40Hp VFD Pump application

[philly]

I am in the process of coming up with a plan for installing a VFD on a 480V 40hp pump which pumps water from a tank. I was planning on using the Allen Bradley Powerflex family and had a couple of questions that came up in my design process.

For this water pump application is there a need to get a 40hp Heavy Duty drive, or will a Normal Duty drive suffice?

Is there a need to order a braking resistor for this type of application? Does the pump cause regen for certain conditions?

What type of drive control mode works best with pumps. Is it better to use sensorless vector or FOC for a pump application?

The speed of the VFD and pump will be controlled by a level in the tank from a transmitter. However is there a minimum speed which you would want to run the VFD/Pump in order to still be able to move water? In other words is there a certain minimum in such an application for a low speed range?

The drive quotes I have show a Filter w/ CM Core. Does anyone know what this is?

I am feeding this drive with a breaker from an MCC. Is there a need to have another breaker or fused disconnect at the drive panel? I will probably put disconnect but wouln't a CB or fuse just be redundant?

 

[Jraef]

I am in the process of coming up with a plan for installing a VFD on a 480V 40hp pump which pumps water from a tank. I was planning on using the Allen Bradley Powerflex family and had a couple of questions that came up in my design process.

A pump is not a pump is not a pump. I'll assume, for expedience, it is a centrifugal pump.

For this water pump application is there a need to get a 40hp Heavy Duty drive, or will a Normal Duty drive suffice?

Normal duty would be OK.

Is there a need to order a braking resistor for this type of application? Does the pump cause regen for certain conditions?

No, you will never need braking / regen on a pump. In fact, it would be a bad thing because you don't want to create 'water hammer" from stopping it fast.

What type of drive control mode works best with pumps. Is it better to use sensorless vector or FOC for a pump application?

Neither. All you want is simple V/Hz control, anything else on a centrifugal pump will just make it complicated and possibly give you trouble.

The speed of the VFD and pump will be controlled by a level in the tank from a transmitter. However is there a minimum speed which you would want to run the VFD/Pump in order to still be able to move water? In other words is there a certain minimum in such an application for a low speed range?

Yes, there is a minimum speed below which the pump will stop working, usually somewhere around 40%. But to be sure, ask the pump guy. They have pump curves that will tell you (if you know how to read them).

FYI, you cannot just feed the raw level signal into the drive, you have to have a PID Loop Controller in there to maintain a level, otherwise it will hunt up and down constantly. The PF4 should have PID built-in, but IIRC, it's a bitch to set up and use, especially if you have never done one before. Have you? If not, I happen to like the ABB ACS 350 or 550 drives as the easiest to set up PID loops in the keypad. They walk you through all the steps right there while you're commissioning the drive using clear layman's terms.

The drive quotes I have show a Filter w/ CM Core. Does anyone know what this is?

CM Core? Maybe "CM (Common Mode) Choke" perhaps? Never a bad idea, they protect the VFD from line problems and the line for VFD problems. Cheap insurance.

I am feeding this drive with a breaker from an MCC. Is there a need to have another breaker or fused disconnect at the drive panel? I will probably put disconnect but wouln't a CB or fuse just be redundant?

Could be, but there are other issues, all code related. I'm not there to see the other details of the installation so it's hard to say.

 

[StephenSDH]

I agree with Jraef. If you call a distributor they will probably recommend PF 40P. The P is actually for pump. It is a simple V/Hz drive.

 

[TxEngr]

jraef gave you good advice. I would like to add a couple of comments:

When buying a drive, you need to specify constant torque or variable torque to the vendor. Your pump (assuming its a centrifigal pump) is a variable torque application as are fans, etc. A constant torque application would be a conveyor. The variable torque correlates to the 'standard' drive and the constant torque to the 'heavy duty' drive.

Depending on the distance of the drive from the motor, you should consider using VFD rated cable. There's another thread on that here now. I always use VFD rated cable.

Make sure your motor is rated for inverter duty. Any IEEE841 spec motor will be rated for this.

Feeding the drive from a breaker is fine. You don't need another disconnect at the drive unless you just want one. Note that the disconnect has to be rated at 115% of the drive input current rating. Also note that some drive manufacturers (ABB) don't recommend the use of a breaker to feed the drive. Read the manual on the A-B Powerflex to see what they recommend.

 

[Besoeker]

I am feeding this drive with a breaker from an MCC. Is there a need to have another breaker or fused disconnect at the drive panel? I will probably put disconnect but wouln't a CB or fuse just be redundant?

As standard, we fit a door interlocked fuse switch padlockable in the off position at the drive panel to prevent access when the drive is live. We also fit, as standard, an output line contactor for the emergency stop function. It is usual to have an emergency stop PB in the vicinity of the pump and motor.
And we always use steel wire armoured cable.

I know practices are different here in UK so this may not me entirely relevant to your application. I just thought the different approach might be interesting.

 

[Jraef]

As standard, we fit a door interlocked fuse switch padlockable in the off position at the drive panel to prevent access when the drive is live. We also fit, as standard, an output line contactor for the emergency stop function. It is usual to have an emergency stop PB in the vicinity of the pump and motor.
And we always use steel wire armoured cable.

I know practices are different here in UK so this may not me entirely relevant to your application. I just thought the different approach might be interesting.

Redundancy is never a problem for anyone except bean counters and grammar teachers.

 

[StephenSDH]

We also fit, as standard, an output line contactor for the emergency stop function.

Allen Bradley 40P, 70, 700 series drives have a safe-off option board which will redundantly disable the output of the drive in hardwire. It is rated for cat 3 safety circuit. Greatly reduces cost and panel space. For most safety applications you would need 2 output contactors and a safety relay to monitor NC contacts and prevent a E-Stop reset when the contactors weld.

http://www.ab.com/drives/powerflex/safe-off.html

 

[philly]

Thanks for the great information guys.

Interesting point on only needing to use V/Hz for control of the pump as opposed to sensorless vector. I think I have an idea why due to the fact that sensorless vector is used to contorol torque during high torque situations or overload torque situations, however with a centrifugal pump the torque is strictly a function of the pump speed so there will never be a high torque condition. Am I right, or can someone explain to me why only a simple V/Hz control is needed for a pump?

With the drive only operating in V/Hz mode, is there even a need then to perform an autotune on the motor?

Is the PF40P a cheaper drive compared to the PF70? I guess I need to talk to my rep and explain I only need a drive with V/Hz capability.

No, you will never need braking / regen on a pump. In fact, it would be a bad thing because you don't want to create 'water hammer" from stopping it fast.
Neither. All you want is simple V/Hz control, anything else on a centrifugal pump will just make it complicated and possibly give you trouble.

What is for some reason you did have a fast ramp down time on this application like maybe a couple of seconds. Wouldn't this quick ramp down cause regeneration, or does the pump not have enough inertia to cause an issue?

What kind of ramp up/down times are typically used in a pump application?

FYI, you cannot just feed the raw level signal into the drive, you have to have a PID Loop Controller in there to maintain a level, otherwise it will hunt up and down constantly. The PF4 should have PID built-in, but IIRC, it's a bitch to set up and use, especially if you have never done one before. Have you? If not, I happen to like the ABB ACS 350 or 550 drives as the easiest to set up PID loops in the keypad. They walk you through all the steps right there while you're commissioning the drive using clear layman's terms.

Both the level transmitter and the drive will be connected to our DCS network, and therefore the PID loop will reside on the network programming. The output of this PID will provide a speed output to the drive.

 

[StephenSDH]

I think I have an idea why due to the fact that sensorless vector is used to contorol torque during high torque situations or overload torque situations, however with a centrifugal pump the torque is strictly a function of the pump speed so there will never be a high torque condition. Am I right, or can someone explain to me why only a simple V/Hz control is needed for a pump?

Both will apply full torque. With sensorless vector the drive will have tighter speed control then v/hz.

V/Hz = Drive simply runs at a fixed clock speed. The load will rotate at a slightly slower speed depending on how loaded it is.

Sensorless Vector = The drive will compensate for motor slip and has tighter speed control regardless of load. Sensorless vector does not have good speed/torque control at low speeds. (Each drive is different)

With the drive only operating in V/Hz mode, is there even a need then to perform an autotune on the motor?

No, and it don't believe it will give you the option.

Is the PF40P a cheaper drive compared to the PF70? I guess I need to talk to my rep and explain I only need a drive with V/Hz capability.

Yes. The PF70 is a full feature drive.

What is for some reason you did have a fast ramp down time on this application like maybe a couple of seconds. Wouldn't this quick ramp down cause regeneration, or does the pump not have enough inertia to cause an issue?

In pump applications it is generally the water that has the inertia. As Jaref said you don't want to stop it quickly or you will get water hammer. You can set the drive to just coast stop, or a long decel rate 10-20 seconds.

What kind of ramp up/down times are typically used in a pump application?

You are not trying to win a race. Set it to a smooth rate say 10-20 seconds ramp up, and then a coast stop, or 10-20 second decel.

 

[philly]

Both will apply full torque. With sensorless vector the drive will have tighter speed control then v/hz.

V/Hz = Drive simply runs at a fixed clock speed. The load will rotate at a slightly slower speed depending on how loaded it is.

Sensorless Vector = The drive will compensate for motor slip and has tighter speed control regardless of load. Sensorless vector does not have good speed/torque control at low speeds. (Each drive is different)

Why is this tighter speed control with sensorless vector not required for a pump application?

 

[Jraef]

Why is this tighter speed control with sensorless vector not required for a pump application?

It provides no added benefit. But a benefit you CAN get from a VFD in V/Hz mode is what is called the "Energy Saver" control feature, which backs off on the voltage at lower speeds to reduce magnetic losses in the motor. In a centrifugal pump, since the load torque requirement varies with the motor speed, you don't need as much torque when reducing the speed, so you can alter the V/Hz pattern to economize the power use. If you are in Vector mode, you are doing exactly the opposite. You are MAXIMIZING torque production at any speed, so you cannot do the energy saver trick.

So bottom line, vector control on a centrifugal pump doesn't hurt, but it disallows the best use of the VFD technology.

 

[QES]

I would go with the AB power 775 40 HP unit, get the optional HIM, and motor Terminator. This package will cost around 7K.

 

[jdsmith]

Why is this tighter speed control with sensorless vector not required for a pump application?

In a pump application you don't need to care about motor or pump speed, you care about flow and pressure. For a pump you could control motor speed very accurately, use an encoder on the shaft for very precise speed feedback, and program the pump curves into your software to translate the speed into flow. While this works fine on paper, there are a lot of small errors that pile up in the process: pump curves are approximations even on a custom unit, pump performance will drift as parts wear, fluid viscosity and specific gravity will change flow, head pressures change. All of these will lead to inaccuracy if your end goal is to control flow but you are measuring speed instead.

It is best to monitor the parameter you care about as directly as possible.
If you want to control flow, then monitor flow and include the instrumentation and mechanical errors inside the control loop where they don't affect the performance of the control system. It is best to have just one control loop - sensorless vector gives you numerous loops inside of other loops. Simpler is better and you can get the same performance with just one loop and not have all of those others to tune.

 

[philly]

It provides no added benefit. But a benefit you CAN get from a VFD in V/Hz mode is what is called the "Energy Saver" control feature, which backs off on the voltage at lower speeds to reduce magnetic losses in the motor. In a centrifugal pump, since the load torque requirement varies with the motor speed, you don't need as much torque when reducing the speed, so you can alter the V/Hz pattern to economize the power use. If you are in Vector mode, you are doing exactly the opposite. You are MAXIMIZING torque production at any speed, so you cannot do the energy saver trick.

So bottom line, vector control on a centrifugal pump doesn't hurt, but it disallows the best use of the VFD technology.

Let me try to understand what you are explaining.

For any given motor there is a rated torque that the motor can produce, and with a vfd, the drive can alter itself to produce this rated torque at any speed throughout the range of the drive. For some applications at low speeds you may need this full rated torque, and sensorless vector control is one way of adjusting the V/Hz ratio at low speeds to ensure that max torque can be produced at low speeds.

However as you said because with a pump the torque required is a function of speed, there is a lower torque requirement at low speeds. Because we do not to produce full torque at low speeds for this application we do not need to adjust the V/Hz ratio to allow max torque at low speeds. Therefore in an energy saver mode the drive can reduce the voltage thus reducing the V/Hz and therefore reducing the avaliable torque from the motor. Because the torque is not required the reduced torque from the motor has no effect however, becuase the voltage to the motor is less, the magnetic losses to the motor are less, and therefore energy will be saved.

But in sensorless vector mode wont the drive be able to recognize that at a low speed the rated torque isn't required, and therefore adjust itself? Or does the drive always try to maintain rated V/Hz to always be capable of producing max torque.

Also even at low speed, for a given required load torque the torque will be a function of both V/Hz and current. So in the energy saver mode even though the toruqe requirement is low wont lowering the V/Hz still increase the current to still maintain this low amount of required torque?

 

[Jraef]

Let me try to understand what you are explaining.

For any given motor there is a rated torque that the motor can produce, and with a vfd, the drive can alter itself to produce this rated torque at any speed throughout the range of the drive. For some applications at low speeds you may need this full rated torque, and sensorless vector control is one way of adjusting the V/Hz ratio at low speeds to ensure that max torque can be produced at low speeds.

However as you said because with a pump the torque required is a function of speed, there is a lower torque requirement at low speeds. Because we do not to produce full torque at low speeds for this application we do not need to adjust the V/Hz ratio to allow max torque at low speeds. Therefore in an energy saver mode the drive can reduce the voltage thus reducing the V/Hz and therefore reducing the avaliable torque from the motor. Because the torque is not required the reduced torque from the motor has no effect however, becuase the voltage to the motor is less, the magnetic losses to the motor are less, and therefore energy will be saved.

But in sensorless vector mode wont the drive be able to recognize that at a low speed the rated torque isn't required, and therefore adjust itself? Or does the drive always try to maintain rated V/Hz to always be capable of producing max torque.

Also even at low speed, for a given required load torque the torque will be a function of both V/Hz and current. So in the energy saver mode even though the toruqe requirement is low wont lowering the V/Hz still increase the current to still maintain this low amount of required torque?

Whoa... my head is spinning...

The PURPOSE of vector control is to maintain tight control of speed and/or torque at any speed. In a centrifugal pump system, that is basically unnecessary. You CAN do it, but the extra capability is lost on the application. Precision provides no significant improvement in final performance, performance being measured by flow and/or pressure regulation.

Vector Control precludes Energy Saver and vice versa; the performance philosophies behind each feature are at odds with each other. When you are in Vector control mode, the priority of the control algorithm is to maintain torque or speed precision. But to get that you give up the ability to do other functions that, in a constant torque application, provide no benefit there, i.e. the energy saver function. When you want the Energy Saver, it would only work in an application where torque and speed precision is unnecessary, i.e. where Vector control would provide no benefit.

So can the drive recognize that torque is no longer necessary at low speed in a variable torque application? No. The drive is not THAT smart. It just tries to do what you tell it to do. If you put it in vector mode, the ASSUMPTION is that you need the benefits it provides, so the energy saver function is also ASSUMED to be irrelevant. If you put the VFD in V/Hz mode, it is ASSUMED you no longer need the benefits that Vector mode can supply, allowing the use of energy saver.

By the way, some VFD manufacturers have a macro installed they call "Pump Mode" which basically does this all for you in the background, i.e. if you select Pump Mode, you are indirectly selecting both V/Hz Mode an Energy Saver. The reason for that is that Energy Saver is usually independent of V/Hz mode and must be specifically enabled (for other reasons) and if you have a centrifugal pump, you should use it.

 

[weressl]

Thanks for the great information guys.

Interesting point on only needing to use V/Hz for control of the pump as opposed to sensorless vector. I think I have an idea why due to the fact that sensorless vector is used to contorol torque during high torque situations or overload torque situations, however with a centrifugal pump the torque is strictly a function of the pump speed so there will never be a high torque condition. Am I right, or can someone explain to me why only a simple V/Hz control is needed for a pump?

With the drive only operating in V/Hz mode, is there even a need then to perform an autotune on the motor?

Is the PF40P a cheaper drive compared to the PF70? I guess I need to talk to my rep and explain I only need a drive with V/Hz capability.



What is for some reason you did have a fast ramp down time on this application like maybe a couple of seconds. Wouldn't this quick ramp down cause regeneration, or does the pump not have enough inertia to cause an issue?

What kind of ramp up/down times are typically used in a pump application?



Both the level transmitter and the drive will be connected to our DCS network, and therefore the PID loop will reside on the network programming. The output of this PID will provide a speed output to the drive.

Make sure the motor is suitable for inverter duty.

I am also still an ABB fan.

Putting the ASD control in the DCS is a waste of good I/O. The drives' internal loop control comes free and more direct control, easier to troubleshoot.

 

[philly]

Whoa... my head is spinning...

The PURPOSE of vector control is to maintain tight control of speed and/or torque at any speed. In a centrifugal pump system, that is basically unnecessary. You CAN do it, but the extra capability is lost on the application. Precision provides no significant improvement in final performance, performance being measured by flow and/or pressure regulation.

Vector Control precludes Energy Saver and vice versa; the performance philosophies behind each feature are at odds with each other. When you are in Vector control mode, the priority of the control algorithm is to maintain torque or speed precision. But to get that you give up the ability to do other functions that, in a constant torque application, provide no benefit there, i.e. the energy saver function. When you want the Energy Saver, it would only work in an application where torque and speed precision is unnecessary, i.e. where Vector control would provide no benefit.

So can the drive recognize that torque is no longer necessary at low speed in a variable torque application? No. The drive is not THAT smart. It just tries to do what you tell it to do. If you put it in vector mode, the ASSUMPTION is that you need the benefits it provides, so the energy saver function is also ASSUMED to be irrelevant. If you put the VFD in V/Hz mode, it is ASSUMED you no longer need the benefits that Vector mode can supply, allowing the use of energy saver.

By the way, some VFD manufacturers have a macro installed they call "Pump Mode" which basically does this all for you in the background, i.e. if you select Pump Mode, you are indirectly selecting both V/Hz Mode an Energy Saver. The reason for that is that Energy Saver is usually independent of V/Hz mode and must be specifically enabled (for other reasons) and if you have a centrifugal pump, you should use it.

Thanks for the explanation. Looking at it another way, can it be said that for a constant torque load, the max torque may be needed across the whole speed range, thus sensorless vector mode will ensure that this constant torque is produced across whole range including low speeds. For a variable torque load since the torqe varies with speed, there wont be as much torque required at low speeds and thus the drive wont need to provide full torque. This is why V/Hz mode and energy saver mode can be used with this variable torque application.

In regards to OCPD, if my breaker in my MCC is sized to protect the feeder cables to the drive, and is rated at 125% of drive input current as required by code then is there a need to put breaker/fuses at the drive? (Besides preference) Does the drive protect itself from short circuits or is this the job of primary OCPD? Would it then make a difference where the OCPD was located? I would think having breaker at MCC and then drive would be redundant?

 

[weressl]

I would go with the AB power 775 40 HP unit, get the optional HIM, and motor Terminator. This package will cost around 7K.

Why? I can buy an ABB ACS550 for $3K.

 

[Jraef]

Why? I can buy an ABB ACS550 for $3K.

But is has one too many "B"s in it...

 

[TxEngr]

...but the extra 'B' means it's the enhanced model.