Big fire has often occurred since the dawn of history in Japan. And Japan has the world worst record even in the scale. The big fire that building burn-out area is over 33,000 m
2 in the 20's of Showa has occurred numerously. After that we were toward one goal called "the fire prevention of a city". As a result, big fire has not been breaking out since the big fire of Sakata in Showa 51 (1976). And the fire prevention measure has moved to the building fire prevention of large-scale/upper layers buildings from the big fire prevention of city area since this time. However, the occurrence of big fire has not been eradicated at all. The danger of big fire occurrence is still remained under the unfavorable conditions like a large earthquake and a strong leeward. Thereupon we analyzed about the fire spread formula of wooden building both in ordinal and earthquake disaster.
The conclusion is as follows.
(1) We changed the fire spread formula as the expression of K city in eq. (3) from eq. (1) with an easier style. It may express in the eq. (3) from in eq. (1).
(2) Fires of K city and M city are shown to Fig. 4 as an example under G=100m
2, v=3 m/sec. If each burn-out area is compared, it is understood that the structure of M city after about 30 years is difficult burning than the building fire of K city.
(3) Applying eq. (3) about the big fire from Showa 21 to Showa 27, Table 2 was obtained. Coefficients a
1 and c
1 are able to indicate each eq. (8) and eq. (9). The correlation figures of measurement value and prediction values of all fire data of Table 2 were shown to Fig. 9. As multiple coefficient of correlation R=0.983, it is agreeing fairly well.
(4) It was shown like Table 4 and also Fig. 10 about the Sakata big fire. It is able to indicate in eq. (3) and R=0.999. It is thought that the Sakata big fire became difficult to burn more previous big fire of the Showa 20's so that shown in Fig. 11.
(5) About the expression at the time of the Kobe earthquake, we re-examined data a part once again. a
1 and c
1 are able to indicate each eq. (11) and eq. (12). a
1 is almost not change the previous paper of ours.
The relation c
1 and x
G is eq. (13) and R=0.899. The correlation between burn-out area of prediction value and measurement is about 0.9. It is agreeing fairly well.
View full abstract