As a student, in the frame of my job at the MPA Stuttgart, I was always fascinated when I could evaluate rupture tests of containers at the Bundeswehr grounds in Meppen. Containers with a thickness of several centimeters busted and were unrecognizable. The tests were carried out with air under the strictest security conditions. The bursting pressure of the several meters long containers was about 170 bar.
This was about 35 years ago.
The requirements for the pressure resistance of the containers always continued to rise. The top pressures are already at an unimaginable 25,000 bar, but with very small diameters.
But also hydrogen tanks for vehicles are tested at up to 1.200 bars. The dimensions lie in the range of approx. 1m. In drill pipes at the oil and gas productions, the pressures are in the range of up to 1000 bar. The test samples here have lengths of up to 15m. They are not just small bombs since there are much larger dimensions than, for example, in a high-pressure injection system.
To make sure that they can safely withstand these high pressures, the pressure resistance, or the compressive strength at which a component may fail, must be determined through tests. In a test chamber, the test sample is impinged with water at a high pressure. As my early experiences in container testing in Meppen confirmed, if air were used, the failure of the test sample would have disastrous effects because the stored energies would be a lot higher due to the compressibility of the air.
But even water is no longer incompressible at such pressures; it is not compressed insignificantly, and thus saves large amounts of energies, which are similarly sized as the stored energy in the container. The total energies with typical uses are in the size of a small car at 120km/h.
Therefore, it primarily must be ensured that no humans are endangered through the tests. There also should be no damage to equipment, buildings or facilities near the test chamber. If this topic is addressed starry-eyed and when testing with water, it may happen that test chambers are completely destroyed or that fixtures as heavy as 100kg can no longer be located after the tests. Sometimes not even sandbags will help.
Construing the test chambers in a way so that they can be operated or easily fixed after the bursting of a test sample (whether is it is a pump casing, a tank, a hose, a pipe, or a drill pipe) is an art in itself. The determination of the size and the direction of the energy of the fragments, without which a useful construction of the test chamber is barely possible, are a challenge.
Merkle & Partner have specialized in the construction of test chambers and are a worldwide leader regarding the simulation, but also the size and the construction of different security concepts. We can simulate the behavior of a high-pressure pump with a weight of 25 tons at the rupture under an internal pressure of 300 bars as well as the behavior of gas tanks. It is immaterial whether we use water or gas as the test medium.
The following video shows a simulation of a cylindrical container burst with water:
The failure of the container, the behavior of the fragments, but also the water hammering on the tank walls are illustrated physically precisely. It is not necessary for us to go to Meppen because we safely test on the computer.
We can individually determine the dimensions of the test chamber and the concept without pointlessly installing materials. This is safe and saves time!
My experience with the recalculation of test chambers, which are installed by companies without the appropriate experience, is that the risks are rather underestimated.
Please contact the Division Director in Structural Mechanics, Dr. Mail Brehm (firstname.lastname@example.org) for any inquiries about containment.
I am looking forward to your feedback
We have specialized in the construction and proof of the pressurized containers, even per various standards (AD 2000, EN 13445, ASME Sect. 8 Div. 1+2, ASME Sect. 3, EN 1591, PD 5500, KTA, RCC-M).
Further topics in containment are fractions of rapidly rotating components. I will address this topic in a separate blog.