Start up of Motor with VSDS 2

[rotw]

Hello,

We are considering a centrifugal compressor motor driven by VSDS.

There s a case where the machine has to start at full start up torque because the process system shall remain fully loaded.

My question is :

The VSDS operates between 60 to 105 percent. Therefore from 0 to 60 percent speed the VSDS wont be available and motor will see the full start up torque. It means the start up is direct on line from 0 to 60 percent and then the VSDS take over.

Is this scenario correct ?

In case the start up torque is very high in the range from 0 to 60 percent, does this impact the sizing of the VSDS ?

Thanks

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[davidbeach]

VSDS?

 

[rotw]

VSDS = Variable speed drive system.

Hope this clarifies this difficult and very unusual acronym.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[ScottyUK]

Sounds like the compressor has some reason - perhaps shaft critical speeds - to keep operate above 60% rated speed. I find it hard to believe that the machine will start DOL and then transfer to VSD operation, it will just accelerate through the critical speeds quickly to avoid setting up a destructive resonance and once it is above 60% rated speed it will not fall below it under normal operation.

 

[itsmoked]

That's crazy! rotw, your understanding must be somehow incomplete. There is no logical reason why a VFD cannot start that compressor as hard as DOL. So why some complex VFD ruining electromechanical time-bomb? I could see a manual switchover for a DOL back-up system with no speed control. Trying to start a motor then halfway thru, during the transient event, switching over to a VFD is... nuts.

Scotty; It's probably more like the centrifugal compressor does virtually nothing below about 60%.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[rotw]

Hi itsmoked and ScottyUK;

correct the compressor does nothing below 60 percent, furthermore critical speed (first one) could become an issue as ScottyUK correctly said.

Actually I am not pushing for starting DOL and then switch on the VFD halfway thru, for instance when the machine hits the 60 percent. Agree it does not look comforting.

As I see it Motor will accelerate from 0 to 60 percent against full start up torque and I guess this has to fast. This phase should not be too much of a concern as generally troubles occur near the 100 percent speed for what concerns stalling the motor (i.e. no sufficient margin between required and available torque).

It is good to hear that VFD is common for start up of the compressor from zero to full speed.

Still I want to investigate a scenario a bit further.

if the compressor starts with the process system fully loaded, you may have to account for dynamic effects and there might be a pick of power when motor accelearte fast from 0 to 60 percent so VFD and Motor in this case will need to be verified that they can get through this acceleration phase or not ? In this will a DOL method bring extra torque/current respective to this extreme case ?

If DOL is virtually used, how does it looks like ? is VFD by passed with direct connection from motor to bus bar is that right ?

does it complicates the design of the VFD?

Thanks again

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[jraef]

On some types of compressors, specifically those with mechanical lubrication systems, acceleration time can be critical because ramping too slowly can mean that lubrication does not reach critical components soon enough.

In the now bygone early days of VFDs that were simply V/Hz control capable, it was often not possible to make the motor create sufficient torque to accelerate it to the minimum required speed quickly enough. However, modern "Vector Control" capability has changed that. A good quality VFD with even just Sensorless Vector Control can now actually make the motor deliver MORE starting torque than you can get with DOL starting. When starting DOL, a typical Design B motor will deliver 160% of Full Load Torque (FLT) for starting, called Locked Rotor Torque (LRT), because a portion of the current going to the motor is reactive, not active, so it is not yet producing work (torque). But because a VFD with Vector Control can determined the EXACT vector of torque producing current vs flux producing current and adjust accordingly, more of the current going to the motor can be dedicated to creating torque. This means you can get the full Break Down Torque (BDT) from the motor almost instantly, so 200-220% of FLT. The only caveat is that you can only do this for as long as the VFD can handle the heat this causes inside of the components, so usually only 3 seconds or so. But in many cases, 3 seconds of BDT is better than 5 seconds of LRT. What you must do if you are concerned is to determine the real accelerating torque that you need to get past the critical speed, then work backward to see what that means to the motor, then how long it will take using that motor at BDT to get there. If it is more than the VFD overload current rating, you may need to work with your VFD supplier to over size the VFD accordingly.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[ScottyUK]

Most medium / large turbocompressors will have a main shaft-driven oil pump plus a smaller auxiliary pump for startup to prevent the lubrication problem jraef described. above.

I can't imagine that a significant power peak will occur to get through the critical speed. As you've already noted, turbocompressors do very little work at reduced speeds so the drive is essentially supplying the kinetic energy input required to accelerate the machine plus a relatively small amount for the work done.

 

[rotw]

Ok guys, many thanks for the input.
Thanks to jraef for very clear explanation.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[rotw]

jraef,

quoted
What you must do if you are concerned is to determine the real accelerating torque that you need to get past the critical speed, then work backward to see what that means to the motor, then how long it will take using that motor at BDT to get there. If it is more than the VFD overload current rating, you may need to work with your VFD supplier to over size the VFD accordingly.
unquoted

We did exactly the same and worked with the VFD supplier; actually we have identified a high power demand approaching 60% speed. Between break away torque to 50% the accelerating torque is quite flat. The time to get to 70% speed is around 5 to 6 seconds, anyway this analysis is being refined. Critical speeds are no an issue as the one of concern is crossed very quickly and is around 30% nominal speed.

The alternatives so far are :
- depressurize the system (process), this is being investigated
- ask vendor to fine tune the VFD and overload current, which may have some economical downside.

From an economical point of view, if we handle the acceleration with Locked Rotor Torque (LRT), It would take a bit longer to start the motor as rightly pointed out above. However if the thermal curves of the Motor can withstand this increased duration ultimately we could avoid oversizing the VFD - however will the scenario of (LRT) as also economical implications (means what kind of equipment is affected or should we include) ?

In this case, how does this compare to the most desired approach which is to work out the VFD sizing and have the Full break down torque ?

Thanks

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[LionelHutz]

Is this VSDS actually a VFD (a VFD varies the voltage and frequency applied to the motor) or is it some other kind of speed controlling system? You first write VSDS and then jump to using VFD only after others mention VFD. I've never seen a VFD application that transferred from DOL to VFD at part speed. Your latest post seems to be suggesting you're using the motor speed vs torque curve for acceleration calculations using a VFD which is wrong. Your posts and what you want to do is confusing and it may be because you're not talking about a VFD but another type of speed controller.

 

[rotw]

Hi LionelHutz,

As we are talking about a centrifugal compressor driven equipment it seems extremly obvious what is really meant by VSDS. No disprepect, but I dont know what else could be a VSDS if not a VFD when it comes to varying the speed of centrifugal compressors. Yes there is the option of using a hydraulic coupling or a Voith vorecon but that is definitely NOT called a VSDS.

quoted
Your latest post seems to be suggesting you're using the motor speed vs torque curve for acceleration calculations using a VFD which is wrong
unquoted

no offense here but seems your understanding of the problem is uncomplete.

There are many example of specifications for VFD converter driving compressors which and specify - when this is the case - that the entire drive system shall be capable to accelerate the driven equipment under full pressurization condition (actually it is about re-start up) ;
So what is wrong in taking into account the speed vs. torque curve acceleration as far as the ramp up from 0 rpm to minimum operating speed is concerned ? are you responsibly suggesting to ignore this requirement ?

See here an example of the consequences :

http://www.eng-tips.com/viewthread.cfm?qid=247962

By the way, the aim is not to start on DOL and switch to VFD; jraef explained this very clearly and sufficently ;
The aim is to understand the impact on sizing of the VFD when it comes to accomodate a stringent starting torque requirement due to pressurized conditions, which may apply of not depending on the application.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[mikekilroy]

rotw,

I suspect you do not fully understand English language, and that is why you replied as you did. Your reply continues to not make total sense.

LionHutz gave you a very sound reply, yet you want to discredit it??

Your link is not apples to apples to this post: it is about an UNDERSIZED vfd - not the same issue you have.

Back to English if you are willing to listen again and try to understand engineering:

A PROPERLY SIZED VFD WILL GIVE YOU THE SAME OR MORE START UP TORQUE AS DOL.

Let that sink in a minute or two....

OK, sunk in?

Now repeat after me in English: what part of that don't you understand? Why in the world would you EVER want to suggest DOL then switch to vfd when plain ol vfd gives more all the way?

http://www.KilroyWasHere<dot>com

 

[LionelHutz]

The motor speed vs torque and speed vs current curves from 0 speed to the breakdown torque speed are completely meaningless when using a motor on a VFD. You writing about the accelerating torque being quite flat from breakaway to 50% speed seems like you're describing the motor torque curve. My definition of accelerating torque is the difference between the torque produced by the motor and the torque required by the load during motor acceleration. Is this what you mean when you write accelerating torque?

A VFD never operates the motor below the breakdown torque peak of the motor curves. You take the current capability of the VFD and determine how much extra slip toward breakdown is possible before reaching that current. The torque you get at that slip is the maximum possible motor torque.

Anything that allows you to vary the speed of the load is a VSDS. A Voith Vorecon is considered a VSDS since it allows you to control the speed of the load.

 

[rotw]

Lionelhultz,

'You writing about the accelerating torque being quite flat from breakaway to 50% speed seems like you're describing the motor torque curve.'

The torque required by compressor is flat during acceleration till about 50% speed;
Motor torque was assumed to give roughly 2 times of the rated torque continously available during acceleration in the study.

What is clearly not understood in this thread is the dynamics of the load.

jraef pointed out
quote
The only caveat is that you can only do this for as long as the VFD can handle the heat this causes inside of the components, so usually only 3 seconds or so. But in many cases, 3 seconds of BDT is better than 5 seconds of LRT. What you must do if you are concerned is to determine the real accelerating torque that you need to get past the critical speed, then work backward to see what that means to the motor, then how long it will take using that motor at BDT to get there. If it is more than the VFD overload current rating, you may need to work with your VFD supplier to over size the VFD accordingly.
unquote

That is exactly what I am trying to find out; means the impact on the VFD and what are the alternatives.

So Lionelhultz, motor speed vs torque curve are completely part of the picture for acceleration calculations certainly in this particular case; The accelerating torque being the difference between the available and required torque curve. We are clear on that.

I will skip the one message above which I found to be very unpolite and not bringing any values to the point which I adressed.
It is by the way quite arrogant style of communicating.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[jraef]

I dont know what else could be a VSDS if not a VFD when it comes to varying the speed of centrifugal compressors.

Just to clear up this part of it, using the term "VSDS" could have been inclusive of a DC drive or a mechanical vari-drive system, such as a variable pitch belt drive.

So LionelHutz was correct in that we (most of us) did make a leap of faith that you were referring to a VFD.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[LionelHutz]

What is clearly not understood in this thread is the dynamics of the load.

So, you realize people need to know the load info yet you can't be bothered to post the load info? Every time you're questioned you post another little crumb of info about the system you're dealing with. Honestly, I really could care less about helping someone, for free, when I have to drag every little bit of useful info out of them.

In my experience, the speed vs torque curve is often unreliable between breakdown torque and synchronous speed. Either there is not enough resolution in the curves or the curves are just plain wrong in that area or both. You can post all you want about the speed vs torque curve being useful, but all that is really useful is the torque you can get from the motor while running from the VFD.

The only thing arrogant here is your responses when questioned. Apparently, you expect everyone to follow your rambling and partial explanations. Good luck.

 

[waross]

Do I have this right Jeff?
For DOL starting the complete torque curve must be considered, from zero speed up to running speed.
For VFD starting the torque related to the slip frequency must be considered. That is, the torque curve is read backwards from full load speed torque down to breakdown torque. With current limiting properly implemented in a VFD, the slip may be determined and may be increased up to the point of breakdown torque, which is greater than the locked rotor torque. This increase over locked rotor torque is available from zero speed up to normal breakdown torque speed.
As an example, if breakdown torque occurs at 50% speed, then we may say that breakdown torque occurs at about 50% slip.
For DOL starting, the entire torque curve is of interest.
For properly applied VFD starting, The top end of the torque curve, from full load torque and speed down to breakdown torque and speed, is of most interest.
Put another way:
For DOL starting we must consider the complete torque curve.
For properly applied VFD starting we are more concerned with the portion of the torque curve from locked rotor torque speed up to full load speed, and with proper application of the slip frequency we may utilize any portion of the torque curve at at speeds below breakdown torque speed.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[rotw]

LionelHutz,

I did not mean to offend you. I have appreciated your help.
My comment was directed towards the post of mikekilroy which I found harsh and I maintain it.

I have to admit that VSDS can be used for hydraulic couplings, conceptually it is not wrong to call this VSDS - although I have never heard it too often really ; sometimes vorecon is refered as VSD or ASD (adjustable speed drive); however on this one I stand corrected.

In my experience, the speed vs torque curve is often unreliable between breakdown torque and synchronous speed. Either there is not enough resolution in the curves or the curves are just plain wrong in that area or both. You can post all you want about the speed vs torque curve being useful, but all that is really useful is the torque you can get from the motor while running from the VFD.

Yes true; but when it is generated through a study its a different story ; it is as accurate as the assumptions used to build the model.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[sebber]

I'm a few months late, but just for whoever reads this thread:

VSD (or sometimes VSDS) is just a general term for a drive that varies speed, no matter what technology, electric (=VFD), hydrodynamic, mechanical...

The thread mentions the Voith Vorecon - one Vorecon type, mainly used for compressors, has limited speed control capability, typically from 60 to 105% speed or so. I.e. while the motor runs up to full speed, the compressor runs up to minimum Vorecon speed, i.e. 60%. This seems to be exactly what rotw describes in the first post of the thread.

Which makes me believe that while this thread discusses mainly VFD, the variable speed drive they will actually use is a Vorecon!