Increasing drive output from 60 Hz to 120 Hz... transformer losses?

[jjustice]

If I increase the output frequency on a given drive from 60Hz on the output to 120Hz, what can I expect as far as output-side transformer loss changes? The output voltage would remain the same. I am pretty sure winding losses will go up, but based on practical experience, by what factor? I suspect that halving the volts/hertz ratio will drop the flux density by an amount that will negate any increases in core eddy and hysteresis losses and give good no-load performance. I am concerned about the cooling on the unit under a loaded condition however. Any help would be greatly appreciated!

 

[ScottyUK]

You're correct regarding how losses will behave. Hysteresis losses would go up because they happen on a cycle-by-cycle basis, but the flux density would be halved because the voltage remains constant while the frequency doubles. Winding losses may increase due to skin effect but, unless this is a really big transformer, skin effect is of little concern at 120Hz. You may find the overall losses don't change greatly, but you'd need to know more about the core material and the operating flux density to say for sure.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jjustice]

Scotty, thanks for your reply! I could post a BH-curve and show you where I ran the particular design if you'd like, but I can assure you at 120Hz its WAY down in the "linear" portion of the graph. Transformer sizes in this application are 600KVA and smaller... nothing "large" by any means, but they do run a 480V input and routinely see heavy loading. That said, I think they are too small to run into a skin effect problem. We usually start worrying about that around the 1000KVA 480V mark. I am much more concerned about the eddy component of the winding stray losses. Frequency squared is in the numerator of that calculation. Its very difficult to measure stray losses and then "accurately" determine what percentage of that number is stray loss due to the tank, various structural components, eddy losses, etc. Perhaps practical experimentation is the way to go here...

 

[7anoter4]

I agree with Scotty the transformer losses will not raise significantly.
The mmf [E=U-Zp*I1] will decrease [2-3% I think], as Zp will raise [Zp=Rp+jXp] Xp=Lp*2*pi()*f will be doubled.
E=B*S*w1*2*pi()*f and as Scotty said if f2=2*f1 then B2=B1/2.
In this case hysterezis losses =n*f*B^1.6 will decrease B2^1.6=B1^1.6/2^1.6=B1^1.6/3 then:
hysterezis losses2= hysterezis losses1*2/3[67%]
The eddy current will stay unchanged [approx.].As you said already, no significant change will be in copper losses due to
skin effect.
The big change will be in voltage drop across the transformer and the secondary voltage could be 4-5% less.