Thermal Hazards Laboratory

BLEVE Research - BLEVE TESTING 1992 - 1996

A. M. Birk, PhD. P.Eng., Professor

Dr. Birk has been conducting fire tests of propane tanks to study the BLEVE (boiling liquid expanding vapour explosion) event. This work was funded by Transport Canada and the Propane Industry.

If a tank holding a pressure liquefied gas (such as propane) fails suddenly, then part of the contained liquid can boil violently producing an explosive effect. Any compressed gas in the tank will also leave the vessel explosively. The rapid phase change can also produce strong thrust forces that can propel the tank over large distances. If the contents are flammable then immediate ignition will result in a fireball or flash fire. Delayed ignition could lead to a vapour cloud explosion.

Dr. Birk and his research team of engineers and students conducted a test program from 1992 to 1996 where forty 100 gallon automotive propane tanks were exposed to various types of fire impingement. This was done to try to trigger a BLEVE event. The tests were done to learn more about BLEVEs. The tests were designed to study how the following factors affect BLEVE events:

  • pressure relief valve set pressure
  • fire type and severity
  • tank strength
  • tank fill

Out of the 40 tests, there were 13 BLEVEs and 24 ruptures with transient jet releases. For a BLEVE to happen there must be a combination of a weakened tank state and enough liquid and vapour energy in the tank to drive the failure the full length of a tank. For very weak tanks the vapour space energy may be enough to cause catastrophic failure. For stronger tanks a violent boiling response by the liquid may be needed to drive the crack the full length of the tank.

The following three figures show the test apparatus. The tanks were 100 gallon (0.61 m in diameter, by 1.52 m long and had nominal wall thicknesses of about 6 mm). The pressure relief valves were nominally set at 2.15 MPag. The tanks were heated from below using a pool fire (to heat the liquid in the tank) and burners were applied to the tank top to heat the vapour space wall to cause tank rupture (i.e. steel looses strength at high temperature).

The box on top of the tank is a radiation shield to hold in the heat. This allowed the test team to get consistent heating of the tank vapour space. The tanks were equipped with instruments to measure the temperatures of the tank wall as well as the liquid and vapour inside the tank. Pressure transducers were included to measure the tank internal pressure and a high speed pressure transducer was included to record the pressure variations at the instant the tank ruptured.

The research team was also interested in the hazards when the tank did BLEVE. To measure this the team included instruments in the field to measure thermal radiation from the fireball and guages to measure the blast shock wave. There were also video cameras and still cameras at various locations.

In the tests the tanks were heated from above and below until the tank ruptured or was emptied by the relief valve. Sometimes the tank failed totally (i.e. it opened up completely) and this resulted in a BLEVE. In other cases a hole was formed in the wall but it did not propagate the entire length of the tank. The following figure shows a BLEVE in the very early stages just as the tank is opening up. The white cloud is the interface between the released propane and the surrounding air. The next figure shows the end result -- as you can see the tank is flattened on the ground.

The next figure shows the case of a finite failure where the tank did not BLEVE. In this case the tank was strong enough to stop the crack from travelling the full length of the tank. Another way to say this is that the energy in the liquid and vapour was not sufficient to keep the crack moving along the tank. In this case the tank contents would have vented rapidly through the large opening. This jet would have been both liquid and vapour and it would have produced a very large but short duration jet fire. This jet produces a large thrust force and in some cases the tank could be propelled over some distance.

The final figure shows a fireball from a tank failure. Amazingly this was not a BLEVE but rather a very large finite failure. You can see the tank flying up in the air due to the jet thrust force. In this case there is both a fireball and a ground fire. This is an example of where the hole size was so large that the tank emptied in less than a second. The fireball duration was about 2 seconds and therefore the fire looks like a fireball from a BLEVE.

for more information contact A. M. Birk at birk@ME.QueensU.CA