KR20090116860A - Method for analyzing fire dangerousness of railway tunnel - Google Patents

Abstract

PURPOSE: A method for analyzing fire risk of a railway tunnel is provided to analyze the fire risk by predicting the probability of accidents, and comparing the predicted information with a standard value. CONSTITUTION: The past statistical data is inputted(S11). The tunnel information and train information of the analysis object are inputted(S12). The accident occurrence frequency about the tunnel length is predicted(S13). The number of deaths and the probability of accident occurrence over the number of deaths are generated. The FN curves express the probability of accident occurrence over the number of deaths and the number of deaths in a graph(S15). The data range with higher value than the standard FN curve is determined as having higher fire risk(S16).

Description

Method for analyzing fire dangerousness of railway tunnel}

The present invention is a risk analysis method of a railway tunnel fire, and more specifically, a fire occurrence scenario in a current railway tunnel is assumed based on the historical distance and fire statistics of the entire railway vehicle, and the number of deaths for each scenario is estimated. And after predicting the probability of occurrence and compare with the reference value to determine the fire risk of the railway tunnel.

Railway tunnel fires have a lower incidence than ordinary fires but, in the past, once a fire can result in a catastrophic disaster resulting in loss of life and property. At present, most of the fire protection disaster design technology level for the domestic railway tunnel fire is designed around the law. Legally oriented designs are designed with mandatory regulations without flexibility due to detailed guidelines and technical standards of the law for the purpose of minimum safety, and safety is not considered.

Today, the construction of railway tunnels continues to increase, and modern rail transportation systems are progressing in speeding up railroad cars and increasing the length of railway tunnels. In view of this, the risk of fire in railway tunnels is increasing.

In particular, the long tunnel has the advantage of reducing the operating distance and increasing the speed due to the straightening of the route.However, the disaster prevention performance such as evacuation is reduced, and the tunnel system becomes very complicated due to the lengthening of the tunnel. Because it affects life safety, safety design considering them is necessary. Therefore, a risk-based design technique is required, which requires quantitative risk analysis. In addition, the design technology for the rescue station, which must be installed in the long tunnel, is not provided in Korea, so it is necessary to develop design technology for it. However, these fire safety technologies are implemented in a state where there is a shortage of foreign fragments or related technologies, rather than through phased R & D, so systematic and active R & D is required. Furthermore, there are no special alternatives to the extent to which the measurement and improvement of the fire safety level of railroad cars and the evaluation of problems have been made.

The present invention has been made in view of the above-mentioned point, and quantifies the factor items for the fire analysis of the railway vehicle to automatically generate a virtual fire occurrence scenario from the user's input information, the number of deaths that occur for each scenario and The aim is to analyze the risk of railway tunnel fires by predicting the probability of occurrence.

According to the risk analysis method of the railway tunnel fire of the present invention for achieving the above object, (S11) receiving and storing the historical statistical data including the operating distance and the number of accidents by year of the entire railway vehicle from the user step; (S12) receiving and storing the tunnel information and the vehicle information of the analysis target from the user; (S13) estimating an accident occurrence frequency (MF: Master Frequency) for the corresponding tunnel length after calculating the number of accident occurrences (count of accidents / train km) per unit driving distance using the historical statistical data; (S14) branching the frequency of accidents into a plurality of predefined scenarios through an event tree defining an accident fire type, and the number of deaths (N) of individual scenarios (Sn) and the probability of occurrence of an accident more than the number of deaths (N) (F) Generating (Sn (F, N)); (S15) generating a scenario-specific FN curve in which the death number (N) of each scenario and the probability of occurrence of an accident (F) equal to or greater than the death number (N) are graphed; And comparing the scenario-specific FN curve with the reference value FN curve, and diagnosing that the scenario-specific FN curve has a high risk of fire for a data section having a higher value than the data interval of the reference value FN curve. do.

According to a preferred feature of the present invention, the event tree, based on the tunnel information and the vehicle information to branch into a plurality of scenarios for each type of accident fire that can occur when a virtual fire occurs, from the cause of the fire and the cause of the fire The event event is generated as a hierarchical node, and the node has a node value defining a probability of occurrence of the event event, and the probability of occurrence of an accident f of each scenario is quantitative by the node value quantified while moving the tree. It is characterized by being predicted and analyzed.

Furthermore, the event tree may include a root node to which the accident occurrence frequency (MF) is assigned; A fire originating node branching from the root node and including a derailment-collision, a derailment, a collision, a mechanical defect and a fire protection; And an event event node which sequentially branches from the fire occurrence node and includes an event event node including fire detection, fire extinguishing, driver identification, announcement, stop in tunnel, and control stop. Reaching an event event node and branching into a plurality of unique scenarios.

Preferably, the step (S14), (S141) generating the accident occurrence probability (f) of the individual scenario using the node value of the individual node in the process of branching on the event tree (MF) It is characterized by including.

Further, the step (S14), (S142) for each vehicle in the individual scenarios by the difference between the allowable evacuation time (Aset) and the required evacuation time (Rset), the PDF (Probability Density) of the survival discrimination function for each vehicle Function) (G _pdf : G <= 0 (death), G> 0 (survival)).

Further, the allowable evacuation time (Aset) is the available time allowed for the evacuation of passengers, the time when the temperature is 80 ℃ or more; Time to be 10 m or less in visibility; And the shortest time among the times when the number Fr is 0.9 or less.

Furthermore, the required evacuation time Rset is an available time required for evacuation of the passenger, and includes a detection time after the occurrence of the fire until the passenger or the engine detects the fire; Reaction time after determining the evacuation behavior after the fire detection and before the actual evacuation movement; And it is characterized in that the time including the total of the travel time to escape the vehicle to the safe evacuation site.

Preferably, the PDF of the survival discrimination function (G _pdf ) is generated by using (combined) various distribution values of the allowable evacuation time (Aset) and the required evacuation time (Rset) using Monte Carlo simulation considering the uncertainty. It features.

Furthermore, (S143) G _cdf = in the CDF of the survival discrimination function for each scenario and vehicle through a CDF (Cumulative Density Function) transformation that accumulates the Y-axis probability value for the x-axis G value of the PDF of the survival discrimination function (G _pdf ). And generating a death probability D for each vehicle and a scenario when 0.

Furthermore, (S144) the number of deaths for each vehicle is obtained using the number of passengers per vehicle corresponding to the death probability D, and the number of deaths of individual scenarios Sn corresponding to the sum of the number of deaths per vehicle (Sn) And N).

Furthermore, in the (S145) individual scenario (Sn), data Sn (f, N) paired with the probability of death (D) and the number of deaths (N) is sorted by the number of deaths, and then the death probability (D) is accumulated. And generating a probability (F) of accident occurrence more than the number (N).

In the risk analysis method of the railway tunnel fire according to the present invention, in order to analyze the fire risk of the railway tunnel, the data such as historical statistics, railway information, train information, fire type and judgment indicators are quantified according to the fire type scenarios in the tunnel. Determine the risk of fire by estimating the number of deaths and the probability of an accident occurring.

Hereinafter, with reference to the accompanying drawings looks at in detail the configuration of a preferred embodiment of the present invention.

<1. Technology concept>

1 illustrates historical statistical data in accordance with an embodiment of the present invention.

In the present invention, in order to analyze the risk of a railway tunnel fire, reference is made to historical statistical data including the distance traveled and the number of accidents by year of the past railway vehicles. 1 illustrates a case where a user writes historical statistical data in a text file. Row 1 represents data for the past seven years. Line 2 represents the individual year. Line 3 represents the total train km by year. Line 4 is the number of passengers in the year. Line 5 is the number of accidents by year. Line 6 is the total distance traveled by three lines divided by the number of accidents in five rows, representing the number of incidents per train km.

In the present invention, it is possible to predict the occurrence frequency (MF: Master Frequency), which means the number of accidents that can occur in the tunnel of the analysis target using the historical statistical data. Equation 1 below defines the expression of the accident occurrence frequency (MF).

Accident Frequency (MF) = (Number of Accidents / Total Mileage) X Tunnel Length

2 illustrates data of tunnel information and train information according to an embodiment of the present invention.

Tunnel information includes tunnel specifications (e.g. tunnel length, width, height), fire protection facilities (e.g. fire detectors, fans, etc.), and tunnel design information (e.g., rescue stations, evacuation gangs, tunnels) Vertical shafts, etc.);

Train information includes total length, specifications of individual vehicles, number of passengers, etc.

In the fire risk analysis of railway tunnels, the size of fire, fire location, fire growth rate, smoke condition, tunnel width and height, location of rescue station, spacing of evacuation pit, etc. affect the survival rate. A design change of train information is required.

3 illustrates an event tree according to one embodiment of the invention.

In the present invention, in order to analyze the fire risk on the basis of historical statistical information, tunnel information, train information, first predict a possible fire scenario and analyze the fire risk for each scenario. In the present invention, the analysis of the fire risk is to analyze the number of deaths (N) and the probability of occurrence of an accident (f) that may occur in each scenario.

The event tree of the present invention branches to possible fire scenarios based on past fire occurrence types and probabilities.

The event tree is composed of a plurality of nodes, and is composed of a fire occurrence node branched from a root node and an event event node sequentially hierarchically branched from a fire occurrence node. Each individual node has a node value that defines an event probability. Node values of individual nodes are predefined as probability values by quantitatively analyzing past fire occurrences and subsequent event events.

In the present invention, a fire occurrence node is composed of a derailment-collision (fire caused by a collision with a derailment), a derailment, a collision, a mechanical defect and a fire branch, and branches from a root node. Of course, the sum of the probability values of the individual fire originating nodes is 1.

The event event node is composed of event event nodes including hierarchical and sequential branching of fire detection, extinguishment, driver identification, announcement or not, stop in tunnel, and control stop. The sum of all probabilities that can occur in an individual event event is one.

The number of end nodes in the event tree refers to the number of scenarios of the present invention, and in the process of branching from the root node to the end node, the probability of occurrence of an accident (Sn (f)) is determined for each scenario and the occurrence of an accident is generated from the event tree. The number Sn (mf) can be predicted. Equation 2 below defines the formula of the probability of occurrence and the number of accidents in each scenario.

Individual scenario incident probability ( Sn (f)) = individual scenario ( Sn ) Included in the path furnace Product of devalues

Individual scenario incidents ( Sn (mf)) = total incident frequency (MF) X probability of incident ( Sn (f))

(By scenario Sn The sum of (f) is 1, for each scenario Sn The sum of (mf) is MF.)

When the probability of accident occurrence (Sn (f)) of each scenario is obtained, the number of deaths is calculated for each vehicle of each scenario. The calculation of the number of deaths uses the probability distribution function PDF (Probability Density Function) of the survival discrimination function G. Equation 3 below defines the probability distribution function G _pdf of the survival determination function G.

G _pdf  = Allowable evacuation time ( Aset )-Required evacuation time ( Rset )

Aset  Available safety egress time

Rset  : Required safety egress time

G _pdf  <= 0: death, G _pdf  > 0: Survival

In Equation 3 above, the allowable evacuation time (Aset) is the available time allowed for the passenger to evacuate, which is the lowest of the following three values.

i) the time at which the temperature becomes 80 ° C or higher;

ii) the time when the visibility is less than 10 m

iii) Time to be Froude 0.9 or less

The required evacuation time Rset is an available time required for the passenger to evacuate, and includes a detection time + reaction time + movement time.

The detection time is the time after which a passenger or engineer detects a fire. The reaction time is the time taken after determining the evacuation behavior after the fire detection and before the actual evacuation movement begins. Travel time is the time to escape a vehicle to a safe evacuation site.

For example, the required evacuation time (Rset) in the case of a stationary fire is as follows. Assuming a passenger or engineer detects a fire, the detection time follows a lognormal distribution. Assuming no passenger or engineer detects, apply a uniform distribution between 60 and 180 seconds. The reaction time is zero if the fire is extinguished on site. In the event of indigestion and emergency evacuation broadcasts, the lognormal distribution (30, 20, 100, 10) is followed. Follow log normal distributions (100, 130, 800, 30) if not digested on site and not even emergency evacuation broadcast. Travel time is the sum of the time it takes a passenger to leave the vehicle and the time it takes to travel from the vehicle exit to a safe evacuation site.

As another example, in the case of a running fire, a breaking time is further added to the above-described necessary evacuation time Rset. The reason why the braking time is further added is that unlike the fire after the above-mentioned accident, a fire occurs during operation and it takes time for the engineer to stop the vehicle. Of course, the case where the vehicle can exit the tunnel is excluded from consideration. In other words, the braking time refers to the time it takes for the engineer to stop the train running when the train vehicle is aware of a fire. Here, if the engineer stops at the desired stop point immediately after detecting the fire, only the general braking time is taken into account. If the driver can stop at the desired stop point, the braking time is included in the braking time. Should be.

FIG. 4 shows a probability distribution function PDF (Probability Density Function) graph of the survival discrimination function G obtained in individual scenarios and individual vehicles using Equation 3. FIG.

In the present invention, Monte Carlo simulation using uncertainty is used to derive the probability distribution function of Equation 3. Monte Carlo simulation generates a probability distribution function (PDF) using any random combination of these distribution values to take into account the various distribution values of the plurality of factors constituting the allowable evacuation time (Aset) and the required evacuation time (Rset). If the value of G _{pdf on the} X axis exceeds 0 (G _pdf > 0), the passenger survives because of the allowable evacuation time (Aset).

FIG. 5 illustrates a cumulative density function (CDF) of the survival discrimination function G obtained by accumulating the Y-axis probability value with respect to the X-axis G value of the probability distribution function (PDF) of FIG. 4.

This is obtained by accumulating the Y-axis probability values when the X-axis G value or more is accumulated. And the probability value D when G _cdf = 0 represents the probability of death in individual scenarios and individual vehicles.

As described above, when the probability of death in the individual scenario and the individual vehicle is obtained, the number of deaths of the vehicle is calculated using Equation 4 below.

Deaths by Vehicle by Scenario = In-vehicle  Number of Passengers X Probability of Death (D)

Deaths by Scenario ( Sn (N)) = total number of deaths per vehicle by scenario

When the number of deaths Sn (N) for each scenario is obtained, Sn (f, N) is generated using the above-described scenario occurrence probability Sn (f). After sorting Sn (f, N) by the number of deaths, Sn (f) is accumulated by generating Sn (f) to define the probability of death of more than N deaths. From this, Sn (F, N) is generated, which is the data of the predictive FN curve.

6 illustrates a case where data of a predicted FN curve is compared with respect to a reference FN curve according to an embodiment of the present invention.

The FN curve defines the probability of occurrence of an accident, F, in which at least N deaths occur, and the reference FN curve is data released for each country in consideration of social risks in the country. Here, social risk is a comprehensive analysis of accidents in various fields such as nuclear facilities, factory facilities, transportation, ports, and aviation. Compared to the predicted FN curve, the death rate of the predicted curve is lower than that of the reference curve for the same number of deaths. Therefore, a section with a higher F value than the reference curve is a dangerous section, and a disaster prevention facility or an evacuation facility is required through a tunnel design change.

Hereinafter, a risk analysis system to which the above-described technical concept of the present invention and a risk analysis method thereof will be described.

<2. System Configuration>

7 shows a schematic configuration of a risk analysis system 1 of a railway tunnel fire according to an embodiment of the present invention.

The risk analysis system 1 according to the present invention includes historical statistical information input means 11 for receiving fire information generated during past railway operation and storing statistical data, and analysis target information for receiving and storing tunnel information of a risk analysis target. The input means 12, the accident occurrence frequency predicting means 13 for predicting the occurrence frequency (MF) of the analysis target in the tunnel by using the historical statistics information, predicting the possible scenarios in the tunnel and in each scenario Per-scenario-specific death prediction means 14 for predicting the number of deaths and the probability of death Sn (F, N) above the number of deaths per scenario, and for each scenario for generating a scenario-specific FN curve using the predicted Sn (F, N) And a fire risk diagnosis means 16 which analyzes and diagnoses the risk of the tunnel by comparing the FN curve generating means 15 and the reference FN curve with the predicted FN curve.

FIG. 8 shows a detailed configuration of the death prediction means 14 for each scenario of FIG.

The scenario-specific death prediction means 14 according to an embodiment of the present invention includes an accident occurrence probability generating means 141 which predicts an accident occurrence probability Sn (f) in each scenario, and survival determination for each vehicle of each scenario. Survival determination function PDF generating means 142 for generating a PDF distribution of the function, death probability generating means 143 for converting the PDF distribution to a CDF distribution and generating a probability of death (D) in the individual vehicles of the individual scenarios, and the probability of death (D) and the number of deaths generating means 144 for predicting the deaths of the individual scenarios using the number of passengers of the individual vehicles and the probability of accident occurrence (Sn (f)) for the deaths of the individual scenarios accumulated by And an accident occurrence probability generating means 145 for predicting the accident occurrence probability Sn (F).

The detailed functions and operations of the individual components constituting the above-mentioned railway tunnel fire risk analysis system 1 will be described through the railway tunnel fire risk analysis method described below.

<3. How to Configure>

The risk analysis method of the railway tunnel fire according to an embodiment of the present invention may be preferably realized through the construction of the system 1 described above.

9 shows a schematic configuration of a risk analysis method of a railway tunnel fire according to an embodiment of the present invention.

The historical statistical information input means 11 receives historical railroad operation data and stores it as statistical information for risk analysis on the tunnel of the analysis target (S11). Historical rail service data includes the annual distance traveled and the number of accidents for all rail vehicles.

The analysis target information input unit 12 receives and stores tunnel information and train information of the risk analysis target (S12). Tunnel information includes specifications such as length, width, height, and data of evacuation and disaster prevention facilities. The train information includes the number of vehicles, specifications of the vehicles, the number of passengers per vehicle, and the like.

1 and 2 described above show an example of a data batch input for creating a data file and inputting a file in advance, but the user may input the screen according to a screen guide using a GUI interface. If the tunnel information is determined to be dangerous according to the result of the risk analysis, safety design through design change is required.

When the user starts the risk analysis after all the historical statistics and analysis target data are input, the accident occurrence frequency predicting means 13 calculates the number of accident occurrences per 1 km train distance obtained from the historical data using Equation 1 The frequency of occurrence of an accident (MF) in the tunnel length is calculated using S13. The frequency of accidents is an estimate of the number of accidents in one year.

Subsequently, the scenario-specific death prediction means 13 predicts a plurality of fire scenarios that may occur for the tunnel of the analysis target, and the number of deaths in each scenario and the probability of occurrence of accidents in which the deaths occur (Sn (F, N)). ) Is generated (S14).

When the step S14 is classified in detail, the accident occurrence probability generating unit 141 visits the event tree and branches into a plurality of scenarios, and the event occurrence probability constituting the individual scenarios defined in Equation 2 is multiplied to cause accidents for each scenario. The occurrence probability Sn (f) is predicted (S141).

The event tree is a binary tree pre-built from past fire types, as already described with reference to FIG. 3, which causes fire originating nodes (derailment-collision, derailment, collision, mechanical defects and arson) branching from the root node and individual fire occurrence causes. It consists of event event nodes (fire detection, extinguishment, engine driver, announcement, stop in tunnel and control stop) that sequentially define fire events generated from the node. That is, in the event tree of FIG. 3, there are five fire event nodes that can be branched from the root node, and six event event nodes sequentially arranged from individual fire cause nodes. Once a branch of an event tree is made, thereafter there is no need to go through the tree unless the probability value corresponding to the individual node value changes, and then store the probability of occurrence of the event (f) of the individual scenario and use the stored value. It is possible.

On the other hand, the survival determination function PDF generating means 142 generates a PDF (probability distribution function) of the survival determination function by using Equation 3 for the individual vehicles of the individual scenarios (S142). This PDF has already been described as a probability distribution function obtained by combining the distribution values of the various individual factors constituting G _pdf using Monte Carlo simulation.

After G _pdf is obtained, the death probability generating means 143 performs a CDF (Cumulative Density Function) conversion that accumulates the probability values of the Y-axis with respect to the X-axis G value with respect to the G _pdf , and then calculates the CDF of the survival discrimination function for each scenario and vehicle. Death probability D for each scenario and vehicle when G _cdf = 0 is generated (S143). That is, D defines the probability of death of passengers in individual vehicles in individual scenarios.

After the probability of death D is obtained, the number of deaths generating means 144 multiplies D by the number of passengers of the individual vehicles using Equation 4 to obtain the number of deaths of the individual vehicles and add the number of deaths of the total vehicles to the number of deaths of the individual scenarios ( Sn (N)) is generated (S144).

If Sn (N) is predicted, the accident occurrence probability generating means 145 generates an accident occurrence probability of death of N or more for the number N of deaths in the individual scenario (S145). To this end, a list Sn (f, N) is generated from the probability of occurrence of an accident (f) of the individual scenario of step S141 and the number of deaths (N) in step S144, and the deaths are arranged in the order of the number of deaths (N). In order to calculate the probability of the number N or more, a corresponding accident occurrence probability f is accumulated to generate Sn (F, N) (S145). That is, Sn (F, N) defines the probability of occurrence of an accident (F) in which more than N deaths occur in individual scenarios.

After Sn (F, N) is obtained, the scenario-specific FN curve generating means 15 generates a scenario-specific FN curve in which a death probability N of the number of deaths N and the number of deaths N or more is expressed in a graph (S15).

Then, the fire risk diagnosis means 16 compares the predicted FN curve and the reference value FN curve for each scenario, and diagnoses that the risk of fire is high for the data section in which the predicted FN curve has a higher value than that of the reference value FN curve (S16). Here, it is possible to guide the design change by presenting an input screen of the corresponding tunnel information to the user for the individual scenario determined as the danger section. Of course, it is also possible to adjust the risk of fire by changing the design of the train information.

As described above, an embodiment of the risk analysis method of the railway tunnel fire according to the present invention is configured. Although the present invention has been described by way of limited embodiments and drawings, the present invention is not limited thereto, and the technical spirit of the present invention and the claims to be described below by those skilled in the art to which the present invention pertains. Of course, various modifications and variations are possible within the scope of equivalents.

The following drawings, which are attached to this specification, illustrate preferred embodiments of the present invention, and together with the detailed description of the present invention serve to further understand the technical spirit of the present invention. It should not be construed as limited to.

1 is an exemplary diagram of historical statistical data according to an embodiment of the present invention.

2 is a diagram illustrating data of tunnel information and train information according to an embodiment of the present invention.

3 is an exemplary diagram of an event tree in accordance with one embodiment of the present invention.

4 and 5 are exemplary diagrams of a survival determination function according to an embodiment of the present invention.

6 illustrates an FN curve in accordance with an embodiment of the present invention.

Figure 7 is a schematic diagram of a risk analysis system of a railway tunnel fire according to an embodiment of the present invention.

8 is a schematic diagram of a scenario-specific death prediction means according to an embodiment of the present invention.

9 is a schematic flowchart of a risk analysis method of a railway tunnel fire according to an embodiment of the present invention.

Claims (11)

(S11) receiving and storing historical statistical data including a yearly driving distance and the number of accident occurrences of the entire railway vehicle from the user; (S12) receiving and storing the tunnel information and the vehicle information of the analysis target from the user; (S13) estimating an accident occurrence frequency (MF: Master Frequency) for the corresponding tunnel length after calculating the number of accident occurrences (count of accidents / train km) per unit driving distance using the historical statistical data; (S14) branching the frequency of accidents into a plurality of predefined scenarios through an event tree defining an accident fire type, and the number of deaths (N) of individual scenarios (Sn) and the probability of occurrence of an accident more than the number of deaths (N) (F) Generating (Sn (F, N)); (S15) generating a scenario-specific FN curve in which the death number (N) of each scenario and the probability of occurrence of an accident (F) equal to or greater than the death number (N) are graphed; And (S16) comparing the scenario-specific FN curve and the reference value FN curve to diagnose that the risk of fire is high for the data section in which the scenario-specific FN curve has a higher value than that of the reference value FN curve. Rail tunnel fire risk analysis method comprising the. The method of claim 1, The event tree, Branching into a plurality of scenarios for each type of accident fire that may occur when a virtual fire occurs based on the tunnel information and the vehicle information, An event event generated from each fire occurrence cause and a fire event are formed into hierarchical nodes, and each node has a node value defining a probability of occurrence of the event event, and is individually determined by the node value quantified while moving the tree. A risk analysis method for a railway tunnel fire, characterized in that the probability of occurrence of an accident in a scenario (f) is quantitatively predicted and analyzed. The method according to claim 1 or 2, The event tree, A root node to which the accident occurrence frequency (MF) is assigned; A fire originating node branching from the root node and including a derailment-collision, a derailment, a collision, a mechanical defect and a fire protection; And An event event node which sequentially branches from the fire causing node and includes fire detection, fire extinguishing, driver identification, guide broadcasting, stop in tunnel, and control stop. And a branching into a plurality of unique scenarios by reaching an event event node of the lowest level from the root note. The method of claim 1, The step (S14), In operation S141, the occurrence frequency f of the railway tunnel fire is generated using the node values of the individual nodes in the process of branching on the event tree. Risk analysis method. The method according to claim 1 or 4, The step (S14), (S142) Probability Density Function (PDF) of the survival discrimination function for each scenario and vehicle according to the difference between the allowable evacuation time (Aset) and the required evacuation time (Rset) for each vehicle of the individual scenario (G _pdf : G <= 0 (Death), G> 0 (survival)). The method of claim 5, The allowable evacuation time (Aset), Available time allowed for passengers to evacuate, Time to become the temperature of 80 degreeC or more; Time to be 10 m or less in visibility; And Time to become Froude 0.9 or less The risk analysis method of the railway tunnel fire, characterized in that the shortest time. The method of claim 5, The necessary evacuation time (Rset), Available time required for evacuation of passengers, Detection time after a fire occurs until a passenger or engineer detects a fire; Reaction time after determining the evacuation behavior after the fire detection and before the actual evacuation movement; And Travel time to escape vehicle to safe evacuation site The risk analysis method of a railway tunnel fire, characterized in that the time including the sum of. The method of claim 5, PDF (G _pdf ) of the survival discrimination function, A risk analysis method for a railway tunnel fire, which is generated by using (combination) various distribution values of allowable evacuation time (Aset) and required evacuation time (Rset) using Monte Carlo simulation considering uncertainties. The method of claim 5, (S143) G _cdf = 0 days in the CDF of the survival discrimination function for each scenario and vehicle through the CDF (Cumulative Density Function) transformation that accumulates the Y-axis probability value for the x-axis G value of the PDF of the survival discrimination function (G _pdf ). A risk analysis method for a railway tunnel fire, comprising the step of generating a death probability (D) for each vehicle scenario. The method of claim 9, (S144) calculating the number of deaths for each vehicle by using the number of passengers per vehicle corresponding to the probability of death (D), and obtaining the number of deaths (N) of the individual scenario Sn corresponding to the sum of the number of deaths per vehicle (Sn). Rail tunnel fire risk analysis method comprising the. The method of claim 1, (S145) After sorting data Sn (f, N) paired with the probability of death (D) and the number of deaths (N) for each scenario (Sn) by the number of deaths, accumulating the probability of death (D) and the number of deaths (N). The risk analysis method of the railway tunnel fire, characterized in that it comprises the step of generating a probability (F) of accident occurrence.

Images