Starting large motors/loads with capacitor assist

[raisinbran]

In the mining industry, motors that can satisfactorily move the load have difficulty starting the load because of voltage drops in the system. These particular motors are mechanically connected to the load using fluid couplers to take advantage of the motor breakdown torque. The motors are started across the line. The reduced voltage (80%) lowers the torque enough that sometimes the load (chain conveyor) will not start and must be partially emptied by hand.
I was thinking about capacitor assisted starting to boost the motor voltage until the load is accelerated. The capacitors would be switched on with the motor and switched off sometime after the load is up to speed. With a fluid coupling, the motors come up to spedd very quickly, and the load is accelerated over a 10 second (approx.) time frame. I was not planning on doing anything too sophisticated, possibly just using a timer to turn the capacitors off. I also would assure that the capacitors were off before the motor was switched off.
From what I read in previous threads, there are concerns about over-voltages and over-excitation. Is this just a concern when the motor is switched off, or could I get into problems when the motor is running?
Thanks in advance,
Raisinbran

 

[Barry1961]

That is a lot of voltage drop. I guess increasing the wire size or having the power company increase the voltage is not an option.

Is the motor never getting up to speed and tripping out on overload? I would have thought the fluid coupling would let the motor get up enough speed to reduce the slip/current so the voltage drop would not be that bad.

I have run into chain conveyors drives that were designed to the dynamic/sliding co-efficient of friction which is right at half of the static. Steel on steel dry goes from .8 to .4 and lubricated goes from .16 to .03 after they start moving. Of course the actual values will change with material condition ect.....but I am confident the conveyors are well maintained.

Barry1961

 

[raisinbran]

Barry,
Typically, there are three motors driving the conveyor - each 800 to 1200 HP. The motors are stagger started to minimize the voltage drop due to motor inrush. The motors get up to speed and then the fluid couplers try to engage the load at or near motor breakdown torque values. The motors are rated at 250 - 300% breakdown torque, so the current is 3 - 4 times full load. This situation lasts for 10 seconds (approx.) until the load is accelerated to full speed. During this period, that is when the voltage is in the 80-85 percent range, and unfotunately, that is when I want the voltage to be as high as possible to allow the motors to deliver as close to rated breakdown torque as possible. At 80% voltage, I only get 64% of what I bought.
Summary of "typical" distribution system.
Utility delivery to surface - 100 to 150 MVAsc
Surface transformer 10 MVA
Typical distribution voltage - 12.47 or 13.8 kV
4/0 Distribution cable 20,000 feet long - length varies as mining progresses.
Utilization transformer - 5 MVA; 4.16 kV secondary
Motor cables - #1 AWG, 1500 feet long.
The system is adequate to "run" the load, but sometimes it is difficult to "start" the load. That is when I want to use the capacitor assist.
I hope that this better explains my problem.
Raisinbran

 

[dpc]

This can be done, but if you run out the numbers, you will find that you need a lot of capacitance to make much difference. The power factor of a motor at locked rotor is probably less than 0.2.

You do have to be a bit concerned about the impact of suddenly switching this amount of caps in and out of your system. When the motor is running and the caps are still energized, the voltage may go high - this is a steady-state effect that can be calculated pretty easily.

But you may see some switching transients such as voltage oscillations due to high frequency ringing.

And whatever you're using to switch the caps has to be rated for this nasty capacitive switching service.

 

[jraef]

Your application is a little different than cap assisted starting from a dead stop. You want to add-in the caps when the fluid couplings engage the load, which is after the motor is already at full speed, not at locked rotor. That's more akin to how a power factor correction cap would work, and it may actually provide some benefit in your situation. It wouldn't really boost the motor torque, but the caps would supply additional VARs to your motor circuit, lowering the burden on your power distribution system. However, all of the power still needs to come from that system, so I'm not sure how much of a benefit you will realize. On top of that, the timing would need to be nearly perfect in order to have the desired effect. You would likely need to somehow monitor slip and switch on the caps when the slip is at maximum but before you fall back behind the breakdown torque point and go into stall. Once it does that, the caps won't help. Add to that the likely transient problems that dpc mentioned and you may be wiser to investigate beefing up your distribution system if possible.

"Venditori de oleum-vipera non vigere excordis populi"

 

[Barry1961]

You have probably done all of this already, but just to be sure.

Adjusted the fluid level in the couplings to evenly distribute the load between motors.

Adjust the flow from the fluid delay fill chamber to compensate for the stagger start.

Calculated if the amount of voltage drop matches the expected with measured current draw. Do you have any extra impeadance somewhere?

The voltage is balanced.

Tested to see if conveyor is overloaded, drive under sized or something is binding.


I can understand them not wanting to run bigger cable that far but it may be the cheapest answer.

Barry1961

 

[raisinbran]

jraef,
I was planning to energize the capacitors when the second motor started, and leave them energized until?? - maybe the motor current came down to below full load. I was thinking about 2000 kVAR total (not per motor).
I agree that "beefing up" the system is probably a better idea, but, as always, it is easier said than done. The "system" changes almost daily as loads are added or moved. The 20,000 foot long cable that I mentioned above will start out at 20,000 and be reduced to 5,000 feet in six months, and then will go to 20,000 feet in three more months.
It is literally like trying to hit a moving target.
Raisinbran

 

[raisinbran]

Barry1961
The fluid levels are adjusted to get breakdown torque current out of the motors. The load is locked and the current is checked to see if it at or near breakdown values. Load sharing is not too critical.

Stagger start is 1-1/2 to 2 seconds, but it is worth re-checking the fluid flow rates to the working chambers.
Voltages are balanced, and no extra cables are stored anywhere. As the system is moved, excess cable is removed from the distribution system.
The drives/motors are about 10 percent oversized for the job. There are no overloads while running. This is why load sharing was not that critical. Loading is also cyclic as the material is deposited at various locations along the chain conveyor.
Larger cables will help, but not too much. As these cables are handled, there is an incentive to make them as small as possible.

Raisinbran

 

[rbulsara]

I am with jraef..as far as effect of caps go.

It is not clear, what happens when you say, load do not 'start', but motor is at full speed before the fluid coupling is engaged. So what happens, in those non-start cases? What trips? Motor OL relay or the branch circuit breaker? This may give more clues.

80% voltage is 20% drop, which is very much expected on a DOL start or even more. It may be that the clutch is trying to engage the load too fast. I am not sure your method of starting is superior to reduced voltage start with all load engaged. Trying to engage full load (conveyor) quickly at rated speed may be worse than trying start the load from zero and slowly buildig up..At lower speed you need less power too for a give torque. I may think little more about this.

Also is this new problem on old system or this is a new system?
It is entireyly possible that the motors are not properly matched..may be this needed wound rotor motors...

 

[Barry1961]

Can you stagger the starts out some more? An extra second or two may help. The fluid coupling fill time for that size should be adjustable to compensate for 4 sec stagger starts.

If both of the motors are up at full speed and engage at the same time the inertia of the combined rotors might be enough to break the conveyor chain loose out of static friction. The fluid couplings of course are going to cushion this effect a lot.

If you can get the conveyor chain into sliding instead of static friction the torque required to pull the chain due to friction is cut in half. But if you produce a little more surge in power to get things going it will also increase your acceleration which will increase the required torque.

Sounds like you are being painted into a corner. Maybe if you tweak everything just right it will work most of the time.

Barry1961

 

[dpc]

jraef is right - I missed the part about fluid couplings and starting unloaded. I don't see the caps doing much to help you - certainly not as much as they would when starting motor from a standstill.

 

[PowerfulStuff]

Hi,
I'm not sure that the caps would have any benefit for the starting. Even with a very large bank the amount of energy stored would be tiny compared to the total energy required.

Is it possible to add a flywheel to the motor so that some extra energy can be stored mechanically?

Also a fair amount of energy during the load connection would go into accelerating the conveyor- if you know the mass of the conveyor and final speed it should be possible to determine how much improvement you might get by reducing the acceleration required, by maybe using a pony motor to start or install a (suitably oversized) vector VSD. I've seen an apron feeder started when overfilled using a VSD size to 160% of motor size. With the short term overload capacity of the VSD it was able to provide locked rotor current for sufficient time to get things going.

 

[jraef]

Aside from the fact that a pony motor would be totally useless in this application (the main motor cannot re-accelerate the load, how would a smaller one do it?), PowerfullStuff does bring up a good point. Your only solution may be to put VFDs on those motors. Expensive, but it is likely that it would solve the problems, especially considering that your power delivery system specs are variable. VFDs would be capable of solving several important aspects of this problem.

1) They can deliver additional acceleration torque in the motor without the corresponding current surge seen with running Across-the-Line, because a properly set up Vector drive can allow the motor to temporarily produce breakdown torque while compensating to slip changes, thereby reducing the shock on the electrical system.

2) With the right VFDs, you can properly load share the conveyor like you do now with the fluid couplings, thereby eliminating the need for the fluid couplings all together. There is a trend in overland conveyor systems to go this way. There are even companies specializing in the complex software and PLC systems to monitor and control them so as to eliminate the "rubber band effect" when part of the load is going down hill while other parts go up.

3) As your power system changes when the equipment moves, adjustments to the VFD operating parameters can be relatively easy, so no redesign would be necessary.

The main hurdle will be capital cost. 4160V VFDs are not cheap, but dont forget, you can also eliminate the need for the starters that you have now. Worht a second look IMHO.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[PowerfulStuff]

jraef,
You may have misunderstood my intention. I was looking at the pony motor to be low speed, high torque to get everything moving and the main motor to then bring it up to speed. It all depends on how much of the starting 'hit' is due to acceleration and how much due to friction.

 

[jraef]

No offense, but read his original posts more carefully. The motor is already at full speed when he engages the fluid coupling. It is at that point that the motor pulls so much power that it drops tha line voltage 20%. A pony motor would only provide benefit when starting from a dead stop to avoid inrush current.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[mc5w]

Using a power factor correction capacitor is not a bad idea, but should be broken up into individual units for each motor. However, switching in more than what it takes to correct full load power factor and then switching some out is a bad idea. There is a tech brief over Powell Electric about flashovers in a reduced voltage motor starter caused by switching in and out excess capacitance during the starting process.

Your best be for starting the load correctly is to install solenoid valves that prevent the fluid couplings from filling until all of the motors are started and up to full speed. The fluid couplings would then simultaneously fill with fluid at the same rate which will help prevent some of the motors from drawing too much current. This would require some control wires connecting all 3 of the motors. Telemecanique make an inductive proximity switch that had a built in motion ( speed ) detector that responds to a toothed wheel or just a bolt in place of a setscrew. When the number of passes per second reaches a threshold the output contacts change state. Once all 3 motion detectors close contacts you would have say a 2 or 3 second time delay before energizing the solenoid valves.

I know that somebody is going to yell at me about a bolt instead of a setscrew causing a rotating imbalance, but for a slow speed shaft that is not critical and the bolt can be held in place with Loctite(R).

You do not want to start the fluid filling process while the motors are accelerating. What you want is a gradual takeup of torque so that any substation voltage regulators have a chance to adjust the voltage. This is a slower starting sequence than what you have now but will cetainly be less brutal - a 5 MVA transformer has too much reative voltage drop to brute start the conveyor.

You also may have to change operating procedures so that the conveyor is allowed to empty out before stopping. This could be implemented with yellow buttons for a time delayed normal stop and red buttons for emergency stop. You would also need red and green lights at each loading point to tell people when they can dump stuff on the conveyor.

 

[jraef]

Aside from mc5w's tangents (no doubt well intentioned), he makes a good point about separately controlling your fluid injection. Is it possible to do that on your couplings?

Right now it appears as though you are "timing", by fluid fill levels, the full injection to coincide with the peak current draw on the motor with the understanding that this represents the breakdown torque point. While technically correct, any very slight mistiming could result in your motor torque sliding back down behind the knee towards pull-up torque, essentially allowing it to stall. A better approach may be to allow them all to get to absolute full speed before coupling the load by fully injecting the fluid, and do it with all three couplings simultaneously. At that point you have the ramp-up of coupled torque happening at the motor's FLT, with breakdown torque available from all 3 motors to accelerate the conveyor once the fluid couplinga fully engage. It would be a cheaper solution than VFDs by far, and if it doen't work, not much invested.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[Marke]

Hello raisinbran

The situation that you describe is not uncommon and there are a few things to ensure to minimise this issues.

1. Use delay fill fluid couplings and try to ensure that all motors are up to speed before any of the couplings begin to engage. This will require a cascaded start of the motors, but these would typically be DOL start and therfore the start time is very short.

2. Limit the maximum torque of the couplings to around 160 - 180%. This is usually enough to start a fully loaded conveyor, and provided that all motors are driving together the over load current should not be higher than 200% and the electrical to mechanical transfer efficiency should be pretty high. The voltage drop should be around twice normal running voltage drop.

3. Power factor correction can be added which will help to reduct the line current and will therefore have some impact on voltage drop, but the correction that should be added would be 80% of the magnetising current of the motors and can be left in while the motors are running. The reactive current through the motors under these conditions would not be significantly higher than normal run and additional capacitors would probably not add any advantage.

The important point is to use delayed fill couplings to ensure that all motors are up to speed before the torque is transferred to the load, otherwise you can get the situation where the motor will not get to full speed and the slip losses are in the motor rather than the coupling. With the slip losses in the motor, it will be drawing a very high current and causing significant voltage drop.

If the coupling transfer through the fluid coupling is higher than the torque developed by the motor during start, the coupling is not able to do it's job.
If you are haveing difficulty getting the motors to full speed, and you are not using delay fill couplings, reduce the fluid level to reduce the torque transfer untill the motors can get ot full speed.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[aolalde]

Hi raisinbran:

Because your problem of torque arises when the motor is working after the breakdown of the speed-torque curve, a VFD will not provide any improvement.

The system works most of the times, and then you need only a little extra torque for critical load conditions.

Your best option is the capacitor bank close to the motor, as you originally posted. Measure the PF and kVA demand when the clutch engages and select a capacitive bank that provides around 80% of the reactive power.

The motor torque will be improved because the reduction of current in the lines will increase the voltage applied to the motor terminals then the torque will be increased with the squared ratio of the improved voltage as compared to the original voltage.

Take the safety precautions and proper interrupters to handle the capacitive load and provide automatic means to discharge the capacitor bank after it is disconnected from the line.

 

[davva]

Raisinbran,

Did you ever follow up your plans to implement a capacitor start during the run up stage?

We have a similar problem in the Marine industry on a vessel in that in a fallback mode we have a weak MV (10MVA 6.6kV) supply which is transformed down to 440V via a 2.5MVA transformer. We are required to start a motor (600kVA) star-delta, whilst the transformer is almost fully loaded. The voltage drop at final consumers around the vessel can fall outside of classification requirements (greater than 10% dip for more than 2secs) during the star acceleration period.

I had thought of moving the motor to be supplied from a dedicated supply transformer, but then thought about capacitor motor starting for the acceleration period to improve the voltage drop and then switching them out when the motor is up to speed(which is when I stumbled across this discussion).

Ignoring all the resonance effects, overvoltage etc.. Am I right in saying the corrected inrush current (seen from the upstream supply breaker) would be reduced (approximately) by the starting power factor, i.e. if the starting power factor of the motor was 25% then the inrush current following the capacitor start modification would be 25% of the original value (on the basis that the capacitors were identically sized to the motor's kVAr starting requirements)