Sorting Cast Steel With Ultrasonics

Date 07-01-2005 17:13

I posted the following question in the Technical forum, but I did not get the type of responses I was looking for. I have been unsuccesful in attempting to sort based on the decay of multiple echoes. Also, a research paper found on the ASNT web site suggested that sorting could be done based on the amount of noise present between the main bang and the first back reflector. The as cast component should have had a significantly higher noise ratio than the heat treated component. My testing found the difference between the two to be negligable. I am sure the technique would be more effective with cast Iron rather than cast steel.

"I know there are some good UT practitioners that hang out here, so I hope someone can help me. I need to sort out some heat treated steel castings from non-heat treated steel castings. If I can do it with Ultrasonics, it could save me alot of time. Does any one here have any experience with this?"

By GRoberts (***)

Date 07-02-2005 01:05

I'm far from being a UT expert, but it looks like you are trying the same things I would have looked for. One question I have though, is what size/thickness are the castings you are working on. I've been told that castings where I work have to be heat treated before we can UT them, but we are normally working with very large castings, and wall thicknesses can be anywhere from around 1" to 20"+. Is it possible that the grain size before heat treatment of your castings is not large enough to cause sound attenuation that is noticable?

By thirdeye (***)

Date 07-02-2005 02:20

Here are just a few more questions....

What type of heat treatment and what type of cooling was used?

Was the heat treatment performed to improve the machineability or for some other reason?

It sounds like you have some confirmed samples of each batch, is there a difference in the hardness?

What is the reference to the article from the ASNT site, I would like to check it out.

What size and frequency of transducers were recommended?

~thirdeye~

By NDTIII

Date 07-02-2005 03:39

I have a lot of experience with UT on cast stainless steel for flaw detection.

What I found is the velocities in each piece vary so much from one area to another and in various directions.

In some cases, the individual grains were as large as or larger than the side drilled holes we used for calibration. I also found that spun cast was even more severe than static cast S/S.

The best setup I found was using a digital UT scope with a strong pulser
and 1" diameter, dual 1.0 MHz. focused probes (pitch-catch). I don't know if you would require focused probes since you are only checking multiple echoes.

A computerized UT imaging system would probably provide better results.
But to be honest, as I stated before, the velocities are just too unpredictable. If you have any success, please let me know.

You may also want to post this question on the forum at www.ndt.net.

By thekoz

Date 07-05-2005 14:02

NDT III

Thanks for the info. I was not able to detect with a great degree of certainty which pieces were as cast and which were heat treated. I thought I had a good technique of sorting with the decay of multiple echoes, but then I found a few pieces which were known to be as cast that had the same signature as the heat treated sample. I stopped testing immediately. The parts are now being sorted using microscopy. That is the only sure way to do it for now. Hardness tests and eddy current testing were attempted and found to be inconclusive.

Regards

thekoz

By thekoz

Date 07-05-2005 14:17

Thirdeye,

I am not a casting expert, but I will do my best to answer the questions.

I was told it was a shell casting and it was surrounded by some sort of steel shot. It was performed to increase the strength and toughness. There was not a difference in the hardness due to type of casting process (so I was told). See the reply to NDT III for how I did with my testing. The link for the article is below. Hope it works.

http://www.asnt.org/publications/materialseval/solution/dec97solution/dec97sol.htm#Table%201

This article does not directly apply to my situation, but my thought was to use some of the theories in it to develop a technique for UT sorting. It did not work for me. Let me know what you think.

Regards

thekoz