Positive and negative VARs and Solar inverter Grid connect schemas

[coop3339]

Im looking for some help understanding positive and negative VARs. I'm also dyslexic so this greatly complicates the matter.

My original understanding was that adding positive VARs was the same as adding capacitance to the circuit and adding negative VARs was the same as adding inductance. I think this was due to my understanding that added capacitance to an AC circuit will cause a leading power factor and adding inductance will cause a lagging power factor.

Is is actually correct that adding capacitance to an AC circuit is adding negative VARs and adding inductance to an AC circuit is adding positive VARs?

If so a 'VAR' (Volt Amps Reactive) could be more accurately be thought of as a 'IRVA' (Inductive Reactive Volt Amps)? When thought of this way, adding capacitance would be the same as adding negative IRVAs. I don't expect that the naming conventions will be or should be changed to match my way of thinking but I thought this could help someone elses understanding of the concept.

Assuming my understanding of the above is correct, adding negative VARs (adding capacitance) would usually have the effect of raising voltage levels due to most grids having some degree of a lagging power factor. By adding the capacitance, the lagging power factor is brought closer to unity. This in turn will raise the voltage levels because of the overall impedance will be lowered and voltage drops will be less in the transmission lines. To bring it all together, in most cases, adding negative VARs is likely to increase voltage levels.

If my understanding of the above is true, why would a utility request that voltage triggered reactive power correction would supply negative VARs when a certain voltage level is exceeded?

To state this another way, they are requesting that when the voltage level exceeds 109% of nominal, -44% VAR is to be injected.

My confusion is, would this not only add to the overvoltage condition?

Wouldn't it make more sense to inject 44% VARs to lower the voltage? I understand this would likely make the power factor worse but it would also work to lower the voltage.

Any feedback on this would be much appreciated.

 

[coop3339]

I just had another thought on this. Maybe by having the inverters move the power factor closer to unity, the overall grid impedance encountered by the inverter will be reduced. This could make it easier for the inverter to push power into the grid and lower the overall voltage required to do so.

The reason why the voltage is high in the first place is likely due to high grid impedance. Looking at it this way, i guess it could make sense to add capacitive power to lower the overvoltage condition.

 

[David Castor]

Inductors consume vars and capacitors create vars. I'd forget the positive and negative concept. By convention, in power systems, var flow is always referring to inductive vars, so these vars are consumed in inductive loads such as motors.

At any particular point in the system, if the real power (kW) is flowing in one direction and the vars (kVAR) are flowing in the OPPOSITE direction the power factor is LEADING. If kW and kVARS are flowing in the SAME direction, the power factor is LAGGING. The "leading" and "lagging" terminology refers to the current phase angle in relation to the voltage phase angle. So a generator that is producing kW and kVars is in a lagging condition. A motor that is consuming kW and kVars is also lagging.

In general, if you add capacitors, the system voltage at that location will increase (up to a point).

 

[coop3339]

Inductors consume vars and capacitors create vars. I'd forget the positive and negative concept. By convention, in power systems, var flow is always referring to inductive vars, so these vars are consumed in inductive loads such as motors.

At any particular point in the system, if the real power (kW) is flowing in one direction and the vars (kVAR) are flowing in the OPPOSITE direction the power factor is LEADING. If kW and kVARS are flowing in the SAME direction, the power factor is LAGGING. The "leading" and "lagging" terminology refers to the current phase angle in relation to the voltage phase angle. So a generator that is producing kW and kVars is in a lagging condition. A motor that is consuming kW and kVars is also lagging.

In general, if you add capacitors, the system voltage at that location will increase (up to a point).

Thanks for the reply David. Unfortunately, the parameters in the inverter are set as a percentage of total power delivered and can be either a -N% VAR, apparently indicating capacitive power injection or +N% VAR, apparently indicating inductive power injection.

I've never heard power factor described in the way you have in the second paragraph. I'm used to seeing power factor described in relation to the voltage, where is the current peak occurs before the voltage peak it is a leading power factor and if it occurs after the voltage peak, it is a lagging power factor or in relation to phase angle as you state later.

I would be interested in learning to look at it in this way if you or someone could elaborate. This may clear up a term I've heard power engineers use, absorption, used instead of negative VAR injection, I think.

Still interested in why the utility wants capacitive power injected if the voltage passes a overvoltage trigger threshold. The only thing I can think of is this will make it easier for the inverter to push power to the grid. Because of this, the same amount of power can be delivered but without pushing the voltage as high.

I guess it would be the same apparent power delivered but less real power due to the percentage that will be capacitive.

 

[David Castor]

I would review the inverter documentation to see what they are using for their sign (+/-) convention. To me a minus sign would indicate absorption of vars (inductive) and positive would imply a source of vars (capacitive). But there is often a lot of confusion in this terminology, especially if the inverter is coming from overseas.

We're both saying the same thing regarding power factor - just using different terms. But often a generator is described as being LEADING if it is supply vars to the system -- this is incorrect.

 

[jaggedben]

Is is actually correct that adding capacitance to an AC circuit is adding negative VARs and adding inductance to an AC circuit is adding positive VARs?

As I learned relatively recently the meaning of positive and negative power factor is a convention and there are two different ones.

https://www.productinfo.schneider-electric.com/pm5100/pm5100-user-manual/PM5100%20User%20Manual/English/BM_PM5100UserManual_EAV15105_0000425845.ditamap/$/C_Measurements_PowerFactor_SignConvention_0000034208

 

[jaggedben]

To state this another way, they are requesting that when the voltage level exceeds 109% of nominal, -44% VAR is to be injected.

...

This sounds generally correct to me, although they are definitely stating it from their point of view. When the voltage is too high they want the inverter to add inductance to the system because this will lower the voltage back down (or keep it in check at least).

I believe that this translates to the inverter outputting a lagging power factor. But only because I remember reading that in a description of the UL1741 volt-var function. (I'm not an engineer and can't show you the math. ) For low voltage the inverter is supposed to have a leading power factor. IIRC

 

[jaggedben]

Maybe by having the inverters move the power factor closer to unity, the overall grid impedance encountered by the inverter will be reduced. This could make it easier for the inverter to push power into the grid and lower the overall voltage required to do so.

This sounds backwards to me. The inverter is trying to push current into the grid, opposite the direction that the grid voltage wants to push current. This is hardest for the inverter to do at unity power factor. If the inverter shifts current wave then the peak current will occur somewhat off the peak grid voltage that the inverter is pushing against, so the inverter should have an easier time. And not raise the voltage as much at the peaks.

The reason why the voltage is high in the first place is likely due to high grid impedance. Looking at it this way, i guess it could make sense to add capacitive power to lower the overvoltage condition.

Also backwards? I'm not at all sure how exactly to explain though. It might be a matter of convention or point of view (i.e. grid's POV or inverter's POV).

 

[David Castor]

You have to separate real power from reactive power. In an ac system, the real power flow is determined by the voltage phase angle between your source and the grid. To export power, the power angle is increased and this increases the kW flow into the grid. If you reduce the angle enough, power will flow into your "generator". In a rotating machine, this is generator motoring. Control of kW flow in an ac system is almost entirely independent of the voltage. The voltages control the reactive power (vars).

 

[GoldDigger]

This sounds backwards to me. The inverter is trying to push current into the grid, opposite the direction that the grid voltage wants to push current. This is hardest for the inverter to do at unity power factor. If the inverter shifts current wave then the peak current will occur somewhat off the peak grid voltage that the inverter is pushing against, so the inverter should have an easier time. And not raise the voltage as much at the peaks.



Also backwards? I'm not at all sure how exactly to explain though. It might be a matter of convention or point of view (i.e. grid's POV or inverter's POV).

But the job of the inverter is not just to push current into the system, it is to deliver meterable power into the grid. For this purpose, only real VA counts. But to make the job of the utility easier they may want some or all of the connected inverters capable of performing network VAR correction which would otherwise require them to install capacitor banks.
Ideally POCOs would like to be able to remotely control the power factor of individual inverters to help stabilize the grid.
I believe that because, in part, of ferro-resonance, excessive capacitive VARs on the grid can cause undesirable grid voltage increases. So a fixed percentage VAR might not be a good thing.

 

[winnie]

But the job of the inverter is not just to push current into the system, it is to deliver meterable power into the grid. For this purpose, only real VA counts. But to make the job of the utility easier they may want some or all of the connected inverters....

If the appropriate billing and control structures are in place, then the inverters can sell VARs in the same way they sell real power.

Not being part of the industry I only know this is possible in principle, not if it is ever done in practice.

 

[GoldDigger]

If the appropriate billing and control structures are in place, then the inverters can sell VARs in the same way they sell real power.

Not being part of the industry I only know this is possible in principle, not if it is ever done in practice.

I have not ever heard of this being done, although as you say it is theoretically possible. It would require a meter specifically configured to be able to measure VAR. And the rate per KVAhour would be significantly less than the rate per Kwh.
It is more likely that an inverter would be configured to offset local motor load VAR to avoid possible POCO penalties.

 

[coop3339]

But the job of the inverter is not just to push current into the system, it is to deliver meterable power into the grid. For this purpose, only real VA counts. But to make the job of the utility easier they may want some or all of the connected inverters capable of performing network VAR correction which would otherwise require them to install capacitor banks.
Ideally POCOs would like to be able to remotely control the power factor of individual inverters to help stabilize the grid.
I believe that because, in part, of ferro-resonance, excessive capacitive VARs on the grid can cause undesirable grid voltage increases. So a fixed percentage VAR might not be a good thing.

In this case the settings in question are the volt/VAR settings. My confusion is whether negative VARs are capacitive or inductive. There is some conflicting information on the naming but it appears most sources indicate that negative VARs are capacitive. Like Jag, I thought the opposite made more sense. The other confusion is, why a utility would want the negative VARs injected when there is a overvoltage condition developing?

My thought is that by bringing the PF closer to unity, the inverter would not have to raise the voltage as high to push power to the grid. This would be due to lower 'apparent impedance' because of the better PF. The same mechanism that can raise the network voltage, the reduction of I2R losses, will also make the impedance seen by the inverter lower. I don't know if this is what the reason is but why else would they want to add capacitance to a system that is already approaching overvoltage?

Also keep in mind that the most common reason for the voltage to rise out of limits in solar feeders is due to grid impedance. The real part of the system impedance can not be easily improved without changing hardware, like conductors or transformers but the reactive part of the impendence can be lowered by correcting power factor. This is why I believe they are adding the negative VARs when the voltage rises above a trigger point. There are also triggers based on frequency but in this case these are not being used.

As for selling VARs, I believe this is possible but to my knowledge its not common with solar inverters. However I would think battery storage systems probably do. They are commonly used for frequency support also.

 

[jaggedben]

But the job of the inverter is not just to push current into the system, it is to deliver meterable power into the grid. For this purpose, only real VA counts. But to make the job of the utility easier they may want some or all of the connected inverters capable of performing network VAR correction which would otherwise require them to install capacitor banks.
Ideally POCOs would like to be able to remotely control the power factor of individual inverters to help stabilize the grid.
I believe that because, in part, of ferro-resonance, excessive capacitive VARs on the grid can cause undesirable grid voltage increases. So a fixed percentage VAR might not be a good thing.

You're not disagreeing with my statement that the inverter should have 'an easier time' pushing the current into the grid at a leading or lagging power factor. Or are you? I think you are raising a side issue that isn't really a problem for our OP. The dynamic volt-var function will allow the inverter to have more uptime when grid voltage is too high and almost out of range. The point of the volt-var function is to keep the inverter from pushing the voltage out of range to where it would be forced to shut down. The inverter therefore gets more up time and delivers more meterable power to the grid.

 

[wwhitney]

You're not disagreeing with my statement that the inverter should have 'an easier time' pushing the current into the grid at a leading or lagging power factor.

I'm not clear on that means, what is 'an easier time'?

The dynamic volt-var function will allow the inverter to have more uptime when grid voltage is too high and almost out of range. The point of the volt-var function is to keep the inverter from pushing the voltage out of range to where it would be forced to shut down. The inverter therefore gets more up time and delivers more meterable power to the grid.

Seems to me that would only work when the feeder between the inverter and the grid consists of wires that are below the hundred kcmil range. Which takes advantage of that fact that below that size, the conductor resistance per unit length is higher than the conductor reactance per unit length. A recent discussion of this phenomenon starts here: https://forums.mikeholt.com/threads/voltage-drop.2579148/page-2#post-2898122

Cheers, Wayne

 

[David Castor]

All power generation revenue is based on selling real power - kilowatts. Vars are simply a measure of the relative power factors within the system and do not require actual energy to create or absorb, other than the losses created by the increased current. If I am running a merchant power plant, such as a PV facility, I want my inverters to produce ONLY kW (unity power factor) because that is what my revenue is based on. I don't want to use up some of the kVA capacity of the inverters producing vars. But the utility needs vars for voltage support, so it is always a negotiated peace and the utility will often require some type of voltage regulation to help maintain the voltage. For large generation facilities, transmission grids often require power system stabilizers and other voltage support equipment be installed.

 

[jaggedben]

In this case the settings in question are the volt/VAR settings. My confusion is whether negative VARs are capacitive or inductive. There is some conflicting information on the naming but it appears most sources indicate that negative VARs are capacitive. Like Jag, I thought the opposite made more sense. The other confusion is, why a utility would want the negative VARs injected when there is a overvoltage condition developing?

It's a direction convention issue? With IEEE convention the vars are negative if the power flow from the solar inverter is considered negative? (That's the utility's point of view.) But the vars are positive if the solar inverter power flow is considered positive? (That's your point of view.) In both cases they are lagging?

Sorry if I'm wrong and just being confusing. But I bet it would help to have the utility clarify their conventions.

 

[jaggedben]

I'm not clear on that means, what is 'an easier time'?

Let's say it's easier if the inverter doesn't have to raise the RMS voltage as high to push out the same current.

Seems to me that would only work when the feeder between the inverter and the grid consists of wires that are below the hundred kcmil range. Which takes advantage of that fact that below that size, the conductor resistance per unit length is higher than the conductor reactance per unit length. A recent discussion of this phenomenon starts here: https://forums.mikeholt.com/threads/voltage-drop.2579148/page-2#post-2898122

I wasn't aware of that. Thanks. Yes, my comments above weren't considering that.

 

[ggunn]

Let's say it's easier if the inverter doesn't have to raise the RMS voltage as high to push out the same current.

? A grid tied PV inverter is a current source, not a voltage source. It doesn't have to raise the voltage at all. The voltage is higher at its terminals than at the service disconnect because of voltage drop in the conductors, but if the conductors were superconductors with zero resistance (no voltage drop) the voltages would be the same and the inverter would still work just fine.

 

[jaggedben]

Our conductors aren't superconductors so I'm not sure what the point of that hypothetical is. My understanding is that with resistive conductors physics says that the voltage has to be higher at the source than at the load. Therefore is a solar inverter is going to cause a load to draw power from the inverter instead of the grid, it has to raise the voltage at the node where all three come together higher than it would be if the load drew power from the grid. Right?