Energy Savings Calculator for VFD applications?

# Thread: Energy Savings Calculator for VFD applications?

1. ## Energy Savings Calculator for VFD applications?

Does anyone have a reference that can be used to estimate energy savings when applying VFDs to electric motors that would otherwise run 95% of the time completely unloaded?

I have an application where a 100HP motor runs 24x7x365 to pump fluid in a loop. A few times a day the fluid is diverted to a high elevation tank using about 90HP to make the elevation change and then when the tank is topped off, the fluid resumes its circulation route consuming only a few HP.

The nature of the application is such that we cannot stop circulation of the fluid. As such, the 100HP motor is running at a fraction of its load the vast majority of the time. As I understand it, an Induction motor running at partial load has a terrible power factor.

If I were to put a VFD in front of the motor, I could slow the motor down when in circulation mode to the minimum velocity required to keep the fluid mixed and then quickly ramp up during the time that pumping is needed.

My thought would be that the VFD has a close to 1.0 Power Factor and I could additionally reduce power consumption during the idle period.

I would like to find an energy savings document or online tool that helps me document what these energy savings might be . . . is anyone aware of a resource that could help me with this?

Thanks in advance . . .

2. Diamond
Join Date
Dec 2008
Location
Vt USA
Posts
6,736
Post Thanks / Like
Likes (Given)
736
2316
Good question

I just googled " VFD energy saving" and got returns that I'm reading now.

It might seem that there is a "free lunch" to be had. That is, input amps can be lower than output amps. ;-)

ETA
A calculator ;-)
Payback - VFD Calculator - Drives - Products & Services - United States - WEG

3. Stainless
Join Date
Aug 2014
Location
bainbridge island
Posts
1,129
Post Thanks / Like
Likes (Given)
215
267
The data I have from my 5hp 3p motor when run on single phase is as follows
power consumption is:
250-350 watts no load across the nominal range of 208-240vac.
48-44 watts no load at at 16 to 20 volts ac. note that below 16vac, power consumption increases until it looses synchronization. 20vac was about "ideal" at no load.

running the motor from true three phase will cut those no load numbers in half, but the ratio would remain nearly the same, about 7.5:1.

Assuming your motor only needs say, 5 hp to circulate at say 20% of nominal rpm:

You may be able to get away with programming a linear volts per hz increase, starting out at say, 1v/hz (6v/6hz), increasing to 8v/hz (480v/60hz) as the frequency increases from 6hz to 60. note that in your case, a 8:1 ratio of volts per hz, starting at 1 and increasing to 8, may be too high a ratio and won't start the pump reliably. but if there is little bearing and seal friction, then yes my example of running a 480v centrifugal pump off of 6 volts at 6hz, that's all it should take to run the pump.

Not only do you get the benefit of the reduction of heat losses into the fluid (pumping at full speed into a recirculating pipe) but you get the benefit of nearly zero copper losses in the motor itself (at the slower speeds)*.

Theoretically setting a vfd to "vector mode" will take care of the volts per hz and the vfd will figure out how to drive the motor at the optimal flux density. however, i find very little literature concerning this. it seems that what most vfds call "vector mode" is only half of it: optimizing the change in volts/hz to optimize acceleration and deceleration times without overshooting the rpm target. But its not going to sit there and experiment with dropping the volts/hz while changing the frequency to keep the shaft at the same speed, while calculating the efficiency of the motor to figure out where the optimum efficiency point is.

However if you could find a vfd that will take an encoder input, in theory a proper "vector mode" control will drop the volts per hz as the torque on the motor shaft is decreased. To some extent this requires an encoder on the motor because the inverter can't (reliably) tell the difference between power leaving the motor through the shaft, and power leaving the motor windings (and magnetic core) in the form of heat. (though there are ways to do so, they may not be reliable)

* Where as if you kept the nominal volts per hz, which would be 48 volts at 6hz, you would have the same current flowing in the motor windings as you would at 480v/60hz at no load (which is usually figured to be 30% of nameplate full load amps). (This is why even inverter pump motors may only advertise a 10:1 speed ratio, because the inverter is still there pumping nameplate v/hz into the motor **, even though it only needs a tenth or a fifth of it, and the fan needs to run at least 200rpm so it can blow away the no load static power loss of the motor, which for my 5hp motor is on the order of (i'm assuming,) 150 watts. (at a tenth of the nominal rpm, the pump won't consume enough power to even matter and the copper loss in the motor will exceed the pump demand, by a factor of about 3. (not including friction in the pump seals which could be significant.)

**or worse, its pumping more than nameplate volts per hz into it because that's the default "boost" setting normally used just to start the motor quickly.

if you're asking why spend all this trouble.. it might be worth 1000\$ a year to you if that 100hp motor sucks up 2000 watts even though it only needs to deliver 1 hp to the pump. by reducing the volts per hz you should be able to get the motor back up to 90% efficiency and you should only see perhaps 1000 watts of power to run the pump at 10% of its nominal rpm, if that works out. but that cubic power ratio of centrifugal pumps is really something.

recently i tore apart a trashed hot tub two speed pump. 1.5hp at 3500 rpm.. 0.18hp at 1750. it had two separate start and run windings.

4. Most VFDs have horrible power factors, due to the rectifier/capacitor input. Since it is a delay of the current until near peak voltage, it probably is not helping much, even though it involves a capacitor.

Power factor gets worse with light loading of the VFD....because the current is short pulses near peak voltage. RMS current is higher than the actual average current would indicate, for the same reason. Only VFDs with a PFC front end will fix that, and they are still both rare and pricey even now.

The benefit of the VFD comes when you can slow the motor, and reduce the output voltage. The VFD acts as a "buck regulator", and in those voltage conditions, output current may be larger than input current. So even the lousy power factor is on a lower current.

A VFD may make a difference if it is tied into the control system as suggested. Of course, there is also the possibility of just using a small motor, that can be more heavily loaded vs capability, for the ordinary use, and using the 100HP ONLY for the "pump-up" condition.

5. Originally Posted by CalG
Good question

I just googled "VFD energy saving" and got returns that I'm reading now.

It might seem that there is a "free lunch" to be had. That is, input amps can be lower than output amps. ;-)

ETA
A calculator ;-)
Payback - VFD Calculator - Drives - Products & Services - United States - WEG
Thanks, I hadnt seen the Weg calculator. I'll give it a go and see what it says. I was hoping to find a general concise white paper so I could develop a spreadsheet solution. Most of the calculators are optimized to the features of the drive being sold by that supplier.

Johansen - generally most AC Vector mode or "field oriented control" algorithms specifically maintain a constant Volts / Hz ratio regardless of speed and load (assuming you have enough line voltage to support the requested motor speed). FOC algorithms reduce torque producing current when lightly loaded but the terminal voltage is varied as a function only of motor speed.

What you are suggesting is field weakening the motor when less than rated torque is required. There is a modified FOC algorithm that does this developed by a company called AC Kinetics and I have some experience with this company and its algorithm. I have written a few parametric FOC routines that do what you suggest in addition to increasing the V/Hz ratio at low speeds for higher torque capacity at lower current draw for short periods of time, but the overhead required to support these kinds of applications generally has not been worth the effort long term.

Given that my focus has always been on performance at the motor shaft and not energy savings, I had not considered the energy savings implications of methods used to alter the flux density during differing operation conditions of the motor.

I'll stick with the simple ramping down/up/down of motor speed for energy savings as this is simple to implement and easily understood by the end user who ultimately needs to maintain the application long term.

6. Stainless
Join Date
Aug 2014
Location
bainbridge island
Posts
1,129
Post Thanks / Like
Likes (Given)
215
267
Originally Posted by motion guru
Johansen - [...]

What you are suggesting is field weakening the motor when less than rated torque is required. There is a modified FOC algorithm that does this developed by a company called AC Kinetics and I have some experience with this company and its algorithm. I have written a few parametric FOC routines that do what you suggest in addition to increasing the V/Hz ratio at low speeds for higher torque capacity at lower current draw for short periods of time, but the overhead required to support these kinds of applications generally has not been worth the effort long term.

Given that my focus has always been on performance at the motor shaft and not energy savings, I had not considered the energy savings implications of methods used to alter the flux density during differing operation conditions of the motor.

Interesting, thanks. I think you'll find the savings significant, though there is a cost associated with reducing the air gap flux, because you can't increase the field strength instantly, it takes time due to the limited voltage available from the vfd, but its on the order of less than a second I suppose for a 100 hp motor.

but you're dealing with a pump, it probably doesn't matter if it takes an extra second to get up to speed.

7. Stainless
Join Date
Jun 2013
Location
Northern Il
Posts
1,279
Post Thanks / Like
Likes (Given)
672
1213
Motion,
I think you will find that a substantial portion of your energy savings will be in the reduced energy transfer to the fluid you are pumping. i.e. heat

To get an accurate number you really need to determine what the real head loss of the piping is at the given fluid velocities. This varies with the square of the velocity so it does make a very big difference, especially if the piping it self is less than optimal which is usually the case.

Without knowing much more about the piping and the fluid characteristics such as viscosity and specific gravity can't give you too much more but a 25% reduction in pump power savings is very common.

Electrical engineers usually attack problems such as this from purely a mechanical side but in reality the fluid dynamics side of the equation is where the savings are really at. This is especially true if a chiller is being used to control fluid temperature, you have a compound savings if that is the case.

8. Hi Ziggy . . . yes, we are looking at that as well. Over 240 feet of piping involved in the system and I have an engineer on staff with a PhD in non-Newtonian fluid dynamics who will work those numbers. I am the Mechanical engineer who shifted to drive design for a living who will figure out the electrical savings . . . the hope is that we can demonstrate enough savings to justify the cost of the drive in short order.

9. Stainless
Join Date
May 2017
Country
FINLAND
Posts
1,634
Post Thanks / Like
Likes (Given)
465
762
Originally Posted by johansen
However if you could find a vfd that will take an encoder input, in theory a proper "vector mode" control will drop the volts per hz as the torque on the motor shaft is decreased. To some extent this requires an encoder on the motor because the inverter can't (reliably) tell the difference between power leaving the motor through the shaft, and power leaving the motor windings (and magnetic core) in the form of heat. (though there are ways to do so, they may not be reliable)
Isn't that sort of what ABB does with their DTC(direct torque control)

10. Stainless
Join Date
Jun 2013
Location
Northern Il
Posts
1,279
Post Thanks / Like
Likes (Given)
672
1213
Originally Posted by motion guru
Hi Ziggy . . . yes, we are looking at that as well. Over 240 feet of piping involved in the system and I have an engineer on staff with a PhD in non-Newtonian fluid dynamics who will work those numbers. I am the Mechanical engineer who shifted to drive design for a living who will figure out the electrical savings . . . the hope is that we can demonstrate enough savings to justify the cost of the drive in short order.
Our company has been doing a lot of retrofits on many of our filter systems that did not originally have VFDS on the pumps. Many these systems had total aggregate pump horsepower approaching 750hp. We initially went through the numbers with the customers and we were both rather skeptical but after the installs reality proved that the initial calculations for cost savings were very accurate. we were happy and the customers were very happy.

In our applications, I think we were around 6mo. payback.

Sounds like you are putting together are rather impressive engineering staff. Congratulations on that.

11. Originally Posted by Ziggy2
Our company has been doing a lot of retrofits on many of our filter systems that did not originally have VFDS on the pumps. Many these systems had total aggregate pump horsepower approaching 750hp. We initially went through the numbers with the customers and we were both rather skeptical but after the installs reality proved that the initial calculations for cost savings were very accurate. we were happy and the customers were very happy.

In our applications, I think we were around 6mo. payback.

Sounds like you are putting together are rather impressive engineering staff. Congratulations on that.
Hi Ziggy . . . yes we are doing some fun stuff in a wide variety of applications . . . getting ready to install a bunch of 450HP drives just north of you in Wisconsin.

I drove through Northern Illinois 2 weeks ago helping my daughter and son-in-law move to Hillsdale Michigan where he will go to school and I'll be back in Chicago in a few weeks for a supplier meeting. We should grab a dinner some time when I am in your neck of the woods.

12. Stainless
Join Date
Jun 2013
Location
Northern Il
Posts
1,279
Post Thanks / Like
Likes (Given)
672
1213
Originally Posted by motion guru
Hi Ziggy . . . yes we are doing some fun stuff in a wide variety of applications . . . getting ready to install a bunch of 450HP drives just north of you in Wisconsin.

I drove through Northern Illinois 2 weeks ago helping my daughter and son-in-law move to Hillsdale Michigan where he will go to school and I'll be back in Chicago in a few weeks for a supplier meeting. We should grab a dinner some time when I am in your neck of the woods.
I would enjoy that very much.