Has anyone out there using the iLearnInterpreter?

I worked on CSI MHM software and it had 'nspectr'. It also was quite good and interpretes the data.

What is the success rate of these tools?

If these tools are so powerfull, than there is no point of us to continue in acquiring vibration analysis knowledge.

Just a little word from the experienced guys out there.

Since the dish ran away with the spoon, did the cow jump over the moon?

Yes it can.

:-)

Well, if the "expert" software and I agree it´s all ok, specially if I wrote the software... It would be a bit scizzy otherwise but who knows, one day.. If not take a look at the rules the software used to find some funny fault and check the data why it gives that ruling. Once in 15 years and 3-4 for various flavors of expert software the software beat me when I turned my back to it, also due to that it repeatedly called the same fault and in 99,9% of the cases it was not the case. So as software is more enduring and capable of repetitive tasks than humans I use it as a tool to extract the machines that need human analysis. Not that it exclude human scanning of the data or human analysis but it cover the data scanning 100% as a extra safety. You can never (so far) beat a human analysis based on experience by software so keep learning and also learn from the analysis that the software do. Software is just another tool in the toolbox just like the yellow handle analyzer (just more expensive). Olov

Training manual on the yellow-handled analyzer? Any firm offer classes? Fast lane to hearing loss, blow-up in your face - of course nothing could go wrong.

I agree with Oli. The software is a good screening device if all setups are correct.

I have decision support software as one of my tools and do use it with success. But as mentioned above we have spent a lot of time tweeking alarm levels and rules. The thing with our software is that it makes calls on things that share frequencies which requires us to look at the Spectrums. Nothing can replace the analyst looking at FFT's IMHO.


Best Regards, Dave Dane

I think Nspector had a rule that if the point began with the letter "P" it had to be cavitating

My problem was that GBX outboard gives locked coupling, and once it was correct! Olov

Hi Zee!
I offer a couple of humble suggestions.
I used a popular software for some years which had the AI analyzer. As I remember, it would identify a "possible resonance" on almost every piece of machinery. But I found, as someone has suggested above, that going through the "rules" of the software, that caused it to flag certain problems, taught me a couple of things.
I had a fast computer, so it only took seconds to run the AI along with my analysis. If I had an alarm flag on a point, I would analyze it. When I got through with my analysis, I would run the AI. I ran the AI after my analysis so it wouldn't cause me to be one track minded.
I remember clearly "one" time the AI software saved me. I was in a hurry and distracted by other things and sped through an analysis on a point and then ran the AI. I was startled by what it said! I went back and looked, and sure enough, a very easily recognized fault that I would have caught 999 times out of a thousand. I wasn't even sure that I had looked at the spectrum before I moved on, I was so distracted.
After a while, I was fortunate enough to become the head of the Reliability group, and although I no longer did the analysis, I always encouraged the analysts that reported to me to run the AI and learn the "rules". It is a great tool to help train a young analyst. But I would never depend on it as the primary analyst, not "no way no how".
I guess what it boils down to is, if you have the time to run it, it can only help you if run in addition to your analysis.
Happy Fault Hunting

Warning: Vendor comment

I am the guy who wrote iLearnInterpreter. I can tell you that nothing will ever replace experience and knowledge, and nothing should stop your pursuit of learning.

If you want to buy Interpreter to study spectra and automatically diagnose faults with 100% accuracy - then save your money; that's not what it does. I would never say that a software tool could look at a single spectrum and diagnose a fault.

Let me tell you why I created Interpreter.

We have a training/reference tool called iLearnVibration. Many years ago I started receiving e-mails from people to say that would diagnose faults by scanning through all the diagnostic 'slides' in iLearnVibration to find sample spectra that looked like the patterns they could see on the screen.

So I created the Reference Center. You could type in a question and find the answer in iLearnVibration, and you could describe the spectrum (go through a tree of fault symptoms, etc.) and it would lead you to information that would help you diagnose the fault.

But there was still a problem. You had to find the harmonics and sidebands in the data. And it was a little time consuming to translate what you saw in the spectrum into the tree of symptoms in the Reference Center.

So that gave me an idea!

I created Interpreter to actually look at the spectrum and find the patterns itself. Interpreter literally looks at the spectrum on the screen (inside your VA software). It then searches for the harmonics, sidebands, major peaks, etc. and highlights them on the screen. With the information it gathers, Interpreter then displays to the user the fault conditions that may exist - but then it takes the analyst by the hand so that he or she can diagnose the fault. It displays reference information and links to iLearnVibration (if owned) to provide the background training.

Enough said.

The point is that it is not intended to diagnose all fault conditions and be a tool that tells you what is wrong with your machine and what you should do about it. It helps you make that decision.

Who is it for? There are three people I had in mind:

1. So many people are new to vibration analysis and need help identifying the key patterns and interpreting them. It is like a big brother looking over their shoulder. If we all had the expertise of some of the contributors on this forum then we would not need Interpreter (or our training courses for that matter!).

2. More experienced people appreciate a little bit of help. It just may identify a pattern and/or make a suggestion that you may have otherwise overlooked.

3. It is a training tool. It helps you to learn more about the analysis and diagnostic process.

So, I have been careful not to oversell the capabilities of the product. If it could diagnose faults with 100% accuracy I would not sell it for such a low price!

I hope this has been helpful.

Jason Tranter
www.mobiusinstitute.com

Jason,

I didn't mean any offense. Anyhow, you really did summed up the whole thread.

I was very much interested to buy an interpreting software because of a couple of resons,

1) That I am relatively new (as compared to you guys out there).
2) Whatever Vibraman said in his comments.

I will use your point of views and will discuss this matter with my top management. I recently got the iLearnVibration and its superb.