ECN Forum Forums Technical Discussion Electrical Theory and Applications Power factor and transformer efficiency

We have installed a single phase 50A 120/240V service to a small building that is about 2500ft from the utility power meter. We have used two new 15 KVA 600V to 120/240V transformers to step the voltage up and then down for voltage drop. We have buried 2500 ft of 3 conductor power cable for the 600V feed to reach the load end.

The customers load normally is very small most of the year except when they need air conditioning or its very cold.

According to utility power meter they are consuming about 1400 KWH per month, much more than we would expect for the loads attached and the weather we have had.

We are not metered or billed for the fairly significant reactive power componant that would be evident in this type of arrangement but I suspect poor power factor could be affecting the overall efficiency of the setup.

My reasoning is that the efficiency of a typical distribution transformer may be affected when the input line has a poor power factor. At this light loading this would be the case.

I am looking for discussion regarding how a transformer input with a poor power factor and thus a distorted waveform could raise its core and winding losses.

If this is found to be a factor than we can mitigate with a small capacitor in the system sending unity power to the load transformer.



We notice that the step down transformer on the load end is quite hot to the touch. Both transformers are new epoxy encapsulated outdoor units by ACME or REX I forget which.

We plan to visit the installation to measure the
power factor of the 600V circuit loaded and unloaded.

If anyone has advice as to if we are on the right track or have any useful suggestions or measurement it would be an informative discussion.

Be mindful of impedance realities when running underground feeders:

IF the ground is saturated with conductors -- think salty soils, saturated ground with ionic soils -- then you're heating the soil.

This is the reason why ASEA ( Sweden ) became the world's expert in DC power transmission. Sweden has endless islands just off her coast -- which have to be powered from the national grid. Ages ago the experts ran in to your headache. The Poco was bleeding energy into heating up the Baltic Ocean. It's purely an AC/ Impedance effect.

By shifting over to DC, the inductance losses went to zero.

Electrical Contractor Magazine wrote up the San Francisco experience. Less than three years ago, a DC link was run under the bay direct to the city. It had to be DC because AC would've bled too much energy -- and also would've cooked the insulation.

It's probably out there on the Web.

Next, encapsulated transformers ALWAYS run hot to the human touch -- even when everything is perfect. The epoxy functions like a double fur coat. So, the only way that ANY heat can be shed is if the exterior is so hot you can't sit on it.

Their insulation is so robust, they can take such temperatures for forty-years.

In all of your post you provide no mention of your conductor sizes.

I'd skip any concern about reactive power issues. If you're not being billed for loading down the Poco, then it's nothing. 15kVA is a joke to the Poco.

You may want to re-calculate your taps. With the long run they should be different, not symetrical.

Lastly, Pocos get away with underground distribution because they wrap their grounded neutrals around their three hots. (Medium voltages, primary distribution, typical) (Each hot is given its own braided wrapper.)

This means that their inductance waves don't travel outside their cable bundles/ buried triplex schemes. This means radio and TV peace -- and supression of inductance waves into the soil.

When you ran your secondary conductors -- they weren't shielded. So your power is interacting with every inductable feature of nature -- such as soils that immitate sea water -- electrically.

If this is an issue, even larger conductors will not have a material impact. The impedance is coming from close association with the wrong kind of soil.

All of the above is opinion from a distance, so double check everything. Your soil conditions may actually be favorable and my supposition entirely misplaced.

The cable run is on a hillside of mostly sand and gravel soil, not sure what the soils magnetic properties could be.
We have run a #2 AWG 3 conductor ACWU aluminum armoured cable using 2 conductors for the 600V and the other #2 conductor along with the bonding conductor and armour in the cable grounded. We have grounding electrode on the 120/240V incoming utility service and another electrode at the load end on the load side neutral of the transformer.

The no load losses due to inductive or capacitive cable properties at 600V
would seem very small in my opinion but that is something we have yet to
measure. I recognize there are significant losses at utility distribution voltages as you mention especially cables of larger size.

We know that the utility will not be metering reactive power in the system but we believe there would have to be reactive power especially in the higher impedence of the lightly loaded 600V windings and I wonder what effect this has on the core efficiencies of the transformer.

It would seem to me that having the voltage and current out of phase like this must have some effect on transformer efficiency but adding a capacitor somewhere would consume some wattage too.

Tesla, I do not understand this advice:

"You may want to re-calculate your taps. With the long run they should be different, not symmetrical."

We have both transformers using the same 100% taps so 240V-600V step up and 600V-240V step down.
We do have taps on the transformers to adjust the turns ratio.
Is having them identical a problem? This is a new consideration to me.

Thanks

One should actually pull some voltage readings at both spots -- loaded and unloaded.

It's a law of physics that you'll have some voltage drop across 2,500 feet.

Consequently, you may wish to compensate -- which can be done at either or both ends -- by using the alternate taps.

At the Service end, you'd step up the voltage... say 2.5% MORE than the standard as set by the factory.

At the panel end, you'd step up the delivered voltage by SHAVING the step-down turns ratio.

And, you can do both.

The point being to make sure that the panel voltage is within norms.

As a practical matter, as I said, I'd test it loaded and unloaded and make sure that the voltage swing stayed within practicality. I wouldn't waste my time trying to perform a math calculation at all.

You don't want the voltage spiking too high under a trivial load -- nor do you want it bogging down when usage is strong.

It's a judgement call.

After 2,500 feet standard factory tap settings are almost certainly wrong.

Adding a capacitor won't change the wattage significantly, but if you over correct the power factor, it will increase the I*I*R losses. I think the capacitor would have to be on the load end to make any difference.

It isn't impossible that you would have a fault in the underground cable.

Seams to me a current measurement at each end is in order. Both loaded and no load. Also disconnect the primary at the load end and measure the current. Make sure you do not have a nick or bare spot in the feeder that is leaking current to the earth. Make sure you also test this when the earth is wet as the leakage will be worse then.

A direct break in the insulation would peg a meter.

I had a <i>very</i> trivial manufacturing flaw in #10 THHN, once.

It tripped the GFCI breaker in a heart beat.

======

I'm going to presume that the feeder was checked -- if not meg 'd.

Hello Bigpapa, and thank you for the interesting post!

We should look at this scenario from the KWHs consumption first, to determine if there is an issue.

The POCO has submitted Billing Statements to the Client, which show the Consumption for a typical Billing Period of 1,400 KWH.

I am _Assuming_ this Billing Period is 30 Days (30 Days between reading of KWH Meter).
I am also _Assuming_ the KWH Meter is a Standard Analog type, with a Rotating Register and 4 or 5 Physical "0-9" Dials, driven by the Rotating Register.

If the Client is billed at a Single Flat Rate, of let's say 12¢ / KWH, this would be $168.00 for a consumption of 1,400 KWH. Not too bad at all!

If the Client is billed per "Basic & Rate Exceeding", then the results may be as follows:

1. Basic Lifeline = 1st 500 KWH @ 12¢ / KWH = $60.00
2. Remaining 900 KWH @ 18¢ / KWH = $162.00
Total for 1,400 KWH Consumption / 30 Day Billing Cycle = $222.00
Still not too bad; close to what I pay.

The values and Breakdowns listed above would be typical for a Standard Analog type KWH Meter.
If the Client has a Digital (TOU) Meter, the Billing Rates will be reflected per the Time Of Use, so anything above Basic Lifeline Consumption may become inflated - possibly as high as 45¢ / KWH for peak Hours.

With that in mind, let's look at the Breakdown of 1,400 KWH per 30 Days:

1,400 KWH for 30 Days = 46.67 KWH per Day.
46.67 KWH per Day = 1.94 KWH per Hour.
Also, 46.67 KWH per Day would equal 5.834 KWH running for Eight Hours.

Could the Client be using this level of Power under normal circumstances? It is very likely.

Reference the Cooling / Heating Equipment.
If the Cooling is Three Tons / Heating is 36,000 BTU, the Peak Power Drawn from the POCO would be somewhere between 6.0 KW and 10.5 KW - depending on efficiency.
Figure an average of 8.25 KW running for a total of Six Hours each Day, and the 46.67 KWH per Day Consumption is easily done!
This does not include Lighting and other Loads.

The best way to address this scenario is to decide if there is a reason for your Client to Consume much less than an average of 46.67 KWH per Day.

This can be a simple Revelation:
Are the Lights always on?
What Appliances are running?
Are the Heating and Air Conditioning "Fighting" for Dominance?

True Power Consumption is what the KWH Meter tallies. Low Power Factor will not show a Readable KWH Consumption, unless there is <.50 PF constantly on the 15 KVA Maximum Apparent Power Rating of the First Transformer.
This would be evident at the First Transformer's Input as Overheated Conductors.
15 KVA @ 240V would equal 62.5A, so if the Primary Feeders are #2 AL. Protected by a 70/2, there would be noticeable Heat with a constant 15 KVA Load (+ 3 Hours at 15 KVA).

Measuring PF at the Transformers with no Load will show the results of the Core Losses and the Transformer's Primary to Secondary "Stabilizing Load" (mostly in VARs, however whatever Heat is produced will be equivalent Wattage).
About the same will be found under Full Load or 50% Load, just with higher Heat - which would improve the Power Factor.

Lastly, if 50% of the Transformers' Maximum Apparent Power Rating was drawn for Eight Hours each Day, with a Power Factor of 80%, this would result in 48 KWH per Day.
Equation:

15 KVA x 0.5 = 7.5 KVA.
7.5 KVA x 0.8 PF = 6.0 KW.
6.0 KW x 8 Hours = 48.0 KWH.

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

I have a few Questions to pose:

1. Is the Client's Meter a TOU Digital Type?
2. Did the Client request investigation regarding excessive Billing Amounts or excessive Consumption?
3. What are the Client's typical Loads?
4. Could the 46.67 KWH per Day be normal?

Any additional information you could add would assist greatly in a much more accurate Guess...

--Scott (EE)