A method for predicting emergency resource demand in major and serious production safety accidents

Abstract

This invention discloses a method for predicting the emergency resource demand for major and serious production safety accidents. The method first collects accident characteristic indicators and accident consequence indicators for major and serious accidents; then, based on each accident characteristic indicator, it matches the corresponding indicator weight coefficients in Appendix 1 to calculate the enterprise's safety entropy coefficient w for major and serious accidents. i ; and then according to w i Based on the accident consequence index parameters, the severity D of the accident is quantitatively calculated. i Finally, based on the severity of the accident, D... i Using multiple linear regression analysis, a prediction method for emergency resource demand M in major and serious production safety accidents was established. resource Model for calculating emergency resource demand D resource This invention enables the calculation of relatively accurate emergency resource prediction results in situations of information asymmetry during emergency rescue for major accidents. This method scientifically and effectively improves the speed of emergency decision-making and response, saves time in allocating emergency resources, and enhances the suitability of emergency resource allocation.

Description

Technical Field

[0001] This invention belongs to the field of production safety technology, specifically relating to a method for predicting the emergency resource demand for major and serious production safety accidents. Background Technology

[0002] In accident emergency rescue, it is necessary to allocate different types and quantities of emergency resources, including emergency rescue personnel such as emergency rescue industry experts, professional emergency rescue teams, fire brigades, and medical teams, as well as emergency supplies such as basic living necessities, emergency equipment and supporting materials, engineering materials, and machining equipment. The dispatch and allocation of emergency resources is a crucial link in the accident emergency rescue process and a key factor in determining the scientific and efficient conduct of emergency rescue. Due to the uncertainty, complexity, unpredictability, and severe destructiveness of accidents, the demand for and dispatch of emergency resources are also dynamic and uncertain. When an accident occurs, emergency rescue forces and resources should arrive at the accident site in the shortest possible time; otherwise, delays will hinder rescue efforts or lead to the escalation of the situation, causing more serious casualties and property damage.

[0003] Analyzing and predicting the quantity of emergency resources needed for rescue after an accident is crucial for shortening rescue time and avoiding waste of emergency resources. Commonly used prediction methods include analogy, trend extrapolation, regression analysis, time series forecasting, grey system model forecasting, and combined forecasting. Multiple regression theory, when studying the influence of multiple factors, can establish a multiple regression model of multi-factor coupling by calculating the correlation between various influencing factors and the degree of regression fit. Therefore, multiple regression analysis is chosen to analyze and predict the quantity of emergency resources required in the emergency rescue process of production safety accidents. When predicting the emergency resource demand for an accident, only the relevant accident characteristic parameters and accident consequence parameters (such as industry type, accident type, number of casualties, etc.) need to be input into the model to obtain the required quantity of emergency resources for that accident. After an accident occurs, using mathematical models to predict the required quantity of emergency resources and arranging the allocation of those resources in a timely manner can shorten emergency rescue time, improve the efficiency of emergency resource utilization, avoid waste of emergency resources, and minimize the potential losses caused by the accident. Summary of the Invention

[0004] To address the above problems, this invention provides a method for predicting the emergency resource demand for major and catastrophic production safety accidents. This invention, considering the differences in industry type, accident type, accident region, accident time, number of casualties, and accident impact range among major and catastrophic accidents, proposes an accident severity coupled assessment method (M). destruction Model. Based on the above influencing factors and accident severity D i Able to determine the demand for emergency resources D in accident emergency rescue.resource The prediction provides a theoretical basis for various industries to improve their emergency response capabilities.

[0005] A method for predicting the emergency resource demand of major and serious production safety accidents includes the following steps:

[0006] (1) Collect accident characteristic indicators and accident consequence indicators for major and serious accidents; the accident characteristic indicators include industry type w industry Accident type w type , accident area w region Accident time (month w) month Weekday week Time period w frame The accident consequence indicators include the number of casualties (D). death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing ), Scope of the accident (Number of employees within the enterprise D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter );

[0007] (2) Based on the various accident characteristic indicators in step (1), match the corresponding indicator weight coefficients in Appendix 1 to calculate the safety entropy coefficient w for major and serious accidents of the enterprise. i ; and then according to w i Based on the accident consequence index parameters, the severity D of the accident is quantitatively calculated. i ;

[0008] (3) Based on the severity of the accident D i Using multiple linear regression analysis, a prediction method for emergency resource demand M in major and serious production safety accidents was established. resource Model for calculating emergency resource demand D resource , can be represented as:

[0009] D resource =b0+b death D death +b serious D serious +b minor D minor +b missing D missing +b employee b employee +b perimeter D perimeter +b i D i

[0010] in,

[0011] Dresource Forecast the emergency resource demand for a certain type of major production safety accident:

[0012] b0、b death b serious b minor b missing b employee b perimeter b i The independent variable is the number of deaths D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i The least squares estimates of the regression coefficients;

[0013] D death For the number of deaths, in people;

[0014] D serious The number of seriously injured persons;

[0015] D minor The number of people with minor injuries;

[0016] D missing The number of missing persons;

[0017] D employee The number of employees within the company;

[0018] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise;

[0019] D i The severity of the accident.

[0020] The method for predicting emergency resource demand in major and serious production safety accidents according to the present invention includes the following steps in its specific operation process:

[0021] (1) When a major or serious accident occurs, collect 12 accident indicators;

[0022] ① Collect accident characteristic indicators. Based on the characteristics of the accident, quickly refer to Appendix 1 for the table of weight coefficients for accident characteristic indicators and select the corresponding indicator: Industry type w industry Accident type w type , accident area w region Accident time (month w) month Weekday week Time period w frameThe weighting coefficients are then used to calculate the safety entropy coefficient w for major and serious accidents of the enterprise according to equation (1-1). i ;

[0023] w i =w industry +w type +w region +w month ·w week ·w frame (1-1)

[0024] in,

[0025] w i The safety entropy coefficient for major and catastrophic accidents in enterprises represents the degree of system safety disorder caused by the accident. i The smaller the value, the better the system's security;

[0026] w industry The weighting coefficient for industry type indicators;

[0027] w type The weighting coefficient for accident type indicators;

[0028] w region The weighting coefficient for the accident area indicators;

[0029] w month The weighting coefficient for the monthly indicators of the accident;

[0030] w week The weighting coefficient for the accident week indicator;

[0031] w frame The weighting coefficients for indicators during the accident period;

[0032] ② Collect indicators of accident consequences, including the number of deaths (D). death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter The specific value;

[0033] (2) Quantitatively calculate the severity D of the accident. i ;

[0034] The specific operation involves establishing an accident severity coupled assessment M. destruction Model:

[0035]

[0036] in,

[0037] Di The severity of the accident;

[0038] w i The safety entropy coefficient for major and serious accidents in enterprises;

[0039] D death For the number of deaths, in people;

[0040] D serious The number of seriously injured persons;

[0041] D minor The number of people with minor injuries;

[0042] D missing The number of missing persons;

[0043] D employee The number of employees within the company affected by the accident (in persons).

[0044] D perimeter The number of permanent residents in the surrounding area of ​​the enterprises affected by the accident is [number].

[0045] (3) Forecast of emergency resource demand for major and serious production safety accidents M resource Model

[0046] Based on the multiple linear regression analysis method, a prediction mechanism for emergency resource demand M in major and serious production safety accidents was established. resource Model;

[0047] First, assume that the independent variable of the model is the number of deaths, D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i To facilitate model construction and simplification, all independent variables can be represented using D. p This indicates that the dependent variable is the predicted emergency resource demand D for major and serious production safety accidents. resource There is a linear relationship between the independent variable and the dependent variable, according to the dependent variable D. resource The equation constructed from the correlations between the variables is called the emergency resource demand forecasting multiple linear regression equation Z. resourse , can be represented as:

[0048]

[0049] in,

[0050] D resourcePredict the demand for emergency resources for a certain type of major production safety accident;

[0051] β0, β death β serious β minor β missing β employee β perimeter β i The independent variable is the number of deaths D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i The regression coefficients, all regression coefficients can be obtained using β. p Substitute; each regression coefficient reflects the effect of the corresponding independent variable on the dependent variable D. resource The degree of linear influence;

[0052] D death For the number of deaths, in people;

[0053] D serious The number of seriously injured persons;

[0054] D minor The number of people with minor injuries;

[0055] D missing The number of missing persons;

[0056] D employee The number of employees within the company;

[0057] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise;

[0058] D i The severity of the accident;

[0059] ε is the random error term of the model's predicted values, which follows a normal distribution N(0, σ). 2 ).

[0060] In Z resourse In the equation, the regression coefficient β p The maximum likelihood estimate can be expressed as the sum of squares of the random error term ε:

[0061]

[0062] By solving for the partial derivative of ε and simplifying it, we can obtain:

[0063] D pT D p β p =D p T D resource (1-5)

[0064] Calculate β using the least squares method. p The least squares estimate b p :

[0065] b p =(D p T D p ) -1 D p T D resource (1-6)

[0066] By applying the maximum likelihood estimation method and simplifying the multiple linear regression equation Z for emergency resource demand forecasting, resourse (1-3) Eliminate the random error term ε in the model to obtain the predicted emergency resource demand M for major and serious production safety accidents. resource Model:

[0067]

[0068] in,

[0069] D resource Forecast the emergency resource demand for a certain type of major production safety accident:

[0070] b0、b death b serious b minor b missing b employee b perimeter b i The Z-variable linear regression equation for predicting emergency resource demand is respectively resourse (1-3) Regression coefficients β0, β death β serious β minor β missing β employee β perimeter β i The least squares estimate;

[0071] D death For the number of deaths, in people;

[0072] D serious The number of seriously injured persons;

[0073] D minor The number of people with minor injuries;

[0074] D missing The number of missing persons;

[0075] D employee The number of employees within the company;

[0076] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise;

[0077] D i The severity of the accident.

[0078] As a preferred technical solution, it can be determined based on the severity of the accident, D. i The numerical values ​​are used to classify accident levels, specifically as follows:

[0079] Level I: D i ≥2.7192

[0080] Level II: 0.3510≤D i <2.7192

[0081] Level III: 0.113832729≤D i <0.3510

[0082] Level IV: D i <0.1138.

[0083] D i The higher the value, the more serious the accident.

[0084] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0085] (1) This invention selects industry type w industry Accident type w type , accident area w region Accident time (month w) month Weekday week Time period w frame The six accident characteristic indicators and the number of casualties (D) death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing ), Scope of the accident (Number of employees within the enterprise D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Six accident consequence indicators were used as influencing factors of accident severity. Safety entropy theory and information entropy theory were introduced, and an accident severity coupled assessment M was constructed based on the entropy weight method. destruction Model for quantitative calculation of accident severity D i .

[0086] (2) This invention utilizes multiple linear regression analysis theory to predict the emergency resource demand M for major and serious accidents in mining, transportation, hazardous chemicals, fireworks and firecrackers, construction, and other industries, combined with the actual rescue process and the demand for emergency resources (emergency rescue personnel and rescue materials). resource The model achieves relatively accurate predictions even under conditions of information asymmetry, improving the speed of emergency decision-making, reducing emergency response time and the time required to allocate emergency resources, providing a theoretical basis for enhancing the industry's emergency response capabilities, and has broad application prospects in the field of accident emergency rescue.

[0087] (3) When the method of the present invention is applied to the prediction of the demand for critical emergency resources in major accidents, the absolute deviation is as low as 0.00% and as high as 16.67%; the average absolute deviation is 6.07%, which is within an acceptable range. Therefore, the prediction model for the demand for emergency resources in major production safety accidents established by the present invention is effective. Attached Figure Description

[0088] Figure 1 This is a flowchart illustrating the construction of the emergency resource demand prediction model for major and serious production safety accidents according to the present invention. Detailed Implementation

[0089] The present invention will be further illustrated by the following examples, but these are not intended to limit the invention.

[0090] Taking one major accident each in the mining, transportation, and hazardous chemicals industries as examples, calculate the emergency resource demand D for major accidents. resource The model construction flowchart is shown below. Figure 1 .

[0091] (I) Establishing a forecasting system for emergency resource demand in major and serious production safety accidents. resource Model

[0092] Step 1: Establish an accident severity coupled assessment M destruction Model

[0093] 1.1 Determination of Indicators

[0094] (1) Background of indicator extraction

[0095] By studying several major and serious accidents that have occurred in my country in recent years, and statistically analyzing the number of deaths, serious injuries, minor injuries, missing persons, accident areas, accident times (month, weekday, time period), number of employees within the enterprises, and number of permanent residents outside the enterprises, the distribution characteristics of the accident data were analyzed (data results are shown in Appendices 2-5). The findings are as follows:

[0096] By industry type, high-risk industries such as mining, hazardous chemicals, transportation, and construction accounted for 86.4% of the total number of major and serious accidents, with the transportation industry having the highest number of accidents and fatalities.

[0097] According to the type of accident, fire-related major accidents have the highest number of incidents and the highest number of deaths. No major accidents have occurred in the categories of mechanical injury, electric shock, blasting, drowning, boiler explosion, and other injuries in recent years.

[0098] Major and serious accidents vary depending on the region and the level of industry development.

[0099] According to the time of the accidents, December had the highest number of major and serious accidents. The week and time period with the highest number of major and serious accidents were Saturdays and between 7-9 am and 13-17 pm.

[0100] (2) Determination of various indicators

[0101] Based on the above analysis, 12 accident indicators were extracted, namely, industry type w. industry Accident type w type , accident area w region Accident time (month w) month Weekday week Time period w frame ), number of casualties (w) death Number of seriously injured (w) serious Number of people with minor injuries (w) minor Number of missing persons w missing ) and the scope of the accident (number of employees within the company w) employee The number of permanent residents in the surrounding area of ​​the enterprise (w) perimeter ), Analyzing the coupling effect of multiple factors on accident severity assessment M destruction The impact of the model is assessed, and the weighting coefficients of each indicator are calculated based on the degree of impact.

[0102] Among them, the six accident characteristic indicators are: industry type w industry Accident type w type , accident area w region Month of the accident w month Accident Weekday week and the time period of the accident w frame Used to calculate the safety entropy coefficient w of a company's major accidents. i The six accident consequence indicators are: number of deaths (w) death Number of seriously injured (w) serious Number of people with minor injuries (w) minor Number of missing persons w missing Number of employees within the enterprise (w) employee and the number of permanent residents in the surrounding area of ​​the enterprise wperimeter As an accident severity coupled assessment M destruction The undetermined constant of the model's information entropy w severity .

[0103] 1.2 Calculation of weight coefficients for six indicators related to accident characteristics

[0104] (1) Industry type w industry Weighting coefficient calculation

[0105] ① Characteristics of industry distribution indicators for major and serious accidents

[0106] Table 1-1 presents statistical data on major and serious accidents in mining, hazardous chemicals, fireworks and firecrackers, construction, metal smelting, transportation and other industries in recent years, based on six evaluation items: number of accidents, number of deaths, number of serious injuries, number of minor injuries, number of missing persons, and number of casualties in each industry.

[0107] Table 1-1 Distribution Indicators of Major and Serious Accidents by Industry Type

[0108]

[0109] ② Entropy weight method for calculating the weight coefficients of accident characteristic indicators

[0110] The original data index matrix X is formed by using the 7 industries and 6 accident indicators in Table 1-1 as evaluation indicators. ij for:

[0111]

[0112] All of the above evaluation indicators are positive indicators, for the original data indicator matrix X ij After positive transformation and non-negative translation, the normalized matrix S is obtained. ij for:

[0113]

[0114] Using the entropy weight method to calculate equation (1-9), we obtain the industry type w. industry The calculation results of the indicator weight coefficients are shown in Table 1-2.

[0115] Table 1-2 Industry Types industry Indicator weight coefficient calculation results

[0116] Serial Number Industry type <![CDATA[Weight coefficient w industry > Serial Number Industry type <![CDATA[Weight coefficient w industry > 1 Hazardous chemicals 0.4111 5 Construction 0.0455 2 Transportation 0.3077 6 Metal smelting 0.0176 3 mine 0.1276 7 Fireworks and firecrackers 0.0014 4 Other industries 0.0891

[0117] Accident type w type , accident area w region Accident time (month w) month Weekday week Time period w frameThe calculation process of the weight coefficients of the 5 indicators and industry type w industry same.

[0118] (2) Accident type w type Weighting coefficient calculation

[0119] Based on the distribution index characteristic data of major and serious accidents in Appendix 2, the accident type w is obtained. type The calculation results of the indicator weight coefficients are shown in Table 1-3.

[0120] Table 1-3 Accident Types w type Indicator weight coefficient calculation results

[0121] Serial Number Accident Types <![CDATA[Weight coefficient w type > Serial Number Accident Types <![CDATA[Weight coefficient w type > 1 Other explosions 0.3126 8 Mistakes and scrambles 0.0335 2 Container explosion 0.2572 9 Vehicle damage 0.0216 3 fire 0.1267 10 permeable 0.0166 4 collapse 0.0771 11 Crane injury 0.0095 5 gas explosion 0.0541 12 Falling from a height 0.0084 6 gunpowder explosion 0.0416 13 Object strike 0.0041 7 Poisoning and suffocation 0.0355 14 Scalding 0.0017

[0122] ⑶Accident area w region Weighting coefficient calculation

[0123] Based on the regional distribution index characteristics of major and serious accidents in Appendix 3, the accident area w is obtained. region The calculation results of the weighting coefficients are shown in Table 1-4.

[0124] Table 1-4 Accident Area w region Indicator weight coefficient calculation results

[0125]

[0126] (4) Accident month w month Weighting coefficient calculation

[0127] Based on the monthly distribution index characteristics of major and serious accidents in Appendix 4, we obtain the accident month w. month The calculation results of the weighting coefficients are shown in Table 1-5.

[0128] Table 1-5 Accident Months (w) month Indicator weight coefficient calculation results

[0129] Serial Number month <![CDATA[Weight coefficient w month > Serial Number month <![CDATA[Weight coefficient w month > 1 8 0.3099 7 5 0.0465 2 3 0.2361 8 12 0.0433 3 6 0.1055 9 4 0.0405 4 7 0.0688 10 2 0.0251 5 11 0.0562 11 10 0.0148 6 9 0.0531 12 1 0.0003

[0130] (5) Accident Weekday week Weighting coefficient calculation

[0131] Based on the weekly distribution index characteristics of major and serious accidents in Appendix 5, the accident week w is obtained. week The calculation results of the weighting coefficients are shown in Table 1-6.

[0132] Table 1-6 Accident Week (w) week Indicator weight coefficient calculation results

[0133]

[0134]

[0135] (6) Accident period w frame Weighting coefficient calculation

[0136] Based on the characteristic data of the distribution indicators of major and serious accidents in Appendix 6, the accident time period w is obtained. frame The calculation results of the weighting coefficients are shown in Table 1-7.

[0137] Table 1-7 Accident Time Period w frame Indicator weight coefficient calculation results

[0138] Serial Number Week <![CDATA[Weight coefficient w frame > Serial Number Week <![CDATA[Weight coefficient w frame > 1 14:00-15:00 0.2086 13 2:00-3:00 0.0269 2 22:00-23:00 0.1975 14 11:00-12:00 0.0267 3 16:00-17:00 0.0699 15 10:00-11:00 0.0265 4 7:00-8:00 0.0658 16 13:00-14:00 0.0167 5 15:00-16:00 0.0459 17 23:00-24:00 0.0160 6 6:00-7:00 0.0423 18 20:00-21:00 0.0152 7 8:00-9:00 0.0410 19 18:00-19:00 0.0143 8 0:00-1:00 0.0383 20 5:00-6:00 0.0100 9 19:00-20:00 0.0333 21 4:00-5:00 0.0060 10 12:00-13:00 0.0314 22 21:00-22:00 0.0051 11 17:00-18:00 0.0299 23 3:00-4:00 0.0036 12 9:00-10:00 0.0280 24 1:00-2:00 0.0009

[0139] 1.3 Calculation of weighting coefficients for six indicators related to accident consequences

[0140] The number of casualties and the scope of the accident include six accident consequence indicators (number of deaths w) death Number of seriously injured (w) serious Number of people with minor injuries (w) minor Number of missing persons w missing Number of employees within the enterprise (w) employee The number of permanent residents in the surrounding area of ​​the enterprise (w) perimeter The original data index matrix X consists of 162 sets of data (the major and serious accident data table in Appendix 7). ij for:

[0141]

[0142] All of the above evaluation indicators are positive indicators. For the original data indicator matrix X in equation (1-10) ij After positive transformation and non-negative translation, the normalized matrix S is obtained. ij for:

[0143]

[0144] The entropy weight method was used to calculate equation (1-11) to obtain the number of casualties and the affected area w. severity The calculation results of the indicator weight coefficients are shown in Table 1-8.

[0145] Table 1-8 Number of casualties and scope of the accident severity Indicator weight coefficient calculation results

[0146] Serial Number Casualties and the extent of the accident <![CDATA[Weight coefficient w severity > Serial Number Casualties and the extent of the accident <![CDATA[Weight coefficient w severity > 1 Number of people with minor injuries 0.4943 4 Missing persons 0.0802 2 Number of seriously injured 0.2166 5 Number of employees within the company 0.0705 3 Death toll 0.1105 6 Number of permanent residents in the surrounding area of ​​the enterprise 0.0279

[0147] 1.4 Model Construction

[0148] Using the entropy weight method, after constructing the original matrix and standardizing it, the weight coefficients of each indicator of accident severity were finally calculated. The calculation results are shown in Tables 1-2 to 1-8, and Appendix Table 1 is a summary table of the weight coefficient values ​​for accident characteristic indicators. From this, the safety entropy coefficient w for major and serious accidents in the enterprise can be obtained. i for:

[0149] w i =w industry +w type +w region +w month ·w week ·w frame (1-1)

[0150] in,

[0151] w i The safety entropy coefficient for major and serious accidents in enterprises;

[0152] w industry The weighting coefficient for industry type indicators;

[0153] w type The weighting coefficient for accident type indicators;

[0154] w region The weighting coefficient for the accident area indicators;

[0155] w month The weighting coefficient for the monthly indicators of the accident;

[0156] w week The weighting coefficient for the accident week indicator;

[0157] w frame The weighting coefficients for indicators during the accident period.

[0158] To quantify the severity of the accident's damage, the weighting coefficients of six accident consequence indicators were analyzed in relation to the accident severity D. i The impact is assessed by employing the "information entropy" theory and combining it with the safety entropy coefficient w of major and serious accidents in enterprises. i Quantitative calculation of accident severity D i :

[0159]

[0160] in,

[0161] D i The severity of the accident;

[0162] w i The safety entropy coefficient for major and serious accidents in enterprises;

[0163] w severityIt is an undetermined constant of information entropy, namely the weighting coefficient of the indicators of the number of casualties and the scope of the accident.

[0164] D severity These are the specific numerical values ​​for the number of casualties and the extent of the accident's impact.

[0165] The weighting coefficients w for the indicators of casualties and the scope of the accident in Table 1-8 are used. severity Substituting these values ​​into equation (1-12) as undetermined constants for information entropy, and combining them with the security entropy coefficient w obtained from equation (1-1), i Construct an accident severity coupled assessment M destruction Model:

[0166]

[0167] in,

[0168] D i The severity of the accident;

[0169] w i The safety entropy coefficient for major and serious accidents in enterprises;

[0170] D death For the number of deaths, in people;

[0171] D serious The number of seriously injured persons;

[0172] D minor The number of people with minor injuries;

[0173] D missing The number of missing persons;

[0174] D employee The number of employees within the company;

[0175] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise.

[0176] 1.5 Accident Severity Classification

[0177] Based on the calculation results, the severity of the accident, D, is... i It is divided into four levels: Level I, Level II, Level III, and Level IV. The results of the level classification are shown in Table 1-9.

[0178] Table 1-9 Accident Severity D i Classification

[0179] Serial Number <![CDATA[Accident severity D i > Level Classification 1 <![CDATA[D i ≥2.719165740]]> Level I 2 <![CDATA[0.351035767≤D i <2.719165740]]> Level II 3 <![CDATA[0.113832729≤D i <0.351035767]]> Level III 4 <![CDATA[D i <0.113832729]]> Level IV

[0180] Step 2: Establish a forecast of emergency resource demand M for major and serious production safety accidents. resource Model

[0181] Based on a multiple linear regression analysis model, a forecast of emergency resource demand M for major and serious production safety accidents is established. resource Model.

[0182] First, let's assume the independent variable is the number of deaths, D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i To facilitate model construction and simplification, all independent variables can be represented using D. p Substitute; the dependent variable is the predicted emergency resource demand D for major and serious production safety accidents. resource There is a linear relationship between the independent variable and the dependent variable, according to the dependent variable D. resource The equation constructed from the correlations between the variables is called the emergency resource demand forecasting multiple linear regression equation Z. resourse , can be represented as:

[0183]

[0184] in,

[0185] D resource Forecast the demand for emergency resources for a certain type of major production safety accident, such as the demand for professional rescue team personnel (D). rescuer Demand for self-rescue devices (D) respirator Combat suit demand D uniform Demand for dry powder fire extinguishers (D) extinguisher Oxygen detector demand D calibrator Demand for hydraulic lifters D hydraulic Demand for demolition tools D demolition Demand for hazard warning signs (D) sign Oxygen respirator demand D breathing Demand for combustible gas detectors D detector wait;

[0186] β0, β death β serious β minor β missing β employee β perimeter β i The independent variable is the number of deaths D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D)employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i For ease of model construction and simplification, all regression coefficients can be represented using β. p Substitution. Each regression coefficient reflects the influence of the corresponding independent variable on the dependent variable D. resource The degree of linear influence;

[0187] D death For the number of deaths, in people;

[0188] D serious The number of seriously injured persons;

[0189] D minor The number of people with minor injuries;

[0190] D missing The number of missing persons;

[0191] D employee The number of employees within the company;

[0192] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise;

[0193] D i The severity of the accident;

[0194] ε is the random error term of the model's predicted values, which follows a normal distribution N(0, σ). 2 ).

[0195] In Z resourse In the equation, the regression coefficient β p The maximum likelihood estimate can be expressed as the sum of squares of the random error term ε:

[0196]

[0197] By solving for the partial derivative of ε and simplifying it, we can obtain:

[0198] D p T D p β p =D p T D resource (1-5)

[0199] Calculate β using the least squares method. p The least squares estimate b p :

[0200] b p =(D p T D p) -1 D p T D resource (1-6)

[0201] Then, by applying the maximum likelihood estimation method and simplifying the multiple linear regression equation Z for emergency resource demand forecasting, resourse (1-3) Eliminate the random error term ε in the model to obtain the predicted emergency resource demand M for major and serious production safety accidents. resource Model:

[0202]

[0203] in,

[0204] D resource Forecast the emergency resource demand for a certain type of major production safety accident:

[0205] b0、b death b serious b minor b missing b employee b perimeter b i The Z-variable linear regression equation for predicting emergency resource demand is respectively resourse (1-3) Regression coefficients β0, β death β serious β minor β missing β employee β perimeter β i The least squares estimate;

[0206] D death For the number of deaths, in people;

[0207] D serious The number of seriously injured persons;

[0208] D minor The number of people with minor injuries;

[0209] D missing The number of missing persons;

[0210] D employee The number of employees within the company;

[0211] D perimeter The number of permanent residents in the surrounding area of ​​the enterprise;

[0212] D i The severity of the accident.

[0213] In accident severity coupled assessment M destructionBased on the model, predict the critical emergency resource requirements D for major accidents in different industries. resource The death toll D death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter and the severity of the accident D i Substitute the predicted emergency resource demand M for major and serious production safety accidents resource Model (1-7) yields the predicted quantity of corresponding emergency resources. Table 1-10 shows the predicted M based on the emergency resource demand for major and serious production safety accidents. resource Demand models for critical relief supplies in several high-risk industries.

[0214] Table 1-10 Forecast of Key Emergency Resource Demand for Several Major and Serious Production Safety Accidents (M) resource Model

[0215]

[0216]

[0217] in,

[0218] D rescuer The required number of personnel for a professional rescue team is [number] people.

[0219] D respirator To meet the demand for self-rescue devices, units;

[0220] D uniform The required number of combat suits is [number] sets.

[0221] D extinguisher The required quantity of dry powder fire extinguishers is [number] units.

[0222] D calibrator The required quantity of oxygen detectors is [number] units.

[0223] D hydraulic The demand for hydraulic lifters is [number] units.

[0224] D demolition To meet the demand for demolition tools, sets are required.

[0225] D sign The required number of hazard warning signs is [number];

[0226] D breathing The demand for oxygen respirators is [number] units.

[0227] D detectorThe demand for combustible gas detectors is [number] units.

[0228] D death For the number of deaths, in people;

[0229] D serious The number of seriously injured persons;

[0230] D minor The number of people with minor injuries;

[0231] D employee The number of employees within the company;

[0232] D i The severity of the accident.

[0233] (II) Collecting data on major and serious accidents in the mining, transportation, and hazardous chemical industries.

[0234] Three major accidents in the mining, transportation, and hazardous chemical industries were collected as Examples 1-3. Accident-related information was compiled, and the results are shown in Table 1-11. The accident severity D was also calculated. i .

[0235] Table 1-11 Statistical Data of Implementation Examples

[0236]

[0237] Example 1

[0238] Death toll D death =15 people;

[0239] Number of seriously injured (D) serious =1 person;

[0240] Number of people with minor injuries (D) minor =8 people;

[0241] Missing Persons D missing =0 people;

[0242] Number of employees in the enterprise D employee =105 people;

[0243] Number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter =1019 people;

[0244] Industry type index weighting coefficient w industry =0.1276;

[0245] Accident type index weighting coefficient w type =0.0541;

[0246] Accident area indicator weighting coefficient w region =0.0269;

[0247] Accident Monthly Indicator Weighting Coefficient w month =0.0562;

[0248] Accident Weekly Index Weighting Coefficient w week =0.0291;

[0249] Accident Period Indicator Weighting Coefficient w frame =0.0167;

[0250] Substituting the above data into equation (1-1), we can obtain the safety entropy coefficient w for major accidents of the enterprise. i =0.2086;

[0251] Combined assessment of accident severity M destruction The model yields the accident severity D. i =1.5526, the severity of the accident is Level II.

[0252] Example 2

[0253] Death toll D death =35 people;

[0254] Number of seriously injured (D) serious =0 people;

[0255] Number of people with minor injuries (D) minor =13 people;

[0256] Missing Persons D missing =0 people;

[0257] Number of employees in the enterprise D employee =57 people;

[0258] Number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter =0 people;

[0259] Industry type index weighting coefficient w industry =0.3077;

[0260] Accident area indicator weighting coefficient w region =0.0621;

[0261] Accident Monthly Indicator Weighting Coefficient w month =0.1055;

[0262] Accident Weekly Index Weighting Coefficient w week =0.0926;

[0263] Accident Period Indicator Weighting Coefficient w frame =0.0265;

[0264] Substituting the above data into equation (1-1), we can obtain the safety entropy coefficient w for major accidents of the enterprise. i =0.3701;

[0265] Combined assessment of accident severity M destruction The model yields the accident severity D. i =3.0690, the severity of the accident is Level I.

[0266] Example 3

[0267] Death toll D death =19 people;

[0268] Number of seriously injured (D) serious =0 people;

[0269] Number of people with minor injuries (D) minor =12 people;

[0270] Missing Persons D missing =0 people;

[0271] Number of employees in the enterprise D employee =98 people;

[0272] Number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter =0 people;

[0273] Industry type index weighting coefficient w industry =0.4111;

[0274] Accident type index weighting coefficient w type =0.1267;

[0275] Accident area indicator weighting coefficient w region =0.0037;

[0276] Accident Monthly Indicator Weighting Coefficient w month =0.0688;

[0277] Accident Weekly Index Weighting Coefficient w week =0.2714;

[0278] Accident Period Indicator Weighting Coefficient w frame =0.0143;

[0279] Substituting the above data into equation (1-1), we can obtain the safety entropy coefficient w for major accidents of the enterprise. i =0.5418;

[0280] Combined assessment of accident severity M destruction The model yields the accident severity D. i =5.9120, the severity of the accident is Level I.

[0281] (III) Calculating the predicted emergency resource demand value D in the embodiment resource .

[0282] Based on the calculated accident severity D i Further calculate the predicted value D of the emergency resource demand for the accident. resource The predicted values ​​were compared and analyzed with the actual amount of emergency resources used to obtain the absolute deviation D between the predicted and actual values. x for:

[0283]

[0284] in,

[0285] D actual This refers to the actual amount of emergency resources used during the rescue process;

[0286] D predicted To predict the emergency resource demand for major and serious production safety accidents (M) resource The predicted values ​​of emergency resources obtained from the model calculation.

[0287] Actual emergency resource usage value D in each embodiment actual Demand forecast D predicted The absolute deviations are shown in Table 1-12.

[0288] Table 1-12 Calculation Results and Absolute Deviation of Examples

[0289]

[0290] Of the 18 emergency resources listed in Table 1-12, the minimum absolute deviation was 0.00%, and the maximum was 16.67%. Eleven resources had an absolute deviation of less than 5%: oxygen detectors and hydraulic lifters in Example 1; four emergency resources in Example 2 excluding combat suits and hydraulic lifters; and five emergency resources in Example 3 excluding dry powder fire extinguishers. Four emergency resources had an absolute deviation between 5% and 15%: the number of professional rescue team members, self-rescue devices, and dry powder fire extinguishers in Example 1; and combat suits in Example 2. The remaining three emergency resources had an absolute deviation between 15% and 20%.

[0291] The average absolute deviation of the emergency resource demand predictions in the above three embodiments is 6.07%, which is within an acceptable range. This verifies the effectiveness of the emergency resource demand prediction method for major and serious production safety accidents established in this invention. resourceThe effectiveness and practicality of the model. This invention enables the generation of relatively accurate emergency resource demand forecasts under conditions of information asymmetry, improves the speed of emergency decision-making, reduces emergency response time and emergency resource allocation time, and provides a theoretical basis for improving the industry's emergency response capabilities.

[0292] The quick reading table, which summarizes the weight coefficients of various accident characteristic indicators in this invention (Tables 1-2 to 1-7), is shown in Appendix 1: Appendix 1 Table of Accident Characteristic Indicator Weight Coefficient Values

[0293]

[0294]

[0295] Appendix 2: Distribution Indicators of Major and Serious Accidents

[0296]

[0297] Appendix 3: Regional Distribution Indicators of Major and Serious Accidents

[0298] Area code Number of accidents Death toll Number of seriously injured Number of people with minor injuries Missing persons Casualties 1 14 159 0 122 14 295 2 13 252 12 187 8 459 3 11 242 3 151 5 401 4 11 169 31 81 0 281 5 11 153 0 118 4 275 6 8 161 8 57 0 226 7 8 135 0 32 0 167 8 7 137 2 40 0 181 9 6 97 29 208 5 334 10 6 97 5 65 4 171 11 5 110 1 85 0 196 12 5 87 41 103 0 231 13 5 69 5 77 0 151 14 5 69 2 45 1 117 15 5 68 1 47 0 116 16 5 62 6 34 0 102 17 5 61 0 144 0 205 18 4 252 76 759 0 1087 19 4 211 58 752 8 1029 20 4 104 3 34 0 141 21 4 67 0 78 2 147 22 3 66 0 84 0 150 23 3 23 4 28 11 66 24 2 55 0 19 0 74 25 2 30 0 12 0 42 26 2 29 0 48 0 77 27 2 29 0 12 0 41 28 1 18 0 0 2 20 29 1 12 10 3 0 25 total 163 3035 297 3463 59 6854

[0299] Appendix 4: Monthly Distribution Indicators of Major and Serious Accidents

[0300]

[0301] Appendix 5: Weekly Distribution Indicators of Major and Serious Accidents

[0302]

[0303] Appendix 6: Characteristics of Time Period Distribution Indicators for Major and Serious Accidents

[0304]

[0305] Appendix 7: Data Table of Major and Serious Accidents

[0306]

[0307] Continued table

[0308]

[0309] Continued table

[0310]

[0311] Continued table

[0312]

[0313] Continued table

[0314]

[0315] Continued table

[0316]

Claims

1. A method for predicting the emergency resource demand in major and serious production safety accidents, characterized in that, Includes the following steps: (1) When a major or serious accident occurs, collect 12 accident indicators; ① Collect accident characteristic indicators. Based on the characteristics of the accident, quickly refer to Appendix 1 for the table of weight coefficients for accident characteristic indicators and select the corresponding indicator: Industry type w industry Accident type w type , accident area w region The weighting coefficient for the accident time, where the accident time is the month w. month Weekday week and time period w frame Then, calculate the safety entropy coefficient w for major and serious accidents of the enterprise according to formula (1-1). i ; In i =in industry +in type +in region +in month ·In week ·In frame (1-1) in, w i The safety entropy coefficient for major and catastrophic accidents in enterprises represents the degree of system safety disorder caused by the accident. i The smaller the value, the better the system's security; w industry The weighting coefficient for industry type indicators; w type The weighting coefficient for accident type indicators; w region The weighting coefficient for the accident area indicators; w month The weighting coefficient for the indicators of the month of the accident; w week The weighting coefficient for the accident week indicator; w frame The weighting coefficients for indicators during the accident period; in, Appendix 1: Table of Weighting Coefficients for Accident Characteristic Indicators ② Collect indicators of accident consequences, including the number of deaths (D). death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter The specific value; (2) Quantitatively calculate the severity D of the accident. i ; The specific operation involves establishing an accident severity coupled assessment M. destruction Model: in, D i The severity of the accident; w i The safety entropy coefficient for major and serious accidents in enterprises; D death For the number of deaths, in people; D serious The number of seriously injured persons; D minor The number of people with minor injuries; D missing The number of missing persons; D employee The number of employees within the company affected by the accident (in persons). D perimeter The number of permanent residents in the surrounding area of ​​the enterprises affected by the accident; (3) Forecast of emergency resource demand for major and serious production safety accidents M resource Model Based on the multiple linear regression analysis method, a prediction mechanism for emergency resource demand M in major and serious production safety accidents was established. resource Model; First, assume that the independent variable of the model is the number of deaths, D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i To facilitate model construction and simplification, all independent variables can be represented using D. p This indicates that the dependent variable is the predicted emergency resource demand D for major and serious production safety accidents. resource There is a linear relationship between the independent variable and the dependent variable, according to the dependent variable D. resource The equation constructed from the correlations between the variables is called the emergency resource demand forecasting multiple linear regression equation Z. resourse , can be represented as: in, D resource Predict the demand for emergency resources for a certain type of major production safety accident; β0, β death β serious β minor β missing β employee β perimeter β i The independent variable is the number of deaths D. death Number of seriously injured (D) serious Number of people with minor injuries (D) minor Missing Persons D missing Number of employees within the enterprise (D) employee The number of permanent residents in the surrounding area of ​​the enterprise (D) perimeter Accident Severity D i The regression coefficients, all regression coefficients can be obtained using β. p Substitute; each regression coefficient reflects the effect of the corresponding independent variable on the dependent variable D. resource The degree of linear influence; D death For the number of deaths, in people; D serious The number of seriously injured persons; D minor The number of people with minor injuries; D missing The number of missing persons; D employee The number of employees within the company; D perimeter The number of permanent residents in the surrounding area of ​​the enterprise; D i The severity of the accident; ε is the random error term of the model's predicted values, which follows a normal distribution N(0, σ). 2 ); In Z resourse In the equation, the regression coefficient β p The maximum likelihood estimate can be expressed as the sum of squares of the random error term ε: By solving for the partial derivative of ε and simplifying it, we can obtain: D p T D p β p =D p T D resource (1-5) Calculate β using the least squares method. p The least squares estimate b p : b p =(D p T D p ) -1 D p T D resource (1-6) By applying the maximum likelihood estimation method and simplifying the multiple linear regression equation Z for emergency resource demand forecasting, resourse (1-3) Eliminate the random error term ε in the model to obtain the predicted emergency resource demand M for major and serious production safety accidents. resource Model: D resource =b0+b death D death +b serious D serious +b minor D minor +b missing D missing +b employee D employee +b perimeter D perimeter +b i D i (1-7) in, D resource Forecast the emergency resource demand for a certain type of major production safety accident: b0、b death b serious b minor b missing b employee b perimeter b i The Z-variable linear regression equation for predicting emergency resource demand is respectively resourse (1-3) Regression coefficients β0, β death β serious β minor β missing β employee β perimeter β i The least squares estimate; D death For the number of deaths, in people; D serious The number of seriously injured persons; D minor The number of people with minor injuries; D missing The number of missing persons; D employee The number of employees within the company; D perimeter The number of permanent residents in the surrounding area of ​​the enterprise; D i The severity of the accident.

2. The method for predicting emergency resource demand in major and serious production safety accidents according to claim 1, characterized in that, According to the severity of the accident D i The numerical values ​​are used to classify accident levels, specifically as follows: Level I: D i ≥2.7192 Level II: 0.3510≤D i <2.7192 Level III: 0.113832729≤D i <0.3510 Level IV: D i <0.1138.

Images