VFD Surging

[jdsmith]

We retrofitted a furnace a few months ago by adding two 125 HP forced draft fans. One fan is driven by a 125# steam turbine and the other has a 1200 RPM IEEE 841 motor with an AB Powerflex 755 drive. They are redundant - only one fan is required to run at around 60% speed to maintain proper airflow into the furnace.

We had an AB field engineer out to start up the drive and check out the control I/O and parameter settings. All drive commands are either hardwired contacts or 4-20 mA - no communications protocols. A 4-20 mA input is set to control the speed with Volt/Hz control - no fancy stuff programmed in the drive.

When we were test running the fan at startup the drive current was surging from about 60 to 85 amps, swinging back and forth once per second or so. The speed command was constant and the RPM reading from the drive stayed fairly constant. The field engineer said the amps surging was due to actual mechanical load changes in the fan, furnace backpressure, etc.

Now that the FD fans have been running a few months, when the turbine driven fan is running the turbine speed is constant, the feedback signal to the turbine controller is constant, airflow is constant, etc. - just what you would expect. The VFD driven fan is still surging like it was during startup. Since the fans are identical within manufacturing variances, all vibration checked out OK, and they are blowing into the same furnace with the same firing conditions, we can no longer buy the story that the driven load is changing once per second causing the drive current to swing around substantially. I don't think the drive cares if the curring is surging, but surging current means surging motor torque which will tear up the motor and coupling prematurely.

Any ideas?

 

[dtonjes]

Many, many, many ideas. First thing I would do is stick a pot on the voltage input and see if the surging goes away. This will tell you whether it is the drive down or the 4-20 signal in. Best off the cuff guess is too slow ramp time, not mating the reaction time of whatever is driving the 4-20. Anyway as troublshooting goes, isolate the problem the squash it. (you definitly need a better AB guy. Do not listen to anyone even if they have the shirt. Verify for yourself. The equipment will never steer you wrong, just your inplant operators and every Tom, Dick, and Harry vendor that walks in your front door; but I digress) mechanical is way down the list if the fan goes around. You could bypass the VS drive and wire directly to the breaker for a short test run at 100%, but that is just one step above mechanical and climbing inside to see if the fan is OK.

 

[cadpoint]

- no communications protocols.

The Tech's make it work and they better be ready for the certification the AB ceritified it. Put them on the fire.

Seems it's the motor that is none responsive to the demand. Truely a layman statement. The motor in respects to a VFD should be able to controled for all aspects of a VFD, and no hick-ups...

Meggar the motor! And call them back!

PS: seems alot going on here...

 

[StephenSDH]

I just finished up a 50hp methane fan that was surging. The load and current are hand in hand. V/Hz drives just run at a fixed clock speed. Sounds like the drive is maintaining speed just fine, but the load varies.

 

[weressl]

We retrofitted a furnace a few months ago by adding two 125 HP forced draft fans. One fan is driven by a 125# steam turbine and the other has a 1200 RPM IEEE 841 motor with an AB Powerflex 755 drive. They are redundant - only one fan is required to run at around 60% speed to maintain proper airflow into the furnace.

We had an AB field engineer out to start up the drive and check out the control I/O and parameter settings. All drive commands are either hardwired contacts or 4-20 mA - no communications protocols. A 4-20 mA input is set to control the speed with Volt/Hz control - no fancy stuff programmed in the drive.

When we were test running the fan at startup the drive current was surging from about 60 to 85 amps, swinging back and forth once per second or so. The speed command was constant and the RPM reading from the drive stayed fairly constant. The field engineer said the amps surging was due to actual mechanical load changes in the fan, furnace backpressure, etc.

Now that the FD fans have been running a few months, when the turbine driven fan is running the turbine speed is constant, the feedback signal to the turbine controller is constant, airflow is constant, etc. - just what you would expect. The VFD driven fan is still surging like it was during startup. Since the fans are identical within manufacturing variances, all vibration checked out OK, and they are blowing into the same furnace with the same firing conditions, we can no longer buy the story that the driven load is changing once per second causing the drive current to swing around substantially. I don't think the drive cares if the curring is surging, but surging current means surging motor torque which will tear up the motor and coupling prematurely.

Any ideas?

It is a more and more common occurence. The problem is most likely mechanical and has to do with the different behavior of the compressable air column at different speeds. (This is why it does NOT show up with non-compressable liquids in pumps.) It is a complex model and has to do with the dimensions of the air-duct, varying turbulences, etc.

We had a similar problem on a much larger fan application and the solution was accomplished through parameter adjustments within the drive! (He actually had to access parameters that are not accessible to the ordinary user.) However the engineer servicing the drive was HIGLY skilled individual, not just a startup 'engineer' who is usually no more than a trained technician. So be persistent!

 

[TxEngr]

Good suggestions from all.

It well could be a mechanical problem. I would suggest you look to see if there is a fixed damper in the duct that may not be tightly fixed. This could cause the changes in load you are seeing as the 'fixed' damper oscilates.

It could also be in the tuning of the drive as weressel suggested. Make sure your AB guy has DriveTools software and you may need to tweak the tuning on the speed controller in the drive. It may be a little to hot for your application. If you look in the parameter book, you can find the info but it's not the easiest to understand so you probably need some additional help.

If you want to consider a different drives person (non-AB) then you can pm me and I'll make a recommendation of some folks I've used in the past that were helpful in this area. I have no vested interest in them other than they've done me a good job.

 

[rcwilson]

Fan surge is a problem when fans run at variable speed or the fan ductwork is changed. It is most likely a mechanical problem. Does it change or go away at different speeds?

Get the fan manufacturer involved.

 

[ptonsparky]

Belt driven fans? Are they slipping? We have had that problem before on a 250hp. Soft start.

 

[Besoeker]

Now that the FD fans have been running a few months, when the turbine driven fan is running the turbine speed is constant, the feedback signal to the turbine controller is constant, airflow is constant, etc. - just what you would expect. The VFD driven fan is still surging like it was during startup. Since the fans are identical within manufacturing variances, all vibration checked out OK, and they are blowing into the same furnace with the same firing conditions, we can no longer buy the story that the driven load is changing once per second causing the drive current to swing around substantially. I don't think the drive cares if the curring is surging, but surging current means surging motor torque which will tear up the motor and coupling prematurely.

Any ideas?

Is it a closed loop control system i.e. the drive is trying to maintain airflow at a set point with feedback from a transducer that measures airflow?
From your description, it sounds like it is.
If so, it might well be a closed loop stability problem.
There are a few things you could look at. The ramp rate might be too slow resulting in the drive response being too slow. Parameters 535 to 538.
If it has been set up as a PID controller, then maybe look at the PID parameters, particularly gain, integral, and derivative (P1086-P1089).
There is also an autotune facility.
The foregoing is from Chapter 3 of the manual - Programming and Parameters
If you do decide to change any parameter make sure you first note the original value(s).

I have not used AB drives but I have come across similar problems with others and tuning some of the parameters as above has fixed them. That said, I have had no joy with the autotune feature. It has always been manual setting of parameters.

Good luck.

 

[Besoeker]

User manual FYI.

http://literature.rockwellautomation.com/idc/groups/literature/documents/um/750-um001_-en-p.pdf

 

[jdsmith]

It is a more and more common occurence. The problem is most likely mechanical and has to do with the different behavior of the compressable air column at different speeds. (This is why it does NOT show up with non-compressable liquids in pumps.) It is a complex model and has to do with the dimensions of the air-duct, varying turbulences, etc.

We had a similar problem on a much larger fan application and the solution was accomplished through parameter adjustments within the drive! (He actually had to access parameters that are not accessible to the ordinary user.) However the engineer servicing the drive was HIGLY skilled individual, not just a startup 'engineer' who is usually no more than a trained technician. So be persistent!

Good suggestions from all.

It well could be a mechanical problem. I would suggest you look to see if there is a fixed damper in the duct that may not be tightly fixed. This could cause the changes in load you are seeing as the 'fixed' damper oscilates.

It could also be in the tuning of the drive as weressel suggested. Make sure your AB guy has DriveTools software and you may need to tweak the tuning on the speed controller in the drive. It may be a little to hot for your application. If you look in the parameter book, you can find the info but it's not the easiest to understand so you probably need some additional help...

Fan surge is a problem when fans run at variable speed or the fan ductwork is changed. It is most likely a mechanical problem. Does it change or go away at different speeds?

Get the fan manufacturer involved.

The normal operating point is around 65% speed. The oscillation occurs at speeds above 35-40%.

I talked with two AB engineers on Friday and I am going to take a data capture for them showing the speed command, motor RPM, DC bus voltage, and motor amps using DriveObserver software. They also requested that we use a hand tach on the motor shaft to see if the speed is changing slightly or staying constant. I'll look into the possibility of any dampers that may have some slight movement in them.

Once I get the data for the AB folks and get some feedback from them I'll share that with the fan manufacturer and the ductwork designers.

Belt driven fans? Are they slipping? We have had that problem before on a 250hp. Soft start.

It is direct coupled. Actually in the ~1200 motors we have in the plant, I'm not sure we have any belted equipment. Most are direct coupled, some are run through speed reducing or increasing gearboxes, and cooling fans are run through 90 degree gearboxes.

Is it a closed loop control system i.e. the drive is trying to maintain airflow at a set point with feedback from a transducer that measures airflow?
From your description, it sounds like it is.
If so, it might well be a closed loop stability problem.
There are a few things you could look at. The ramp rate might be too slow resulting in the drive response being too slow. Parameters 535 to 538.
If it has been set up as a PID controller, then maybe look at the PID parameters, particularly gain, integral, and derivative (P1086-P1089).
There is also an autotune facility.
The foregoing is from Chapter 3 of the manual - Programming and Parameters
If you do decide to change any parameter make sure you first note the original value(s).

I have not used AB drives but I have come across similar problems with others and tuning some of the parameters as above has fixed them. That said, I have had no joy with the autotune feature. It has always been manual setting of parameters.

Good luck.

There is closed loop control but it is in the Honeywell DCS, not the drive. There is airflow instrumentation going into the DCS and the DCS is sending the drive a 4-20 mA speed command. I think we can rule out loop tuning issues in this case because if I put the drive in local mode and set a fixed speed from the front panel of the drive the behavior stays the same.

Thanks for all of the input, I'll update when I know more.

 

[Besoeker]

There is closed loop control but it is in the Honeywell DCS, not the drive. There is airflow instrumentation going into the DCS and the DCS is sending the drive a 4-20 mA speed command. I think we can rule out loop tuning issues in this case because if I put the drive in local mode and set a fixed speed from the front panel of the drive the behavior stays the same.

Thank you for reading what I posted.
You clearly have a very good understanding of the system.

I had one drive that was a bit unstable even on open loop speed control. It was on a pump rather than a fan. It had a setting for better efficiency on a centrifugal fan or pump (cube law). This reduces the V/f ratio at lower speeds. Given that the VSD was there to improve efficiency, enabling this mode of operation seemed appropriate. I did and the drive went unstable.
I eventually sorted it out with some parameter changes. But it did surprise me that it could go unstable in open loop control. I still don't understand how.
Stuff happens.

 

[jdsmith]

I had one drive that was a bit unstable even on open loop speed control. It was on a pump rather than a fan. It had a setting for better efficiency on a centrifugal fan or pump (cube law). This reduces the V/f ratio at lower speeds. Given that the VSD was there to improve efficiency, enabling this mode of operation seemed appropriate. I did and the drive went unstable.
I eventually sorted it out with some parameter changes. But it did surprise me that it could go unstable in open loop control. I still don't understand how.
Stuff happens.

The more knowledgeable of the two AB field engineers I had in repeated the same story. The Powerflex drives have a similar setting - in V/Hz mode you can select "standard" or "pump/fan" curves. I don't remember if he had problems running the pump/fan curves on pumps or fans, but he had enough bad experiences that he won't use that setting any more.

 

[winnie]

On an experimental drive, consisting of a fixed frequency PWM bridge with capacitor bank, DC supply via a battery bank, we saw 'surging' at no load if the excitation was high enough. Looking at the shaft with a strobe, we could see the shaft accelerating and decelerating, at a frequency well below the drive frequency. I could only assume that we were seeing an oscillation between energy stored in the magnetic field and energy stored in the capacitor bank. We never had a chance to follow that up completely, although now that you'all have reminded me of it, I might take a look again since we now have better instrumentation.

-Jon

 

[Besoeker]

The more knowledgeable of the two AB field engineers I had in repeated the same story. The Powerflex drives have a similar setting - in V/Hz mode you can select "standard" or "pump/fan" curves. I don't remember if he had problems running the pump/fan curves on pumps or fans, but he had enough bad experiences that he won't use that setting any more.

Thank you kindly for your reply.
You clearly have a very good grasp of the situation.
And very articulate.

 

[weressl]

The normal operating point is around 65% speed. The oscillation occurs at speeds above 35-40%.

I talked with two AB engineers on Friday and I am going to take a data capture for them showing the speed command, motor RPM, DC bus voltage, and motor amps using DriveObserver software. They also requested that we use a hand tach on the motor shaft to see if the speed is changing slightly or staying constant. I'll look into the possibility of any dampers that may have some slight movement in them.

Once I get the data for the AB folks and get some feedback from them I'll share that with the fan manufacturer and the ductwork designers.



It is direct coupled. Actually in the ~1200 motors we have in the plant, I'm not sure we have any belted equipment. Most are direct coupled, some are run through speed reducing or increasing gearboxes, and cooling fans are run through 90 degree gearboxes.



There is closed loop control but it is in the Honeywell DCS, not the drive. There is airflow instrumentation going into the DCS and the DCS is sending the drive a 4-20 mA speed command. I think we can rule out loop tuning issues in this case because if I put the drive in local mode and set a fixed speed from the front panel of the drive the behavior stays the same.

Thanks for all of the input, I'll update when I know more.

The additional data supports the problem we experienced and I described. It is the load end behavior that forces the drive to behave the way it does. It is akin to (natural) resonance and as I said it has to do with the sum of ariflow, turbulence and ductwork flexing. Carrier frequency adjustment, V/f ratio control tailored to the load profile, current/flux control were the areas where the engineers worked. If you have ANY dampers for 'backup' service or leftover form the previous control mode, try to remove it. The simpler the flow profile is the elast problem will you experience.

 

[jdsmith]

...I think we can rule out loop tuning issues in this case because if I put the drive in local mode and set a fixed speed from the front panel of the drive the behavior stays the same.

We fixed the surging problem - I was partially correct and partially wrong. In the post above I stated that I could rule out loop tuning problems since running the drive at a fixed speed from the front panel still caused surging. I was assuming that the drive was operating in open loop V/Hz mode. We found out that the "Slip RPM at FLA" (parameter #621) was populated with a value, which actually turned on a speed control loop when I thought the drive was in open loop mode. I set the slip RPM to zero, effectively disabling the slip compensation loop, and the surging disappeared. Detailed block diagrams of the drive's control system are available through the DriveTools help menus and these block diagrams are not in the manual. We went through the block diagrams to verify that no other loops were controlling speed directly.

There were a few thoughts on why this happened. I'm guessing that AB has data on slip vs. speed for maybe 2 and 4 pole machines. This motor is a 6 pole and maybe the slip vs. speed relationship is different for a 6 pole motor. The V/Hz block diagram doesn't show which parameters are used internal to the slip comp block to calculate the gain. This could indicate that they are only using % of full speed in the slip comp loop, or they just don't want to share that info in the manual.