accelerometer pick up for natural frequency test of rotor blades

[steamturbine]

During natural frequency test of rotor blades of ulility turbine, we have to mount accelerometer pick up on blade.

I want to know whether there is any empirical formulae for calculating what should be the weight of accelerometer pick up with respect to weight of rotor blade.

or in other words, whether there is any defined ratio of weight of accelerometer pick up to the weight of rotor blade.

Kindly also tell where we should mount the accelerometer pick up on rotor blade during natural frequency test, for reliable results (distance from the blade tip to accelerometer pick up)

What are the best methods for fixing accelerometer pick up on blade surface.
(steamturbine)

 

[Hatch]

I don't know of any defined formula for the weight of the transducer compared with the test object. I try to keep it at 1/20 maximum, or I use the smallest accelerometer that I have. Since you are talking about a low pressure turbine blade (mentioned in your other post), a standard accelerometer should meet the limit, since the blades of most low pressure stages are over 6" in length and are relatively heavy. If you have access or are purchasing an accelerometer, then buy the smallest and lightest unit.

One test I would do in your case is to mount the accelerometer on the blade and do the test. Then repeat the test with the acclerometer mounted at the root of the blade (space permitting). While the signal at the root will be much weaker, you can still compare the frequencies and if they change too much the accelerometer is probably too heavy.

I would mount the accelerometer at 3/4 blade length from the root. Then you will not be at a node in the first mode and at an anti-node in the second mode.

Since you can't drill into a blade, use "crazy glue" or expoxy. They can be cleaned off afterwards. I prefer "crazy glue" since it is very fast acting and very stiff. C. Hugh (

http://www.Hatch.ca)

 

[ANTOINE]

To my knowledge the pure way to look at this is to say that the added mass of the accelerometer should not be more than 10% of the modal mass of the structure. To have the modal mass you have to perform a modal analysis and thus make measurements on the blade ... if you can do those measurements with a non contact laser vibrometer... That's for the "heavy" way of looking at this..

More practically, you can run reciprocity test (impact at blade tim, measurement at blade root and then reverse and check the modifications of the FRF by plotting them on a single plot) which will give you the linear behavior of the structure and show possible peak drifts due to added mass (in the case of a linear structure)...Or you can measure natural frequencies with a non contact system (laser, proximity tranducer) with and without the accelerometer.

Check out ENDEVCO 2222C models (1.2gr)

Thomas.

 

[gunnarhole]

Just a few comments:

For mounting accelerometers, there are waxes available that are great at room temperature (beeswax and others).

As far as transducer mass effects go I remember seeing a tech sheet from PCB Piezotronics some years ago that rigorously covered this issue. Try contacting them for a copy. (they have the wax too)

As a last thought. Have you considered using a sound level meter as your vibration transducer? I have used this on axial compressors -it's fast and repeatable.

Regards,

Gunnar

 

Have you tried using a microphone instead of an accelerameter? A good B&K with the frequency response matched to your blade works great. You do not need to calculate for the added accelerometer mass. If you are familiar with the blade modes you can place the microphone at the right position to measured the desired response. This works great on large LP turbine blades.

 

[GregLocock]

I answered this in another forum, and since no-one else has suggested it -

Measure the FRF with one accelerometer.

Stick another accelerometer alongside.

Repeat the measurement.

Assuming that your accelerometer is only a small proportion of the modal mass then you now have a graphic illustration (offtopic : what the hell is a non-graphic illustration?) of the effect of the extra mass. If it is enought o invalidate your results then get the FE model you are correlating with to add the accelerometer mass for the correlation runs. Cheers

Greg Locock

 

[Franko]

The effect of root fixation can be much greater than that for a couple gram accelerometer attached near the base of the blade profile section. For typical turbines the longer the blade, the less reduction from looseness in the root and is a function of a particular design. At speed the blade will be held much tighter at the root lands; there is also centrifugal stiffening from blade stresses and blade untwisting that must be calculated for an accurate evaluation. Special tests and FEA may be required if you are trying to make sure each blade is "tuned" from major resonance in a turbine generator application. Accelerometer will also show predominant disk modes. Contact the turbine manufacturer - he may have test data to compare to for values in the disk at standstill.

 

During NFT of turbine blade having tie wires, I find many fundamentals in FFT spectrum but some fundamentals are not seen in other blades. How to recognise which is the 1st and 2nd mode natural frequency.Will missing frequency have any affect in interpretation or healthiness of the blade.

 

During NFT of turbine blade having tie wires, I find many fundamentals in FFT spectrum but some fundamentals are not seen in other blades. How to recognise which is the 1st and 2nd mode natural frequency.Will missing frequency have any affect in interpretation or healthiness of the blade.