Transformer Heat Loss 2

[whycliffrussell]

Is there a standard way/formula to calculate heat loss (in watts) of a transformer?

I'm assuming you'd have to take into accountthe power factor and the efficeny of the transformer?

Is it just the real power throughput of the transformer multiplied by the 1-effy?

Also...does anyone know of a table(s) of standard transformer efficiencys and power factors?

Thanks in advance!

 

[davidbeach]

Nope, you have to have specific information on the transformer, though similar transformers generally have similar losses. There are two types of losses in a transformer - core and winding (often known as iron and copper). The core loss is a (nearly) constant loss associated with the transformer being energized and is a function of the voltage across the magnetizing branch of the transformer equivalent circuit. The winding loss is a result of the resistance of the windings and is proportional to the square of the current through the transformer. Typically the more winding material, the lower the winding losses, but that additional winding material requires a larger core and higher core losses.

Depending on winding material, temperature rise, and construction type, the losses can vary by more than a factor of 2 for the same size transformer. When considering large transformers installed indoors it can be less expensive to buy the lowest loss transformer and save the cost of the air conditioning and the ongoing operational losses than to buy the lowest cost transformer and pay for the losses and the air conditioning necessary to remove the heat of the losses.

 

[tinfoil]

To add to davidbeach's good comments:

Another factor to consider is maximum fault current. You may have to spec a minimum impedance (%Z) to ensure a particular worst-case fault level. This in turn has a dramatic effect on the amount of steel core and copper windings used in the xmfr design, which affects the heat losses.


BTW, less heat generated is another way of saying 'less electrical energy lost to transformation', which means reduced electricity purchases on top of reduced ventilation/AC (if the revenue metering is on the source side of the xmfr). Don't forget to include this in your cost model.

 

[whycliffrussell]

Thanks for your comments...but, I'm not specifying a transformer. I was just asked by another engineer to give them an idea of how much heat would be contributed by a 75kVA 600/120/208VAC non-force cooled transformer. Not every problem is a 'textbook' problem...I do not have the shunt resistence 'model' of the core or the series resistance of the windings. I was just wondering if there were a general formula or rule of thumb to get an idea of how much heat would be generated by the transformer (I assumed full load conditions and calcualted the heat generation based on both the core and winding loss)...using the 'Westinghouse Electric Utility and Distribution Reference' and standard no load and full load loss data I was able to estimate the heat generation by the transformer.

 

[jghrist]

If you know the efficiency, the loss in pu is 1-efficiency in pu as in your OP. Loss in kW at full load would be rated kVA times loss in pu.

 

[dpc]

As jghrist says, all transformer losses end up as heat rejected into the surrounding area. So if you make a rough estimate of transformer efficiency, the lost energy is the heat added to the room. There isn't anywhere else for it to go.

Power factor isn't going to have any impact on the heat generated, except to cause more current to flow and increase the kVA load on the transformer.

For a 75 kVA, 150 deg rise dry-type, Cutler-Hammer gives an efficiency of 97.2% at 1/4 load and 96.7% at full load. So figure 3% loss at 75 kVA which would be 2250 watts.

If you're looking for greater precision, you need to know the amp load on the transformer and separate out the core losses (which are constant) and the winding losses (which vary as the square of the current).

 

[davidbeach]

Even at the 75kVA level, you will find rather different losses for standard (150^oC rise) dry type, 115^oC rise, 80^oC rise, and energy star transformers. Loading becomes highly critical as they differ as to how much is core loss and how much is winding loss. Comparing the 150^oC rise and 80^oC rise transformers, the total losses of the 150 at full load are much higher than the total losses of the 80 at full load, but at no load the losses of the 150 are actually significantly lower than the no load losses of the 80.

If all you are looking for is how much cooling does the transformer need, dpc's 3% of rating is a good rule of thumb, but it may be rather over sized if the transformer is never fully loaded.

The energy star transformers are another matter; they are designed for maximum efficiency at 35% (or there 'bouts, I don't have the literature at hand) and you can find transformers that are more efficient at full load or at no load.