JP2002296362A - System and method of evaluating damages due to wind by typhoon composited using stochastic technique - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately and precisely estimate wind damages by a future typhoon by using a stochastic technique. SOLUTION: On the basis of statistically analyzed data on the atmospheric pressure field of a past typhoon, a stochastic distribution model, which gives parameters used for deciding the atmospheric pressure field of the typhoon, is constructed, Monte Carlo simulation is executed, and many imaginary typhoons are composited stochastically. In the composited imaginary typhoons, a condition is changed, and wind damages which are estimated, regarding many cases which are not comparable with conventional cases, can be evaluated. A stochastic model, which can actually execute the evaluation of the wind damage, is constructed as a computer system. When the computer system is used, a wind disaster premium rate which is more rational than that in conventional cases is calculated, and a maximum estimable damage amount can be estimated with higher accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system and a method for evaluating wind damage caused by a typhoon. More specifically, the present invention relates to a system and method for evaluating wind damage expected to be caused by a typhoon synthesized using a probabilistic method based on data on past typhoons, particularly in relation to non-life insurance designs. .

[0002]

2. Description of the Related Art When a typhoon strikes, very serious natural disasters, including storm and flood damages, can occur. Therefore, preventive measures are taken from various viewpoints. As a long-term measure, it is necessary to consider the maximum instantaneous wind speed during a typhoon when calculating the strength of buildings and bridges as a physical precautionary measure. It is necessary to establish a system and design an appropriate non-life insurance system. In the shorter term, it is also necessary to instantaneously predict the maximum instantaneous wind speed in a region where there is a possibility of an attack when a typhoon is actually approaching.

[0003] Non-life insurance is a system in which the danger of accidental wind and flood damage caused by, for example, a typhoon is distributed among a large number of individuals who have paid a certain premium in advance. Even if a natural disaster is an event that is completely accidental and difficult to predict for each individual, the empirical rule is that when considering a large number of people as a whole, the percentage of those who actually encounter the event approaches a certain value. Known (large number law). However, in order to satisfy the law of large numbers, it is necessary to consider a sufficiently large group, and in that case, a statistical and stochastic method is used. In particular, when properly designing property and casualty insurance, it is important to ensure that each policyholder has a maximum amount of insurance (A) that can be received if a policyholder suffers an economic loss due to a disaster. It is important to reasonably determine the amount of premium (B) to be paid in advance by the policyholder.
In other words, when designing property and casualty insurance, it is necessary to determine how much premium must be collected from each policyholder to pay a particular claim. Here, the ratio of the insurance premium to the insurance amount (= B /
A) is called the insurance rate.

[0004] In order to accurately calculate this insurance rate,
Generally, a large amount of stable data is required. It is also important in the non-life insurance business to assess the maximum possible loss. However, natural disasters such as wind disasters caused by typhoons fluctuate greatly from year to year, so it is difficult to obtain necessary data by short-term observation, and long-term observation is required. In the past, past damage was estimated using past statistical data, such as the wind speed observed during a past actual typhoon.

For example, in Japanese Patent Application Laid-Open No. 10-161993,
There has been proposed a system for evaluating typhoon damage using past weather observation data. In this system, the damage is evaluated by using the past typhoon information as it is, or by slightly changing the parameters such as the course and power in the past typhoon information, assuming the use of a personal computer. .

The present applicant has also filed Japanese Patent Application No.
In 92434, in order to solve this problem, a system and method for evaluating a typhoon-induced wind damage that can be predicted in the future by using a computer simulation based on objective data on past typhoons was proposed. The purpose of this system and method is to accurately and accurately predict wind damage caused by a typhoon in the future even when past statistical data is insufficient.

[0007]

SUMMARY OF THE INVENTION The applicant has filed a Japanese Patent Application 200
In the system and method proposed in 1-92434, based on the data on the past typhoon, and by reproducing the typhoon using computer simulation, what kind of typhoon is substantially the same as or similar to the previous one. Rationally evaluated whether a severe wind damage could occur.

[0008] However, statistical data on past typhoons has been maintained for at most 60 years at most, and even if past typhoons are reproduced, damages due to actual past typhoons and virtual typhoons similar thereto are limited to some extent. Although evaluation can be performed, it is difficult to evaluate a typhoon that has never occurred before and has a reappearance period longer than the observation period, although there is a possibility of occurrence. In addition, there were typhoons that did not cause any damage because they happened to pass through the route, but if the route deviated even slightly, there was a possibility that severe damage could occur. But,
With the conventional method, for example, when a different typhoon hits beyond the range that can be directly inferred from past data, it is possible to accurately determine what kind of wind damage will occur in a specific area It is difficult to predict.

[0009]

Means for Solving the Problems The inventors of the present invention set the pressure field of the past typhoon in order to suppose an accurate damage situation for various typhoons expected to hit in the future. We constructed a probability distribution model that gives parameters to determine the pressure field of the typhoon from the statistical analysis data, and found that it is effective to stochastically synthesize a number of imaginary typhoons by executing Monte Carlo simulations. By combining many assumed typhoons in Monte Carlo simulation and changing the conditions, it is possible to evaluate the expected wind damage for a large number of cases that are incomparable to the conventional cases. The inventors constructed a probabilistic model capable of actually executing such a wind damage evaluation as a computer system, and by using this computer system, a more reasonable wind disaster By calculating the insurance premium rate, it has become possible to estimate the maximum expected damage amount with higher accuracy.

Specifically, according to the present invention, a typhoon-based wind damage evaluation system using a probabilistic method is provided. Virtual typhoon data synthesis means for synthesizing data relating to virtual typhoons including a large number of virtual typhoon barometric pressure fields and traveling paths to the extent that stable evaluation is possible, and the virtual typhoon data obtained by Monte Carlo simulation Virtual data storage means for storing data, data input means for inputting a plurality of data relating to the virtual typhoon obtained from the virtual data storage means, data storage means for storing the plurality of input data, Using the data obtained from the data storage means, the pressure distribution from the center of the typhoon with time and the A pressure field determining means for determining a pressure field given as a function of the heart of the distance, the pressure field storage means for storing the pressure field obtained by the pressure field determining means,
A gradient wind speed calculating means for calculating a gradient wind speed using the pressure field acquired from the pressure field storing means, a gradient wind speed storing means for storing the gradient wind speed obtained by the gradient wind speed calculating means, and the gradient wind speed storing means To get the gradient wind speed from
Surface wind speed calculating means for calculating the surface wind speed using the gradient wind speed, surface wind speed storing means for storing the surface wind speed obtained by the surface wind speed calculating means, and acquiring the surface wind speed from the surface wind speed storing means, Maximum instantaneous wind speed calculating means for calculating the maximum instantaneous wind speed from the surface wind speed, maximum instantaneous wind speed storing means for storing the maximum instantaneous wind speed obtained by the maximum instantaneous wind speed calculating means, and maximum instantaneous wind speed from the maximum instantaneous wind speed storing means. Acquiring and calculating a morbidity and damage rate from the maximum instantaneous wind speed, and using the morbidity and damage rate obtained by the morbidity and damage rate calculation means to generate the typhoon. A stochastic wind damage evaluation system by a typhoon, comprising: a wind damage evaluation means for performing a wind damage evaluation; and a means for outputting a wind damage evaluation obtained by the wind damage evaluation means. There is provided. In particular, in the system according to the present invention, the data defining the state of the typhoon include:
Four parameters of central pressure drop, maximum rotation wind speed radius, traveling speed and traveling direction are used, and these parameters relating to the virtual typhoon are stochastically reproduced using statistical analysis data of past typhoon pressure fields. What is important is that.

Actually, each of the above-mentioned calculation means is a CPU built in a personal computer or the like, and each storage means is a memory.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, in the system of the present invention whose configuration is shown in FIG. 1, an outline of a procedure for performing wind damage evaluation by a typhoon will be described.

This system uses four types of data: typhoon data, meteorological office data, municipal data, and insurance contract data. Typhoon data is the most recent data from 1932 (Showa 7) (1999 [March 1, 2000]
Typhoon center position (latitude, longitude) for each of the major typhoons up to the current day) and the atmospheric pressure at that center position. Typhoon data will be added each time a typhoon occurs. The meteorological office data is the locations (latitude and longitude) of over 160 meteorological offices nationwide in Japan, and sea level pressures at each time observed by each meteorological office during past typhoons. The municipal data is the position (latitude, longitude) of the municipal government office where there are more than 3,300 nationwide. In the system according to the present invention, the positions of these government offices are used as representative points of municipalities which are units of wind speed calculation. In addition, since the municipalities are used as units in the insurance contract, the number of households is also included in the municipalities data in order to convert the wind speed calculated in municipalities into municipalities. Insurance contract data is data on insurance contracts of insurance companies that calculate insurance claims using the system according to the present invention.
The contents by building structure etc. are described for each city, ward and county.

In the system according to the present invention, the atmospheric pressure field is analyzed for each typhoon using past typhoon data, and parameters representing the atmospheric pressure field are calculated. Of the typhoons analyzed,
The typhoon that landed on the coastline defined in this system is subject to statistical analysis. As shown in FIG. 2, the coastline is divided into four areas by several straight lines on the Pacific coast of the Japanese archipelago where the typhoon lands. The reason for such area division is that when considering the characteristics of a real typhoon, the influence of the actual terrain cannot be ignored.

As a result of statistical analysis of past typhoon data, the parameters for describing the barometric pressure field centered on the typhoon are as follows: It has been empirically found that can be expressed by the function form in Table 2. For example, in the case of a probability distribution describing parameters at the time of landing, the system according to the present invention uses a Poisson distribution widely used in a discrete data stochastic model for the annual number of landings of a typhoon. Further, regarding the amount of decrease in the central pressure, the maximum rotation wind speed radius, the traveling speed, and the traveling direction, accurate description can be made by using a lognormal distribution. However, the present invention is not limited to these particular probability distributions, and there are other options as to which specific probability distribution to use.

[0016]

[Table 1]

[0017]

[Table 2] In the system according to the present invention, for the parameters in Tables 1 and 2, random numbers are generated in accordance with statistical analysis results (mean values, standard deviations) using respective probability distributions and function forms, and Monte Carlo simulation is performed. Create 10,000 years of virtual typhoon data (from landing to disappearance). As a result, about 25,000 virtual typhoons can be synthesized. Specifically, for such a large number of typhoons that are considered to be sufficient to provide stable statistics, a pressure field and a path over time are obtained. In order to place more importance on the stability of the results, it is desirable to create several sets of 10,000 years if possible. FIG. 3 shows a flowchart when creating a virtual typhoon.

Based on the parameters created for each typhoon at each time, the wind speed at each city, town, and village at that time is calculated. The wind speed at a certain time at a certain point during a typhoon is
Basically, it can be obtained from the state of the typhoon pressure distribution, the speed of the typhoon, the distance from the center of the typhoon to the calculation point, and the like.
Here, the maximum wind speed among the wind speeds calculated for each time is taken as the wind speed in the municipalities at the time of the typhoon. on the other hand,
An equation for calculating the relationship between wind speed and damage is prepared, and the damage rate for each municipality is calculated from the calculated wind speed. The amount of the loss is obtained by applying this to an insurance contract. Further, by arranging the damage amount for each typhoon in descending order, it is possible to perform a stochastic evaluation of wind damage caused by the typhoon.

In order to calculate the pressure field of the typhoon, that is, the parameter indicating the state of the pressure distribution, (1) the position of the typhoon, (2) the central pressure, (3) the position of each meteorological office, and (4) Four data, the observed sea level pressure, and the Schloemer pressure distribution formula (below, which is considered to be most suitable for objective analysis of the pressure field)

The pressure distribution in each typhoon is estimated by using (1). In FIG. 4, a curve giving the distribution of the atmospheric pressure in a certain typhoon is estimated as a function of the distance from the typhoon center. This shows the pressure distribution of the typhoon

An equation for estimating the atmospheric pressure distribution is created for each typhoon for each typhoon using (Equation 1) (Schloemer's equation). As illustrated in FIG. 4, an equation indicating the pressure distribution from the center of the typhoon to the outer edge is obtained.

[0020]

(Equation 1) Determine the probability distribution of each parameter of the typhoon pressure field obtained in this way (number of landings per year, central pressure drop, maximum rotation wind speed radius, traveling speed, traveling direction, etc.) and generate random numbers according to it A virtual typhoon is created.

Next, the wind speed at each point is determined. FIG.
5 shows the positional relationship of wind speeds to be obtained, such as a gradient wind speed, a surface wind speed, and a maximum instantaneous wind speed. The gradient wind speed, which is the wind in the sky, can be calculated by a physical calculation (see below).

(Equation 2)). Further, the calculated gradient wind speed is converted into a surface wind speed in the case of being affected by the ground surface (

(Equation 3)). Finally, from the surface wind speed obtained in this way, the maximum instantaneous wind speed that is closely related to building damage is obtained (

(Equation 4)).

[0022]

(Equation 2)

[0023]

(Equation 3)

[0024]

(Equation 4) FIG. 6 shows the relationship between the point where the wind speed is calculated and the typhoon. Since the parameters for describing the typhoon are functions of time, the wind speed is calculated for each time. The wind speed at the calculation target point is calculated based on the relationship between the point at which the wind speed is determined and the distance from the typhoon center. In the system according to the present invention, the wind speed is obtained at each time (for example, every 20 minutes), and the maximum value among the wind speeds is set as the wind speed at the point in the typhoon. In this system, the local government office position is used as the wind speed calculation point. A part of the wind speed obtained in this way is shown in FIG.
Is shown in

Next, the damage rate to the wind speed is determined. The damage rate is divided into the damage rate, which is the average number of damages to a certain group at a certain wind speed, and the damage rate, which indicates the average damage that would occur if a damage was received. Can be In this system, when calculating the morbidity rate and damage rate from the wind speed, an equation created based on the damage data of Typhoon 19 in 1991 (below)

## EQU5 ## and

Equation 6) is used. The constants included in these formulas vary depending on the type of insurance, building structure, total loss, half loss, and partial loss. 8 and 9 are graphs showing the relationship between the maximum instantaneous wind speed and the morbidity and damage rates. It should be noted that the constants in these formulas will be reconsidered using damage data that can be accumulated in future typhoons.

[0026]

(Equation 5)

[0027]

(Equation 6) Finally, a damage assessment is performed. The damage rate is calculated for each municipality by insurance type, building structure, and damage level (total loss, half loss, partial loss), and multiplying this by insurance contract data to calculate insurance claims paid. The insurance claims are calculated for each of the past typhoons, summed up, and then divided by the observation period to calculate the annual average insurance claims, and basic data for calculating the insurance premium rate is obtained. Further, by combining the path and scale of the past typhoon, damage can be evaluated for a virtual typhoon. Thereby, it is possible to calculate the wind disaster insurance rate in consideration of the abnormal disaster.

[0028]

As described in detail above, according to the system and method according to the present invention, a virtual potential that may occur in the future far exceeds the range directly predicted from past statistical data. Wind damage caused by a typhoon can be evaluated effectively and accurately. Since a large number of assumed typhoons are synthesized by Monte Carlo simulation using a stochastic method and the wind damage is evaluated,
It is possible to evaluate under much more various conditions compared to the wind damage prediction according to the prior art. Therefore, by using the wind damage prediction obtained by the present invention for various wind damage measures including bridge construction and damage insurance design, it is possible to take more appropriate measures than a conventional wind damage prediction system.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a configuration of a wind damage evaluation system according to the present invention.

FIG. 2 shows an area division in which the shoreline of the Pacific coast of the Japanese archipelago where a typhoon hits is divided into four areas by straight lines.

FIG. 3 shows a flowchart when a virtual typhoon is created by Monte Carlo simulation.

FIG. 4 is a graph showing a pressure distribution in a certain typhoon;
It shows how it is estimated as a function of the distance from the typhoon center.

FIG. 5 shows a positional relationship between wind speeds to be obtained, that is, a gradient wind speed, a surface wind speed, and a maximum instantaneous wind speed.

FIG. 6 shows a relationship between a point where a wind speed is calculated and a typhoon.

FIG. 7 illustrates a part of the wind speed obtained by using the position of a government office in a municipality as a wind speed calculation point.

FIG. 8 is a graph showing the relationship between the maximum instantaneous wind speed and the morbidity rate.

FIG. 9 is a graph showing the relationship between the maximum instantaneous wind speed and the damage rate.

FIG. 10 shows a stochastic evaluation of wind damage caused by a typhoon in which damages by typhoon are arranged in descending order of damage amount.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Masaaki Kawaguchi 2-31-19 Shiba, Minato-ku, Tokyo Within the non-life insurance rate calculation meeting (72) Inventor Takeshi Ken Fujii 36 Kamigamomotoyama, Kita-ku, Kyoto-shi, Kyoto Kyoto Sangyo University

Claims (8)

[Claims] 1. A typhoon-based wind damage evaluation system using a stochastic method, comprising:
Virtual typhoon data synthesis means for synthesizing data relating to virtual typhoons including a large number of virtual typhoon barometric pressure fields and traveling paths to the extent that a statistically stable evaluation is possible by executing a simulation; Virtual data storage means for storing the data relating to the virtual typhoon obtained by simulation; data input means for inputting a plurality of data relating to the virtual typhoon obtained from the virtual data storage means; A data storage unit for storing a plurality of data; and a pressure for determining a pressure field that gives a pressure distribution from the center of the typhoon as a function of time and a distance from the center using the data obtained from the data storage unit. Field determination means; pressure field storage means for storing the pressure field obtained by the pressure field determination means; A gradient wind speed calculation unit that calculates a gradient wind speed using a pressure field acquired from a pressure field storage unit, a gradient wind speed storage unit that stores a gradient wind speed obtained by the gradient wind speed calculation unit, and a gradient wind speed storage unit. A surface wind speed calculating unit that obtains a gradient wind speed and calculates a surface wind speed using the gradient wind speed; a surface wind speed storing unit that stores a surface wind speed obtained by the surface wind speed calculating unit; and a ground surface from the surface wind speed storing unit. A maximum instantaneous wind speed calculating means for acquiring a wind speed and calculating a maximum instantaneous wind speed from the surface wind speed; a maximum instantaneous wind speed storing means for storing a maximum instantaneous wind speed obtained by the maximum instantaneous wind speed calculating means; and Obtain the maximum instantaneous wind speed from the means,
Damage rate and damage rate calculation means for calculating the damage rate and damage rate from the maximum instantaneous wind speed, and using the damage rate and damage rate obtained by the damage rate and damage rate calculation means, wind damage evaluation that can be caused by the typhoon And a means for outputting a wind damage evaluation obtained by the wind damage evaluation means. 2. The system according to claim 1, wherein the plurality of input data are: a position of the typhoon including latitude and longitude, a central pressure of the typhoon, a position of each meteorological office,
And a measured sea level pressure, wherein the pressure field is determined using a Schloemer pressure distribution equation. 3. The system according to claim 1, wherein, during the Monte Carlo simulation, the annual number of landings of the typhoon is represented as a Poisson distribution,
The parameters of the virtual typhoon to be synthesized, the central pressure drop, the maximum rotation wind speed radius, the traveling speed, and the traveling direction are expressed as a lognormal distribution at the time of landing, and for the change after landing, the central pressure drop is an index Functionally, the system is characterized in that the maximum rotational wind radius, the traveling speed and the traveling direction are regarded as changing as a linear function. 4. A method for stochastically evaluating wind damage caused by a typhoon, wherein a statistically stable evaluation is possible by executing a Monte Carlo simulation using statistical analysis data of a past typhoon pressure field. Synthesizing a plurality of data relating to the virtual typhoon including as many virtual typhoons as the atmospheric pressure field and the traveling path; and inputting the plurality of data relating to the virtual typhoon obtained by Monte Carlo simulation. Determining a pressure field that gives a pressure distribution around the typhoon as a function of time and a distance from the center by using the plurality of input data; and using the determined pressure field. Calculating the gradient wind speed; calculating the surface wind speed using the calculated gradient wind speed; Using the top wind speed, calculating the maximum instantaneous wind speed, Using the calculated maximum instantaneous wind speed, calculating the morbidity and damage rates, Using the calculated morbidity and damage rates, Performing a wind damage assessment that may be caused by the typhoon; and outputting the wind damage assessment. 5. The method according to claim 4, wherein the plurality of input data are the position of the typhoon, which is composed of latitude and longitude, the central pressure of the typhoon, the position of each meteorological office, and the observed sea level pressure. And determining the pressure field using a Schloemer pressure distribution equation. 6. The method according to claim 4, wherein during the Monte Carlo simulation, the annual number of landings of the typhoon is represented as a Poisson distribution, and the central pressure, which is a parameter of a virtual typhoon to be synthesized. The amount of decrease, the maximum rotation wind radius, the traveling speed, and the traveling direction are expressed as a lognormal distribution at the time of landing, and for the change after landing, the central pressure drop is exponential, the maximum rotation wind radius, the traveling speed. And the direction of travel is assumed to vary as a linear function. 7. A computer-readable storage medium storing a computer program for realizing each step in the method according to claim 4. Description: 8. A computer program for implementing each step of the method according to claim 4. Description: