# Power Factor Correction: Capacitor Bank

• rollingstein
In summary, utilities typically charge a power factor penalty for both lagging and leading power factors. This can cause issues such as voltage oscillations and additional wear and tear on equipment. Utility companies may also use methods such as cumulative KVAR hours or peak VAR demand to bill for power factor differences from unity. However, in some cases, utilities may have agreements with suppliers to provide reactive power, which can improve overall efficiency.
rollingstein
Gold Member
Typically the power factor penalty that a utility charges is only for inductive lagging power factors or leading ones too? i.e. is the penalty for all PF differences from unity or only on the inductive side?

e.g. Say I have a lot of motors which bring the PF down & I add a static capacitor bank to correct the PF & now say, at night, 30% of the motors are idle but the capacitor bank is still fully online and hence the PF is now leading. Is that a problem?

I can imagine it is a problem for capacitor life but are there other concerns? Is overcorrection as bad as having an uncorrected inductive PF?

rollingstein said:
30% of the motors are idle but the capacitor bank is still fully online and hence the PF is now leading. Is that a problem?
Yes it is:

Motors, transformers and other reactive components consume reactive power, while a capacitor is producing reactive power.

The PF = cos φ, where φ is the phaseshift between voltage and current. The optimal correction of the PF is when φ = 0 → cos φ = 1. When producing more reactive power than is consumed, φ will be negative → cos φ < 1.

In practice PF is corrected to about 0.85. If it raises overnight to 1.0 it's not a problem, but if it falls again due to that too much reactive power is produced, some of the capacitors will simply be switched off.

Last edited:
Do utilities penalize for only lagging power factors or for both lead / lag?

rollingstein said:
Do utilities penalize for only lagging power factors or for both lead / lag?
The power factor is a result of leading/lagging phase shift. The power factor itself is just a scalar(?) with no direction.

In practice the power grid will always consume reactive power, due to motors/transformers connected, so that the phase of the current will lag the phase of the voltage. Capacitors will produce reactive power, and will thus in pratice always increase the power factor.

In case of leading/lagging phase shift by turn, a synchronous motor could be used ( and are used ) for compensation: Magnetizing it too much, it will produce reactive power and vica versa, when connected to a grid. ( As a stand alone unit, it'll just produce more or less emf ). So in practice you can see a small building along a powerline, containing a synchronous motor with its shaft sawn of. It's doing no mechanical work, it's just there to produce ( or consume ) reactive power. This positive/negative production can be adjusted completely continuous(?).

Not good models, but anyway:

http://www.engr.sjsu.edu/ustrasil/spring2006/ee175_sp2005/WebProjects_Sp2005/MatLab_Chen_Howard/EE175WP_Chen_files/image002.jpg

http://www.engr.sjsu.edu/ustrasil/spring2006/ee175_sp2005/WebProjects_Sp2005/MatLab_Chen_Howard/EE175WP_Chen_files/image004.jpg

Last edited:
In general the leading PF can cause local system stability issues (Voltage oscillations, flickering lights) , and increases wear and tear on the systems ( switches, contactors and motors) - the condition of excess capacitance (leading pf) causes large inrush currents during switching events.
So while I have not heard of a Utility Penalizing for excessive capacitance (in the billing for example) - they will come out to investigate as this will typically cause problems for the neighboring customers.
The Lagging PF charge is typically handled (I believe - I have been out of this world for 10+ years ) in two ways. Cumulative KVAR Hours - measured just like Watt Hours - but with Vars, VARs (Volt*Amps Reactive) have a polarity ( + leading and - Lagging ) so if the metering is summing + and - the times the PF is leading would be deducting from the times it is lagging - they will not allow this and just like the typical Watt Hour Meter will not "run backwards" the typical VAR Meter will not as well. The other case is Peak VAR demand - in this case the customer is billed a surcharge for the maximum 15 minute period for the month - that seems harsh right? But the whole reason is that the utility has to supply all of the Current to make the VARs - and they are not delivering power. So all of the systems ( generators, cables, transformers etc) need to be sized to accommodate the total current - even though the delivered current is reactive - it does create a voltage drop in the supplying system.
Note : there are cases where the utility has an agreement with a suppler to provide VARs - similar to an agreement to provide energy (watts) as from a customer's solar array. A lot of the energy storage systems have some form of VAR control where the utility can command the system to export VARs ( look capacitive to the grid). By providing local capacitance the overall utility runs more efficiently and the generators can be maximized for Energy -- not just current.

@rollingstein
There is some confusion as to #4 and #5 ( Leading/lagging , positive/negative ). But don't care about that.

It's because I'm from Europe/Denmark where we (could) state PF as:

Cos φ = 0.8reactive by positive reactive power consumption.
Cos φ = 0.8capacitive by negative reactive power consumption.
. . . when we want it to be expressed exact.

It's just a matter of terminology.

Last edited:
I prefer the Reacive as negative, why? the current follows the voltage. Most people have a voltage based mindset. Also "lagging" is similar issue. -- while agree PF is typically defined as apparent power / real power. In most real world applications there is some "sign" ( + or - ) applied. As an engineer this is a matter of conversation, you should learn to "think" about power in a 4 quadrant concept. It took me years to adopt that!

you should learn to "think" about power in a 4 quadrant concept. It took me years to adopt that!
I have learned it. I'm an electric power engineer.

rollingstein said:
Do utilities penalize for only lagging power factors or for both lead / lag?

At least one charges by power factor irrespective of lead/lag.
A good friend bought an old industrial building . His electric bill showed a premium for "low power factor" .
Thinking it was the fluorescent lights he bought some PF correction capacitors. They made PF worse.

Finally he found when the previous owners had removed the big motors and transformers they'd left the PF correction capacitors behind.
Finding and disconnecting most of those solved his trouble.

You'd have thought the utility would be happy to have some free capacitive vars out there.

Hesch
jim hardy said:
You'd have thought the utility would be happy to have some free capacitive vars out there.

That's what I'd have thought too. Shouldn't they be happy about the free capacitive vars? When the entire system is so chock full of inductive loads?

Hesch said:
Cos φ = 0.8reactive by positive reactive power consumption.
Cos φ = 0.8capacitive by negative reactive power consumption.

Yes. Me too. I find the signs confusing. And I don't deal with this stuff on a daily basis. So I always preface my power factor with inductive / capacitive.

rollingstein said:
Yes. Me too. I find the signs confusing. And I don't deal with this stuff on a daily basis. So I always preface my power factor with inductive / capacitive.

Where i worked we called them "Leading or Lagging" megavars,
leading(capacitive) we called negative and lagging(inductive) positive. edit i'd flipped those at first - mild dyslexia jh
I remember re-arranging the megavar meter and the voltage regulator null meter on our control board
so that when operator raised excitation , everything moved up
field amps , generator volts, megaver meter, and null meter should all move same direction in response to an action.
I had to invert the megavar meter. it had leading on the top half so we flipped it.

That made it more intuitive which way to go during a grid upset.

If i can find an old photo will post it. Maybe Dan has one.

Last edited:
jim hardy said:
Where i worked we called them "Leading or Lagging" megavars,

Leading and lagging is very clear too. It's when someone uses positive or negative that I get confused.

The Lagging PF charge is typically handled (I believe - I have been out of this world for 10+ years ) in two ways. Cumulative KVAR Hours - measured just like Watt Hours - but with Vars, VARs (Volt*Amps Reactive) have a polarity ( + leading and - Lagging ) so if the metering is summing + and - the times the PF is leading would be deducting from the times it is lagging - they will not allow this and just like the typical Watt Hour Meter will not "run backwards" the typical VAR Meter will not as well.

How does a VAR meter work internally? I mean, Wattmeter are common & I sort of know how they are made internally. They have been around since pre-electronics era.

But in a non-electronic era was there a trick to measuring VARs? How did one design a purely electro-mechanical device to measure only the reactive component?

rollingstein said:
But in a non-electronic era was there a trick to measuring VARs? How did one design a purely electro-mechanical device to measure only the reactive component?

Our plant was a 1960's design.
The megawatts were measured by a "Lincoln Thermal Converter" , a quite accurate device from the 1930's
that uses voltage and current from the metering transformers to heat a thermopile in a box with precise thermal characteristics.
That thermopile's millivolt output is proportional to power,
and its thermal mass gives it a few second time constant low pass response..(old adage "Encourage Mother Nature to help your design work well")
Realizing that any device that multiplies VIcosθ will report watts
if we fool such a device by handing it a current signal 90 degrees out of phase
Our megavar meter was an inexpensive little Hall effect analog multiplier.
Less accurate than the thermal converter but faster..
It received phase AC voltage and phase B amps
which if you draw your phasor diagram, are 90 degrees out of phase.

Megawatts are the holy grail in a power plant so megavars were not measured with equal precision.
Megawatts were the prominent display in the control room, displayed on a huge recorder high on the panel.
Megavars only got a solitary meter near the generator voltage and ammeters.

hope that helps.

old jim

Last edited:
rollingstein
jim hardy said:
At least one charges by power factor irrespective of lead/lag.

I checked the fine print of my utility. It too penalizes irrespective of lead / lag. Sad.

jim hardy said:
Finally he found when the previous owners had removed the big motors and transformers they'd left the PF correction capacitors behind.

I've an interesting situation. My utility gives a 5% discount on the bill if power factor is maintained 0.985 or above.

Say I add a capacitor bank. How much is the parasitic resistance of a correction capacitor bank? I wonder if those kWs will eat up a large chunk of this 5% incentive.

Commercially / Practically you cannot get a purely capacitive capacitor right?

rollingstein said:
I've an interesting situation. My utility gives a 5% discount on the bill if power factor is maintained 0.985 or above.

Say I add a capacitor bank. How much is the parasitic resistance of a correction capacitor bank? I wonder if those kWs will eat up a large chunk of this 5% incentive.

Commercially / Practically you cannot get a purely capacitive capacitor right?

I doubt that the resistance of the capacitors will eat up much of that discount. If they did, two things would result. They would get very warm and perhaps need some form of external cooling. Nobody would make them for this application because there would be no market.

BTW what were you thinking using any form of the word "pure" in an engineering forum?

anorlunda said:
I doubt that the resistance of the capacitors will eat up much of that discount. If they did, two things would result. They would get very warm and perhaps need some form of external cooling. Nobody would make them for this application because there would be no market.

Cooling fans seem common. e.g. See here:

https://library.e.abb.com/public/a8eacbc0eb7634be852578df0067d127/1SXP981002D0202.pdf

OTOH, you are right about low losses. ABB quotes 6 Watts per kVAR. That seems approx. 0.5% power loss.

rollingstein said:
ABB quotes 6 Watts per kVAR. That seems approx. 0.5% power loss.

Capacitor total losses are less than 0.5 watts per kvar. Auto-
bank total losses (without reactors), including accessories such
as Power Factor Controller and contactors are less than 1.5
watts per kvar.

So it's not due to losses in the capacitors, that cooling is needed.

The significant losses, due to reactive power, are in the grid. That's why someone will give you 5% discount.

Of financially reasons, the PF is in practice only compensated to about 0.85. Capacitor banks are expensive and are worn due to short circuits inside the capacitors. They are made self-healing, but they loose capacitance every time such a short circuit/healing takes place. So within some amount of years you will have to replace them. Your 5% discount must pay this replacement.

Hesch said:
Of financially reasons, the PF is in practice only compensated to about 0.85. Capacitor banks are expensive and are worn due to short circuits inside the capacitors.

Interesting. Thanks. If so, I need to recalculate our payback. Because I mostly see our corrected PF in the neighborhood of 0.98.

Maybe we are over-correcting? We do have to keep replacing capacitors every so often but I'm not sure someone has does a cost benefit evaluation. Our load is predominantly pumps, compressors, agitators & general lighting loads.

0.85 sounds too low to me for a corrected PF. In fact, at 0.85 PF my utility tacks on a 6% surcharge over the base bill as a penalty.

rollingstein said:
Maybe we are over-correcting? We do have to keep replacing capacitors every so ofte

Can you measure current through your correcting capacitors with maybe a clamp-around meter ?

Remember the current gain of a parallel resonant circuit...

rollingstein
rollingstein said:
at 0.85 PF my utility tacks on a 6% surcharge over the base bill as a penalty.
I don't how you are doing "over there", but here (Denmark) electricians subscribes these magazines wherein a recommended PF is calculated every month on basis of coal prices and whatever.

rollingstein
Ok, on deeper thought here's what I suspect is happening in our system. I wanted to get opinions on whether this scenario seems plausible.

Our monthly power factors are pretty good. Averaged above 0.98 and hence we get a good reward incentive from our utility. So far I don't have data on the hour by hour PF trend only the average calculated by the utility in making our monthly bill. This PF I suspect is only a ratio of the cumulative reading on the kWh meter divided by the kVAh meter. So far so good.

Now, we do get assessed another penalty based on how high our peak kVA load is (with reference to some previously agreed kVA load benchmark; I'm not sure how that was computed).

Unfortunately we have had a high penalty on that count: Our peak kVA has been too high. Now one figure I don't know right now is when this peak kVA happened and if it coincided with a peak kW too or not.

What I do know is that we have a primitive manual bank of capacitors for PF correction (of our mainly inductive load of motors and florescents). The operators manually take as many online / offline as needed to bring the PF close to unity. What I am suspecting is that on some occasions, e.g. an impending shutdown a lot of plant loads get shed. But the capacitor bank stays online.

Would that cause a very high peak kVA demand? If the kW component of the load fell and the inductive loads were shed too but the compensating capacitive kVARs stayed online?

Does this scenario make any sense? If so then I have to think of empirically verifying this. But if my supposition itself is flawed then no point digging further.

Hesch said:
I don't how you are doing "over there", but here (Denmark) electricians subscribes these magazines wherein a recommended PF is calculated every month on basis of coal prices and whatever.

Ha! We have a static schedule for the penalty that gets periodically revised. Utilities are a dinosaur. The last time the PF-reward / penalty schedule was revised was 2012 I think. :)

rollingstein said:
Would that cause a very high peak kVA demand? If the kW component of the load fell and the inductive loads were shed too but the compensating capacitive kVARs stayed online?
Well, yes it would result in a high peak kVAr, but the PF would get a lot worse. But maybe you could think of this idea:
Hesch said:
a synchronous motor with its shaft sawn of. It's doing no mechanical work, it's just there to produce ( or consume ) reactive power. This positive/negative production can be adjusted completely continuous(?).
The motor could at the same time be used to drive your air condition, whatever. Some controller could adjust the magnitizing current, so that the PF would be exactly 1 at any time. The bearings would need some oil once a month.

rollingstein
Hesch said:
Well, yes it would result in a high peak kVAr, but the PF would get a lot worse.

Well, the instantaneous PF would, you are right. But since the billing is done on an average basis for the whole month I don't think the PF will be affected by much.

Basically, what I'm saying is that the peak demand charges are computed on any instantaneous kVA whereas the PF penalty is calculated on an average basis. So if over-correcting even for a few minutes, your PF charge will be unaffected but OTOH your peak demand charge will reflect your error.

Am I getting this right?

Hesch said:
The motor could at the same time be used to drive your air condition

Can it be used in this way? I thought the motor used for PF correction is a no-load motor?

rollingstein said:
Can it be used in this way? I thought the motor used for PF correction is a no-load motor?

Sure, any synchronous motor can be a motor or generator or synchronous condenser (0 power VAR only). They respond better than operators, but their response is not instantaneous. You need some clarification about the time period defined for "instantaneous" PF. It can't be zero.

rollingstein
rollingstein said:

Am I getting this right?
Yes.
rollingstein said:
Can it be used in this way? I thought the motor used for PF correction is a no-load motor?
It can be mechanically loaded at the same time. The generator at your utility is (negative) mechanically loaded while producing kVAr's.

rollingstein
Hesch said:
It can be mechanically loaded at the same time. The generator at your utility is (negative) mechanically loaded while producing kVAr's.

Thanks! I found a great article describing the use of a synchronous motor in a process plant here in case it helps anyone else:

https://www.site.uottawa.ca/~rhabash/project-synversusinduction.pdf

Hesch said:
It can be mechanically loaded at the same time. The generator at your utility is (negative) mechanically loaded while producing kVAr's.

Out of curiosity, are there practical systems that have a natural leading power factor & hence cannot be corrected by a capacitor bank?

rollingstein said:
Out of curiosity, are there practical systems that have a natural leading power factor & hence cannot be corrected by a capacitor bank?

Power transmission lines, especially underground power cables. We correct those with shunt reactors.

## 1. What is power factor correction?

Power factor correction is the process of improving the power factor of an electrical system by adding capacitors to the circuit. This helps to reduce the amount of reactive power and improve the overall efficiency of the system.

## 2. Why is power factor correction important?

Power factor correction is important because it helps to reduce energy losses, improve voltage levels, and increase the capacity of the electrical system. This can lead to cost savings and improved performance of electrical equipment.

## 3. How does a capacitor bank improve power factor?

A capacitor bank is a group of capacitors that are connected in parallel and installed in an electrical system. These capacitors help to absorb reactive power, which in turn reduces the amount of reactive power flowing through the system and improves the power factor.

## 4. How do you determine the size of a capacitor bank for power factor correction?

The size of a capacitor bank for power factor correction is determined by calculating the reactive power of the system and the desired power factor. This calculation takes into account the type of load, voltage level, and power factor correction goal.

## 5. What are the benefits of power factor correction?

The benefits of power factor correction include reduced energy losses, improved voltage levels, increased capacity of the electrical system, and cost savings. It can also help to improve the lifespan and performance of electrical equipment.

• Electrical Engineering
Replies
2
Views
764
• Electrical Engineering
Replies
9
Views
7K
• Electrical Engineering
Replies
3
Views
1K
• Electrical Engineering
Replies
8
Views
1K
• Electrical Engineering
Replies
8
Views
1K
• Electrical Engineering
Replies
3
Views
2K
• Electrical Engineering
Replies
5
Views
1K
• Electrical Engineering
Replies
2
Views
1K
• Electrical Engineering
Replies
2
Views
1K
• Electrical Engineering
Replies
55
Views
7K