Power factor correction experiment

[Besoeker]

I am afraid you did not read post #244.The current taken by the example induction motor with full load current of 18A driving a fan type load does not exceed 7A at 0.55 p.u voltage.

That calculation is invalid.

http://books.google.co.uk/books?id=osa7ldPgUpEC&pg=PA275&lpg=PA275&dq=steinmetz+motor+equivalent+circuit&source=bl&ots=S5NwRKdPmE&sig=uv74AQ-YNxXlZrmgrHu0Tc9HDNo&hl=en&ei=bhLOTvv1L9SR8gP2w-TbDw&sa=X&oi=book_result&ct=result&resnum=1&ved=0CB8Q6AEwAA#v=onepage&q=steinmetz%20motor%20equivalent%20circuit&f=false

Look at Fig 6.6 on page 278 and you might be able to work out why.

The increase in rotor resistance was meant to drive the rated load at much reduced speed.But here we are discussing about fan type load whose torque requirement is much less at reduced speed.So your objection does not seem relevant.

What mivey wrote is pretty like the example you gave in post #275 to support your contention.

 

[Besoeker]

111203-1939 EST

Besoeker:

I can measure resistance, voltage, and current relatively accurately for my purposes.

This I have done and over the measurement range of the Kill-A-Watt I have determined that it is reasonably accurate. Then I have used low D capacitors to show that the Kill-A-Watt reads near 0 watts with the capacitors as the only load. Note: how bad the TED performance is when subjected to a nearly pure 90 deg shift in the current.

For intermediate PFs my motor test shows stability of the resistive component as the PF is changed. This is a fairly good test that the Kill-A-Watt is fairly accurately handling changing power factor.

An instantaneous power measurement on the whole house for proof that PFC won't save any energy is much easier, than trying to use the utility kWh meter to accomplish the same objective. But I am going to have to follow that route with my outside meter if I can not get a kWh meter from the utility. Currently we have wind, rain, snow, cold as obstructions to using the outside meter. My plan is to set a video camera in front of the meter to get accurate timing information of disk rotation, simultaneously do the TED whole house monitoring, and cycle the PFC on and off over maybe 5 to 10 minute periods. This time is partly controlled by the furnace blower on-off period. Blower on time is about 11 minutes. Refrigeration equipment has long time constants, and there are three of those, and some miscellaneous other items that cycle on and off. So to get a clean steady load may take a while, especially during the day.

This is clearly not something I can tell an average consumer to do and expect much success.

I agree but I think you and I and the posters on this forum are not the average consumer.

For my investigation I turned off all cyclic loads and ran each test for 30 minutes.
I did that a number of times with both the magic device in circuit and without it.
The energy readings on the metering to my house was up or down by no more than 0.01kWh on any of the tests. The least significant digit on our energy meter and, as I said, that is the final arbiter on what you pay.

I did also download currents and voltages in *.csv format from the 'scope to drop into a spreadsheet so that I could present them as waveforms and put them in my report.
Also, with them in numerical format I could, and did, use that data to calculate instantaneous and average power over a few cycles.

In short, the device yielded neither power savings nor energy savings.

 

[gar]

111204-1640 EST

I am now evolving a procedure for the average homeowner to use to prove that the PFC capacitor saves no energy. As follows:

Force furnace fan to stay on. Turn off air conditioner. Turn off other items that might produce a cycling load. For example, copy machines, or laser printers. Leave refrigeration equipment on. Refrigerators and freezers have long on times. If dehumidifier tends to run most of the time leave it on. Use the TED instantaneous power reading and the one second plot to determine that nothing changes within the test time.

We want inductive loads to be on during the experiment.

Use a West Bend kitchen timer or a regular stopwatch as the time measuring device. Use 8 revolutions of the spinning disk power company kWh meter to determine load power.

When load power seems to be stable turn on the PFC capacitor. Go to the kWh meter and wait until the black mark reaches something you will use as your reference point. Start the timer. On the eighth revolution when the black mark reaches your reference point stop the timer. Before starting this test record the TED instantaneous reading, and at the test end. Both should be essentially the same, maybe +/- 10 W. If not something changed during the test and you have to start over.

Turn the PFC capacitor off. There will be a change in the TED reading based on my experience. Record this new value. And start a new 8 revolution time measurement. Record TED at the end again and the 8 revolution time.

I expect your two time measurements to be quite close. If not something changed just at the time you switched the PFC capacitor off.

The actual load power can be calculated from the equation:
Pave = Kr * N * 3600 / Tn
where
Kr is the meter constant
N is the number of revolutions, and
Tn is the time in seconds for N revolutions.

My just completed experiment used 4 revolutions because it is raining out. Also I had to restart the experiment because something turned off in the middle of the test.

Results:
Capacitor ON. Ted before 1280-1270, and after 1280, time 131 seconds.
Capacitor OFF. TED before 1420, and after 1420, time was 132 seconds.

The timer only records in increments of 1 second. Thus, essential the same energy use with or without the PFC capacitor.

For my meter the power was:
Pave = 12 * 4 * 3600 / 131 = 1319 W
Probably less than 1 % error between the two tests.

.

 

[mivey]

111204-1640 EST

The actual load power can be calculated from the equation:
Pave = Kr * N * 3600 / Tn
where
Kr is the meter constant
N is the number of revolutions, and
Tn is the time in seconds for N revolutions.

Kr is the dial multiplier and is not the number they should look for. Kh is the secondary constant and tells you what each turn of the dial represents for the meter (watt-hours per revolution). With no PTs or CTs, the secondary and primary Kh are the same. If there are CTs and PTs, we have to find the primary constant by Pri Kh = Sec Kh * CT ratio * PT ratio.

The formula for the average power is:

kW = 3.6 * Pri Kh * # revolutions / # seconds

 

[gar]

111205-2116 EST

mivey:

Are any residential meters using PTs and CTs?

.

 

[texie]

111205-2116 EST

mivey:

Are any residential meters using PTs and CTs?

.

In my neck of the woods we use CT metering on residential above 200 amps. I don't know why any of the utilties around here don't use direct metering in the bigger sizes. Where I am originally from the utility gave up CT metering below 600 amp a long time ago. One of those mysteries of the trade I guess.

 

[mivey]

111205-2116 EST

mivey:

Are any residential meters using PTs and CTs?

.

CTs for sure. I don't recall any PTs at the moment but included it in the formula for completeness.

 

[mivey]

CTs for sure. I don't recall any PTs at the moment but included it in the formula for completeness.

Just remembered a block of retirement homes that were primary metered. Not quite the same as one residence. We had some rather large homes but I just don't remember if any had a 480 volt service.

 

[gar]

111205-2233 EST

texie and mivey:

Thanks. That means I may have to indicate that some residential customers may require the more complex equation. However it would have no effect on the proving whether or not the PFC capacitor being in or out of the circuit had any change effect on the kWh reading. It only amounts to a scale factor.

But now the issue arises, if Cts are used, then to what extent does this degrade the performance of the total meter system at a non-unity power factor? I have not run experiments on CTs, but I would expect tape wound, high permeability cores would be very good. I suspect that TED uses a powdered iron or ferrite split core and this might be part of their problem. It is clearly obvious it is a split core, but I do not know the core material.

.

 

[T.M.Haja Sahib]

That calculation is invalid.

That calculation is based on law of conservation of energy:Input=output+losses.Do you say conservation law of Physics is invalid?:lol:

What mivey wrote is pretty like the example you gave in post #275 to support your contention.

He made misapplication of it.

 

[T.M.Haja Sahib]

111205-2233 EST

texie and mivey:

Thanks. That means I may have to indicate that some residential customers may require the more complex equation. However it would have no effect on the proving whether or not the PFC capacitor being in or out of the circuit had any change effect on the kWh reading. It only amounts to a scale factor.

But now the issue arises, if Cts are used, then to what extent does this degrade the performance of the total meter system at a non-unity power factor? I have not run experiments on CTs, but I would expect tape wound, high permeability cores would be very good. I suspect that TED uses a powdered iron or ferrite split core and this might be part of their problem. It is clearly obvious it is a split core, but I do not know the core material.

gar:
Assuming that you would welcome my input,I want to say for low power factor loads,the inductance of the voltage coil of the watt meter introduces error.So the introduction of PFC capacitor across the load may change its reading. Similarly,CT and PT introduce phase angle errors at any load.So the watt meter reading should be multiplied by a suitable factor to arrive at the real power.More details are available in 'Standard Handbook for Electrical Engineers' by Fink.

 

[Besoeker]

That calculation is invalid.

That calculation is based on law of conservation of energy:Input=output+losses.Do you say conservation law of Physics is invalid?:lol:

It's wrong, a slip up on your part. And nothing whatsoever to do with contravening the laws of physics.
Check again the link I posted in #310.
You then just might see why your calculation is wrong.

He made misapplication of it.

Are you saying that your hoist example is also a missapplication.

 

[Besoeker]

111204-1640 EST

Go to the kWh meter and wait until the black mark reaches something you will use as your reference point. Start the timer. On the eighth revolution when the black mark reaches your reference point stop the timer. Before starting this test record the TED instantaneous reading, and at the test end. Both should be essentially the same, maybe +/- 10 W. If not something changed during the test and you have to start over.

For me, it was simpler. Our kWh meter is digital, the last digit being 0.01kWh.
I could have worked out power from that rather than using the numbers from the oscilloscope but the objective was to determine energy consumption rather than power.

Turn the PFC capacitor off. There will be a change in the TED reading based on my experience.

TED?
I'm sure you have already explained what that is but to save me scrolling back, maybe you could do so again?

 

[USMC1302]

TED = "The Energy Detective", a family of monitoring/measurement devices.

 

[T.M.Haja Sahib]

It's wrong, a slip up on your part. And nothing whatsoever to do with contravening the laws of physics.
Check again the link I posted in #310.
You then just might see why your calculation is wrong.

This is unfair.I made a calculation in post#244.Instead of showing the mistakes in the calculation,you simply referring me to some web page to guess.Is it not possible to point out in your own words how the calculation is wrong?

I think you did not see section 6.14.1 of that web page.Please see it now.

You might reverse your stand.

Are you saying that your hoist example is also a missapplication.

That was put forward just to show you that an induction motor speed can be varied greatly by change in stator voltage only.

 

[gar]

111206-0847 EST

T.M.Haja Sahib:

In response to your post #311. Yes, such factors do introduce some error, the magnitude is what is important. I need quantitative values from real world devices. Broad qualitative comments are not much use.

Standard USA utility kWr meters are quite accurate. Within about +/-0.3% on 10% thru 100% resistive load, and at 100% load current at 50% lagging about +/-0.7%. From p22 of US Department of Interior report "Watt-Hour Meter Maintenance and Testing" that I referenced in an earlier post.

.

 

[T.M.Haja Sahib]

111206-0847 EST

T.M.Haja Sahib:

In response to your post #311. Yes, such factors do introduce some error, the magnitude is what is important. I need quantitative values from real world devices. Broad qualitative comments are not much use.

The book I referred also gives the equation relating the watt meter reading and actual power.My point is you can determine the error of your own watt meter to a reasonable degree of accuracy,given you have necessary equipments to measure the phase angles of CT,PT and that of voltage coil of watt meter.

Standard USA utility kWr meters are quite accurate. Within about +/-0.3% on 10% thru 100% resistive load, and at 100% load current at 50% lagging about +/-0.7%. From p22 of US Department of Interior report "Watt-Hour Meter Maintenance and Testing" that I referenced in an earlier post.

Those utility Kw meters might be low power factor compensated.Does yours also do?

 

[gar]

111205-0919 EST

Besoeker:

The Kill-A-Watt is a meter with a three prong plug and a three prong socket. This meter is for individual plug-in loads. In between is the meter and I suspect a resistor shunt for current measurement.

Maximum current is 15 A and it complains above this. For the EZ model the voltage range is spec'd at 85 to 125 V with 0.1 V resolution. It doesn't complain on any overvoltage to 135 V (I have not taken it higher). It is RMS reading and measures or calculates --- voltage, current, power, VA, frequency, and power factor.

Below about 1 A it resolves 0.1 W in the newer EZ units, doesn't work too well below about 2 W on some units. Above something around 1 A it switches to 1 W resolution.

For about $30 at Home Depot it is a rather good value and useful with knowledge of its limitations.

Handles reactive loads very well as indicated by data from measurements that I have made. A limited useful device for the average consumer. No data collection except by the eyeball.


The TED Systems consist of two different designs, but similar. TED stands for "The Energy Detective". These systems are basically power and energy monitors with data collection possible. Voltage and power are directly measured, albeit, with substantial error on power of non-resistive loads. Time is measured with a Dallas real time clock chip, a rather good clock. Current is measured but only available as VA. On the 5000 series power factor is obtained somehow. For low power factor they apparently fudge VA to equal about 15 possibly to avoid a divide by zero problem.

Spec'd for 100 to 130 V with 0.1 V resolution. Maximum current per current transformer is 200 A. Two current transformers are supplied, one for each phase of the center tapped single phase system. Thus, maximum power is about 200 * 240 = 48,000 watts.

Power Line Communication is used and is a major problem. Software and circuit design are also major problems. There is much to be disliked about the TED systems, but for the money it can provide useful information.

This is also a useful system for the average consumer at $165 to $240 for a minimum system. It is intended for whole house monitoring. But it is worthless for evaluating a PFC capacitor.

Will $250 plus spent of the TED system save the average consumer that amount. Maybe, if it helps them keep things turned off that do not need to be on. In other words change their life style.


webmaster:

What was wrong with the software for this site some 6 months ago? This present software is junk.

.

 

[Besoeker]

This is unfair.I made a calculation in post#244.Instead of showing the mistakes in the calculation,you simply referring me to some web page to guess.

I seem to recall that you have also linked to external internet content. Is that also unfair?
Anyway, the web page was so that you wouldn't have to guess.

Is it not possible to point out in your own words how the calculation is wrong?

Hey ho, here we go.....have you heard of the KISS principle? I'll give it a try....

Your 2200W comes from operation at 60% speed or 0.4pu slip.
At such a high slip, the motor current at rated voltage would be nearly the same as locked rotor current which is typically around 6 pu.
The current is approximately proportional to the supply voltage so would be a bit over 3 pu at 0.55V pu.
So, roughly three times the current of the operation at 10kW.

I think you did not see section 6.14.1 of that web page.Please see it now.
You might reverse your stand.

Nope. Not one iota.
See above in relation to current.
And the curves Fig 6.27 are not entirely credible for an actual real life motor.
This from my post #177:

The curves in Figure 27 would be very unusual for a real life machine. It would require a high resistance rotor or, in the case of a wound rotor machine, external rotor resistance. In either case it would be very inefficient. I suspect that the curves were drawn with exaggerated slip to provide clarity rather than accurately reflect what happens in real life.

That was put forward just to show you that an induction motor speed can be varied greatly by change in stator voltage only.

But it isn't voltage only.
It's a wound rotor machine plus an external rotor resistor.

 

[T.M.Haja Sahib]

I seem to recall that you have also linked to external internet content. Is that also unfair?
Anyway, the web page was so that you wouldn't have to guess.

I tried my level best.I am not able to see the wrong in the calculation.Can't you do any thing about it?

Hey ho, here we go.....have you heard of the KISS principle? I'll give it a try....

Your 2200W comes from operation at 60% speed or 0.4pu slip.
At such a high slip, the motor current at rated voltage would be nearly the same as locked rotor current which is typically around 6 pu.
The current is approximately proportional to the supply voltage so would be a bit over 3 pu at 0.55V pu.
So, roughly three times the current of the operation at 10kW.

If you do not mind,please give a description how an auto-transformer starter works and compare the current it delivers to motor with those you mentioned.You will then see those are starting currents of motor and not its running currents.

But it isn't voltage only.
It's a wound rotor machine plus an external rotor resistor.

That rotor resistor is an optional item.