SEM-based chemical enterprise environmental influence risk evaluation method

Abstract

The invention belongs to the technical field of evaluation of the influence degree of chemical enterprises on the surrounding environment, and particularly relates to an SEM-based chemical enterprise environmental influence risk evaluation method, which comprises the following steps: constructing chemical enterprise environmental influence risk evaluation indexes; variables are set, and a chemical enterprise environment influence risk evaluation value structure equation mechanism model is constructed; evaluating the advantages and disadvantages of the mechanism model; and correcting the mechanism model and evaluating the environmental influence risk of the chemical enterprise by using the mechanism model. When the chemical enterprise environment influence risk is evaluated, the method has the advantages of processing complex causal relationships, analyzing latent variables, comprehensively evaluating the model adaptation degree, combining path analysis and factor analysis, accurately estimating parameters, carrying out statistical verification based on a large amount of data and the like.

Description

Technical Field

[0001] The invention belongs to the technical field of evaluation of the degree of impact of chemical enterprises on the surrounding environment, and specifically relates to a SEM-based environmental impact risk evaluation method for chemical enterprises. Background Art

[0002] In the normal production and operation activities of chemical enterprises, many drugs that are harmful to the environment or people are involved, which may cause environmental pollution and damage to human health. At present, in the existing technology, the risk assessment method for the impact of chemical enterprises on the environment is the analytic hierarchy process (AHP), which has the following problems:

[0003] (1) Subjectivity and judgment bias

[0004] Subjectivity of judgment matrix: The AHP method relies on the subjective judgment of decision makers or expert teams when constructing judgment matrices. Even if multiple experts are involved, individual experience, knowledge level, and bias may lead to different judgments on the same issue, thus affecting the objectivity of the overall result.

[0005] Inconsistency problem: Although AHP introduces a consistency verification mechanism (CR value), in actual operation, the consistency of the judgment matrix is ​​often difficult to fully guarantee, which can easily lead to irrationality in the decision-making process.

[0006] (2) Complexity and hierarchy

[0007] Difficulty in constructing a hierarchical structure: When environmental risk issues involve many factors, it may be very complicated to establish a reasonable hierarchical structure. Especially for large and complex industrial enterprises, the hierarchical model may be too large, making it difficult to understand and analyze.

[0008] Risk of information loss: In the hierarchical process, some details and relationships may be simplified or ignored, resulting in the loss of important information, which may affect subsequent decision-making.

[0009] (3) Limitations of consistency checking

[0010] CR value limit: The AHP method stipulates that the CR value should be less than 0.10. However, in practical applications, it is not always feasible to achieve this consistency standard. Once the CR value is unqualified, the judgment matrix needs to be corrected, which may lead to unstable and unreliable results.

[0011] Error amplification: In a multi-level structure, low consistency may cause error accumulation in the entire model, reducing the reliability of the final evaluation results.

[0012] (4) Limitations of indicator selection

[0013] Insufficient comprehensiveness of indicators: Although multiple indicators are selected in the AHP method, whether these indicators fully represent all aspects of environmental risks of industrial enterprises still needs further discussion. Industry characteristics and environmental changes may result in the selected indicators not covering all potential risks.

[0014] Lack of dynamic adjustment: Environmental risk is a dynamically changing system, while the indicators used in AHP are often static and lack the flexibility to adapt to environmental changes. This may make the indicators lose their effectiveness in long-term applications.

[0015] (5) Computational and processing complexity

[0016] Computational burden: For large decision-making problems, AHP requires a large number of matrix operations and consistency checks, which will create a certain computational burden, especially when dealing with multiple levels and multiple indicators, increasing the consumption of time and resources.

[0017] Numerical sensitivity: For some small changes, especially in the judgment matrix, the numerical sensitivity is high, which may lead to significant changes in weight distribution, thus affecting the final decision result. Summary of the invention

[0018] In order to solve the above technical problems, the present invention provides a SEM-based environmental impact risk assessment method for chemical enterprises.

[0019] The present invention is implemented by providing a method for environmental impact risk assessment of chemical enterprises based on SEM, comprising the following steps:

[0020] Step 1: Construct the environmental impact risk assessment indicators of chemical enterprises, including several first-level indicators η and several second-level indicators Y under each first-level indicator η;

[0021] Step 2: Set the first-level indicator η as the first-order internal latent variable, the error term of the first-order internal latent variable is ζ, set the second-level indicator Y as the internal observed variable, set the error term of the internal observed variable as ε, the load factor between the internal observed variable and the first-order internal latent variable is λ, the environmental impact risk assessment value of the chemical enterprise is the second-order internal latent variable ξ, the load factor between the second-order internal latent variable and the first-order internal latent variable is γ, and construct the structural equation mechanism model of the environmental impact risk assessment value of the chemical enterprise:

[0022] ξ=γ*η+ζ (1)

[0023] η=λ*Y+ε (2);

[0024] Step 3: The chi-square test, chi-square freedom ratio, root mean square error of approximation, fit index, adjusted fit index, regular fit index, irregular fit index, comparative fit index and incremental fit index were used to evaluate the quality of the mechanism model obtained in step 2;

[0025] Step 4: According to the evaluation results of step 3, the mechanism model is modified, and the environmental impact risk of chemical enterprises is evaluated using the modified mechanism model.

[0026] Preferably, in step 1, the first-level indicators include chemicals and production methods η1, environmental receptors η2, and safety management methods η3. The second-level indicators under chemicals and production methods η1 include inherent hazards of chemicals Y1, chemical reserves Y2, and production methods Y3. The second-level indicators under environmental receptors η2 include surrounding population density Y4, natural disasters Y5, and environmental risk receptors Y6. The second-level indicators under safety management methods η3 include equipment safety management methods Y7, establishment of environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y10. 10 , Production site selection Y 11 and management experience Y 12 .

[0027] Further preferably, in the step 2, the primary indicator chemicals and production methods η1, environmental receptors η2, and safety management methods η3 are set as first-order internal potential variables, and the error terms of the three first-order internal potential variables are ζ1, ζ2, and ζ3, respectively; the secondary indicators of inherent hazards of chemicals Y1, chemical reserves Y2, production methods Y3, surrounding population density Y4, natural disasters Y5, environmental risk receptors Y6, equipment safety management methods Y7, establishment of environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y1 are set as ... environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y1 are set as secondary indicators of inherent hazards of chemicals Y1, chemical reserves Y2, production methods Y3, surrounding population density Y4, natural disasters Y5 10 , Production site selection Y 11 and management experience Y 12 Assuming them as internal observed variables, the error terms of the twelve internal observed variables are ε1, ε2, ε3, ε4, ε5, ε6, ε7, ε8, ε9, ε 10 , ε 11 , ε 12 The load factors between the internal observed variables and the first-order internal latent variables are λ 11 , 12 , 13 , 24 , 25 , 26 , 37 , 38 , 39 , 310 , 311 , 312, the environmental impact risk assessment value of chemical enterprises is the second-order internal latent variable ξ1, and the loading factor between the second-order internal latent variable and the first-order internal latent variable is γ 11 , γ 21 , γ 31 .

[0028] Further preferably, in step 2, the structural equation mechanism model of the environmental impact risk assessment value of chemical enterprises is specifically:

[0029] ξ1=(γ 11 *η1+ζ1)+(γ 21 *η2+ζ2)+(γ 31 *η3+ζ3) (3)

[0030] η1=(λ 11 * Y1+ε1)+(λ 12 * Y2+ε2)+(λ 13 * Y3+ε3) (4)

[0031] η2=(λ 24 * Y4+ε4)+(λ 25 * Y5+ε5)+(λ 26 * Y6+ε6) (5)

[0032] η3=(λ 37 *Y7+ε7)+(λ 38 *Y8+ε8)+(λ 39 *Y9+ε9)+(λ 310 *Y 10 +ε 10 )+

[0033] (λ 311 *Y 11 +ε 11 )+(λ 312 *Y 12 +ε 12 )(6).

[0034] Further preferably, the step 3 specifically comprises:

[0035] Step 301: Evaluate the quality of the mechanism model through the chi-square test, according to the following formula:

[0036] T=(N-1)F min (7)

[0037] Among them, T represents the test value of model fit, which is regarded as the chi-square value, N is the number of samples, and F minIt indicates the minimum function estimate of the fitting function obtained by various parameter estimation methods; when the T value reaches a significant level, the mechanism model has a poor fit, and when the T value does not reach a significant level, the mechanism model has a good fit;

[0038] Step 302: Evaluate the quality of the mechanism model through the chi-square freedom ratio, according to the following formula:

[0039] X 2 / df(8)

[0040] Among them, df is the chi-square freedom ratio; the smaller the chi-square freedom ratio, the higher the fit of the mechanism model, and vice versa, the worse the fit of the mechanism model;

[0041] Step 303: Evaluate the quality of the mechanism model through the root mean square error of the approximation, according to the following formula:

[0042]

[0043] N is the number of samples. The smaller the RMSEA index is, the better the fit of the mechanism model is;

[0044] Step 304: The method of evaluating the quality of the mechanism model by the fitting index is based on the following formula:

[0045]

[0046] Among them, S represents the covariate of the sample, GFI is a ratio less than 1, the closer the GFI value is to 1, the higher the fit of the mechanism model is, and when the GFI value is smaller, the difference between the numerator and the denominator is larger, which means the fit of the mechanism model is lower;

[0047] Step 305: The method for evaluating the quality of the mechanism model by adjusting the fitting index is based on the following formula:

[0048]

[0049] Among them, t is the number of estimated parameters, DP is the number of observation data collected for environmental impact risk assessment of chemical enterprises; the AGFI value is between 0 and 1, the closer the value is to 1, the better the fit of the mechanism model is, and the closer it is to 0, the worse the fit is;

[0050] Step 306: The method of evaluating the quality of the mechanism model by using the regular fit index is based on the following formula:

[0051]

[0052] Among them, χ 2 indep represents an independent model, Indicates a comparative model; the closer the NFI value is to 1, the greater the improvement of the theoretical model over the baseline model; when the NFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model;

[0053] Step 307: The method of evaluating the quality of the mechanism model by using the informal fitting index is based on the following formula:

[0054]

[0055] The closer the NNFI value is to 1, the greater the improvement of the theoretical model over the baseline model. When the NNFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model.

[0056] Step 308: The method of evaluating the quality of the mechanism model by comparing the fitting index is based on the following formula:

[0057]

[0058] The closer the value of the CFI index is to 1, the more ideal the mechanism model is;

[0059] Step 309: The method of evaluating the quality of the mechanism model by the incremental fitting index is based on the following formula:

[0060]

[0061] IFI ranges from 0 to 1, and a larger value indicates a better fit of the mechanism model.

[0062] Compared with the prior art, the advantages of the present invention are:

[0063] (1) Dealing with complex causal relationships:

[0064] Multi-cause and multi-effect model: SEM can handle more complex causal relationships and can simultaneously analyze the relationships between multiple independent and dependent variables. This makes SEM particularly suitable for studying complex environmental risk systems and can reveal the deep connections and potential structures between variables.

[0065] (2) Latent variables and measurement model:

[0066] Observed variables and latent variables: SEM allows the use of observed variables to measure latent constructs (latent variables). These latent variables can be concepts that are difficult to observe directly (such as certain risk perceptions). Through this, SEM can more comprehensively capture the potential dynamics within complex systems.

[0067] (3) Overall fit evaluation:

[0068] Model fit index: SEM provides a series of model fit indexes (such as CFI, TLI, RMSEA and χ2 These indices can comprehensively evaluate whether the model as a whole matches the data, which helps researchers verify the rationality of the model.

[0069] (4) Processing multiple model relationships simultaneously:

[0070] Combination of path analysis and factor analysis: SEM can perform path analysis and factor analysis at the same time, so it can deal with multiple aspects such as measurement errors, structural relationships, causal chains, etc. in one comprehensive model.

[0071] (5) Accuracy of parameter estimation:

[0072] Complex model parameter estimation: SEM is capable of performing sophisticated parameter estimation, including estimation of direct effects, indirect effects, and total effects, which makes the results of environmental risk assessment more accurate and detailed.

[0073] (6) Data-driven analysis:

[0074] Data-based verification: The parameter estimation and hypothesis testing of the SEM model are based on statistical analysis of a large amount of data, which can provide highly accurate quantitative analysis results. This data-driven approach is particularly suitable for environmental risk assessments that require detailed and quantitative research conclusions.

[0075] The advantages of structural equation model (SEM) in environmental risk assessment are mainly reflected in the processing of complex causal relationships, analysis of latent variables, comprehensive model fitness evaluation, combination of path analysis and factor analysis, accurate parameter estimation, and statistical verification based on large amounts of data. These advantages give SEM a significant advantage in dealing with complex environmental risk systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 The invention provides a structural equation mechanism model for the environmental impact risk assessment value of chemical enterprises. DETAILED DESCRIPTION

[0077] 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 accompanying drawings and 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.

[0078] The present invention provides a method for evaluating the environmental impact risk of a chemical enterprise based on SEM, comprising the following steps:

[0079] Step 1: Construct the environmental impact risk assessment indicators of chemical enterprises, including several first-level indicators η and several second-level indicators Y under each first-level indicator η;

[0080] The first-level indicators include chemicals and production methods η1, environmental receptors η2, and safety management methods η3. The second-level indicators under chemicals and production methods η1 include inherent hazards of chemicals Y1, chemical reserves Y2, and production methods Y3. The second-level indicators under environmental receptors η2 include surrounding population density Y4, natural disasters Y5, and environmental risk receptors Y6. The second-level indicators under safety management methods η3 include equipment safety management methods Y7, establishment of environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y10. 10 , Production site selection Y 11 and management experience Y 12 .

[0081] Step 2: Set the primary indicator chemicals and production methods η1, environmental receptors η2, and safety management methods η3 as first-order internal latent variables. The error terms of the three first-order internal latent variables are ζ1, ζ2, and ζ3 respectively. Set the secondary indicators of inherent hazards of chemicals Y1, chemical reserves Y2, production methods Y3, surrounding population density Y4, natural disasters Y5, environmental risk receptors Y6, equipment safety management methods Y7, establishment of environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y1 as the first-order internal latent variables. 10 , Production site selection Y 11 and management experience Y 12 Assuming them as internal observed variables, the error terms of the twelve internal observed variables are ε1, ε2, ε3, ε4, ε5, ε6, ε7, ε8, ε9, ε 10 , ε 11 , ε 12 The load factors between the internal observed variables and the first-order internal latent variables are λ 11 , 12 , 13 , 24 , 25 , 26 , 37 , 38 , 39 , 310 , 311 , 312 , the environmental impact risk assessment value of chemical enterprises is the second-order internal latent variable ξ1, and the loading factor between the second-order internal latent variable and the first-order internal latent variable is γ 11 , γ 21 , γ 31 ,like Figure 1 , construct the structural equation mechanism model of environmental impact risk assessment value of chemical enterprises:

[0082] ξ=γ*η+ζ (1)

[0083] η=λ*Y+ε (2);

[0084] Specifically:

[0085] ξ1=(γ 11 *η1+ζ1)+(γ 21 *η2+ζ2)+(γ 31 *η3+ζ3) (3)

[0086] η1=(λ 11 * Y1+ε1)+(λ 12 * Y2+ε2)+(λ 13 * Y3+ε3) (4)

[0087] η2=(λ 24 * Y4+ε4)+(λ 25 * Y5+ε5)+(λ 26 * Y6+ε6) (5)

[0088] η3=(λ 37 *Y7+ε7)+(λ 38 *Y8+ε8)+(λ 39 *Y9+ε9)+(λ 310 *Y 10 +ε 10 )+

[0089] (λ 311 *Y 11 +ε 11 )+(λ 312 *Y 12 +ε 12 )(6).

[0090] Step 3: The chi-square test, chi-square freedom ratio, root mean square error of approximation, fit index, adjusted fit index, regular fit index, irregular fit index, comparative fit index and incremental fit index were used to evaluate the quality of the mechanism model obtained in step 2;

[0091] Step 3 specifically includes:

[0092] Step 301: In SEM analysis, the most commonly used model evaluation method is the chi-square test. In SEM, the chi-square value is a statistic converted from the fitting function, which reflects the difference between the derived matrix and the observed matrix of the SEM hypothesis model. The chi-square test is used to evaluate the quality of the mechanism model according to the following formula:

[0093] T=(N-1)F min (5)

[0094] Among them, T represents the test value of model fit, which is regarded as the chi-square value, N is the number of samples, and F minIt indicates the minimum function estimate of the fitting function obtained by various parameter estimation methods; when the T value reaches a significant level, it means that the null hypothesis is not established and the mechanism model has a poor fit; when the T value does not reach a significant level, it means that the null hypothesis is established and the mechanism model has a good fit;

[0095] Step 302: In the SEM analysis, a chi-square freedom ratio, called the normed chi-square value, can be calculated to compare the fit between models. The chi-square freedom ratio is used to evaluate the quality of the mechanism model according to the following formula:

[0096] X 2 / df(6)

[0097] Among them, df is the chi-square freedom ratio; the smaller the chi-square freedom ratio, the higher the fit of the mechanism model, and vice versa, the worse the fit of the mechanism model;

[0098] Step 303: Evaluate the quality of the mechanism model through the root mean square error of the approximation, according to the following formula:

[0099]

[0100] N is the number of samples. The smaller the RMSEA index is, the better the fit of the mechanism model is;

[0101] Step 304: GFI is the abbreviation of goodness-of-fit index, which is similar to the explained variance in regression analysis, indicating the proportion of variance and covariance of observed data that can be explained by the hypothesized model. The method of evaluating the quality of the mechanism model through the fit index is based on the following formula:

[0102]

[0103] Among them, S represents the covariate of the sample, GFI is a ratio less than 1, the closer the GFI value is to 1, the higher the fit of the mechanism model is, and when the GFI value is smaller, the difference between the numerator and the denominator is larger, which means the fit of the mechanism model is lower;

[0104] Step 305: AGFI is similar to the adjusted explained variance in regression analysis. When calculating the GFI coefficient, the model fit index is calculated after taking the degrees of freedom into consideration. When there are more parameters, the larger the AGFI index value, the more conducive it is to obtain an ideal fit conclusion. The method of evaluating the quality of the mechanism model through the adjusted fit index is based on the following formula:

[0105]

[0106] Among them, t is the number of estimated parameters, DP is the number of observation data collected for environmental impact risk assessment of chemical enterprises; the AGFI value is between 0 and 1, the closer the value is to 1, the better the fit of the mechanism model is, and the closer it is to 0, the worse the fit is;

[0107] Step 306: NFI is a relative index calculated using the nested model comparison principle, which reflects the degree of difference between the hypothesized model and an independent model without any covariation assumption between the observed variables. The method of evaluating the quality of the mechanism model through the formal fit index is based on the following formula:

[0108]

[0109] Among them, χ 2 indep represents an independent model, Indicates a comparative model; the closer the NFI value is to 1, the greater the improvement of the theoretical model over the baseline model; when the NFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model;

[0110] Step 307: NNFI is a relative index calculated using the nested model comparison principle, which reflects the degree of difference between the hypothesized model and an independent model without any covariation assumption between the observed variables. The method of evaluating the quality of the mechanism model by the informal fit index is based on the following formula:

[0111]

[0112] The closer the NNFI value is to 1, the greater the improvement of the theoretical model over the baseline model. When the NNFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model.

[0113] Step 308: The CFI index reflects the degree of difference between the hypothesized model and the independent model without any covariation relationship, and also takes into account the discreteness of the tested model and the central chi-square distribution. The method of evaluating the quality of the mechanism model by comparing the fit index is based on the following formula:

[0114]

[0115] The closer the value of the CFI index is to 1, the more ideal the mechanism model is;

[0116] Step 309: IFI is used to deal with the problem of NNFI fluctuation and the impact of sample size on NFI index. The method of evaluating the quality of the mechanism model through the incremental fit index is based on the following formula:

[0117]

[0118] IFI ranges from 0 to 1, and a larger value indicates a better fit of the mechanism model.

[0119] Step 4: According to the evaluation results of step 3, the mechanism model is modified, and the environmental impact risk of chemical enterprises is evaluated using the modified mechanism model.

[0120] In this embodiment, after correction, γ 11 , γ 21 , γ 31 , specific values ​​of ζ1, ζ2, ζ3, and the three indicators of inherent hazards of chemicals Y1, chemical reserves Y2, and production methods Y3 were revised into comprehensive indicators of production process hazards P1 and comprehensive indicators of pollutant emissions P2. The three indicators of surrounding population density Y4, natural disasters Y5, and environmental risk receptors Y6 were revised into the environmental quality function index of the receptor P3 and the comprehensive index of residents' harm P4. The three indicators of equipment safety management methods Y7, environmental emergency plan establishment Y8, environmental emergency disposal Y9, and major accidents Y1 were revised into comprehensive indicators of environmental quality function of the receptor P3 and comprehensive indicators of residents' harm P4. 10 , Production site selection Y 11 and management experience Y 12 The six indicators were revised into three indicators: environmental risk management system P5, accident emergency plan P6, and employee safety training effect P7, and the corresponding load factors λ and error terms ε of P1-P7 were obtained, specifically:

[0121] ξ=(0.14n1+0.98)+(0.42n2+0.82)+(0.96n3+0.08)

[0122] η1=(0.09P1+0.99)+(4.34P2+0.28)

[0123] η2=(0.85P3+0.27)+(0.93P4+0.14)

[0124] η3=(0.86P5+0.26)+(0.9P6+0.03)+(0.98P7+0.04)

[0125] So we get:

[0126] ξ=0.013p1+0.61p2+0.83p3+0.86p4+0.94p5+0.36p6+0.39p7+2.55

[0127] Example:

[0128] 500 questionnaires were distributed to the public in and around chemical enterprises, and 300 valid questionnaires were collected. The 300 valid questionnaires were classified and analyzed according to the revised model. The total value of environmental risk assessment after modeling was 100. The different risk levels are shown in Table 1.

[0129] Table 1 Risk level distribution

[0130] General environmental risks Greater environmental risk Major environmental risks 0-29 30-60 61-100

Claims

1. The SEM-based environmental impact risk assessment method for chemical enterprises is characterized by: The following steps are involved: Step 1: Construct the environmental impact risk assessment indicators of chemical enterprises, including several first-level indicators η and several second-level indicators Y under each first-level indicator η; Step 2: Set the first-level indicator η as the first-order internal latent variable, the error term of the first-order internal latent variable is ζ, set the second-level indicator Y as the internal observed variable, set the error term of the internal observed variable as ε, the load factor between the internal observed variable and the first-order internal latent variable is λ, the environmental impact risk assessment value of the chemical enterprise is the second-order internal latent variable ξ, the load factor between the second-order internal latent variable and the first-order internal latent variable is γ, and construct the structural equation mechanism model of the environmental impact risk assessment value of the chemical enterprise: ξ=γ*η+ζ (1) η=λ*Y+ε (2); Step 3: The chi-square test, chi-square freedom ratio, root mean square error of approximation, fit index, adjusted fit index, regular fit index, irregular fit index, comparative fit index and incremental fit index were used to evaluate the quality of the mechanism model obtained in step 2; Step 4: According to the evaluation results of step 3, the mechanism model is modified, and the environmental impact risk of chemical enterprises is evaluated using the modified mechanism model.

2. The SEM-based environmental risk assessment method for chemical enterprises according to claim 1 is characterized in that: In step 1, the first-level indicators include chemicals and production methods η1, environmental receptors η2, and safety management methods η3. The second-level indicators under chemicals and production methods η1 include inherent hazards of chemicals Y1, chemical reserves Y2, and production methods Y3. The second-level indicators under environmental receptors η2 include surrounding population density Y4, natural disasters Y5, and environmental risk receptors Y6. The second-level indicators under safety management methods η3 include equipment safety management methods Y7, establishment of environmental emergency plans Y8, environmental emergency disposal Y9, and major accidents Y10. 10 , Production site selection Y 11 and management experience Y 12 .

3. The SEM-based environmental risk assessment method for chemical enterprises according to claim 2 is characterized in that: In step 2, the first-level indicator chemicals and production methods η1, environmental receptors η2, and safety management methods η3 are set as first-order internal potential variables, and the error terms of the three first-order internal potential variables are ζ1, ζ2, and ζ3, respectively. The second-level indicators of inherent hazards of chemicals Y1, chemical reserves Y2, production methods Y3, surrounding population density Y4, natural disasters Y5, environmental risk receptors Y6, equipment safety management methods Y7, environmental emergency plan establishment Y8, environmental emergency disposal Y9, and major accidents Y1 are set as the first-order internal potential variables. 10 , Production site selection Y 11 and management experience Y 12 Assuming them as internal observed variables, the error terms of the twelve internal observed variables are ε1, ε2, ε3, ε4, ε5, ε6, ε7, ε8, ε9, ε 10 , ε 11 , ε 12 The load factors between the internal observed variables and the first-order internal latent variables are λ 11 , 12 , 13 , 24 , 25 , 26 , 37 , 38 , 39 , 310 , 311 , 312 , the environmental impact risk assessment value of chemical enterprises is the second-order internal latent variable ξ1, and the loading factor between the second-order internal latent variable and the first-order internal latent variable is γ 11 , γ 21 , γ 31 .

4. The SEM-based environmental risk assessment method for chemical enterprises according to claim 3 is characterized in that: In step 2, the structural equation mechanism model of the environmental impact risk assessment value of chemical enterprises is specifically: ξ1=(γ 11 *η1+ζ1)+(γ 21 *η2+ζ2)+(γ 31 *η3+ζ3) (3) η1=(λ 11 * Y1+ε1)+(λ 12 * Y2+ε2)+(λ 13 * Y3+ε3) (4) η2=(λ 24 * Y4+ε4)+(λ 25 * Y5+e5)+(λ 26 * Y6+e6) (5) η3=(λ 37 *Y7+e7)+(λ 38 *Y8+ε8)+(λ 39 *Y9+e9)+(l 310 *Y 10 +e 10 )+(λ 311 * Y 11 +e 11 )+(λ 312 *Y 12 +e 12 (6).

5. The SEM-based environmental risk assessment method for chemical enterprises according to claim 3 is characterized in that: The step 3 specifically includes: Step 301: Evaluate the quality of the mechanism model through the chi-square test, according to the following formula: T=(N-1)F min (7) Among them, T represents the test value of model fit, which is regarded as the chi-square value, N is the number of samples, and F min It indicates the minimum function estimate of the fitting function obtained by various parameter estimation methods; when the T value reaches a significant level, the mechanism model has a poor fit, and when the T value does not reach a significant level, the mechanism model has a good fit; Step 302: Evaluate the quality of the mechanism model through the chi-square freedom ratio, according to the following formula: X 2 / df(8) Among them, df is the chi-square freedom ratio; the smaller the chi-square freedom ratio, the higher the fit of the mechanism model, and vice versa, the worse the fit of the mechanism model; Step 303: Evaluate the quality of the mechanism model through the root mean square error of the approximation, according to the following formula: N is the number of samples. The smaller the RMSEA index is, the better the fit of the mechanism model is; Step 304: The method of evaluating the quality of the mechanism model by the fitting index is based on the following formula: Among them, S represents the covariate of the sample, GFI is a ratio less than 1, the closer the GFI value is to 1, the higher the fit of the mechanism model is, and when the GFI value is smaller, the difference between the numerator and the denominator is larger, which means the fit of the mechanism model is lower; Step 305: The method for evaluating the quality of the mechanism model by adjusting the fitting index is based on the following formula: Among them, t is the number of estimated parameters, DP is the number of observation data collected for environmental impact risk assessment of chemical enterprises; the AGFI value is between 0 and 1, the closer the value is to 1, the better the fit of the mechanism model is, and the closer it is to 0, the worse the fit is; Step 306: The method of evaluating the quality of the mechanism model by using the regular fit index is based on the following formula: Among them, χ 2 indep represents an independent model, Indicates a comparative model; the closer the NFI value is to 1, the greater the improvement of the theoretical model over the baseline model; when the NFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model; Step 307: The method of evaluating the quality of the mechanism model by using the informal fitting index is based on the following formula: The closer the NNFI value is to 1, the greater the improvement of the theoretical model over the baseline model. When the NNFI value is close to 0, it means that there is little difference between the theoretical model and the baseline model. Step 308: The method of evaluating the quality of the mechanism model by comparing the fitting index is based on the following formula: The closer the value of the CFI index is to 1, the more ideal the mechanism model is; Step 309: The method of evaluating the quality of the mechanism model by the incremental fitting index is based on the following formula: IFI ranges from 0 to 1, and a larger value indicates a better fit of the mechanism model.

Images