We are manufacturing Babbeted Bearings used in Steem Turbine & Generatores up to 110MW.In order to reduce the rejection rate due to improper bond between the Shell of the Bearing and the Babbet material,we are designing a experiment with the following parameters.
1.Dehydrogenation treatment for Shell of the Bearing.
2.Surface finish on the Internal Diameter of the Shell.
3.Cleaning of the Shell Internal Diameter.
4.Tinning of the Internal Diameter of Shell
5.Pre-heating of the Shell before babbetting.
The Bond between the Shell & Babbet is checked by U.T(NDT).

I request for the brief method of doing the above and method of ensuring with acceptance norms.

sarasheed.

There may be some experts visiting here in the Cove who can answer this question, but I think we might serve you best by helping you find a website more focused on surface treatments for metals. I'll do some checking in my old files.

In the meantime, anyone got an instant answer or suggestions where else to look?

It appears to me Sarasheed is looking for advice on designing an experiment. If that's the case, it doesn't matter what the process is. The determination would have to be made on what type of DOE was required. I see from the post there are 5 factors trying to determine the effect on bond. We would need to know how many levels were in each factor. Without more information, I would think a Factorial ANOVA would probably be the best for the purpose. I would like to know what DOE package they are using as well. There is a small program called DOE KISS that would probably work nicely.

Just some thoughts. Any DOE experts, please jump in with suggestions.

Dave

My thought was to borrow ideas from other operations who had already resolved the problem - I detest reinventing the wheel, even to the extent of designing and running experiments, expecially the cost and expense of a full factorial DOE. I recall going through some agonizing efforts in the 70's and 80's, only to discover weeks later (no internet back then for speedy knowledge transfer) that someone had written an article covering exactly the same problem a year earlier and had come to the same conclusion as we did.

Good point - there is nothing wrong with some good research before jumping into a complicated experiment.

Dave

Sarasheed, did you get an answer to your question? I have a very similar situation and need to figure out how to handle it...

Thanks.