Variable Freq. Drives

[TxEngr]

We've had a great deal of theoretical discussion here but did Rey-Man get his question answered?

To summarize -

• Yes, you can apply a VFD to this application to help control the flow. This has the advantage of energy savings over other options.
• An alternative is to trim the pump impeller, but you can only trim the impeller so much and the effeciency losses become large.
• A VFD has a higher up-front cost, but it's long term cost is smaller with typical paybacks in the 1-3 year time frame on 500 HP and below applications
• personal experience on this number).

A few other suggestions as you look at this:

• Don't let the speed of the motor drop below 33% of rated RPMs for any extended period of time. The decreased airflow can adversely affect the motor.
• If you must run an 1800 rpm motor for extended periods at slower speeds, have your motor shop install a fan for a 1200 rpm motor to improve the air flow.
• Watch you cable lengths if you go with a VFD. If you have really long cable runs, it can adversely affect the motors life.
• Have someone who can read pump curves look at the application with you to make sure you're making the best choice.

Now on with the theoretical discussions!

TxEngr

 

[Jraef]

We've had a great deal of theoretical discussion here but did Rey-Man get his question answered?

To summarize -

...snip

Now on with the theoretical discussions!

TxEngr

Nicely done.

 

[Besoeker]

So if you are telling me that adding a vfd to reduce load, rpm, temperature to get an increase in life-cycle - I'm going to suggest that the system is really poorly designed.

I don't think anyone with any knowledge of VSDs would suggest installing them solely for all such benefits. Some of them just come as a bonus.
Reduced mechanical shock, motor starting stress, projected insulation life etc.

http://ecmweb.com/mag/electric_hot_issue_motor/

As a general rule of thumb, insulation life doubles for each 10?C of unused insulation temperature capability. For example, if you design a motor to have a total temperature of 110?C (including ambient, rise, and hot spot allowance), but build it with a Class B (130?C) system, an unused capacity of 20?C exists. This extra margin raises the expected motor insulation life from 20,000 hr to 80,000 hr.

http://books.google.co.uk/books?id=8mQewRR_3q4C&pg=RA1-PA66&lpg=RA1-PA66&dq=direct+online+motor+starting+and+mechanical+shock&source=bl&ots=TdkNoFI4hX&sig=OFlK_yqAsXKGESs_y3IZpN_n0TU&hl=en&ei=GbJ5Soq3DIqRjAeJt-mnBg&sa=X&oi=book_result&ct=result&resnum=2#v=onepage&q=&f=false
Page 66

 

[weressl]

Don't let the speed of the motor drop below 33% of rated RPMs for any extended period of time. The decreased airflow can adversely affect the motor.

Not quite true. Variable torque aplication can be safely run to the 1:10 turndown ratio, since the current and hence the generated heat decerease is nonlinear.

If you must run an 1800 rpm motor for extended periods at slower speeds, have your motor shop install a fan for a 1200 rpm motor to improve the air flow.

That is not just for one specific RPM, but accross the board for variable torque loads. Constant torque load requires the external fan in almost all cases.

Watch you cable lengths if you go with a VFD. If you have really long cable runs, it can adversely affect the motors life.

Manufacturers instructions specify the maximum length for each type cables and none of the values are defined as 'too long'.
It can also effect the actual avilable shaft horsepower and increased interference. The length on some models also vary with the chosen carrier frequency.

Have someone who can read pump curves look at the application with you to make sure you're making the best choice.

Seldom will you have curves available for varying speed, so you have to request that separately from the manufacturer and intrepolate it between the ranges. Make sure you get the corresponding HP and efficiencies plotted, not just the head/pressure. Make sure you specify the minimum speed you want to run, or ask for confirmation from the pump manufacturer.

On large motors watch out for the forced oil flow quantities if it is a shaft driven pump.

 

[TxEngr]

weressl,

My comments on the 33% speed had to do with motor cooling, not the turndown of the drive/motor/pump combination. Yes, if you use a TEAO motor, you can get the full range including 100% torque at zero speed. But for a standard motor with no additional cooling, I use the 1/3 rated rpm as a rule of thumb and often put a minimum rpm/hz limit at that point on the drive to prevent going slower. Also, the installation of the 1200 rpm fan on the 1800 rpm motor is a 'poor man's' method of getting improved cooling on an existing motor without the cost of a separate cooling fan.

I think we're both in agreement here that a VFD is a great application for what rey-man is asking. I was concerned that we had gotten a little too technical and was trying to condense the discussion down to key points that he could use to ask his vendors better questions.

TxEngr

 

[weressl]

weressl,

My comments on the 33% speed had to do with motor cooling, not the turndown of the drive/motor/pump combination. TxEngr

So did my comment. 1:10 turndown ratio on a TEFC motor with variable torque application has proven to be more than adequate cooling.

The 30 or 33% perhaps comes from the fact that centrifugal pump and fan curves become very flat below that general area and become an inefficent region to use it for control. So for those common applications some people choose the lover limt of the controller to be set, but not because of the inadequacy of cooling.

If you look at motor manufacturers instructions they also cite 1:10 ratio for VT applications without modification.

 

[Cold Fusion]

Reducing load and speed will increase the bearing life BEYOND it's rated speed/load design.

Please explain the physics how the reduction of load an rpm will shorten the bearing life. ---

This is getting way out of topic. I'll explain this one but no more.

Here is a quote from SKF:
Dynamic bearing loads - Requisite minimum loadThe correlation between load and service life is less evident at very light loads. Other failure mechanisms than fatigue are determining.
To achieve satisfactory operation, ball and roller bearings must always be subjected to a given minimum load. A general "rule of thumb" indicates that minimum loads corresponding to 0,02 C should be imposed on roller bearings and minimum loads corresponding to 0,01 C on ball bearings. The importance of applying this minimum load increases where accelerations in the bearing are high, and where speeds are in the region of 50% or more of the limiting speeds quoted in the product tables, see section "Speeds and vibration". If minimum load requirements cannot be met, NoWear? bearings could be considered.
Recommendations for calculating the requisite minimum loads for the different bearing types are provided in the product sections.

The correlation between load and service life is less evident at very light loads. Other failure mechanisms than fatigue are determining.
To achieve satisfactory operation, ball and roller bearings must always be subjected to a given minimum load. A general "rule of thumb" indicates that minimum loads corresponding to 0,02 C should be imposed on roller bearings and minimum loads corresponding to 0,01 C on ball bearings. The importance of applying this minimum load increases where accelerations in the bearing are high, and where speeds are in the region of 50% or more of the limiting speeds quoted in the product tables, see section "Speeds and vibration". If minimum load requirements cannot be met, NoWear? bearings could be considered.
Recommendations for calculating the requisite minimum loads for the different bearing types are provided in the product sections.


In short, the balls skid instead of roll.

cf

 

[hillbilly]

In short, the balls skid instead of roll.

cf

Good analogy.:smile:

steve

 

[weressl]

This is getting way out of topic. I'll explain this one but no more.

Here is a quote from SKF:
Dynamic bearing loads - Requisite minimum loadThe correlation between load and service life is less evident at very light loads. Other failure mechanisms than fatigue are determining.
To achieve satisfactory operation, ball and roller bearings must always be subjected to a given minimum load. A general "rule of thumb" indicates that minimum loads corresponding to 0,02 C should be imposed on roller bearings and minimum loads corresponding to 0,01 C on ball bearings. The importance of applying this minimum load increases where accelerations in the bearing are high, and where speeds are in the region of 50% or more of the limiting speeds quoted in the product tables, see section "Speeds and vibration". If minimum load requirements cannot be met, NoWear? bearings could be considered.
Recommendations for calculating the requisite minimum loads for the different bearing types are provided in the product sections.

The correlation between load and service life is less evident at very light loads. Other failure mechanisms than fatigue are determining.
To achieve satisfactory operation, ball and roller bearings must always be subjected to a given minimum load. A general "rule of thumb" indicates that minimum loads corresponding to 0,02 C should be imposed on roller bearings and minimum loads corresponding to 0,01 C on ball bearings. The importance of applying this minimum load increases where accelerations in the bearing are high, and where speeds are in the region of 50% or more of the limiting speeds quoted in the product tables, see section "Speeds and vibration". If minimum load requirements cannot be met, NoWear? bearings could be considered.
Recommendations for calculating the requisite minimum loads for the different bearing types are provided in the product sections.

If that would be true in the case of motors - not a theoretical aplication - would the manufacturer include a warning that the motors should not be run without a load? Of course not since the rotor itself provides this base loading, not to mention the coupled load on the shaft in the REAL LIFE case we are discussing. See my other post concerning how bearing life is calculated. Both reduced speed and reduced load will INCREASE the life expectancy of the bearing.

In short, the balls skid instead of roll.

cf

Is this your conclusion? As the narrative above, which is not separated by quotation marks, does not contain any logical link to draw that conclusion.

 

[winnie]

The bearing life equation given in post 40 is an approximation which will fall apart in various corner cases. My understanding is that these corner cases have to do with various forms of lubrication failure.

Cold Fusion posted one of these cases, but I agree with Laszlo that this situation is very unlikely to apply to an electric motor, VSD or no. In addition to the weight of the rotor itself, the radial magnetic forces in a motor are 50-100x greater than the tangential forces that make the motor turn. While the symmetry of the motor means that these radial forces are essentially balanced out, it is very unlikely that they will _perfectly_ balance out. The net result is that the 'self loading' of the motor bearings will avoid the minimum load problem.

The only real world situation that I am aware of, where reduced speed causes problems with standard ball and roller bearings, has to do with motors in storage. If motors are allowed to sit with no rotation at all, then vibration (from the surrounding plant) can cause the rolling elements to push through the lubricating film and produce direct metal to metal contact with the bearing races. When these motors are then placed in service, they suffer significantly reduced bearing life. Many manufacturers recommend regular manual rotation of motor shafts when the motors are in storage for extended periods.

The VSD versus bearing life issue that I am aware of is capacitive coupling of current to the rotor; this circuit is closed via the bearings. Micro arcing in the bearings can greatly reduce bearing life. Proper design (either grounding or insulation) can eliminate this problem, but it is another thing to consider.

-Jon

 

[Besoeker]

The only real world situation that I am aware of, where reduced speed causes problems with standard ball and roller bearings, has to do with motors in storage. If motors are allowed to sit with no rotation at all, then vibration (from the surrounding plant) can cause the rolling elements to push through the lubricating film and produce direct metal to metal contact with the bearing races.

Sometimes called "brindling" as I recall.

The VSD versus bearing life issue that I am aware of is capacitive coupling of current to the rotor; this circuit is closed via the bearings. Micro arcing in the bearings can greatly reduce bearing life. Proper design (either grounding or insulation) can eliminate this problem, but it is another thing to consider.

I agree. GAMBICA and REMA (rotating electrical machines association) produced a fairly good guide on this in 2002 with a range of mitigation measures.
In my opinion, a good output reactor or filter takes care of the many real or apocryphal issues.
The reduced dv/dt takes incipient partial discharge failure out to much beyond expected insulation life and reduces capacitive discharge currents which sometimes afflict bearings if known routine procedures are not adopted.

 

[JWCELECTRIC]

18 Pulse VFD's & Bearing Failure

18 Pulse VFD's & Bearing Failure

This post has turned into a great topic on VFD's, motors, Bearings etc, I think the OP has more info than he needs. My question would be has anyone come across a situation using 100HP + motor with an 18 pulse VFD where the bearings were failing and pits were developed on them? and has anyone come across a soulution?

 

[Besoeker]

This post has turned into a great topic on VFD's, motors, Bearings etc, I think the OP has more info than he needs. My question would be has anyone come across a situation using 100HP + motor with an 18 pulse VFD where the bearings were failing and pits were developed on them? and has anyone come across a soulution?

I assume that the 18-pulse refers to the input and that the output is the usual 3-phase PWM waveform?

The first few mitigation measures from the GAMBICA guide:

1. Ground potential equalisation (i.e. high frequency bonding) between VSD, motor and load. The principle is to ensure the lowest possible impedance path on the shield connection to avoid currents travelling through the bearings and back to ground. Rigorously follow the manufacturers installation procedures.
2. Reduce the PWM switching frequency if practical.
3. Use insulated bearings.
4. Use an inverter output choke or filter.

 

[weressl]

I assume that the 18-pulse refers to the input and that the output is the usual 3-phase PWM waveform?

The first few mitigation measures from the GAMBICA guide:

1. Ground potential equalisation (i.e. high frequency bonding) between VSD, motor and load. The principle is to ensure the lowest possible impedance path on the shield connection to avoid currents travelling through the bearings and back to ground. Rigorously follow the manufacturers installation procedures.
2. Reduce the PWM switching frequency if practical.
3. Use insulated bearings.
4. Use an inverter output choke or filter.

Shaft grounding: http://www.inpro-seal.com/CDR/mgs.html

Shaft circulating currents develop due to the unequal phase voltages, unequal winding and rotor resistances. It exist at 60/50Hz constant speeds. It is greatly exacerbated by the pseudo-sinewaves, leading spikes, standing wave resonances and harmonics as generated by the ASD's.

 

[Besoeker]

Shaft grounding: http://www.inpro-seal.com/CDR/mgs.html

Shaft circulating currents develop due to the unequal phase voltages, unequal winding and rotor resistances. It exist at 60/50Hz constant speeds. It is greatly exacerbated by the pseudo-sinewaves, leading spikes, standing wave resonances and harmonics as generated by the ASD's.

Yes, that was the next item on the GAMBICA list.
"7.5 Install a shaft grounding system
This technique has been applied for a number of years to reduce the low frequency currents associated with the monopolar effects of stator/rotor magnet asymmetry. The application of such devices has typically been limited to multi-megawatt high voltage machines and is very infrequently applied in Europe to standard low voltage motors although is is reportedly more widely used in the USA"

 

[Jraef]

Sometimes called "brindling" as I recall.

The word is "brinelling" (no "d") and technically, the phenomenon is called "false brinelling" for those who want to do a search and learn more. Brinelling is just exceeding the stress capacity of a bearing or other mechanical component. "False brinelling" is where there are similar effects, but are caused by different issues, one of which is a motor sitting idle for a long time while undergoing vibration from nearby machinery.

One often un-realized potential benefit of having a VFD on a motor is the ease with which you can prevent this. By virtue of the VFD's ability to rotate the motor very slowly, you can set up a control system to do so automatically every so often just to move the lubricant around in the bearings and races. A few VFDs even have timer programmability that can do this for you automatically (i.e. without another external timer/controller).

But if your motor runs at least once per day, this is not usually an issue worth worrying about unless it is subject to very high external vibration. I used to see it a lot on portable machinery (large truck mounted refrigeration compressors) because the vibration of traveling down the highways for a week at a time without being operated would cause false brinelling. We had to implement a small hydraulic pony motor powered by the trailer's brake pressure to rotate the motors once per day to solve it.

 

[Besoeker]

The word is "brinelling" (no "d") and technically, the phenomenon is called "false brinelling" for those who want to do a search and learn more.

Yes, I thought something didn't look right about the word. Correction appreciated, thanks.

 

[weressl]

The word is "brinelling" (no "d") and technically, the phenomenon is called "false brinelling" for those who want to do a search and learn more. Brinelling is just exceeding the stress capacity of a bearing or other mechanical component. "False brinelling" is where there are similar effects, but are caused by different issues, one of which is a motor sitting idle for a long time while undergoing vibration from nearby machinery.

One often un-realized potential benefit of having a VFD on a motor is the ease with which you can prevent this. By virtue of the VFD's ability to rotate the motor very slowly, you can set up a control system to do so automatically every so often just to move the lubricant around in the bearings and races. A few VFDs even have timer programmability that can do this for you automatically (i.e. without another external timer/controller).

But if your motor runs at least once per day, this is not usually an issue worth worrying about unless it is subject to very high external vibration. I used to see it a lot on portable machinery (large truck mounted refrigeration compressors) because the vibration of traveling down the highways for a week at a time without being operated would cause false brinelling. We had to implement a small hydraulic pony motor powered by the trailer's brake pressure to rotate the motors once per day to solve it.

Clever solution!

 

[weressl]

Yes, I thought something didn't look right about the word. Correction appreciated, thanks.

How about fluting, cratering?

http://www.reliance.com/pdf/motors/whitepapers/raps868.pdf elaborates on for those more interested.

 

[Besoeker]

How about fluting, cratering?

http://www.reliance.com/pdf/motors/whitepapers/raps868.pdf elaborates on for those more interested.

Fluting is also mentioned.
My home printer scanner fax failed about a week ago.
I took this from the reports with Mrs B's wee camera:


It isn't a problem with which I have first hand experience.
Maybe our good installation practises play a part.