Large Boilover experiments in Japan Hiroshi Koseki National Research Institute of Fire and Disaster, Japan Tomakomai, Japan.

Large Boilover experiments in Japan

Hiroshi KosekiNational Research Institute

of Fire and Disaster, Japan

Tomakomai, Japan

Abstracts

• In order to know boilover phenomenon in large tanks, large scale boilover experiments in a 5 m diameter pan using crude oil were conducted. Initial fuel thickness was 0.7m.

• To know repeatability we tested twice in the same conditions.

• About 80 minutes after ignition, large boilover was occurred. Radiation from fire increased about twenty times as much as that of steady burnings.

• Times to boilover were within our expecting which we obtained with smaller pans and other reports.

Background

Past boilover accidents in Japan

We had twice boilover accidents in Japan

(1) Ｈｅａｖｙ　ｏｉｌ C, Ｙｏｋｋａｉｃｈｉ，Ｏｃｔｏｂｅｒ　１５ , 1954

(2) Crude oil, Ｎｉｇａｔａ， June 　 1 ６ , 1964

Ｎｉｇａｔａ , June 1964

Earthquake damages in Oil facilities

• In Japan we have many earthquakes in Japan

• Sometimes, oil tanks were damaged, and some were caught fires

• Two Tomakomai fires in 2003 were reported by Mr. Hirai

• Nigata earthquake damaged tanks and continued more than two weeks in 1964, and made boilover in crude oil tanks

Recent Fire Tests in Japan

• Joint work among NRIFD, U. Tokyo and Oil company 20m pan fire test

• Large crude oil fire tests in 20 m pan(1997)

• Boilover tests(1999)

Tests were conducted in

Tomakomai, Hokkaido, Japan

20m pan fire test

Kinds of boilover

• Boilover: Fuel layer thickness is 0.55 m

Clear hot zone (isothermal zone) is made

Our target: 1) normal boilover

2) Effects of water existence

Water emulsion crude oil was burnt• Slopover• Frothover• Thin-layer boilover: Fuel layer is very thin

Mechanism of boiloverHigh temperature zone was made after long bur

n, and then increase its width,Details were shown in many text books

Boilover test #1

Boilover test #2

1m pan boilover tests were done in NRIFD large test facilities using several oils

Crude oil, Heavy oil made boilover

１ｍ pan test using Arabian crude oilHot zone was made in NRIFD

Temperature inside the liquid, using keroseneNo hot zone was made

Fuel property

Test pan

Temperature measurementThermocouples were bedded along with pan pole

Crude oil

• Crude oil we used, is equivalent to Arabian light crude oil supplied by Idemitsu Oil Co., which is most popular crude oil.

• That is, we bought from Idemitsu Hokkaido Refinary, in which large oil tank fires occurred in 2003.

Distillation curveequivalent to Arabian light crude oil

Summary of test results

Results-Radiation

Results-temperature along with pan pole (axis)

Results-Fuel and water temperatureHot zone was made and its temperature increase

Hot zone regression rate, thickness, temperature vs time

ConclusionTime to boilover

1 10

1

10

50

5

5

(hr)

時間

(m)燃 焼 層 厚 さ

Time to boilover can be estimated with fuel thickness because hot zone regression rate is known from many our data and real accidents. According to API data, we obtain these lines, and our data were within these lines.

Time (hour)

Fuel thickness (m)

Conclusion

• About 70 minutes after ignition, we had boilover, and external radiation increased about 22 times as much as steady burning.

• We measured temperature inside the fuel, and found that there might be three steps

1) no isothermal zone was observed 2) Isothermal zone was made, but not high temperatur

e 3) Then, temperature of isothermal zone increased, an

d reached water layer

Conclusion

• Our data support results of our small test and other group did

• Data is open, anybody can use our data, because large boilover test is difficult to do so often

• Publication: Fire safety Journal vol.41, 2006

• Other information is available

Contact me, [email protected]