I built this pressure tank a few years ago and it works great, but I’m interested in automating it so I can test pressure cycles. I know Blue Robotics has a 1000 dive rating on some of their fittings. Is this something they’ve tested in house? and if so, what kind of equipment set up did they use?
I know there are a number of ways to do this, but I’m looking for some examples of pressure tank builds that have worked for others in the past to start from.
Thanks,
Brad
Are you looking at doing live testing while under pressure or what appears on your vessel … just subjecting something to pressure?
Right now there is a manual pump, so I’m able to see how parts react under hydrostatic pressure. In the future i want to simulate repetitive dives. The tank I have is rated for 500 PSI, (~350m depth). so i want to be able to dive to ~300m approximately 1000 times, while monitoring for leaks in a pressure housing.
I have a small hydraulic pump and some solenoid valves that I can configure to do this, but I’m looking for other examples of similar tank set ups that have done this already to learn from.
Brad
Hi Brad,
We have a few pressure testers here at BR. Our main workhorses are our two “Pressurizers” which allow us to test down to about 1500 MSW. each one has a couple selectable pressure tanks, with lids that allow us to test up to 18 M10 components at a time, or fill the lids with blanks for testing things like our water tight enclosures. They also have heating and cooling coils for allowing testing across temperature ranges.
Each pressurizer uses a single piston driven by two ball screws with high torque servo motors to generate the pressure (since water is nearly incompressible you don’t need much displacement). A set of limit switches and a pressure transducer allows the system to run pressure testing automatically. This lets us very easily do both pressure hold and pressure cycling testing on our components. Our exact procedures have changed over time but typically all of our products have gone through many hundreds of cycles at each stage of development since it only take a couple hours to run a few hundred cycles.
We also have a new tester known as “Big Squish” which has a bigger volume for testing some of our bigger enclosures (Up to our 6" WTE), and should be able to test deeper than the Mariana trench! It has seem some good use testing components for DORIS down to 6600m. Currently it’s driven by a manually controlled pneumatic pump, which makes cycle testing a challenge, but we have some ideas to change that!
It’s worth noting that while these systems look a little scrappy after years of use, we’ve put a lot of engineering hours into making sure they are safe. It probably goes without saying, but please be careful when working with these types of high pressure equipment.
Let me know if you have any questions!
This is fantastic! Exactly the kind of info I was looking for.
For your control system, do you know if the feedback loop relies solely on pressure transducer, or does it just drive the stepper motors to a set position to reach a desired pressure?
I have a safety margin and a pressure relief valve built into my system as well, in addition to having someone with a pressure certification doing the welding, and a third party hydrostatic pressure test afterwards. Safety First.
Happy to help! and glad to hear you are keeping things safe!
We use closed loop control based on the pressure transducer for reaching the target pressure (I think the control logic might just be “stop when current pressure >= target pressure”, but I haven’t dug into the code personally)
I think you would really struggle to do reliable testing with just fixed displacement. The size and flexibility of whatever you are testing can change the required travel a decent bit, so you would definitely need to manually set it each test, and even then sometimes things can change a bit as things settle throughout the cycles.
I’d recommend having some safety checks as part of such a loop, to stop (and either wait for the operator to signal continuation, or auto-back off the pressure) if there’s a sudden drop in pressure (from the test object imploding or the setup springing a leak), and also to stop if the pressure isn’t changing (in case the sensor is broken or the motor has failed). For convenience you may also want some notification method (a chime or alarm, or email / text / online message) for when something like that occurs, given you likely won’t be actively watching the tank during hours of pressure cycling.
For extra safety and data accuracy you can have 3 pressure sensors that ‘vote’ on the actual pressure (in which case it’s obvious when one has failed), but that may be excessive if you have a physical guarantee the pressure is released before you get any kind of physical access to the chamber (e.g. as opposed to unscrewing something on the lid while pressurised and ending up with high pressure water squirting out). That may be your pressure relief valve, but if it’s possible for it to unknowingly fail closed you’ll want some other mechanism or sensor(s) as confirmation.
Thanks for the advice Eliot. I think adding a buzzer and a watchdog for a pressure leak should be implemented as well. In addition to the pressure release, I also have a manual valve to release water, and a mechanical pressure gauge.
I’ll post some more pictures here when I get something working.
Brad
