Methods of leak testing casting? 1

[cbrf23]

Hello,

I'm having trouble specifying non-destructive examination for inspecting castings (which are used to manufacture high-pressure hydraulic cylinders) prior to machining.

I have a casting that has four cylinders machined into it - cylinder is about 2" in diameter by 4.5" deep, and each houses a piston that is hydraulically powered at normal operating pressures up to 10,000psi. During prototyping, we had a problem with a handful of castings leaking out through the wall of the cylinder (creating a very effective oil mister). All of these castings were inspected 100% via xray and also visually inspected - no porosity was detected in any of these tests, and the castings did hold pressure up to around 5,000 psi or so, so the porosity is microscopic enough that the castings would hold fairly significant pressure, just not enough.


Right now, my testing consists of hydrostatic testing of finished machined products - we pressurize to working pressure with the normal hydraulic fluid used and hold for a period of time.
Our casting manufacturer wants some specification to work to that they can test the castings prior to machining, but I'm not sure what (if any) inspection method would accurately predict the servicability of the castings after machining.

I don't know of a good way to seal a cast (unmachined) surface that would allow us to test at 10,000psi, and I'm not familiar enough with this type of testing to know if there is a lower pressure alternative. We did hook up a pneumatic line from shop (≈100 psi) but at this pressure we were unable to detect leaks even in castings which were known leakers (identified via high-pressure hydrostatic tests).

Any advice on possible testing methods or standards would be very welcome! I'm sorry I can't share pictures or drawings.

 

[btrueblood]

You might consider changing to a forged blank for the part, rather than a casting.

Short of that, I don't know of any good techniques to identify flaws that open up under pressure loads. Somebody will likely suggest helium leak detection, but as you found with the pneumatic test, the leak may not appear at low pressure (and, you still have the problem of sealing against the rough as-cast surface).

 

[btrueblood]

Sorry, got interrupted.

We use castings for valve bodies, albeit at lower pressures than yours (typical max. hydro at 900-100 psi). We chase leaks in cast ductile iron, and some types of bronze, parts all the time, and on some larger and more expensive castings, often have them impregnated after machining before we ever see them. I'm not sure if impregnation can work for your pressures and alloys, but it might be worth investigating.

 

[1gibson]

Machine the sealing surfaces but not the bores until after successful hydro? Better question is, will it be cost effective to push this risk onto the casting manufacturer, or can you alter your processes and order extra castings expecting some leakage and scrap?

 

[EdStainless]

I would rough machine and then air test at 150psi under water. You can use a clamping type of fixture for sealing. You will find any leaks that oil at 10k can find.
I have never had good luck testing as cast surfaces, often the porosity is closed and the surfaces don't allow good sealing.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[cbrf23]

Couple of good suggestions.
@EdStainless - we did test with air at ≈100 psi (110-120 typical shop air here) and the castings which did leak oil at 10,000 psi did not leak air. This test was done by sealing the cylinder and attaching a digital pressure gauge. We then monitored the gauge for 60 minutes after removing the air supply to measure any change in pressure. This is a commonly accepted leak testing practice, and I find preferential to the submersion tests because 1) don't need a giant tub to submerge 2) won't cause ferrous metals to rust 3) quantifiable as far as leak rate goes, if you check pressure at regular intervals.

I do agree, in that I also question the value of any testing prior to machining, as the machining is going to open up new material and potentially create new leak paths, and I don't think there is really any good way to do test for this.

Just thought I'd check and see if anyone had any ideas we hadn't considered.

 

[Compositepro]

A bubble test is probably 100 to 1000 times more sensitive than a loss of pressure test. A helium leak detector might be a million times more sensitive.

 

[EdStainless]

You can't do static pressure test with air (any gas) unless you control the temp to +/-1C, and the part and gas are at the same temp. You also need to have a pressure gage that will read 100.00 psi. A steady stream of small bubbles will hardly move a pressure gage.
There is a better way, hook up two parts, and pressurize them together. Shut off the source and wait a bit, then isolate them from each other. This way you can use a differential pressure gage with very high sensitivity to see if the test piece is changing wrt the reference piece. They still need to be at the exact same temp. But using delta P you can detect 0.001psi change.
I would also suggest that you use a clean gas, no water (-40DP) and no oil, so shop air is out. Nitrogen works well.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[cbrf23]

Hmmm. Good information - thank you both! I'll have to convey that about the temperature to the chief engineering here.
I know I asked about pressure dropping on a hydrostatic leak test once and was told it was likely due to temperature (forcing the hydrualic fluid through the pump orifice generates some heat) and to just wait for it to equalize before starting the test.

Maybe we can do some more tests using different setups if we still have the old (known bad) castings.
Appreciate the help!