Measure tool wear using vibration or cutting force measurement?

[AX3L]

Hi everyone!
I'm looking to introduce a system for fault detection in bearings and misalignments using vibration analysis (thanks for all the help with that in the vibration forum, by the way!). I found out that using the same equipment (a DAQ - Data Acquisition unit) one can measure tool wear very accurately, not only regular "dulling" wear but things like chipping and irregularities that leads to bad surface finish and defect products. I'm not quite sure about how this works in real life though, and I'm therefore wondering if anyone here have any experience with this? Today were measuring cutting force using the built in systems in the machines, but I'm thinking of this as an easy way of extend and complement those measurements to find even more problems.

(Sorry if I posted this wrong by the way, I thought it's not really a question about vibrations but more of a general machine question but please move it if appropriate).

Thanks in advance!
Axel

 

[ornerynorsk]

Power requirement is generally going to be your most reliable indicator of tool wear. All other aspects such as heat generated, pulling down of rpm's, edge dulling, poor finish, will effectively translate back to power requirement.

If your machine system has the resolution to gather meaningful data, and not just a needle-type analog meter with 5% graduations, I would say that this is still the best tool.

My 2 cents!

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

 

[AX3L]

Orneynorsk: Thanks a lot for your answer! That's a bit different answer than what I expected though. To me it seems like vibration analysis would be more accurate but maybe there's a reason why tool wear is measured using force and not vibrations.

While where at it, we have some problems with drill breaking in deep hole drilling of metal (large trucks engine blocks). From what I understand vibration is a key factor to this, do you think vibration analysis would help detect those kind of problems?

Just out of curiosity by the way, are you from Norway or is it just a coincidence you have "Norwegian" in the last part of your nick name?

 

[ornerynorsk]

Axel,

I should elaborate by saying that, yes, vibration analysis can be a useful tool as well, if you can isolate the pick-up to the workpiece or toolholder. Once natural harmonics in the drivetrain of the machine, the spindle, the feeding mechanisms, etc are graphed and can be excluded from the cutting signature, VA should show distinct trends in tool wear. As an exercise in pure science, I think it could be quite informational and lead to better tool management. Trending the power consumption, by its nature, will readily show the short term issues such as tool wear, but I think VA will perhaps show the long term trends better, such as bearing issues, motor winding degradation, misalignments of machinery (feeders or stock stands, ballscrew mounts, etc) etc, which actually can all contribute to premature tool wear.

Regarding my handle, I am actually in the US, but my father's lineage is from Norway. I'm first generation to use English as my native language, but I still retain the heavy accent of my older relatives, hence the nickname.

About the deep hole drilling, I don't have a tremendous amount of experience in this, but chip evacuation and heat build up are two of the more predominant issues, provided feeds and speeds are correct. VA could certainly be used to dial in correct feeds and speeds, the "sweet spot", if you will.

Good luck in your vibration analysis project. It would be interesting to hear some results as you continue on.

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

 

[AX3L]

Orneynorsk: That's the plan, I'm thinking of placing the accelerometer(s) at the tool holder itself (I'm now referring to turning operations) and analyse it using FFT and wavlet transform in some Matlab and Simulink script at the computers next to the machines. This way, one should easily be able to see any strange behavior in the vibration compared to recent measurements. At least this works in theory, and being a young inexperienced engineer that's unfortunately the only world I know about at the moment.
But as I said this is more of a bonus function really, my main goal is to automate the vibration analysis of bearings and find faults and imbalances that leads to defect products and production stop in an early stage. And since an extra accelerometer isn't expensive at all today and I already have the equipment to measure installed I could easily add another one to measure tool wear by vibration, if it makes any sense that is.

I'm glad to hear people showing interest in their Scandinavian heritage, I'm from Sweden though and as you probably can tell by my grammar still lives here .

If there's anyone who is interested in this tool wear by vibration analysis thing I recommend this report from an European university:

http://upetd.up.ac.za/thesis/submitted/etd-12212006-110642/unrestricted/dissertation.pdf

I've not been able to read how effective the method is compared to the force measurement from this though, but one have obviously been able to use this vibration technique for quite a while now.

 

[Zinov]

We have already developed system to do that for high speed machine tools. It works very well.
It works for other machinery (pumps, compressor, gear boxes etc..)
We have portable kit where you can install the acc. on the non-rotating parts of the spindle.
Collect the signals analyse them and have word reports.
We have developed a ramp-up ramp down test that helps to detect any early pbs in the bearings (mostly completed).
The vibrational analysis can be supported by oil analysis, to confirm findings.
Write me if you have a budget to buy our analysis kit.

 

[AX3L]

Zinov: Portable analyse equipment is something we have a lot of already. And while they're accurate and all, they obviously require some manual handling for the measurements and since we have many machines to be monitored we can't do it as often as I would like to in order to see trends and find problems in an early stage. The question now is whether or not adding another accelerometer to monitor toolwear would give us some more information about the wear than the force measurement we do today.

 

[Zinov]

We had devolopped a new system to monitor wind turbines installed in the ocean. Once the accelero are in place no manual handling is needed. So for a machine shop it is much easier.
For your tool brekage in engine blocks,a classic pb, you should start with stub drills or short series drills, ask your tool vendor.
Mazak has already included tool breakage detection system in their controler.
Siemens allows you to access built-in accelero to monitor the spindle
etc...
My advice, do not spend years of your age re-inventing the wheel...
Good luck.

 

[Viktor]

Power measurements and power limits set on the machine controller - this is the right answer. Anything else that far is a toy.

Viktor

http://viktorastakhov.tripod.com

 

[AX3L]

Zinov and Viktor: Thank you both for your answers! I now realize vibration analysis for tool measurements at least for ordinary machines isn't worth doing. I would like to try it sometimes though, mostly to see if it's possible, but for now I'll concentrate at the bearing fault detection.

 

[Viktor]

AX3L

To select the right direction, you have to know the application. It is first of all the drilling system you are dealing with. In the automotive industry, new machine come with powerful high-speed spindles having built-in censors and Siemens controllers that cleanup the signal so that the power related only say a single drilling operation can be seen on the screen as a graph starting from the beginning of the hole till its end. The whole history of such a drill starting when it is new and finishing at the end of the tool life can be visualized on the monitor so the system learns how to set the limits for the power/drilling torque for the future applications. It is rather accurate and reliable. The whole history is stored so you can see the variation within a manufacturing lot of drills or the differences between the manufacturing lots or even compare the tool from different suppliers. As I see it, possibilities are endless. unfortunately, not many specialists use this available option :-(

Viktor

http://viktorastakhov.tripod.com