One reason for torsional vibration issue:
Torque ripple
In this post we look at torque ripple, its background, character and impact on the shaft string. After reading through you will have a brief idea what the torque ripple is and how to deal with it in a variable speed drive system.
Torque ripple is often believed to be the most common root cause of torsional vibration. True or myth? You can make your own opinion at the end of this series. First we shall explain what a torque ripple is and how it affects the rotordynamics of machinery and its torsional performance.
From mechanical perspective the torsional behavior of VFD fed drive system is one of the most important aspects. Excessive torsional vibration might lead from occasional trips, increased noise, accelerated mechanical wear up to a mechanical failure.
Torque ripple: How much is acceptable?
When talking about torque ripple many people are concerned about the percentage value. Is 5% okay? Is 10% okay? Well, the ripple (referring to the mean torque value) can range from 2% to about 12%. Certain VFD soft starters may have higher ripple – up to 20%. However, we shall keep in mind that these VFDs are not intended for continuous operation.
While the value of torque ripple enjoys lot of attention, it is not the most important quantity. In fact, much more relevant is the spectrum of motor torque.
Remark: When talking about torque ripple as a percentage value it shall be clearly specified what the reference is. Most often we consider the rated torque as our base (rated torque = 100%). However, sometimes the ripple can be expressed as percentage of the mean value for given operating point. The latter approach naturally gives higher percentage value of torque ripple for load points with reduced torque. Many algorithms, especially in frequency domain, would use the latter approach as the algorithm usually does not know the value of rated torque.
Torque spectrum
Instead of judging the VFD based on its percentage value of torque ripple it may make much more sense to focus on the spectrum of motor torque.
You may ask your rotordynamicist about his preference:
A) VFD X has overall torque ripple of 2% and there is a torque component matching exactly the first torsional natural frequency of the shaft string. Magnitude of this torque component is 0.5%.
B) VFD Y has somewhat higher torque ripple – approx. 7%. However, none of the torque components interfere with the torsional natural frequencies of the shaft string.
My bet is that most rotordynamicist would choose case B. Why? It is because the motor (air gap) torque is less relevant. What matters is the alternating torque that the sensitive shaft component is exposed to.
Key is to understand that we deal with resonances. Whether it is an electrical resonance in the grid or a mechanical resonance inside the shaft system they all have one important factor. This is the amplification factor (AF).
An example of motor air gap torque is shown above. Figure 1a depicts motor torque in time domain with its characteristic ripple. Figure 1b visualizes the same motor torque as a spectrum in frequency domain.
In order to accurately predict the torsional behavior of a variable speed drive system it is important to understand the frequency spectrum of motor torque. Especially, one need to know how the torque spectrum evolves with motor speed.
Hint:
Most of the VFDs have certain “characteristic” torque components plus other “non-characteristic” ones. The “non-characteristic” components (e.g. interharmonics) are lower in magnitude, but often tend to be more dangerous (especially when not properly identified). While characteristic harmonics usually have straight forward correlation to the machine speed, the relationship between frequency of non-characteristic torque component and machine speed may be more complicated.
Amplification factor and damping
Damping of the resonance is crucial. It defines how the response of the system looks like. Amplification factor is inverse proportional to the damping ratio.
AF = 1/(2⋅d)
AF … Amplification Factor
d … damping ratio
Unfortunately, most rotordynamic systems have relatively low damping and corresponding high amplification when the torsional mode gets excited.
In turbomachinery, the damping ratio may be much less than 1%. Therefore, even a small excitation torque causes large response.
Campbell diagram
A visual representation of the torsional excitation and an interaction with torsional modes provides so called Campbell diagram. In such a plot the torsional excitation is depicted as function of machine speed. Torsional natural frequencies are depicted as well. Any crossing of a torsional natural frequency with one of the excitation components indicates a potential critical speed.
Figure 2 displays an example of a Campbell diagram for typical 4-pole electric motor supplied from a VFD with a speed range from 70% to 105% of nominal motor speed. Green vertical lines indicate extended speed range taking an additional separation margin into account. That is a common practice defined in various standards (in this example API standards frequently used in petrochemical industry).
Ideally, the shaft string is aimed to be designed in a way that there are no intersections of motor torque components with relevant torsional natural frequencies inside the operating speed range. Yet, this is sometimes not possible and an intersection within the operating speed range is unavoidable. VFD software typically allows to parameterize one or more restricted speed ranges. That is one way how to deal with such issue. Of course, the drawback is more limited operation where the basic advantage of a variable speed system cannot be fully utilized. Another way is to allow operation at the critical speed if the torsional response remains within the mechanical design limits. Situations like that are mostly evaluated case by case.
Hungry to know more? This post had been shortened. Subscribe to our premium content and get access to full article and more.
Upcoming rotordynamics training
Do you want to expand your knowledge about rotordynamics? To understand the vibration problems you are facing and hopefully also find a solution?
The 2021 Spring Rotordynamics Training will take place in March.
The event is jointly organized by No Bull Engineering and Xdot Engineering.
Check out this website to learn more:
https://www.rotordynamicscourse.com/
The training is organized by recognized professionals:
Malcolm Leader, P.E. (President – Applied Machinery Dynamics)
Dr. Erik Swanson, P.E. (President – Xdot Engineering and Analysis)
Mark A. Corbo, P.E. (President – No Bull Engineering)
Each of the experts is covering one major topic of the training course.
Summary
Pulsating torque ripple is a “natural part” of inverter duty motors. Many users and consultants try to limit the torque ripple in faith that it will eliminate or at least reduce the torsional vibration problems and mechanical failures. While this approach is somewhat understandable, it makes much more sense to look at the frequency spectrum of the pulsating torque. Therefore, you shall check the frequency components in the motor torque and correlate them with the torsional natural frequencies of the shaft string.
How do we continue?
As stated at the beginning of this article, torque ripple is often believed to be the primary reason for torsional vibration issues. But what if we told you that there is another root cause that is even more relevant and definitely more dangerous? The other reason associated with torsional vibration problems will be revealed in our next post. Stay tuned!
References
[1] Introduction to electro-mechanical interaction, https://mb-drive-services.com/elmech-intro/
[2] Growing trend of torsional vibration issues, https://mb-drive-services.com/trend-of-vibration-issues/
[3] Medium voltage AC drives, https://new.abb.com/drives/medium-voltage-ac-drives
