Detecting cracks in metal

[treez]

Hello,

(This is a double post and i apologise though the other post is in the electric motors section and i considered that the audiences would be likely different for these two places.)

We are having problems with cracks sometimes occurring in the welds in our stators. We would be most grateful if you had any equipment which could help us to detect these cracks.

We are designing and manufacturing electric car motors. The stator for these motors houses many power electronic inverter (H-Bridge) boards which help to drive the rotor.

These stators have like a water jacket in them which is for cooling.

Unfortunately, the stator metal has been welded in places next to where the inverter boards go and the cooling water is escaping through minute cracks in the welds and getting to the inverter boards. As you can imagine this has bad consequences.

Not all the welds leak in this way, but some do.

What we need is some way of detecting these minute cracks.

At the moment we are putting water with washing-up liquid in it over the weld and then pressurising the cooling water chamber with a car tyre foot-pump. -If there are cracks present then we should see bubbles in the water.

This is not a good way to detect the cracks however, and we need some better way.
I would be most grateful if you had any equipment that could detect these cracks for us. Preferably something hand-held, quick and convenient.
Our production Engineer believes that a good way to detect cracks would be for him to pump air into the cooling water chamber and then for him to “listen” for air hissing out of the crack using a microphone.....-either with him listening for the amplified output of the microphone or putting the received audio “hissing” signal into a detector which will detect the special “hissing audio waves” that are characteristic of air hissing out of a crack in the metal.
Kind Regards,

 

[rb1957]

research "dye penetrant"

 

[MotorVib]

X-rayng the welds would be the best but not welding would be better. If you are going in to production you don't want to do that. Can you look at different ways of holding the stator in place like a interference fit or pinning. I have used that on some generators and I have seen it on some motors.

Dye penetrant works good if you have a clean surface with a good finsh. It might not work all that well with a weld.

The air test sould work but I don't know how well your tester will be able to hear. What it the coolent preassure at?

Chris

"In this house, we obey the laws of thermodynamics." Homer Simpson

 

[eromlignod]

Take a look at eddy current crack detectors.

http://www.ibgndt.com/eddydector-crack-detector.htm

Don
Kansas City

 

[MrMyers]

You should also think about why these 'cracks' are present. Questions to answer are:
1. Are they welding flaws - lack of fusion, slag etc?

2. Are they corrosion related (all sorts of water quality/material selection issues open up)

3. Is the design in need of attention to avoid features such as crevices?

4. What is the liklihood of these cracks propagating with time/service...

5. Could you have a big problem at hand in 5-10 years time when a fleet of currently non-through thickness flaw pops through or pinholes to a leak?

 

[handleman]

Cost is a bit higher, but you may want to look into helium or hydrogen mass spec leak testing with a sniffer type arrangement. Check out

http://www.adixen.co.uk/leak_detectors_ H1000.html

Never used this before, but I did investigate it somewhat. Due to process requirements we were not able to switch to this from Helium.

-handleman, CSWP (The new, easy test)

 

[Rjeffery]

As the stators must be made of material that is not diamegnetic, magnetic particle (MT) would be the fastest method for detecting and locating cracks. The "acoustic emissions" test that your production engineer suggests will only detect leakage and not pinpoint the actual orientation, length and severity of the flaw.

The best way to test something is to squeeze it, slowly, until it breaks!

 

[metengr]

After reading the post and responses there are two separate problems. I think using a pressure test is best most effective method to check for through-wall cracks (leakers)in production welds. A higher air pressure, if possible, at 50 psig would be more effective for detecting leaks. I would also have surface NDT performed as a QC spot check for production welds that contain cracks but are not through wall. These cracks can become through wall in service.

Now, once you evaluated your production welds you need to investigate (maybe some destructive examination) as to what is occurring during the welding process to cause the defects, and eventually reduce the number of reject welds to zero.

 

[treez]

Thankyou for these replies.

The holes are thought to be so small that they cannot be detected with the naked eye as through-holes. -But we do know pretty much exactly where the hole will be if it is there.
We really need something quick that can be used on our production line.

Since the holes are so small, i am wondering if this means that Ultrasonic Leak Detectors are unsuitable (Please forgive me if i am wrong, but i have the impression that sonic methods are for scanning large areas to quickly locate leaks which are from relatively large holes) Ultrasonic detectors detect the ultrasound emitted by air hissing outof a leak-hole.

A company has phoned us and told us we need to use hydrogen or helium.....pump the gas into the chamber and then try and detect it in the suspect leak area.

-to do this we would have to get a special fixture made up so that helium leaking from the seal to the chamber didnt give false positives.
-Also, we would need to have a special draft-free room so that the helium didnt blow away before we could detect it coming out of the hole.

In short, an ultrasonic detector sounds the cheapest, quickest and most convenient method -though we are not sure if it has the sensitivity to detect air leaking out of the pinhole with a pressure of 30psi.


we have not managed to detect pin-holes with our washing-up liquid method -but we must still assume that holes are there as we do not feel that the washing up liquid method is sensitive enough,
-we are now also looking for a vacuum pump, so that we can create a vacuum in the chamber, -this should then suck glue into any holes that are there (we first smear the glue over the area where the hole might be)

 

[mappel]

Have you looked at using Magnaflux to detect cracks in your part?

http://www.magnaflux.com/penetrant/penetrant.stm