are missed by traditional methods can be identified, so that preventive measures can be taken in advance, and the increase in security protection costs caused by assessment bias can also be reduced.

[0092] (2) By comparing historical accidents with the results of the evaluation of the present invention, it can be proved that the present method can identify potential high-risk areas in advance during the evaluation process. For example, in some refining and chemical plants, by identifying potential loopholes in path risk propagation in advance, major safety accidents caused by equipment failures were successfully avoided. Many industry experts participated in the implementation of the present invention. During the implementation process, they gave positive feedback on the visualization effect of the safety knowledge graph and the personalized adjustment of expert weights, and believed that this evaluation method can effectively improve the accuracy of risk prediction and help optimize production operation processes.

[0093] Example 1: Risk assessment of protective layer applied to petroleum cracking unit

[0094] An oil company has a catalytic cracking unit, which includes a reaction regeneration unit, a fractionation unit, and an absorption stabilization unit. Due to long-term operation, the unit has the risk of equipment aging and operation errors, and a protective layer risk assessment is required to prevent safety accidents caused by unit failure.

[0095] 1.HAZOP analysis (step S1):

[0096] The evaluation target was determined to be the catalytic cracking unit, and the analysis scope covered the reaction regeneration, fractionation, and absorption stabilization units.

[0097] Nodes are divided according to P&IDs, for example, the reactor temperature is too high (deviation) due to a heater failure or a stuck control valve.

[0098] Through risk matrix assessment, deviations with high risk levels (such as reactor overheating leading to equipment explosion) are determined.

[0099] 2. Build a knowledge graph (S2 step):

[0100] The HAZOP analysis results are converted into "entity-relationship-entity" triples to construct the risk knowledge graph of the device.

[0101] For example, the entities include "reactor overtemperature" and "control valve failure", and the relationship is "cause-effect".

[0102] 3. Expert weight and consensus calculation (step S3):

[0103] The experts’ work experience and educational background are comprehensively considered, assigned different weights, and the consistency is calculated through the scoring range to ensure the objectivity of the expert evaluation results.

[0104] 4. Path risk sorting and optimization (S4 step):

[0105] For paths with higher RPN values ​​(such as heating furnace failure-reactor overheating-equipment explosion), protection measures such as adding automatic alarm systems and multiple redundant controls for heating furnaces are proposed.

[0106] Through this evaluation method, the company optimized the protective layer design of the catalytic cracking unit, effectively reduced the risk probability of reactor overheating, and improved the safety of the overall unit.

[0107] Example 2: Optimization of protective layer design applied to natural gas processing plants

[0108] The main units of a natural gas processing plant include desulfurization, dehydration and fractionation units. Due to the high concentration of hydrogen sulfide in the gas, there is a high risk of personnel poisoning after leakage, and a comprehensive assessment of the protection layer of the device is required.

[0109] 1.HAZOP analysis (step S1):

[0110] The assessment target was determined to be the natural gas desulfurization unit, and the analysis scope included the desulfurization tower, regeneration tower and gas transmission pipeline.

[0111] Analyze nodes, such as the desulfurization tower gas flow is too low (deviation), the direct cause is pipeline blockage or valve closure.

[0112] After risk matrix assessment, scrubber gas leakage was determined to be a high-risk event.

[0113] 2. Build a knowledge graph (S2 step):

[0114] The HAZOP analysis results of the desulfurization tower are converted into a knowledge graph, with nodes including "abnormal gas flow" and "valve closure" and the relationship being "cause-effect".

[0115] Building a graph database facilitates querying the propagation path and consequences of gas leaks.

[0116] 3. Expert weight and consistency evaluation (step S3):

[0117] Invite equipment engineers and safety experts to participate in the assessment, assign weights based on the consistency of the scores, and quantify the risks of gas leakage paths based on expert opinions.

[0118] 4. Path risk sorting and optimization (S4 step):

[0119] Identify the risk paths with the highest RPN values, such as “valve closed-insufficient gas flow-hydrogen sulfide leakage”.

[0120] Specific suggestions are put forward for adding real-time gas detection alarms, emergency shut-off devices and personnel evacuation channels.

[0121] Through this assessment method, the natural gas processing plant has added protection measures for hydrogen sulfide leaks to ensure that once an abnormality occurs, an alarm can be issued in time and emergency actions can be taken, effectively reducing environmental and personnel safety risks.

[0122] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0123] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A quantitative risk assessment method for the protective layer of a petroleum refining device, characterized in that: include: S1, conduct hazard analysis on the system and form a system HAZOP analysis table; S2, based on the system HAZOP analysis table, integrates system process knowledge such as operating procedures to build a system safety knowledge graph; S3, comprehensively consider the impact of the expert's personal background and the evaluation process on the expert's weight, determine the weight coefficient, and assist in the formulation of quantitative risk assessment rules; S4, according to the structural characteristics of the security knowledge graph, taking into account the impact of path risk propagation and the reduction effect of the protection layer on the impact, the RPN calculation method is supplemented with rules to obtain quantitative risk assessment rules, and the system risk assessment is completed based on the rules.

2. The quantitative risk assessment method for the protective layer of a petroleum refining device according to claim 1, characterized in that: The step S1 comprises: (1.1) Define the analysis objectives and scope of petroleum refining units, including reaction regeneration units, fractionation units, and absorption stabilization units, form an expert team, and identify potential risks through HAZOP analysis; (1.2) Collect relevant information such as process flow charts, process equipment diagrams, etc., determine the record table format and plan the analysis schedule; (1.3) Divide nodes through P&IDs, analyze node deviations in the form of "parameters + guide words", identify the direct causes of deviations, including equipment failure, instrument failure and operational errors, and use risk matrices to evaluate the semi-quantitative risk level of accident consequences; (1.4) For high-risk deviations, propose corresponding protective measures to detect deviations in a timely manner or mitigate the consequences of failure.

3. The quantitative risk assessment method for the protective layer of a petroleum refining device according to claim 1, characterized in that: The step S2 comprises: (2.1) Analyze the elements in the HAZOP analysis table, design the safety knowledge graph ontology, including classes, attributes, relationships and rules, and build a network topology diagram based on the assessment requirements; (2.2) Convert HAZOP information into "entity-relationship-entity" triple data, fill the knowledge graph through manual extraction, and use the Neo4j graph database to store structured data, where entities are stored as nodes and relationships are stored as edges; (2.3) Use graph query language to quickly retrieve HAZOP information for failure path extraction and subsequent assessment.

4. The quantitative risk assessment method for the protective layer of a petroleum refining device according to claim 1, characterized in that: The step S3 comprises: (3.1) Comprehensively consider the expert's position, years of work experience and education level, formulate the expert's personal background scoring standard and calculate the background weight; (3.2) Calculate the average distance and consistency of the expert evaluation intervals, and determine the final weight of the expert by combining the background weight and consistency; (3.3) The relative importance of the two weights is adjusted through parameters to achieve a dynamic balance between expert consensus and evaluation weights.

5. The quantitative risk assessment method for the protective layer of a petroleum refining device according to claim 1, characterized in that: The step S4 comprises: (4.1) An evaluation table is developed based on the HAZOP analysis results and expert opinions. Experts score the pathways based on severity (S), probability of occurrence (O), and detectability (D); (4.2) Perform logical accumulation calculation on the risk propagation in the path, and modify the occurrence probability (O) value by combining the protection layer and the alarm failure probability; (4.3) Calculate the RPN value based on the evaluation results, sort the nodes and paths, and make protection layer optimization suggestions for high-risk paths.

6. A quantitative risk assessment system for a protective layer of a petroleum refining device implementing the quantitative risk assessment method for a protective layer of a petroleum refining device as claimed in any one of claims 1 to 5, characterized in that: include: Hazard analysis module: performs hazard analysis on the system and forms a system HAZOP analysis table; The knowledge graph construction module builds the system's safety knowledge graph based on the system HAZOP analysis table and the system process knowledge such as operating procedures; The weight coefficient determination module comprehensively considers the impact of the expert's personal background and the evaluation process on the expert's weight, determines the weight coefficient, and assists in the formulation of quantitative risk assessment rules; The risk assessment module, based on the structural characteristics of the security knowledge graph, takes into account the impact of path risk propagation and the reduction effect of the protection layer on the impact, supplements the RPN calculation method with rules, obtains quantitative risk assessment rules, and completes system risk assessment based on the rules.

7. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the quantitative risk assessment method for the protective layer of a petroleum refining device as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to execute the steps of the quantitative risk assessment method for a protective layer of a petroleum refining device as claimed in any one of claims 1 to 5.

9. An information data processing terminal, comprising the quantitative risk assessment system for a protective layer of a petroleum refining device as claimed in claim 6.

Images