A regional risk prediction method and system based on toxic gas diffusion analysis

Abstract

The present invention discloses a regional risk prediction method and system based on toxic gas diffusion analysis. On the basis of source term tracing and prediction of leakage sources, a regional risk prediction method based on toxic gas diffusion analysis is proposed. The main contents of the invention mainly involve the development and improvement of gas diffusion models, data collection and processing, construction of risk prediction models, uncertainty analysis, research on emergency response strategies, case analysis and verification, as well as system integration and demonstration applications. By comprehensively considering factors such as gas properties, meteorological conditions, and terrain features, the prediction accuracy is improved, providing a scientific basis for risk management, and ultimately realizing the practical application and popularization of the prediction system.

Description

Technical Field

[0001] The present invention relates to the field of toxic gas risk prediction, and particularly to a regional risk prediction method and system based on toxic gas diffusion analysis. Background Art

[0002] The regional risk prediction method based on toxic gas diffusion analysis mainly simulates the diffusion process after toxic gas leakage, calculates the gas concentration in each area, and then predicts the risk level. This method focuses on quantitative analysis, inputs parameters such as accident source intensity, meteorological data, and topography into the atmospheric diffusion model to predict the gas diffusion range and concentration distribution, and combines factors such as population density and environmentally sensitive points to comprehensively predict the toxic gas exposure risk faced by each area, providing a scientific basis for formulating targeted risk prevention and control measures and emergency plans, ensuring effective risk reduction in sudden environmental events, and protecting people's lives and property safety.

[0003] The existing regional risk prediction methods based on toxic gas diffusion analysis have certain limitations. For example, the initial parameters of the diffusion model for actual scenarios cannot be directly measured and obtained; the real-time data integration ability is weak, affecting the timeliness of prediction results; it is highly dependent on the initial parameters of the gas leakage source, and parameter errors are likely to amplify prediction uncertainties; in addition, the prediction results lack pertinence when guiding specific emergency measures and are difficult to meet diverse emergency needs. There is an urgent need to optimize the method and improve the accuracy in combination with actual application scenarios. Summary of the Invention

[0004] The purpose of the present invention is to provide a regional risk prediction method and system based on toxic gas diffusion analysis to solve the problems raised in the above background art.

[0005] To achieve the above invention purpose, one aspect of the present invention provides a regional risk prediction method based on toxic gas diffusion analysis. For the case where the number of sensors monitoring the toxic gas concentration is sufficient, a method for the full spatio-temporal influence range of leakage based on the inversion algorithm is adopted, including the following steps:

[0006] Step S1, obtain real-time gas concentration data;

[0007] Step S2, input real-time meteorological parameters, and use the gas diffusion model to calculate the predicted concentration value at the location of the gas sensor;

[0008] Step S3, use the grey wolf optimization algorithm to calculate the source term parameters, and repeatedly iterate steps S2 and S3 until the convergence condition is met; the source term parameters include the three-dimensional leakage position, leakage rate, and leakage time;

[0009] Step S4, determine the influence range after 1 hour according to the gas diffusion model and the toxicity threshold of hazardous chemicals;

[0010] Step S5: Identify the risk sources of leakage disasters according to the influence scope, mainly including the number of people in the affected area after 1 hour, the road construction level in the affected area, the road congestion situation in the affected area, the number of emergency shelters in the affected range, and the number of major hazard sources in the affected range;

[0011] Step S6: Establish a regional risk prediction index system for hazardous chemical leakage disasters based on the results of risk source identification and combined with expert knowledge;

[0012] Step S7: Have experts score the weights of the regional dynamic risk prediction index system for hazardous chemical leakage disasters. The system constructs a judgment matrix according to the weight calculation and builds an AHP risk prediction model;

[0013] Step S8: Score and assign values to each prediction index according to the influence scope of the gas diffusion model;

[0014] Step S9: Calculate the risk values at different future time periods according to the AHP risk prediction model and combined with the gas diffusion model;

[0015] Step S10: Analyze the main risks at different time periods and put forward suggestions for disposal measures according to the risk values of each item predicted in real time.

[0016] Furthermore, the meteorological parameters described in Step S1 include emission parameters, diffusion parameters, and wind speed, where the horizontal diffusion parameter σ y The calculation formula is as follows:

[0017]

[0018] Calculate the vertical diffusion parameter σ based on the atmospheric stability and the downwind distance x according to the Briggs equation z , and the formula is as follows:

[0019]

[0020] where x is the downwind distance, t is the time, Γ is the vertical temperature lapse rate, and the calculation formula of Γ is as follows

[0021]

[0022] where g is the acceleration due to gravity, c p is the specific heat capacity at constant pressure of air, ΔT is the temperature decrease, and Δz is the height difference.

[0023] Furthermore, the gas diffusion model is the Gaussian plume model, and the formula is:

[0024]

[0025] Where x is the abscissa, corresponding to the longitude of the geodetic coordinate system after conversion, y is the ordinate, corresponding to the latitude of the geodetic coordinate system after conversion, z is the height coordinate, corresponding to the elevation coordinate of the geodetic coordinate system after conversion, Q is the leakage rate, U is the wind speed, and z 0 is the height of the leakage position, and δ y is the horizontal diffusion parameter, and δ z is the vertical diffusion parameter, and C(x, y, z, t) is the concentration of toxic gas at each position point and each time.

[0026] Furthermore, the objective function of the gray wolf optimization algorithm described in step S3 is expressed as the following formula:

[0027]

[0028] Where C cal is the calculated value, C ob is the observed value, M is the total number of sensors monitored, x 0 is the abscissa of the leakage source position, y 0 is the ordinate of the leakage source position, z 0 is the height coordinate of the leakage source position, Q is the leakage rate, t is the time, and the tracing objective is to use the gray wolf optimization algorithm to find the x that minimizes f(x). The tracing objective is to use the gray wolf optimization algorithm to find the x that minimizes f(x). The gray wolf optimization algorithm includes the following steps:

[0029] Step S301, initialize the population number and randomly initialize the positions of the wolf pack, X i =(x i1 , x i2 , …, x in ), where i = 1, 2, …, N, and N is the size of the wolf pack;

[0030] Step S302, for the position X of each wolf i , calculate its fitness value f(X i );

[0031] Step S303, according to the fitness value, select the best three solutions as alpha, beta, and delta wolves, where X α is the position of the optimal solution, X β is the position of the second-best solution, and X δ is the position of the third-best solution;

[0032] Step S304, calculate the coefficients A and C, and the formulas are as follows:

[0033] A = 2·a·r 1 -a;

[0034] C = 2·r 2 ;

[0035] where a linearly decreases from 2 to 0, r 1 is a random number within [0, 1], r 2 is a random number within [0, 1];

[0036] Step S305, for each wolf X i , update its position according to the following formula:

[0037] D α = |C1 * (X α - X i )|;

[0038] D β = |C2 * (X β - X i )|;

[0039] D δ = |C3 * (X δ - X i )|;

[0040] X 1 = X α - A1 * D α ;

[0041] X 2 = X β - A2 * D β ;

[0042] X 3 = X δ - A3 * D δ ;

[0043] X i (t + 1) = (X 1 + X 2 + X 3 ) / 3;

[0044] where A1, A2, and A3 are random values of coefficient A in step S304, and C1, C2, and C3 are random values of coefficient C in step S304;

[0045] Step S306, check the boundaries to ensure that the updated position X i (t + 1) of the wolf does not exceed the domain of the problem;

[0046] Step S307, check whether the maximum number of iterations is reached or a predetermined fitness threshold is reached, and obtain the global minimum of the objective function; otherwise, return to step S302 and repeat the above steps S302 to S307.

[0047] Furthermore, step S4 includes the following steps:

[0048] Step S401: Determine the three - level toxicity threshold of the leaked hazardous chemicals;

[0049] Step S402: Calculate the three - level threshold concentration contour lines;

[0050] Step S403: Determine the geodetic coordinates of the leakage source;

[0051] Step S404: Combine the wind direction information and map the three - level concentration contour lines to the geodetic coordinate system;

[0052] Step S405: Determine the affected area on the map.

[0053] Furthermore, the calculation formula for the three - level threshold concentration contour lines in Step S402 is as follows:

[0054]

[0055] where x is the abscissa, corresponding to the longitude of the converted geodetic coordinate system, y is the ordinate, corresponding to the latitude of the converted geodetic coordinate system, z is the height coordinate, corresponding to the elevation coordinate of the converted geodetic coordinate system, Q is the leakage rate, U is the wind speed, σ y is the horizontal diffusion parameter, σ z is the vertical diffusion parameter σ z .

[0056] Furthermore, in the regional hazardous chemical leakage disaster risk prediction index system, the regional hazardous chemical leakage disaster risk factors include source factors, environmental factors, social factors and emergency response capabilities. Combining the identification reasons analysis of safety factors, secondary indicators are set up to establish a safety evaluation hierarchical structure.

[0057] Furthermore, Step S7 includes the following steps:

[0058] Step S701: Construct the judgment matrix A=(a ij )n*n;

[0059] Step S702: Conduct hierarchical single - ranking and consistency test;

[0060] Calculate the weight value according to the root - extraction method, and the calculation formula is as follows:

[0061] Calculate the product of each element in the matrix row, and the formula is as follows:

[0062]

[0063] where a ij is the element of the judgment matrix in Step S701;

[0064] Calculate the n - th root of M i , and the formula is as follows:

[0065]

[0066] The calculation of vector normalization is as follows:

[0067]

[0068] W i =[W 1 ,W 2 ,…,W n T is the eigenvector. Calculate the maximum eigenvalue according to the eigenvector. The maximum eigenvalue λ max is calculated as follows:

[0069]

[0070] Check the consistency of the judgment matrix. The formula is as follows:

[0071]

[0072] When CI = 0, it indicates that the matrix is completely consistent; the larger the CI, the worse the consistency of the matrix; use the random consistency ratio CR of the judgment matrix to indicate whether the judgment matrix has a satisfactory consistency result. The calculation is as follows:

[0073]

[0074] When CR < 0.10, the result of the judgment matrix is better and the consistency is strong; when CR > 0.10, the matrix needs to be adjusted until the consistency result is satisfactory;

[0075] Step S703, conduct the overall ranking and consistency check of the hierarchy;

[0076] Step S704, determine the evaluation level of each index to form a risk evaluation set;

[0077] Step S705, calculate the membership degree of each factor to its upper - level parent node to obtain the fuzzy relation matrix R representing the relationship between the risk evaluation set and the factor index set, which is expressed as the following formula:

[0078]

[0079] Step S706, calculate the multi - level comprehensive evaluation set B. The calculation method is to perform mutual operations on the eigenvector W i and the fuzzy relation matrix R. The calculation formula is as follows:

[0080]

[0081] where b j is an element of the multi - level comprehensive evaluation set B.​

[0082] Furthermore, for the case of continuous leakage, it is considered that the influence range of leakage diffusion has entered a dynamic stable state, and its influence range will not change significantly with time. The calculation formula for the concentration at a spatial position of the method is as follows:

[0083]

[0084] where x is the abscissa, corresponding to the longitude of the transformed geodetic coordinate system; y is the ordinate, corresponding to the latitude of the transformed geodetic coordinate system; z is the height coordinate, corresponding to the elevation coordinate of the transformed geodetic coordinate system; Q is the leakage rate; U is the wind speed; z 0 is the height of the leakage position, and δ y is the horizontal diffusion parameter, and δ z is the vertical diffusion parameter.

[0085] The source term parameters include the two-dimensional leakage position and the leakage rate. The plane coordinates and the calculation formula for the leakage rate of the leakage position are as follows:

[0086]

[0087] where C cal is the calculated value, C ob is the observed value, M is the total number of sensors for monitoring, x 0 is the abscissa of the leakage source position, y 0 is the ordinate of the leakage source position, and Q is the leakage rate.

[0088] Another aspect of the present invention provides a regional risk prediction system based on toxic gas diffusion analysis, including a gas concentration module, a meteorological parameter module, a source term parameter module, an influence range module, a risk source identification module, a prediction index module, a prediction model module, a scoring and assignment module, a risk calculation module, and a disposal suggestion module, wherein:

[0089] The gas concentration module is used to obtain real-time gas concentration data;

[0090] The meteorological parameter module is used to input real-time meteorological parameters and calculate the predicted concentration value at the position of the gas sensor by using the gas diffusion model formula;

[0091] The source term parameter module is used to calculate the source term parameters by iterative convergence by using the grey wolf optimization algorithm; the source term parameters include the three-dimensional leakage position, the leakage rate, and the leakage time;

[0092] The influence range module is used to determine the influence range after 1 hour according to the gas diffusion model and the toxicity threshold of hazardous chemicals;

[0093] The risk source identification module is used to identify the risk sources of leakage disasters according to the influence scope, mainly including the number of people in the affected area after 1 hour, the road construction grade in the affected area, the road congestion condition in the affected area, the number of emergency avoidance facilities in the affected range, and the number of major hazard sources in the affected range;

[0094] The prediction index module is used to establish a regional risk prediction index system for hazardous chemical leakage disasters according to the risk source identification results and combined with expert knowledge;

[0095] The prediction model module is used for experts to score the weights of the dynamic risk prediction index system for regional hazardous chemical leakage disasters. The system constructs a judgment matrix according to the weight calculation and builds an AHP risk prediction model;

[0096] The scoring and assignment module is used to score and assign each prediction index according to the influence scope of the gas diffusion model;

[0097] The risk calculation module is used to calculate the risk values at different future time periods according to the AHP risk prediction model and combined with the gas diffusion model;

[0098] The disposal suggestion module is used to analyze the main risks at different time periods and put forward suggestions on disposal measures according to each risk value predicted in real time.

[0099] Due to the adoption of this system and method, compared with the prior art, it has the following advantages:

[0100] 1. In order to improve the calculation accuracy of the leakage diffusion model, based on the monitoring data of real-time gas sensors and meteorological sensors, the grey wolf optimization algorithm is adopted to inversely calculate the source term parameters of the leakage source. The grey wolf optimization algorithm is a meta-heuristic algorithm with strong global search ability, low computational complexity and fast convergence speed, which can quickly and accurately calculate the tracing results and improve the emergency rescue efficiency.

[0101] 2. On the basis of tracing and predicting the source term of the leakage source, a regional risk prediction method based on the analysis of toxic gas diffusion is proposed. Based on the tracing calculation results, considering factors such as gas properties, meteorological conditions, and terrain features, a gas diffusion model is used to calculate the spatio-temporal influence scope of the diffusion and evolution of toxic gases. The AHP method is used to predict and evaluate the risk. Through hierarchical analysis, the AHP method comprehensively considers multiple factors and criteria in risk prediction and evaluation, making the prediction and evaluation results more comprehensive and objective. The AHP method is easy to operate, the results are intuitive, which is convenient for decision-makers to understand and adopt, thus improving the accuracy of prediction and evaluation, providing a scientific basis for risk management, and finally realizing the practical application and popularization of the prediction system.

[0102] 3. By comprehensively considering factors such as gas properties, meteorological conditions, and terrain features, the present invention improves the prediction accuracy, provides a scientific basis for risk management, and ultimately realizes the practical application and popularization of the prediction system. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 It is a flowchart of a regional risk prediction method based on toxic gas diffusion analysis.

[0104] Figure 2 It is a schematic diagram of the calculation process of the leakage full-time and full-space influence range based on the inversion algorithm.

[0105] Figure 3 It is a schematic diagram of the calculation process of the influence range of continuous leakage based on the inversion algorithm. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0106] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0107] As Figure 1 shown in the flowchart of the method of the present invention, the embodiments of the present invention provide a regional risk prediction method based on toxic gas diffusion analysis.

[0108] In the case where the number of effectively monitored on-site sensors deployed is sufficient, that is, the number of sensors that detect the concentration of toxic gas is sufficient, the method of the leakage full-time and full-space influence range based on the inversion algorithm can be adopted. The specific steps are as follows:

[0109] Step S1, access the real-time meteorological sensor data and obtain the real-time gas concentration data.

[0110] The accessed sensor data includes:

[0111] (a) Gas concentration data C ob , and PID or electrochemical sensors can be used.

[0112] (b) Meteorological data, including wind speed U and wind direction WD.

[0113] (c) Positioning data and timestamp, and the GPS position data of each monitoring sensor needs to be known.

[0114] Sensors can be connected to the transmission network using ZigBee, Wi-Fi, Bluetooth, LoRa, NB-IoT, etc. The ZigBee protocol is suitable for sensor networks with low power consumption and low data transmission rates and is suitable for large-scale deployment. LoRa and NB-IoT are suitable for long-distance, low-power Internet of Things applications, especially in wide-area network coverage scenarios. Sensor data formats support modes such as OPC, ModBus, MQTT, and WebSocket.

[0115] Determine the emission parameters of the leakage source, including the toxic gas emission rate Q of the pollution source, in kg / s, and the effective emission height z of the pollution source 0 , in m.

[0116] According to the atmospheric stability, calculate the vertical temperature lapse rate. First, it is necessary to determine the vertical temperature lapse rate Γ. Γ is a coefficient related to atmospheric stability, and the calculation formula is as follows:

[0117]

[0118] Among them, g is the acceleration due to gravity, c p is the specific heat capacity at constant pressure of air, ΔT is the temperature decrease, Δz is the height difference, and its value of Γ is usually between 0.75 (very unstable) and 0.15 (very stable).

[0119] Then calculate the horizontal diffusion parameter and the vertical diffusion parameter. According to the atmospheric stability and wind speed U, calculate through the Briggs equation. As time increases, the diffusion range of pollutants will increase, and the horizontal diffusion parameter δ y The calculation formula is as follows

[0120]

[0121] x is the downwind distance (in kilometers). t is the time (in hours).

[0122] According to the atmospheric stability and the downwind distance x, calculate the vertical diffusion parameter δ based on the Briggs equation z , in m.

[0123]

[0124] x is the downwind distance (in kilometers). t is the time (in hours).

[0125] Step S2, input real-time meteorological parameters, and use the Gaussian plume model to calculate the predicted concentration value at the location of the gas sensor.

[0126] Substitute the above-determined emission parameters, diffusion parameters, wind speed, and calculation point coordinates into the following Gaussian plume model equation:

[0127]

[0128] Calculate the concentration C(x, y, z) of toxic gas at each position point, with the unit of μg / m 3 . X is the abscissa, corresponding to the longitude of the transformed geodetic coordinate system, y is the ordinate, corresponding to the latitude of the transformed geodetic coordinate system, z is the height coordinate, corresponding to the elevation coordinate transformed into the geodetic coordinate system, Q is the leakage rate, U is the wind speed, and z 0 is the height of the leakage source location.

[0129] Step S3, use the Grey Wolf Optimization Algorithm to perform the inversion calculation of source term parameters. The source term parameters include the three-dimensional leakage position, leakage rate, and leakage time. Iterate steps S2 and S3 repeatedly until the convergence condition is met.

[0130] For the method of tracing the source of poisonous gas leakage based on the optimization algorithm and gas diffusion model, it is first necessary to construct an objective function. In the present invention, the variance between the observed value C ob and the calculated value C cal is constructed as the objective function of the optimization problem. Assuming the number of monitoring points is M, the objective function is expressed as follows:

[0131]

[0132] The calculated value C cal can be calculated according to the coordinates (x, y, z, t) of the monitoring points using the Gaussian plume model equation. The observed value C ob adopts the method of gas diffusion model simulation, and adds the limiting conditions of sensor accuracy and Gaussian white noise to be closer to the real data. M is the total number of monitored sensors, x 0 is the abscissa of the leakage source location, y 0 is the ordinate of the leakage source location, z 0 is the height coordinate of the leakage source location, Q is the leakage rate, and t is the time.

[0133] The input parameters of the objective function x = (x 0 , y 0 , z 0 , Q, t) is a 5-dimensional vector. The tracing objective is to use the Grey Wolf Optimization Algorithm to find the x that minimizes f(x). The specific steps are as follows:

[0134] Step S301, randomly initialize the positions of the wolf pack, X i = (x i1 , x i2 , …, x in ), where i = 1, 2, …, N, and N is the size of the wolf pack.

[0135] Step S302, for the position X of each wolf i , calculate its fitness value f(X i ).

[0136] Step S303, update the alpha, beta, and delta wolves.

[0137] According to the fitness values, select the best three solutions as the alpha, beta, and delta wolves. Among them, X α is the position of the optimal solution, X β is the position of the second-best solution, and X δ is the position of the third-best solution.

[0138] Step S304, calculate the coefficients A and C using the following formula:[[]]

[0139] A = 2·a·r 1 - a;

[0140] C = 2·r 2 ;

[0141] where a linearly decreases from 2 to 0, r 1 is a random number within [0, 1], and r 2 is a random number within [0, 1].[[]]

[0142] Step S305, for each wolf X i (except for alpha, beta, and delta), update its position according to the following formula:[[]]

[0143] D α = |C1 * (X α - X i )|;

[0144] D β = |C2 * (X β - X i )|;

[0145] D δ = |C3 * (X δ - X i )|;

[0146] X 1 = X α - A1 * D α ;

[0147] X 2 = X β - A2 * D β ;

[0148] X 3 = X δ - A3 * Dδ ;

[0149] X i (t + 1) = (X 1 + X 2 + X 3 ) / 3;

[0150] Where A1, A2, and A3 are random values of coefficient A, and C1, C2, and C3 are random values of coefficient C.

[0151] Step S306, boundary check, to ensure that the updated position X i (t + 1) of the wolf does not exceed the domain of the problem.

[0152] Step S307, check whether the maximum number of iterations is reached or a predetermined fitness threshold is reached, and obtain the global minimum of the objective function; otherwise, return to Step S302 and repeat the above Steps S302 to S307.

[0153] Step S4, according to the Gaussian plume model of toxic gas and the toxicity threshold of hazardous chemicals, determine the influence range after 1 hour. The specific process is as follows:

[0154] Step S401, determine the third-level toxicity threshold of the leaked hazardous chemical.

[0155] Step S402, calculate the third-level threshold concentration contour line, that is, the influence range.

[0156] The Gaussian plume model is a mathematical model used to describe the diffusion of pollutants under stable atmospheric conditions. This model assumes that the pollutant concentration distribution conforms to a Gaussian distribution. Assume that the third-level toxicity threshold of the toxic gas is C 1 , C 2 , C 3 , then the contour map will show the areas that meet the following conditions:

[0157] C(x, y, z) = C 1 ;

[0158] C(x, y, z) = C 2 ;

[0159] C(x, y, z) = C 3 .

[0160] In practical applications, it is usually assumed that the concentration of pollutants is the highest at the emission source height z 0 . Therefore, the model can be simplified by only considering the case of z = z 0 . In this way, the above formula can be simplified to:

[0161]

[0162] Obtain the basic calculation formula for drawing contour maps.

[0163] Step S403: Determine the geodetic coordinates of the leakage source.

[0164] Step S404: Combine the wind direction information and map the third-level concentration contour lines to the geodetic coordinate system.

[0165] Use the geographical coordinates of the emission source as a reference point. According to the average wind speed direction (wind direction), convert the calculated horizontal distance into displacements in the east-west and north-south directions. Using the earth's geographical coordinate system, convert the displacements into longitude and latitude changes. This usually involves spherical trigonometry and is calculated using the following formula:

[0166]

[0167] Where, Δλ is the longitude change, Δφ is the latitude change, Δx and Δy are the displacements in the east-west and north-south directions, R is the radius of the earth, and φ is the latitude of the reference point.

[0168] Step S405: Determine the affected area on the map.

[0169] Step S5: According to the affected area, conduct identification of leakage disaster risk sources, mainly including the number of people in the affected area after 1 hour, the road construction grade in the affected area, the road congestion situation in the affected area, the number of emergency avoidance facilities in the affected area, and the number of major hazard sources in the affected area.

[0170] Steps S1 to S5 are the calculation process of the leakage full-time and full-space affected area based on the inversion algorithm, and its process schematic diagram is as Figure 2 shown.

[0171] Step S6: According to the risk source identification results and combined with expert knowledge, establish a dynamic risk prediction index system for regional hazardous chemical leakage disasters. The main index system is shown in Table 1.

[0172] Determine the influencing factors of the thing to be evaluated, establish a factor index set U, and the risk prediction index set U is a multi-level tree structure. The risk factors of regional hazardous chemical leakage disasters include source factors, environmental factors, social factors, and emergency response capabilities, etc. The prediction index system is a system used to predict the risk degree of regional hazardous chemical leakage disasters, including indicators in multiple aspects. Further combined with the identification reason analysis of safety factors, conduct in-depth analysis of the first-level indicators and set the second-level indicators. Establish a safety evaluation hierarchical structure. Compared with traditional risk prediction, the risk of hazardous chemical leakage disasters evolves over time and is dynamically changing. The various parameters in the prediction system change based on the scope affected by the toxic gas diffusion model.

[0173] Table 1:

[0174]

[0175] Step S7: The expert scores the weights of the dynamic risk prediction index system for regional hazardous chemical leakage disasters. The system constructs a judgment matrix based on the calculated weights and builds an AHP risk prediction model. First, perform hierarchical single sorting and inspection, and use the analytic hierarchy process to calculate the weights of the evaluation indicators for this evaluation unit. Construct matrices for each single-element indicator and calculate their consistency ratios. When the consistency ratio CR < 0.10, it indicates that the matrices constructed among the indicators can achieve satisfactory results, and the sorting and consistency ratio of the total hierarchy can be calculated continuously. Establish a judgment matrix to calculate the AHP weights, and the calculated weight values are shown in Table 1. Then consult the expert group to determine the relative importance among the vertical indicators. The scale is based on the 1-9 scale method. Establish a pairwise comparison matrix and calculate the weight values of each indicator.

[0176] Step S701: Construct a judgment matrix.

[0177] The judgment matrix is for the elements in the upper layer of the hierarchical structure, and determines the relative importance among the elements in the hierarchy. Suppose there are n elements. According to the expert scoring method, compare each pair of elements in each layer to obtain the judgment matrix A = (a ij )n*n, as shown in Table 2:

[0178] Table 2 AHP Judgment Matrix Table

[0179] Ak B1 B2 ... Bn B1 a11 a12 ... aln B2 a21 a22 ... a2n ... ... ... ... Bn an1 An2 ... Ann

[0180] Among them, each element (a ij )n*n uses the 1-9 scale rule and its reciprocal as the judgment scale, as shown in Table 3 specifically:

[0181] Table 3 AHP Judgment Scale Values

[0182] Scale value aij Meaning 1 Factor i is equally important as factor j 3 Factor i is slightly more important than factor j 5 Factor i is more important than factor j 7 Factor i is very important compared to factor j 9 Factor i is absolutely more important than factor j 2，4，6，8 The importance of factor i and factor j lies in the median of the above adjacent judgments Reciprocal The judgment value aij = 1 / dij obtained by comparing factor j with factor i

[0183] Step S702: Perform hierarchical single sorting and consistency inspection.

[0184] Hierarchical single sorting is to determine the weight values of the importance of each element in a certain hierarchy for an element in the upper hierarchy. Calculate the weight values according to the root method.

[0185] First, calculate the product of each row element of the matrix. The calculation formula is as follows:

[0186]

[0187] Then calculate the nth root of M i , and the formula is as follows:

[0188]

[0189] The calculation formula for vector normalization is as follows:

[0190]

[0191] Wi = [W1, W2,..., Wn]T, that is, the eigenvector. Calculate the maximum eigenvalue according to the eigenvector. The calculation formula is as follows:

[0192]

[0193] Check the consistency of the judgment matrix. The calculation formula is as follows:

[0194]

[0195] When CI = 0, it indicates that the matrix is completely consistent; the larger the CI, the worse the consistency of the matrix. To determine whether the judgment matrix has a satisfactory consistency result, the random consistency ratio CR of the judgment matrix is often used. The calculation formula is as follows:

[0196]

[0197] When CR < 0.10, the result of the judgment matrix is better and the consistency is strong; when CR > 0.10, the matrix needs to be adjusted until the consistency result is satisfactory. The values of the average random index RI are shown in Table 4.

[0198] Table 4 Values of the average random index RI

[0199] Order 1 2 3 4 5 RI 0 0 0.58 0.9 1.12 Order 6 7 8 9 ... RI 1.24 1.32 1.41 1.45 ...

[0200] Step S703: Conduct the overall hierarchy ranking and consistency check.

[0201] The overall hierarchy ranking refers to the weight values of the importance indicators of all elements in this hierarchy with respect to the previous hierarchy. After the single hierarchy ranking passes the consistency check, there may be non - consistency situations between different hierarchies. Therefore, it is necessary to conduct a consistency check on the result of the overall hierarchy ranking.

[0202] Step S704: Determine the risk evaluation set.

[0203] For each indicator, give its score with V = (0, 1, 2, 3, 4, 5) as the comment set to form the risk evaluation set. Among them, 5 points represent the highest weight, indicating the greatest impact; 0 points represent no impact.

[0204] Step S705: Calculate the membership degree matrix.

[0205] After the evaluation set V and the index set U are determined, according to the fuzzy mathematics theory, calculate the membership degree of each factor to its upper-level parent node, and obtain a fuzzy relation matrix representing the relationship between the risk evaluation set and the factor index set. This matrix is the membership degree matrix. First, construct a fuzzy mapping f: U→F(V), where F(V) represents all fuzzy sets on the evaluation set V, and is represented by the matrix R:

[0206]

[0207] Step S706, calculate the comprehensive evaluation set.

[0208] The calculation of the multi-level comprehensive evaluation set B is a ranking of the mutual operation of the eigenvector W i and the evaluation membership degree matrix R of each index. The calculation process is as follows:

[0209]

[0210] In the formula, first multiply w j and r ij and then sum to obtain the element b j of the evaluation set. The advantage of the weighted average model is that while retaining the single-index evaluation information, it comprehensively considers the influence of all indicators, effectively integrates the comprehensive evaluation of the indicators, and realizes the overall comprehensive evaluation.

[0211] Step S9, score and assign values to each prediction index according to the influence range of the gas diffusion model.

[0212] Step S10, according to the AHP risk prediction model, combined with the gas diffusion model, calculate the risk values at different future time periods.

[0213] Step S11, the system analyzes the main risks at different time periods based on the risk values of each item predicted in real time and proposes suggestions for disposal measures.

[0214] For the continuous leakage situation, the influence range of leakage diffusion has entered a dynamic stable state, and its influence range will not change significantly with time. The differences between the calculation method of the influence range of toxic gas diffusion based on the inversion algorithm and the leakage full-time and full-space influence range method include the following parts:

[0215] In step S2, simplify the Gaussian plume formula, and then calculate the predicted concentration value at the location of the gas sensor. The simplified spatial position concentration calculation formula becomes:

[0216]

[0217] In step S3, since the number of effective monitoring sensors is small, in order to ensure the effective convergence of the traceability algorithm, it is necessary to simplify the source term parameters of the inversion calculation. The simplified source term parameters only calculate three dimensions of x = (x 0 , y 0 , Q), that is, the plane coordinates of the leakage location and the leakage rate. The objective function is simplified as follows:

[0218]

[0219] As Figure 3 shown is a schematic diagram of the calculation process of the continuous leakage influence range based on the inversion algorithm.

[0220] Based on the source term traceability prediction of the leakage source, the present invention proposes a regional risk prediction method based on the analysis of toxic gas diffusion, which mainly involves developing and improving the gas diffusion model, data collection and processing, constructing a risk prediction system, conducting uncertainty analysis, studying emergency response strategies, case analysis and verification, and system integration and demonstration applications. By comprehensively considering factors such as gas properties, meteorological conditions, and terrain features, the prediction accuracy is improved, providing a scientific basis for risk management, and ultimately realizing the practicality and popularization of the prediction system.

[0221] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A regional risk prediction method based on toxic gas diffusion analysis, characterized in that: In the case where there are enough sensors monitoring the concentration of toxic gases, the method of leak full-time and space impact range based on inversion algorithm is adopted, which includes the following steps: Step S1, obtaining real-time gas concentration data and meteorological parameters; the meteorological parameters include emission parameters, expansion parameters, wind speed, among which the horizontal diffusion parameter σ y The calculation formula is as follows: The vertical diffusion parameter σ is calculated based on the Briggs equation according to the atmospheric stability and the downwind distance x. z , the formula is as follows: Where x is the downwind distance, t is the time, and Γ is the vertical temperature lapse rate. The calculation formula for Γ is as follows: Where g is the acceleration due to gravity, c is the p is the constant pressure specific heat capacity of air, ΔT is the temperature decrease, and Δz is the height difference; Step S2, using a gas diffusion model to calculate the predicted concentration value at the location of the gas sensor; the gas diffusion model is a Gaussian plume model, and the formula is: Where x is the horizontal coordinate, corresponding to the longitude of the geodetic coordinate system after conversion, y is the vertical coordinate, corresponding to the latitude of the geodetic coordinate system after conversion, z is the height coordinate, corresponding to the height coordinate converted into the geodetic coordinate system, Q is the leakage rate, U is the wind speed, z0 is the height of the leakage location, σ y is the horizontal diffusion parameter, σ z is the vertical diffusion parameter, C(x,y,z,t) is the concentration of toxic gas at each location at each time; Step S3, using the Grey Wolf Optimization Algorithm to calculate source item parameters, repeatedly iterating steps S2 and S3 until convergence conditions are met; the source item parameters include leakage three-dimensional position, leakage rate and leakage time; Step S4, determining the impact range after 1 hour based on the gas diffusion model and the toxicity threshold of hazardous chemicals; Step S5, identifying the leakage disaster risk sources according to the impact range, including the number of people in the affected area after 1 hour, the road construction level in the affected area, the road congestion in the affected area, the number of emergency shelters in the affected area, and the number of major hazard sources in the affected area; Step S6, based on the risk source identification results and combined with expert knowledge, establish a regional hazardous chemical leakage disaster risk prediction index system; Step S7, experts weight and score the regional hazardous chemicals leakage disaster dynamic risk prediction index system, and the system constructs a judgment matrix based on the weight calculation to build an AHP risk prediction model; Step S8, assigning a score to each prediction indicator according to the influence range of the gas diffusion model; Step S9, calculating the risk values ​​of different time periods in the future according to the AHP risk prediction model and the integrated gas diffusion model; Step S10, based on the real-time predicted risk value of each item, analyze the main risks in different time periods and propose disposal measures.

2. The method for regional risk prediction based on toxic gas diffusion analysis according to claim 1 is characterized in that: The objective function of the gray wolf optimization algorithm in step S3 is expressed as follows: Among them C cal is the calculated value, C ob is the observed value, M is the total number of monitored sensors, x0 is the horizontal coordinate of the leak source location, y0 is the vertical coordinate of the leak source location, z0 is the height coordinate of the leak source location, Q is the leak rate, t is the time, and the tracing goal is to use the gray wolf optimization algorithm to find the leak source. f(x) The smallest x, the gray wolf optimization algorithm includes the following steps: Step S301, initialize the population number and randomly initialize the position of the wolf pack, X i =(x i1 ,x i2 ,…,x in ), where i = 1, 2, ..., N, N is the size of the wolf pack; Step S302: for each wolf's position X i , calculate its fitness value f(X i ); Step S303, according to the fitness value, select the best three solutions as alpha, beta and delta wolves, where X α is the optimal solution position, X β is the position of the second best solution, X δ is the location of the third optimal solution; Step S304, calculate coefficients A and C, the calculation formula is as follows: A=2·a·r1-a; C＝2·r2； Where a decreases linearly from 2 to 0, r1 is a random number in [0,1], and r2 is a random number in [0,1]; Step S305: for each wolf X i , and update its position according to the following formula: D α =|C1*(X α -X i )|; D β =|C2*(X β -X i )|; D δ =|C3*(X δ -X i )|; X1=X α -A1*D α ; X2=X β -A2*D β ; X3=X δ -A3*D δ ; X i (t+1)=(X1+X2+X3) / 3; Where A1, A2 and A3 are random values ​​of coefficient A in step S304, C1, C2 and C3 are random values ​​of coefficient C in step S304; Step S306, check the boundary to ensure that the updated wolf position X i (t+1) does not exceed the definition domain of the problem; Step S307, check whether the maximum number of iterations is reached or the predetermined fitness threshold is reached to obtain the global minimum value of the objective function; otherwise, return to step S302 and repeat the above steps S302 to S307.

3. The regional risk prediction method based on toxic gas diffusion analysis according to claim 1 is characterized in that: Step S4 includes the following steps: Step S401, determining the third-level toxicity threshold of the leaked hazardous chemicals; Step S402, calculating the three-level threshold concentration contour lines; Step S403, determining the geodetic coordinates of the leakage source; Step S404, mapping the three-level concentration contour lines to the geodetic coordinate system in combination with the wind direction information; Step S405: determine the impact range on the map.

4. The regional risk prediction method based on toxic gas diffusion analysis according to claim 3 is characterized in that: The calculation formula for the three-level threshold concentration contour line in step S402 is: Where x is the horizontal coordinate, corresponding to the longitude of the geodetic coordinate system after conversion, y is the vertical coordinate, corresponding to the latitude of the geodetic coordinate system after conversion, z is the height coordinate, corresponding to the height coordinate converted into the geodetic coordinate system, Q is the leakage rate, U is the wind speed, σ y is the horizontal diffusion parameter, σ z is the vertical diffusion parameter σ z .

5. The regional risk prediction method based on toxic gas diffusion analysis according to claim 1 is characterized in that: The regional hazardous chemicals leakage disaster risk factors in the regional hazardous chemicals leakage disaster risk prediction index system include source factors, environmental factors, social factors and emergency response capabilities. Combined with the identification and cause analysis of safety factors, secondary indicators are set up to establish a safety evaluation hierarchy.

6. The regional risk prediction method based on toxic gas diffusion analysis according to claim 1 is characterized in that: Step S7 includes the following steps: Step S701, construct a judgment matrix A = (a ij )n*n; Step S702, performing hierarchical single sorting and consistency check; The weight value is calculated according to the square root method. The calculation formula is as follows: Calculate the product of each row of the matrix using the following formula: where a ij is the element of the judgment matrix in step S701; Calculate M i The nth root of is as follows: Calculate vector normalization, the formula is as follows: W i =[W1,W2,…,W n ] T is the eigenvector, and the maximum eigenvalue is calculated based on the eigenvalue. max The calculation is as follows: Check the consistency of the judgment matrix. The formula is as follows: When CI = 0, it means that the matrix is ​​completely consistent; the larger the CI, the worse the consistency of the matrix; the random consistency ratio CR of the judgment matrix is ​​used to indicate whether the judgment matrix has a satisfactory consistency result, which is calculated as follows: When CR<0.10, the judgment matrix result is good and the consistency is strong; when CR>0.10, the matrix needs to be adjusted until the consistency result is satisfactory; Step S703, performing total sorting and consistency check on the levels; Step S704, determining the evaluation level of each indicator to form a risk evaluation set; Step S705, calculate the degree of membership of each factor to its upper parent node, and obtain the fuzzy relationship matrix R between the risk assessment set and the factor index set, which is expressed as the following formula: Step S706, calculate the multi-level comprehensive evaluation set B, the calculation method is to transform the feature vector W i The calculation formula is as follows: where b j is an element of the multi-level comprehensive evaluation set B.

7. A regional risk prediction system based on toxic gas diffusion analysis, characterized in that: It includes gas concentration module, meteorological parameter module, source parameter module, impact range module, risk source identification module, prediction index module, prediction model module, scoring module, risk calculation module and disposal suggestion module, among which: The gas concentration module is used to obtain real-time gas concentration data; the meteorological parameters include emission parameters, expansion parameters, and wind speed, among which the horizontal diffusion parameter σ y The calculation formula is as follows: The vertical diffusion parameter σ is calculated based on the Briggs equation according to the atmospheric stability and the downwind distance x. z , the formula is as follows: Where x is the downwind distance, t is the time, and Γ is the vertical temperature lapse rate. The calculation formula for Γ is as follows: Where g is the acceleration due to gravity, c is the p is the constant pressure specific heat capacity of air, ΔT is the temperature decrease, and Δz is the height difference; the meteorological parameter module is used to input real-time meteorological parameters and use the gas diffusion model formula to calculate the predicted concentration value at the location of the gas sensor; the gas diffusion model is a Gaussian plume model, and the formula is: Where x is the horizontal coordinate, corresponding to the longitude of the geodetic coordinate system after conversion, y is the vertical coordinate, corresponding to the latitude of the geodetic coordinate system after conversion, z is the height coordinate, corresponding to the height coordinate converted into the geodetic coordinate system, Q is the leakage rate, U is the wind speed, z0 is the height of the leakage location, σ y is the horizontal diffusion parameter, σ z is the vertical diffusion parameter, C(x,y,z,t) is the concentration of toxic gas at each location at each time; The source item parameter module is used to calculate the source item parameters by iterative convergence using the Grey Wolf optimization algorithm; the source item parameters include the leakage three-dimensional position, leakage rate and leakage time; The impact range module is used to determine the impact range after one hour based on the gas diffusion model and the toxicity threshold of hazardous chemicals; The risk source identification module is used to identify the risk sources of leakage disasters according to the impact range, including the number of people in the affected area after 1 hour, the road construction level in the affected area, the road congestion in the affected area, the number of emergency shelters in the affected area, and the number of major hazard sources in the affected area; The prediction index module is used to establish a regional hazardous chemical leakage disaster risk prediction index system based on risk source identification results and expert knowledge; The prediction model module is used by experts to weight and score the dynamic risk prediction index system of regional hazardous chemical leakage disasters. The system constructs a judgment matrix based on the weight calculation and builds an AHP risk prediction model. The scoring module is used to score and assign values ​​to each prediction indicator according to the influence range of the gas diffusion model; The risk calculation module is used to calculate the risk values ​​of different time periods in the future based on the AHP risk prediction model and the integrated gas diffusion model; The disposal suggestion module is used to analyze the main risks in different time periods and propose disposal measures based on the real-time predicted risk value of each item.

Images