Vibratory stress relief question 2

[Pakou]

Hi,

I have been doing some searching about Vibratory stress relief. As I am quite new in this area I wonder if you could help me. The first thing that shocked me is that some articles I read recommend sub-harmonic vibration in order to stress relief the component. I always thought that shaking at resonance would introduce more energy to the system, so it will be better relieved. Some other articles, however, introduce resonant frequencies. My first doubt to start with, is:

what do you think it´s better? Sub-harmonic or harmonic frequency?

Second:

If I have a piece supossed to have residual stresses with "X" natural frequency. If I apply vibration and the frequency changes, is it due to stress relieve, or may be affected by other factors?

Every information you can give will be well received.

Regards.

 

[metengr]

Some threads to review after conducting a search for vibratory stress relief;

thread330-199377
thread330-166052
thread330-131445

 

[bklauba]

In the VSR On-Line Research Library (

http://www.vsrtechnology.net/library-articles.html

), there are three papers that specifically address your question:

1. Hahn (5th entry), Alfred University

2. Shankar (12th entry), Oregon Graduate Center (now part of the University of Oregon)

3. Yang, Jung & Yancey (15th entry), Edison Welding Institute

All three reach the same conclusion: Both resonant and non-resonant (sub-resonance) vibration can relieve stress, but resonance indeed (as your OP states) is more effective.

BK

 

[Pakou]

thanks to all for your posts.
bklauba, I knew this website and have read some of the papers. Thanks.
Now my concern is if the natural frequency change is only due to stress relieveing or somone else can affect.
I want to do a test in a piece, take its natural freq., apply vibration, measure again until it stops changing. Then i will know that all the stresses are relieved.
Do you think this procedure is reliable?

Regards

 

[bklauba]

During vibratory stress relief, both the resonance frequency AND amplitude change: The frequency decreases while the amplitude increases. In fact, in the vast majority of applications, the change in amplitude, percentage-wise, is far greater than the change in resonance frequency. I have attached a VSR Treatment Chart of a 150 ton capacity lifting yoke, as an example. (The full report for this job can be seen at:

http://www.vsrtechnology.net/pdf/The_Rose_Corporation.pdf

)

VSR Treatment Charts have two plots: An upper plot of workpiece acceleration (g's) and a lower plot of vibrator input power (100% = 3 HP, ~ 2.2 kW) ,both of these are vertical axis, vs. a common horizontal axis of vibrator RPM, data generated from the motor's resolver data. Speed regulation is about 0.03%, and fine-tuning increment is one RPM. The Pre-Treatment Scan is recorded in green (since the part's is "green"), and the Post-Treatment Scan is recorded in red.

Vibrator power data is used to help the operator make informed judgements on setup details, such as vibrator placement/orientation and unbalance settings. Note the large peak in the power curve, coincident with one of the major resonances.

Both plots are made by sweeping the vibrator speed at a rate of 10 RPM/sec, which we have found produces a high-resolution data set.

Note that there is a change in resonance frequency, but it is rather subtle, compared with the change in amplitude. There are some exceptions to this general pattern, but they are exceptions: More than 90% of the charts we see, whether generated by ourselves for subcontract work, or sent in by customers for review/support, show peak growth as being the dominant change during stress relief.

BK

 

[bklauba]

Addressing further the resonant amplitude vs. frequency parameter; which is the best able to monitor the progress of vibration treatment? The attached paper (published by the ASM, May 2005) shows an interesting phenomenon: see Figs. 5, 6 and 7. The reason that the resonance frequency undergoes subtle shifting might well be CAUSED by peak growth.

There appears to be a subtle, but measurable, shift of a resonance peak, in the direction of lower frequency, as resonance amplitude is increased. The data shown was generated by taking a series of VSR Charts (single sweep), and incrementally increasing the vibrator unbalance. As the resonance amplitude increases, the resonance frequency decreases.

The test part was a decades old T-slotted cast iron base, which we use for testing vibrators and performing in-house stress relief work.

The speed regulation and accuracy of the equipment used for this work was about 100 times tighter than the shift that took place, so we have confidence that the effect we are seeing is real.

Bottom line: Resonance amplitude is, in the vast majority of cases, a better parameter to monitor than resonance frequency, to monitor the progress of vibratory stress relief, but make sure that any equipment you use has good speed regulation, so that the data is meaningful.

BK

 

[Pakou]

thanks everybody for your help, specially bklauba. Very interesting info. Now I can understand VSR much better