Correcting voltage dip of XFMR during motor starting

[PJG-FES]

Have an application with a 440 vac, 450 hp, 3600 rpm, motor that will not accellerate an unloaded screw compressor. FLA is 511. LRA is 2964.
This is a large installation utilizing a 3000kva XFMR with a 22.5 kv primary and 440 vac secondary.
Prior to attempting to start the 450 hp motor, the XFMR secondary voltage is 442 vac average, 1200 amps existing burden, and 0.91 lagging power factor.
During the attempted starting of the compressor/motor set, the XFMR secondary voltage dips to 408 vac average, total burden is 3982 amps (of which 2782 is the 450 hp motor), and system power factor is 0.40 lagging.
The motor/compressor set acellerates for 18 seconds before the starter overload protection shuts it down. At no time does this motor exhibit a stall condition. All calculations point that the set will come up to full speed, and it did once during testing, but hasn't been able to since.
The compressor/motor set turns over freely by hand, the motor wiring has been confirmed to be correct, and the motor has been run solo and draws the correct amount of current.
The motor starter is of the soft variety (SCR's) and has a bypass contactor for running. We have attempted to start the set across-the-line using the bypass contactor with no success.
There are presently no power factor correction capacitors installed on the 440volt power system. Voltage drop thru the 440 volt cables, switches, etc. is 25 vac during 450 hp motor acelleration (XFMR to leads at motor) with a 48 volt dip at the XFMR secondary. XFMR primary voltage actually increases by 300 vac during motor starting!!!
We think we can get this started if the voltage dip at the XFMR can be reduced. Cable sizing and switchgear have been confirmed both in proper sizing and integrity.

 

[Jraef]

Well you already have a soft starter there, so you should be able to use that to get it started. Have you tried adjusting the starting parameters? If it has current limit, try lowering the limit. If not, try extending the ramp time.

It is also permissible to bypass the OL during start if you have proven that it cannot start otherwise. But I would be concerned with a nearly 10% voltage drop with regards to the effects on other connected loads. Try to solve the problem first by adjusting the soft starter. An unloaded compressor should start with about 350% current in under 10 seconds, so low enough that even a Class 10 OL relay should not trip out.

Other issues:

• Are you waiting a long time between starts? If too close together, your motor may be starting out too hot and the OL relay knows that. On big motors you may have to wait up to 40 minutes between starts.
• What brand / model of soft starter is it? Some use what is called "2 phase" starting which only uses SCRs in 2 of the 3 phases and a direct line connection on the 3rd phase. While this isn't usually a problem, it can be if the load is too heavy because it can cause the motor to heat up faster than normal if not dealt with properly (some manufacturers have ways to deal with that issues, most do not but never mention it). Still other soft starters use 3 SCRs and 3 diodes, which again works OK on small motors with light loads, but can be really bad on large motors like this.
• Are you absolutely sure the unloader valve is working correctly? One of the most common problems on compressor starters is that the unloader valve is stuck closed; not venting to atmosphere. Just because you are energizing the solenoid does not mean the valve is opening.
• Do you have a Class 10 OL relay on a motor that should (or could) have a Class 20? A lot of soft starters come with Calss 10 OLRs because that is what IEC motors need, but most NEMA design motors (and Above NEMA like yours) can have Class 20 protection.

 

[coulter]

Have an application with a 440 vac, 450 hp, 3600 rpm, motor ... FLA is 511. LRA is 2964.
This is a large installation utilizing a 3000kva XFMR with a 22.5 kv primary and 440 vac secondary.... XFMR secondary voltage is 442 vac average, 1200 amps existing burden, and 0.91 lagging power factor.
During the attempted starting of the compressor/motor set, the XFMR secondary voltage dips to 408 vac average, total burden is 3982 amps (of which 2782 is the 450 hp motor), and system power factor is 0.40 lagging.... All calculations point that the set will come up to full speed, and it did once during testing, but hasn't been able to since....

I can see about three issues:
1. This 440 motor voltage and xfm voltage have me bugged. Unless you are 50 hz, the xfm nominal voltage should be 480V and the motor nameplate should be 460V. If this is true, jack up the xfm taps to at least 480V. That will really help out your motor starting.

2. Motor starting current is 2782A. That is way too high for a motor FLA of 511A. You are right next to LRA. as jraef suggested, look at the parameters, get the current down, ramp rate up.

3. I suspect the once it started was the day the utility voltage was up and the xfm output was closer to 480V.

Repeat: Check the nameplates. Are you sure the motor and xfm are 440V?

carl

 

[bob]

There are presently no power factor correction capacitors installed on the 440volt power system. Voltage drop thru the 440 volt cables, switches, etc. is 25 vac during 450 hp motor acelleration (XFMR to leads at motor) with a 48 volt dip at the XFMR secondary. XFMR primary voltage actually increases by 300 vac during motor starting!!!
We think we can get this started if the voltage dip at the XFMR can be reduced. Cable sizing and switchgear have been confirmed both in proper sizing and integrity.

I agree with the other comments The voltage at the transformer should be
at least 480 volts if your equipment will allow. You will have to get the utility
to change the taps in the transformer. Another option is to install the proper
size capacitors on the motor feeder.

 

[templdl]

There is also an assumption that there is not voltage sag on the primary because a value wasn't provided. One can't assume that the primary voltage will be stiff.
I just had an application question for a 4160v 400hp motor w/51 FLA with an7.8xLRA on 2500 kva transformer that the customer said would start by would drop the voltage to about 3500.
I asked him if the voltage on the primary bus was dropping and he replied, we are the power generating plant generating 650megawatts of power and I don't think that?s our problem. However he did have a significant load on the transformer besides the motor. If we determine that the xfmr is not of a large enough capacity I'm recommending an autotransformer that is less expensive them a isolating type and with lower impedance.
This is a fun application question because it requires gathering more information that was provided at first in an effort to determine what the dynamics of the system is.
The customer did reply that the 4160v bus actually dilled to 3900v when the motor was started with an existing 300a on the transformer. Our computations that we made when taking the impedance of the transformer into account and adding the LRA of the motor our figures resulted in about an 8.8-8.9% voltage drop which is veru close to what they actually measured. 10% or less voltage drop is usually acceptable. Where they got the 3500V at first I have no idea. Now I have to determine what would an acceptable voltage drop be so we can design an autotransformer for them to supply just the motor.

 

[bob]

PJG
How about responding to these guys who took their time to answer your question.
There is a lot of information in these responses.

 

[Jraef]

I can see about three issues:
1. This 440 motor voltage and xfm voltage have me bugged. Unless you are 50 hz, the xfm nominal voltage should be 480V and the motor nameplate should be 460V. If this is true, jack up the xfm taps to at least 480V. That will really help out your motor starting.

...

carl

Lots of older pre-NEMA motor designs were 440V, and so were the transformers that fed them in some areas of the country. If you put 480V to a motor designed for 440V, you are already at +9%. Then if the utility climbs to just +5% of 480V (half of the allowed tolerance), you are at 504V and now far outside of the maximum tolerance for the 440V motor (+15%). Might help for starting torque but the motor will fry in short order.

NEW installations are now going to be 480V, but he never said if it was new or old. I would expect it is old. In that case even if he had a 480V transformer, he should use at least 5% FCBN taps to feed that motor if the nameplate says 440V.

 

[coulter]

... All calculations point that the set will come up to full speed, and it did once during testing, ...

...The motor starter is of the soft variety (SCR's) ...

... but he never said if it was new or old. I would expect it is old. ...

Interesting you would guess old. I'll wait for the OP to answer.

carl

 

[PJG-FES]

Sorry about delay in replying to all that responded to my query.
Some answers:
This is a new installation in Korea.
The motor is wound for their standard current of 440 vac, 60 hz. 3 phase.
We have adjusted the soft starter all over the map with no improvement.
As I am writing this, we have a PE on-site that will be connecting a torque meter to the compressor/motor set in order to determine the motor output and compressor requirement.
We also have dispatched a new 450 hp motor to the site. This motor has more starting torgue than the original and will be installed after the actual starting torque output of the original motor has been verified.
There is no voltage sag on the 12.5 KV/440 vac XFMR primary during accelleration. As a matter of fact, the primary voltage actually increases.
Probably due to the power factor going low.
Stay tuned!

 

[LarryFine]

Just my gut reaction when reading the OP:

Is it possible there is a transformer problem? Remember, an open-Delta system can measure like a Delta, especially when unloaded.

The primary voltage rising seems strange.

 

[PJG-FES]

Outcome following:
Thru the use of a torque meter, we found that the original 450 hp motor was delivering rated starting torque (corrected for reduced voltage) and the compressor requires more starting torque than expected (not certain as to why).
We ended up installing a different 450 hp motor of a different design that has substantially more starting torque than the original motor.
The compressor/motor set was then successfully started and was up to full speed within 8 seconds.
There is an identical compressor set at the same site that has identical starting issues, and we plan on replacing it's motor, as well.
Jeesh!!!