Cockroach,
I understand your point, but my point is that you did not have a "properly designed test" if one malfunctioning part could overpressure the fabrication being tested to a factor of 2 of your test pressure. "Proper test design" would have called for at least a PSV in the test rig and I don't know what else. The air tests I do are pressurized from compressors and at no time do I have a source of pressure connected that is higher than the current soak point.
The times I've done a test with high-pressure natural gas, I test PSVs prior to the test and carefully control the rate of pressure change with explicit proceedures (add WAY more soak periods than really necessary). My biggest fear in that case is always that rather than a purge, some genius will just open the gas source into a dead-ended line. I have never been successful in engineering stupid-proof, but when there is a risk of someone being killed by stupid I attend the test and monitor proceedures very closely.
This topic comes up often and in every discussion people point to specific cases where an air test launched steel across the landscape. I've never seen one of those examples that could be called a properly designed test. Many of the examples are brittle failure of parts tested to 150% of MAWP at ambient temperatures below freezing. A good test should specify a minimum ambient temperature for a test
People that have had a bad experience with an air test are unlikely to ever do another air test or to allow their subordinates to do one. I contend that air tests are exactly as safe as the design-engineer is competent to do the task.
Bingopin,
I know that the OP was asking about a vessel test, but many of the issues are the same for pipeline tests and I've formulated most of my opinions on pipeline tests. For pipeline tests it is getting very difficult and expensive to dispose of hydrotest water. Current NPDES regulations proscribe just letting it run down the bar ditch, many UIC disposal wells are refusing to take it, and drillers won't have anything to do with it anymore.
Determining what test pressure exists in a hydrotest at each point is a major problem in hilly country (I did a test on a 300 psi line with 1,000 ft of elevation change, a gauge a the top would give me 744 psi at the bottom, a gauge at the bottom would have the top empty and at atmospheric pressure). Actually getting the water out of the pipe in hilly terrain can require pushing a pig with nearly the test pressure behind it.
Either kind of test has its own strengths, weaknesses, and limitations. It is our job to pick a test media, test conditions, and test proceedure that are appropriate for the specific test underway--that is what we get the big bucks for.
David
