Estimating Power from slip -> temperature correction 5

[electricpete]

I have always thought that motor power could be estimated based on linear relationship with slip: 0 power at zero slip; nameplate power at nameplate slip.

I just saw IEEE 939-1995 (Energy Efficiency) section 6.15 which suggests that slip will increase by approx 0.35% per degrees C above 25C (winding temperature).


They go on to state "Most motor manufacturers, when specifying the full-load RPM, take their readings at 25C or on a cold motor".

That's the first I've ever heard of it. What use is a ficticous number based on 25C winding temp? Winding temp is sure to increase above that during starting and running.

Am I reading this right?

 

[rhatcher]

Hi electricpete,

I have my head straight again (for now) and wanted to comment more. First, I especially liked your post of 7/17/01 and gave you a star there. Your thought exercise of the effect to the stator of substituting a different rotor was right on in it's assumptions. The current seen by the stator will be equal in magnitude and power factor.

The idea that I want to add to your discussion of the equivalent circuit is that it is "ideal". It does not take into consideration the effect of rotor bar slot leakage reactance (and skin effect) with changing slip frequency. To model the motor for S > 20% (guess-timated value, varies according to rotor design), the resistance R₂ and the inductance X₂ are variable with slip. Note that this is not referring to the '1/S' multiplyer applied to R₂ in the equivalent circuit. Other posts in the forum have discussed the effect of slot depth and bar shape on resistance and inductance of the rotor as slip frequency changes. Anyway, the general shape of an induction motor's speed torque curve will be determined by the rotor bar shape and depth as well as the rotor's DC resistance.

With respect to the question of nameplate speed rating and ambient temperature, I called a friend who used to work at one of the Reliance motor manufacturing plants. His answer was that the ratings were based on 40C. However, when I explained the situation to him he became less confident of his answer. He has promised to call one of the design engineers at the plant to find out. I will let you know when I hear from him.

 

[rhatcher]

Hi nbucska,
Sorry, I keep forgetting to answer you. I haven't heard of specmanship. What is it?

 

[rhatcher]

Hi electricpete,

I heard from my friend at Reliance and was a bit surprised. Again, his source is a motor design engineer at one of the plants and this is based on actual practice. I would have liked to talk to the guy myself but it is probably better that I didn't since I may have never let him get off the phone.

Anyway, according to this source and at this facility, the motors are tested at room temperature which is close to 25C but nowhere near 40C. Then, the recorded speed is rounded to the nearest 5 rpm. The reasoning was that (1) the speed would only change by "1 or 2" rpm between that temperature and 40C and (2) even with rounding to the nearest 5 rpm the resulting namplate values are well within NEMA specifications for nameplating motors. He was careful to note that this was for induction motors and that for DC motors the ambient temperature plays a much more critical role in speed.

I went back and looked at your IEEE calculations to see if this made sense and discovered that it probably does. Going back and looking at the whole thing a little closer, it is somewhat misleading based on the wording and the example used (although perhaps not intentionally). I will post more later.

 

[jbartos]

Suggestions to electricpete (Electrical) Jul 12, 2001 marked by:
It is appropriate to consider all relevant variables when it comes to analysis of relationships.
///Yes.\\ If one considers motor torque- speed approximate equation for the speed with low slip, then there will be several potential parameters/variables that could be considered beside temperature (implied in some parameters/variables).
///Yes.\\Are you saying that important variables have been overlooked? The approach outlined above is that
S(P,Twdg,V)=Snp*(P/Pnp)*0.0035(T?-25C)*(Vnp/V)^2
where:
P = actual load (SHP)
Pnp = Nameplate load
Twdg = winding temp
T? = winding temp per IEEE 739 (although I would use ambient temperature here).
V = actual terminal voltage
Vnp = nameplate voltage
S(P,Twdg,V) = actual slip under actual conditions measured for this calc.
Snp = nameplate slip = syncronous speed minus nameplate speed.
///Generally, if you use two states, e.g. one state corresponding to nameplate data set (or variables in general), and the actual data set (or variables in general), e.g. from the test measurements, then each related data (or variable), as you used them in ratios (except for the slip) has to be addressed, and in its ratio form. That holds true for the temperature too, unless there is some reason to force the temperature values to be equal (which implies the ratio equal to one). Therefore, your concern about the temperature values was legitimate. One temperature value should be rated (corresponding to the rated (nameplate or spec sheet) set of data or variables) and the other temperature value measured to make the slip result (measured or adjusted for temperature) accurate. If you still work on it, you will see the difference.\\\

 

[electricpete]

jbartos - a lot of words, but no meaning to me. The model underlying the equation S(P,T,V)=Snp*(P/Pnp)*0.0035(T-25C)*(Vnp/V)^2 makes good sense to me. I only question which temperature should be used for T in light of the fact that IEEE 739-1995 suggests to use winding temperature while I believe I should use ambient temperature (or substitute the entire delta-T term with my best estimate of increase in rotor temperature during my field measurement above the rotor temperature which existed during factory test at nameplate load and 25C ambient).

My July 12 message was in response to your statement that "it is appropriate to consider all relevant variables...".

So I asked the simple question: "are you suggesting we have missed a significant variable?"

You provided no direct response, only your latest message which talks for awhile about the functional form of the equation and then ends with "if you work on it, you will see the difference". But I have no quarrel with the functional form of the equation (do you?). There is no ratio of temperatures (do you expect there to be?). The term 0.0035(T-25C) is an approximation to a ratio of resistances. So let me come out and ask you, do you have any suggested improvements to the above equation, other than the selection of which T to use? If not, what is it you're trying to say?

 

[rhatcher]

hey electricpete, another star...some things just have to be said. I also have lost patience before and reacted similarly, but to no avail. He just keeps on going. He may be the most prolific poster in the group, but in my opinion he only has about a 20% 'hit' rate (at best). The rest consists of lists of useless web links, quotes taken out of context from references that are obviously not understood, and, as in this case, unintelligable arguments presented in an attempt to prove that he is right when he isn't.

 

[jbartos]

Suggestions and answers to electricpete (Electrical) Aug 17, 2001 marked ///\\\:
jbartos - a lot of words, but no meaning to me.
///How true. It somewhat depends on the last grade that one finishes.\\The model underlying the equation S(P,T,V)=Snp*(P/Pnp)*0.0035(T-25C)*(Vnp/V)^2 makes good sense to me.
///I indicated in my posting: "That holds true for the temperature too, unless there is some reason to force the temperature values to be equal (which implies the ratio equal to one). This means that the 0.0035(T-25C) is questioned. The reason why it is questioned is that you use two different sets of data. One nameplate or specified and the other measured for a different state or set of values. The temperature factor must be considered accordingly else, the temperature could be adjusted the same somehow and the (T-25C) in ratio of normally different temperatures (T1-25C)/(T2-25C) would become (T-25C)/(T-25C)=1 for T1=T2=T, similarly as if Vnp=V, then (Vnp/V)^2 = 1, naturally.\\I only question which temperature should be used for T in light of the fact that IEEE 739-1995 suggests to use winding temperature while I believe I should use ambient temperature (or substitute the entire delta-T term with my best estimate of increase in rotor temperature during my field measurement above the rotor temperature which existed during factory test at nameplate load and 25C ambient).
///Incidentally, there is Form E in IEEE Std 112-1984 which addresses all necessary temperatures:
Stator winding Temperature, Tt in °C
Ambient Temperature, not designated, so define it Ta in °C
Specified Temperature for Resistance correction (see 5.3.1 therein)
and there come temperature rises that have to be appropriately selected according to your motor nameplate temperature rise value.
There are no other temperatures stated. So, this appears to be very clear to very many users of the IEEE Stds, else it would have been elaborated on / revised since this standard has been around for long time (17 years). So where is the problem? Perhaps, if more date from specs and nameplates been posted, then the solutions would be clearer.\\My July 12 message was in response to your statement that "it is appropriate to consider all relevant variables...".
So I asked the simple question: "are you suggesting we have missed a significant variable?"
///It appears that a significant variable has been misinterpreted or maltreated.\\You provided no direct response, only your latest message which talks for awhile about the functional form of the equation and then ends with "if you work on it, you will see the difference".
///You are the only one who could work on this since you did not post the nameplate and measured variables values. So what is this statement all about?\\ But I have no quarrel with the functional form of the equation (do you?).
///I posted my interpretation and potential treatment without insisting on something based on guessing and question marks.\\ There is no ratio of temperatures (do you expect there to be?).
///Yes. That what my previous post generically addressed.\\The term 0.0035(T-25C) is an approximation to a ratio of resistances. So let me come out and ask you, do you have any suggested improvements to the above equation, other than the selection of which T to use?
///It is addressed second time by this posting. The temperature ratio is appropriate unless the temperatures stay the same, then the ratio will be equal to one, similarly as the terminal voltage V equal to rated voltage Vnp, V=Vnp resulting in ratio equal to one.\\ If not, what is it you're trying to say?
///Not applicable question.\\

 

[electricpete]

jbartos It would appear that no further constructive discussion is likely between us.