Dynamic Correlation Analysis Method for System Risk Identification of Petrochemical Product Production Lines

Abstract

The present invention relates to a dynamic correlation analysis method for system risk identification of petrochemical product production lines. The method uses a sliding window method to obtain observed values at relevant sampling moments, calculates the time lag between variables in the petrochemical product production process according to different correlation relationship calculation methods, and determines the correlation coefficient between variables. Secondly, according to the correlation coefficient between variables and the sliding window method, a minimum spanning tree model is dynamically and real-time drawn, and according to the change trend of the minimum spanning tree, that is, whether the correlation between variables changes, the system risk of the petrochemical production process is predicted, and according to the changes of different variables, the location where the risk source appears is identified. The present invention uses the change trend of the correlation-based minimum spanning tree to judge the system risk of the petrochemical production process, which can effectively improve the safety of the process and has high economic value.

Description

Technical Field

[0001] The present invention relates to the field of system risk analysis of petrochemical industry production lines, and specifically to a dynamic correlation analysis technical method for identifying risk sources of petrochemical product production line systems. Background Art

[0002] System risk was initially used in the fields of finance and business, mainly referring to the possibility of suffering economic losses due to uncontrollable fluctuations in the overall market price caused by multiple complex factors. After long-term research, a series of mature theories and methods for system risk assessment have gradually been formed. In recent years, the theories and methods of system risk assessment have also been widely applied to the safety assessment and fault prediction of complex equipment.

[0003] However, due to the large demand for petroleum products and the complexity of the production environment, the physical and chemical reaction mechanisms in the production process of petrochemical products are extremely complex. As a complex production process that requires monitoring of multiple different variables, it is difficult for conventional methods to accurately predict risks caused by equipment failures, production declines, or reaction failures. Therefore, there is an urgent need for an analysis and prediction method to reasonably predict the health status of petrochemical product production lines, minimize the occurrence probability of accidents, ensure the personal safety of operators, and prevent large-scale property losses. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a dynamic correlation analysis method for identifying risk sources of petrochemical product production line systems, which solves the problems of low accuracy and poor real-time performance in system risk prediction during the petrochemical production process.

[0005] The present invention utilizes the theories and methods of system risk assessment, and through objective analysis of data, realizes system risk prediction and risk source identification of petrochemical product production lines, quantitatively evaluates the impact on the safety of petrochemical product production lines when faults occur or performance declines, thereby realizing system risk prediction and risk source identification of petrochemical product production processes based on the dynamic correlation analysis method.

[0006] During the production process of petrochemical products, the correlation between certain variables is not real-time and one-to-one, and there may be a certain time lag in the mutual influence between different variables. Therefore, when calculating the correlation between variables, the time lag of process variables needs to be considered. The Pearson correlation calculation method mainly targets variables with a linear correlation relationship. The Spearman correlation coefficient analysis is based on the rank correlation between variables, and the maximum information coefficient is used to measure the degree of association between two variables, the strength of linear or non-linear. Different correlation coefficient calculation methods focus on different directions, and the variable characteristics in the petrochemical product production process are diverse. Therefore, when analyzing the correlation between variables, the existing connections between variables need to be accurately considered from different angles.

[0007] The technical solution adopted by the present invention to achieve the above object is: a dynamic correlation analysis method for risk identification of petrochemical product production line systems, including the following steps:

[0008] Correlation analysis: Obtain the sampling observation values of the petrochemical product production line, use the sliding window method to select the sampling observation values for a period of time, calculate various correlation coefficients between variables in the production process of a certain petrochemical product during this period, and draw various correlation curves according to the sliding time unit of the sliding window; Based on the sampling moment with the largest correlation in various correlation curves, determine the time lag between variables; According to the determined time lag between variables, use the calculation of correlation coefficients to obtain the maximum value of the correlation coefficient between variables, and use it as the final correlation coefficient between variables of a certain petrochemical product.

[0009] Risk identification: Use the correlation coefficient between variables obtained from the correlation analysis as the weight of the edge in the minimum spanning tree, aiming at the largest correlation, dynamically select the corresponding sampling observation values based on the sliding window method, and draw the minimum spanning relationship tree between variables of a certain petrochemical product; According to the change trend of the minimum spanning relationship tree, predict the probability of the occurrence of system risks in the petrochemical product production line, and identify the location of risk sources according to the change of the correlation relationship between different variables.

[0010] The sampling observation variables include at least one of raw material weight, auxiliary material weight, temperature, air volume, humidity, stirring speed, stirring intensity, and pressure.

[0011] The calculation of the correlation coefficient includes three methods: Pearson correlation coefficient, Spearman correlation coefficient, and maximum information coefficient.

[0012] The calculation of various correlation coefficients between variables in the production process of a certain petrochemical product during this period is as follows:

[0013] Pearson correlation coefficient:

[0014]

[0015] where ρ X,Y is the Pearson correlation coefficient between variable X and variable Y, and E(·) refers to the expectation of the variable;

[0016] Spearman correlation coefficient:

[0017] where ρ s is the Spearman coefficient between variables, is the rank difference of variables X and Y after reordering, and n refers to the number of variables;

[0018] Maximum information coefficient:

[0019]

[0020] Wherein, I[X:Y] refers to the mutual information between variable X and variable Y, p(X,Y) is the joint probability between variable X and variable Y, p(X) and p(Y) are the probability density functions of variable X and variable Y respectively, mic[X:Y] refers to the maximum information coefficient between variable X and variable Y, B is a preset parameter. Generally, B = N 0.6 , where N is the total amount of data, which is the sum of the sampling observations of all variables at all sampling times.

[0021] Determining the time delay size based on the sampling time with the largest correlation relationship among multiple correlation relationship curves includes the following steps:

[0022] Draw the correlation relationship curves for several hours in advance and delay according to the calculation methods of multiple correlation relationship coefficients;

[0023] If in a certain correlation relationship curve, the correlation relationship between variable X and variable Y is the largest at the delayed t moment, then it is determined that the time delay size between these two variables is t;

[0024] Obtain the time delay sizes of variable X and variable Y in three correlation relationship curves respectively, and comprehensively consider according to expert knowledge to finally determine the correlation time delay size between the two variables.

[0025] Dynamically select the corresponding sampling observations based on the sliding window method and draw the minimum spanning relationship tree between variables of a certain petrochemical product as follows:

[0026] In the form of a sliding window, select the sampling observations for a period of time, calculate the correlation relationship between any variables within this period of time, and dynamically draw the minimum spanning tree of variables of a certain petrochemical product.

[0027] Predicting the probability of the occurrence of system risks in the petrochemical product production line system according to the change trend of the minimum spanning relationship tree, and identifying the location of the risk sources according to the change of the correlation relationship between different variables includes the following steps:

[0028] When the minimum spanning tree of a certain petrochemical product variable shows a change that is considered to be intermediate aggregation over time, the correlation between variables becomes stronger, and the sum of the weights of the edges in the minimum spanning tree increases, it is determined that a systematic risk may occur in the petrochemical product production process at this time, and timely investigation is required; the location of the risk source is at the variable corresponding to the change in the correlation relationship in the minimum spanning tree.

[0029] According to the petrochemical product variable and the risk source location corresponding to the occurrence of the systematic risk, control the reaction process of the petrochemical product production line by adjusting the parameters of the petrochemical product variable of the petrochemical product production line to avoid the occurrence of systematic risks.

[0030] A dynamic correlation analysis device for system risk identification of petrochemical product production lines, comprising a memory and a processor; the memory is used for storing a computer program; the processor is used for implementing a dynamic correlation analysis method for system risk identification of petrochemical product production lines when executing the computer program.

[0031] A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, a dynamic correlation analysis method for system risk identification of petrochemical product production lines is implemented.

[0032] The present invention has the following beneficial effects and advantages:

[0033] Based on the complex environment and safety requirements of petrochemical production, with the petrochemical product production line as the research object, the present invention predicts system risks and locates risk sources based on the correlation analysis method, qualitatively analyzes the evolution trends of various types of risks caused by different factors, proposes evaluation methods and theories, and analyzes the impact of correlation changes on system risks. Specifically, the present invention uses a variety of correlation coefficient calculation methods to solve the problem of diverse variable characteristics in the petrochemical production process; uses the correlation relationship between variables and the sliding window method to accurately calculate the time lag between two variables; based on the expert knowledge that variables will show aggregated changes in the early stage of system risk occurrence, uses the method of the minimum spanning relationship tree to identify the correlation changes between variables, and accurately predicts system risks and identifies the positions of risk sources.

[0034] Considering the time lag relationship between variables, various characteristics of variables and the current situation of system risk prediction time for petrochemical product production lines, the present invention uses the correlation relationship between variables, the sliding window and the minimum relationship generation tree method to finally establish a high-risk event warning model for petrochemical product production lines under complex production conditions, realizing dynamic and proactive risk prediction. Description of the Drawings

[0035] Figure 1 Time lag correlation trend diagram between the target (risk source) subsystem and influencing factors;

[0036] Figure 2a Schematic diagram of the change of the minimum spanning tree between variables at different sampling times Figure 1 ;

[0037] Figure 2b Second schematic diagram of the change of the minimum spanning tree between variables at different sampling times;

[0038] Figure 3 Method flow chart of the present invention;

[0039] Figure 4Schematic diagram of system component indicators or the change trend of the correlation degree between components with a high degree of correlation with the risks of high-risk systems. Detailed implementation manners

[0040] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0041] The present invention takes the production process of petrochemical products as the research object, and the correlation analysis between variables as the main research means, integrating analysis methods such as data mining, correlation analysis, and graph theory to predict system risks and identify risk sources according to the changes of variables. The specific process of the present invention is divided into two steps: aiming at the problems of complex petrochemical production process principles, different variable characteristics, and time lags in the mutual influence between variables, a system risk prediction method based on a sliding window and correlation analysis is proposed. The sliding window method is used to obtain the observed values at relevant sampling moments, and according to different correlation relationship calculation methods, the time lag size between variables in the petrochemical product production process is obtained, and the correlation coefficient between variables is determined; secondly, according to the correlation coefficient between variables and the sliding window method, a minimum spanning tree model is dynamically and real-time drawn, and according to the change trend of the minimum spanning tree, that is, whether the correlation between variables changes, the system risk of the petrochemical production process is predicted, and according to the changes of different variables, the position where the risk source appears is identified. The present invention will clarify the evolution law of system risks in the petrochemical product production process and provide valuable reference for the safe and stable production of petrochemical products.

[0042] As Figure 3 shown, a dynamic correlation analysis technical method for identifying system risk sources of a petrochemical product production line includes the following steps:

[0043] Step 1: Correlation analysis. The sliding window method is used to select the sampling observed values for a period of time, and the Pearson correlation coefficient, Spearman correlation coefficient, and maximum information coefficient of the measurement variables in the production process of a certain petrochemical product during this period are calculated respectively. According to the sliding time unit of the sliding window, three correlation relationship curves are drawn, and based on the sampling moment with the largest correlation relationship, the time lag size is comprehensively considered and determined. According to the determined variable time lag size, using the three correlation relationship calculation methods, the maximum value of the correlation coefficient between variables is obtained and used as the final correlation relationship coefficient between variables of a certain petrochemical product;

[0044] Step 2: Risk identification. According to the correlation coefficients between variables calculated in Step 1, using the generation principle of the minimum spanning tree, taking the correlation coefficients between variables as the weights of the edges in the minimum spanning tree, with the purpose of obtaining the maximum correlation, dynamically select corresponding sampling observation values based on the sliding window method, draw the minimum generation relationship tree between variables of a certain petrochemical product, and predict the probability of system risk occurrence according to the change trend of the minimum generation relationship tree, such as the change in the sum of the weights of all edges, and identify the location of risk sources according to the change in the correlation relationship between different variables.

[0045] The determination of time lag in correlation analysis is specifically as follows:

[0046] In the form of a sliding window, continuously select sampling observation values for a period of time (the value for a period of time can be based on expert knowledge, and for the petrochemical production process involved in this patent, it is taken as one hour), where the sliding time unit is one minute. Draw the correlation relationship curves of three different correlation relationship calculation methods for the entire sampling period, that is, the abscissa is the time lag size and the ordinate is the corresponding correlation coefficient. Figure 1 It is a schematic diagram.

[0047] Figure 1 What is drawn in is the schematic diagram of the change of time lag and correlation coefficient of a certain correlation coefficient calculation method. From Figure 1 it can be seen that at the 350th sampling moment of these two variables, the correlation relationship coefficient obtains the maximum value, and it can be approximately regarded that the time lag relationship between these two variables is a difference of 350 minutes.

[0048] The three different correlation relationship calculation methods are: Pearson correlation coefficient, Spearman correlation coefficient, and maximum information coefficient (MIC) correlation relationship analysis method. Among them, the Pearson correlation coefficient is a statistic reflecting the similarity degree of two variables, and the specific calculation formula is as shown in (1).

[0049]

[0050] Among them, ρ X,Y is the Pearson correlation coefficient between variable X and variable Y, and E(·) refers to the expectation of the variable.

[0051] Spearman can be used to measure the strength of the connection between variables, and the specific calculation formula is:

[0052]

[0053] Among them, ρ s is the Spearman coefficient between variables, is the rank difference of variables X and Y after reordering, and n refers to the number of variables.

[0054] The maximum information coefficient (MIC) can be used to measure the degree of association between two variables X and Y, and its specific calculation formula is:

[0055]

[0056] Among them, I[X:Y] refers to the mutual information between variable X and variable Y, p(X,Y) is the joint probability between variable X and variable Y, p(X) and p(Y) are the probability density functions of variable X and variable Y respectively, mic[X:Y] refers to the maximum information coefficient between variable X and variable Y, B is a preset parameter. Generally, B = n 0.6 , where n is the total amount of data.

[0057] Using the calculation methods of three kinds of association relationships, draw the association relationship curves three hours in advance and three hours later. If in a certain association relationship curve, the correlation between variable X and variable Y is the largest at the delayed t moment, then the time lag between these two variables is determined to be t. Respectively obtain the time lag sizes of variable X and variable Y in the three association relationship curves, and comprehensively consider according to expert knowledge to finally determine the association time lag size between the two variables.

[0058] The determination of the correlation coefficient in association analysis is specifically as follows:

[0059] According to the calculated time lag size, calculate the correlation coefficients between the measurement variables of a certain petrochemical product under the three methods, and select the largest value among them as the correlation coefficient between the two variables.

[0060] The specific method of drawing the minimum spanning tree in risk identification is as follows:

[0061] According to the correlation coefficients between the variables of a certain petrochemical product calculated, set the weights of the edges between different variables as their correlation coefficients, and draw the minimum spanning tree according to the principle of the largest weight.

[0062] The specific method of drawing risk identification in risk identification is as follows:

[0063] In the form of a sliding window with a sliding time unit of 30 minutes, a period of sampling observations is selected. In the present invention, a period is one hour. Calculate the correlation relationship between variables within this one hour, dynamically draw the minimum spanning tree of a petrochemical product variable, and judge whether there is a system risk in the production line for producing this petrochemical product according to the change trend of the minimum spanning tree, so as to realize the dynamic correlation analysis of the identification of system risk sources in the petrochemical product production line. Specifically, it can be stated that according to existing expert knowledge, when a petrochemical product has a system risk, the variables will show the same change and the correlation relationship will increase. Therefore, if the minimum spanning tree of a petrochemical product variable shows an intermediate clustering change over time, the correlation between variables becomes stronger, and the sum of the weights of the edges in the minimum spanning tree increases, it can be determined that there may be a systematic risk in the production process of the petrochemical product at this time, and timely investigation is required. The schematic diagram of the identification of system risk in the petrochemical production line based on the minimum spanning tree is as Figure 2a 、 Figure 2b shown. Compared with Figure 2a , Figure 2b the variables in the minimum spanning tree are more clustered, the correlation relationship between variables is enhanced, and the sum of the weights of the edges becomes larger. Therefore, it can be determined that there may be a system risk in the system. In addition, since the variables with relatively large changes in the correlation relationship are PVO, PC, and PP, it can be concluded that the risk in the production process may be more related to these three variables. Therefore, when excluding system risks, the part associated with these three variables can be considered first, so as to realize the identification of system risk sources.

[0064] The present invention calculates the correlation relationship between variables through three methods, which can effectively explore the potential relationship between variables and eliminate the limitations of the calculation of a single correlation relationship algorithm. In addition, considering that the mutual influence between variables in the petrochemical production process is not real-time corresponding and there will be some lag, the sliding window method is used to calculate the correlation relationship between variables at different delay times, which can effectively improve the accuracy of the calculation of the correlation relationship between variables. The present invention uses the change trend of the correlation-based minimum spanning tree to judge the system risk in the petrochemical production process, which can effectively improve the safety of the process and has high economic value.

[0065] The present invention first uses the detection data X∈R collected by sensors and controllers n×m, the detected data matrix X contains n sampling moments, and each sampling moment contains m variables, such as raw materials and auxiliary materials weight, temperature, air volume, humidity, stirring speed, stirring intensity, pressure, etc. Based on three methods, the correlation relationship between the collected variables in the production process of a petrochemical product is calculated, and the sliding window method is used to draw the correlation relationship curve with misaligned time delay. By synthesizing the correlation relationship result curves of the three methods, the time delay relationship between variables is determined; according to the determined time delay relationship between variables, the correlation relationship between variables is calculated based on three methods, and the maximum value among them is used as the final variable correlation relationship; using the correlation relationship between variables, based on the sliding window with 30 minutes as the time unit, the minimum spanning tree is dynamically and real-time drawn, where the weight of the edge in the minimum spanning tree is the correlation relationship between variables, and the drawing principle is based on obtaining the maximum weight; according to the change trend of the minimum spanning tree and the total weight of the edges, the system risk and the identification of risk sources in the production process of petrochemical products are predicted. The present invention utilizes the data collected in the petrochemical process and the correlation relationship between variables in the production process of petrochemical products, improving the accuracy and real-time performance of the system risk prediction of the petrochemical product production line. In addition, the system risk prediction and the risk source analysis are of great significance to the safe production of the petrochemical process.

[0066] Embodiment

[0067] Petrochemical industry mainly includes the following three major processes: basic organic chemical production process, organic chemical production process, and polymer chemical production process. The basic organic chemical production process takes petroleum and natural gas as the starting raw materials, and through refining and processing, basic organic raw materials such as three olefins (ethylene, propylene, butadiene), three benzenes (benzene, toluene, xylene), acetylene, and naphthalene are obtained; the organic chemical production process is based on "three olefins, three benzenes, acetylene, and naphthalene", and through various synthesis steps, organic raw materials such as alcohols, aldehydes, ketones, acids, esters, and ethers are obtained; the polymer chemical production process is based on organic raw materials, and through various polymerization and condensation steps, final products such as synthetic fibers, synthetic resins, and synthetic rubbers are obtained. If the systematic high risk in the production process of petrochemical products is not controlled, it is very easy to cause catastrophic consequences. For example, if a sub-device fails in the production process of a single petrochemical product, it may directly lead to a decline in product quality. If this petrochemical product is the raw material for the production of other petrochemical products, it may lead to quality problems in all petrochemical products. The core of traditional risk analysis lies in the prevention and control of the risk source itself, and the focus is on passively reducing the failure risk of a single petrochemical product production process. In fact, before the abnormal change of the risk source that causes the risk, there are often abnormal fluctuations in itself or other related variables, such as Figure 4As shown. Therefore, based on the established system risk assessment system and key influencing factors, the present invention fully considers the correlation characteristics between the risk sources and other potential relevant factors, and through the moving time window technology, conducts time-delay correlation analysis on the detection data of sensors and controllers of the targeted (risk source) subsystem (production equipment), components (petrochemical products) and other subsystems / equipment (key influencing factors), searches for correlation characteristics (such as time-delay, periodicity or monotonicity, etc.), identifies the response pattern between variables before the abnormal change of the risk source, and establishes a dynamic early warning response model with time attributes. And at the same time when the risk is predicted to occur, the corresponding operators are reminded in time to make appropriate adjustments to production parameters, prevent the occurrence of abnormal situations, restore stable production at the fastest speed, and reduce the economic loss and safety risk coefficient. Thus, in the production process of petrochemical products, based on the measured variables collected in the petrochemical production process, this patent uses the analysis method of the correlation relationship between variables to obtain the time-delay relationship between variables, predicts the risk degree of the petrochemical product production line, and improves the safety and economic value of production.

Claims

1. A dynamic correlation analysis method for system risk identification of petrochemical product production lines, characterized in that: Correlation analysis: Obtain the sampling observation values of the petrochemical product production line, use the sliding window method to select the sampling observation values for a period of time, calculate various correlation coefficients between variables in the production process of a certain petrochemical product during this period, and draw various correlation curves according to the sliding time unit of the sliding window; Based on the sampling moment with the largest correlation in various correlation curves, determine the time lag between variables; according to the determined time lag between variables, use the calculation of correlation coefficients to obtain the maximum value of the correlation coefficient between variables, and use it as the final correlation coefficient between variables of a certain petrochemical product; Risk identification: Use the correlation coefficient between variables obtained from the correlation analysis as the weight of the edge in the minimum spanning tree, aiming at the largest correlation, dynamically select the corresponding sampling observation values based on the sliding window method, and draw the minimum spanning relationship tree between variables of a certain petrochemical product; According to the change trend of the minimum spanning relationship tree, predict the probability of the occurrence of system risks in the petrochemical product production line system, and identify the location of risk sources according to the change of the correlation between different variables; The dynamic selection of the corresponding sampling observation values based on the sliding window method to draw the minimum spanning relationship tree between variables of a certain petrochemical product is specifically as follows: In the form of a sliding window, select the sampling observation values for a period of time, calculate the correlation between any variables during this period, and dynamically draw the minimum spanning tree of variables of a certain petrochemical product; The prediction of the probability of the occurrence of system risks in the petrochemical product production line system according to the change trend of the minimum spanning relationship tree, and the identification of the location of risk sources according to the change of the correlation between different variables include the following steps: When the minimum spanning tree of variables of a certain petrochemical product shows a change considered to be intermediate aggregation over time, the correlation between variables becomes stronger, and the total weight of the edges in the minimum spanning tree increases, it is determined that a systematic risk may occur in the petrochemical product production process at this time, and timely investigation is required; the location of the risk source is at the variable corresponding to the change in the correlation in the minimum spanning tree; According to the petrochemical product variables and the location of the risk source corresponding to the occurrence of the systematic risk, control the reaction process of the petrochemical product production line by adjusting the parameters of the petrochemical product variables of the petrochemical product production line to avoid the occurrence of systematic risks.

2. The dynamic correlation analysis method for risk identification of petrochemical product production line system according to claim 1, characterized in that The sampling observation values include at least one of raw material weight, auxiliary material weight, temperature, air volume, humidity, stirring speed, stirring intensity, and pressure.

3. The dynamic correlation analysis method for risk identification of petrochemical product production line system according to claim 1, wherein, The calculation of the correlation coefficient includes three methods: Pearson correlation coefficient, Spearman correlation coefficient, and maximum information coefficient.

4. The dynamic correlation analysis method for risk identification of petrochemical product production line system according to claim 1, wherein, The calculation of various correlation coefficients between variables in the production process of a certain petrochemical product during this period is specifically as follows: Pearson correlation coefficient: where ρ X,Y is the Pearson correlation coefficient between variable X and variable Y, and E(·) represents the expectation of the variable; Spearman correlation coefficient: where ρ s is the Spearman coefficient between variables, is the rank difference of variables X and Y after reordering, and n refers to the number of variables; Maximum information coefficient: Among them, I[X:Y] refers to the mutual information between variable X and variable Y, p(X,Y) is the joint probability between variable X and variable Y, p(X) and p(Y) are the probability density functions of variable X and variable Y respectively, mic[X:Y] refers to the maximum information coefficient between variable X and variable Y, B is a preset parameter, B = N 0.6 , where N is the total amount of data, which is the sum of the sampling observations of all variables at all sampling times.

5. The dynamic correlation analysis method for risk identification of petrochemical product production line system according to claim 1, characterized in that, The determination of the time lag between variables based on the sampling moment with the largest correlation in various correlation curves includes the following steps: According to the calculation methods of various correlation coefficients, draw the correlation curves several hours in advance and several hours later; If, in a certain correlation relationship curve, the correlation between variable X and variable Y is the greatest at the delay time t, then the time lag between these two variables is determined to be t; Obtain the time lag sizes of variable X and variable Y in three correlation relationship curves respectively, and finally determine the correlation time lag size between the two variables.

6. A dynamic correlation analysis device for petrochemical product production line system risk identification, comprising a memory and a processor; the memory is used to store a computer program; the processor is used to implement the dynamic correlation analysis method for petrochemical product production line system risk identification according to any one of claims 1-5 when executing the computer program.

7. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by a processor, the dynamic correlation analysis method for petrochemical product production line system risk identification according to any one of claims 1-5 is implemented.

Images