Induction motor protection

[veritas]

Hi all

I am dealing with a 2MW induction motor where the manufacturer's data gives the starting time as 18s at 100%Un (33s at 80% Un) with the cold locked rotor time being 14s and the hot locked rotor time being 10s. As it stands it looks like the motor can never be started then?

I figure that the starting current is not 5.2 p.u. for the complete startup time but must taper off at a certain rate when approaching rated speed so that the thermal energy used remains just below the damage level? But I'm not sure if I am right?

Secondly is there any other way of providing proper thermal protection Besides a scheme that switches from one curve when starting to a faster curve when running with the aid of a speed switch?

Thanks for any assistance with this.

Veritas

 

[AMBMI]

I think you simply need a thermal image element (F49).
The equation implemented in such element is taking care of the effect of the current previously "seen" by the motor.
You should be able to fine the thermal time constant to use as relay setting using the manufacturer max allowed time data
All moderns motor protection devices are including such feature

 

[burnt2x]

veritas,

Please check if there are embedded thermal sensors on your motor windings.
Ours are 1 MW induction motors, DOL starting and the scheme is to allow trip blocking on starting of the thermal protection as long as the thermal sensors do not send trip signal due to heat build-up above the specified damage levels. All the winding thermal sensor trip contacts arranged to trip on any detected over-temperature. At synch speed, normal thermal protection take over.

Respectfully.

 

[dpc]

It can probably be started, but it may require some special protective features. The locked rotor safe stall time is based on locked rotor. During a start, the rotor is actually turning, albeit slowly at first, so the heat transfer and electromagnetics are not the same as locked rotor condition.

I believe there is some discussion of this in the GE Multilin 469 manual and probably the SEL-710 manual.

 

[slavag]

Hi.
Veritas, you have two option. Please see first option with speed switchfunctionality
1. You use for start current function I^2xts (start current supervision).
When protecting motors of ExE-type, for instance, the safe stall time may be shorter than the start-up time of the motor, which is why a speed switch on the motor shaft is required to give information on whether the motor is accelerating during motor start up. In this case, the start-up time is set to slightly below the safe stall time.
The speed switch should be open at standstill and closed during acceleration. When the input is activated, stage
Is^2 x ts/ Is> will be blocked. If the motor does not start to accelerate, stage Is^2 x ts will trip when the reference value, Is^2 x ts, is exceeded. If start-up supervision is based on overcurrent protection, stage Is> will trip on
expiration of the set operate time
However, if the safe stall time is longer than the start-up time of the motor when no prior load, a speed switch will not be required.
2. Thermal image ( based on current) function 49.
this function have cold and hot curves, it's separate function.
3. SC overcurrent based on the doubling option, in this case, set of current will choice 70-90% of Is current.
Second option based on the several curves. I will prepare some answer and will add to the thread.
Best Regards.
Slava

 

[dpc]

slava,

The speed switch will work, but the newer relays, e.g. SEL-710 claim they can do it without a speed switch. The SEL-710 actually computes the motor slip from the Steinmetz motor model and so it "knows" the speed (assuming the motor parameters entered are fairly accurate). Based on this it can adjust the safe stall time.

The GE Multilin 469 relay can be programmed with the voltage-dependent Overload curve and has other custom programming capability to deal with high-intertia motors (so they say).

 

[veritas]

Hi all

AMBMI - yes I want to use the 49 element but how to set it? Id the start time is 18s, and the cold LR withstand time is 14s do I set the cold 49 curve below 14s? My fear here is tripping on startup.

burnt2x - (I like this handle) yes there are sensors in the motor but these the customer has taken to a PLC which is in no way connected to the relay. Thus this is not an option.

DPC - thanks, I'll have a look to see what GE and SEL have to say.

slavag - I am also looking at the speed switch option. But to first get the theory right, I believe that is the start time is 18s and the cold LR time is 14s, then the motor has been designed in such a way that the thermal capacity used during startup should not exceed the LR condition. But I would guess that there is not much safety margin.

I think the SC option also requires a speed switch. In anu case I usually have a two stage SC approach. Stage 1 = 8 X FLA, 300ms time delay and Stage 2 = 2 x LRA, 0s time delay.

Thanks all. Now to think about this some more.

Regards.

 

[slavag]

Heh DPC, again formulas and formulas. Models in the protective relay.
Small things:assuming!!! the motor parameters entered are fairly accurate. Based on this it can adjust the safe stall time.
I'm prefer something more simple.
I^2xts for the 5.2In 18sec and stall 5.2In for the 14sec as definite time overcurrent protection w/o any speed switch.
Voltage depended overload????
Old field guy was right. how are calculated those parameters?
Regards.
Slava

 

[dpc]

For the SEL-710 you have to enter some basic motor data (FLA,LRA, safe stall time, etc and it calculates it. For large motors, it is data that is often available - for small motors, often NOT.

But SEL provides some rules of thumb if data is not available. Plus the relay will know the LRA and acceleration time based on actual start data, assuming the motor survives the first start! I'm not promoting the SEL-710, but it does appear to offer some interesting features and a good way of computing the motor thermal model. You should take a look.

I agree the speed switch is the most direct way of knowing, but speed switches are not without their own set of problems as well.

 

[slavag]

O.K. DPC.
Now it's more clear and better with rules of thumb.
Of course I will look it. Learn new it's always good.
Veritas, 8FLA 300ms for my pinion it's so much. Doubling option is good solution.
Regards.
Slava

 

[slavag]

Veritas. For my pinion: first of all you need choice relay.
2MW it's serious motor and you need one good relay, like to:
SEL710, SR469, P243, REF543, etc.
I'm sure those relays have solutions for all type of motors and will provide good protection.
BTW, think also about diff protection 87M.
Good Luck.
Slava

 

[veritas]

burnt2x - the customer is willing to route the RTD outputs to the relay if it will solve the problem. It is an interesting solution and I will ponder it some more.

slavag - I do not think I would go for thermal = 5.2In in 18s and definite time OC at 5.2In for 14s. What if the starting current is just a bit higher than 5.2p.u.? The relay used is SEPAM M87. It has diff.

dpc - has the SEL710 not been replaced with the SEL-749M? Can the latter do the same as the 710? Anyway, it is still something worthwhile persuing though I doubt whether the customer is willing to replace his new SEPAM just for the sake of LR protection at startup. He is quite willing to consider the speed switch route but as you mentioned the reliability of the speed switch makes us hesitant.

But what no-one has answered yet is how is it possible that the LR withstand time is shorter than the startup time? I have heard mention of double caged rotor?

Regards.

 

[LionelHutz]

Part of the reason is that the current tapers off as the motor accelerates and another reason is that once the rotor begins to rotate it is moving air around which helps it to cool a bit. The current may drop 100% fairly linearly up to 80% speed and then it will drop quicker from there to 100% speed. You should have been given a set of curves that show the speed vs current when you got the motor information.

The manufacturer should have given you a time vs current curve. It's quite possible that entering this curve in the relay is all that is required. As the motor accelerates and the current drops the trip time should be stretched out enough to allow the start.

The RTD's should be routed to the relay on a motor that large. Not using them could cost the customer a large motor. I'm not really a believer in the RTD's being capable of protecting a motor during a start. The temperature is changing so quickly that it's possible the RTD will not show a temperature problem until it is too late.

 

[slavag]

Hi Veritas.
I'm not familar with SEPAM M87, but I'm think it's good level of motor protection.
1. Diff protection is good, of course if you have CT's on the motor star point.
2. Start-up of motor 5.2In for the 18sec.
Possible work according to thermal stress I^2*t or according to DT, I'm prefer according to thermal stress.
5.2In for the 18sec isn't mean that at start of motor you have 5.2In within 18sec, but for lock rotor, yes, it's possible.
You need check this option according to the SEPAM manual.
I'm sure, SEPAM have all needed options.
And, Lionel is totally right, RTD for long overload protection only, not for start or LR situation. But check manual, maybe possibel combination of current thermal image with RTD inputs.
Regards.
Slava

 

[burnt2x]

Hi!
Embedded RTD's are far more reliable than temp calculated internally by electronic relays or thermal bimetallic strips trying to simulate motor temp! As long as the lines don't get get cut, you get motor temp at controls. Response time is a concern though. Additional speed monitor prevents runaway temp rise on these big motors.
We did well with our 1 MW FDF motors, I don't see you'll fail on 2 MW ones. It's been over 20 years with no MBO's, believe me!
rspectfully.

 

[electricpete]

I am dealing with a 2MW induction motor where the manufacturer's data gives the starting time as 18s at 100%Un (33s at 80% Un) with the cold locked rotor time being 14s and the hot locked rotor time being 10s. As it stands it looks like the motor can never be started then?

I figure that the starting current is not 5.2 p.u. for the complete startup time but must taper off at a certain rate when approaching rated speed so that the thermal energy used remains just below the damage level? But I'm not sure if I am right?

You are certainly right that current decreases as speed increases. It decreeases very slowly at first (low slip) and then faster as operating speed is approached. And of course heating is proportional to square of current, so a 10% decrease in current means a 20% decrease in heating. And also change in rotor current is slightly different than change in stator current due to magnetizing branch although qualitative conclusion of that ratio is difficult since there are two competing/canceling effects (one due to phase which favors lower ratio rotor/stator current during locked rotor and one due to magnitude of total current against relatively constant exciting current which favors higher ratio rotor/stator current during locked rotor)

Another effect is cooling. The locked rotor has no cooling while the accelerating rotor has cooling. Although I was under the impression the bulk of heat generated during start must be accomodated by thermal capacitance rather than heat dissipation, so this should not be a large factor.

Part of what it means is that you may find it advantageous to differentiate between locked rotor start and normal start by various means. Speed sensor, distance relay etc.



=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

I apologize, I did not read the other response... I see now others mentioned the same things. Sorry.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[veritas]

electricpete - no need to apologise. It is always good if someone reinforces what has been stated before by someone else.

burntx2 - my concern with RTD's is the sluggishness of their response. I would think that a good relay models the thermal capacity within the motor by means of current and RTD input - allows being made for the RTD's time delayed response.

Regards.

 

[burnt2x]

Hi!
I am a believer in doing things things better than the last time. Having said that, I guess we don't have to stay stucked on a solution that did good in the past!
Another solution I was thinking of is that of a motor fitted with a separate cooling fan that will not be dependent on the motor start! By having this setup, you can have fixed cooling air to your motor! Calculate the possible heat that can be generated on your motor and then size your fan CFM to cool your motor appropriately! By that you dont have to worry much on toasting you big motor! This may sound crazy at first but we do have big motors with integrated cooling fans.
My $0.02

 

[veritas]

burnt2x - not a bad idea I must admit. It is worth persuing - at the very least you'll not end up burnt3x!

Thanks and regards.