It hasn’t been that long since humans figured out how to create power grids that integrated multiple generators and consumers. Ever since AC won the battle of the currents, grid operators have had to deal with the issues that come with using AC instead of the far less complex DC. Instead of simply targeting a constant voltage, generators have to synchronize with the frequency of the alternating current as it cycles between positive and negative current many times per second.
Complicating matters further, the transmission lines between generators and consumers, along with any kind of transmission equipment on the lines, add their own inductive, capacitive, and resistive properties to the system before the effects of consumers are even tallied up. The result of this are phase shifts between voltage and current that have to be managed by controlling the reactive power, lest frequency oscillations and voltage swings result in a complete grid blackout.
Flowing Backwards
We tend to think of the power in our homes as something that comes out of the outlet before going into the device that’s being powered. While for DC applications this is essentially true – aside from fights over which way DC current flows – for AC applications the answer is pretty much a “It’s complicated”. After all, the primary reason why we use AC transmission is because transformers make transforming between AC voltages easy, not because an AC grid is easier to manage.
What exactly happens between an AC generator and an AC load depends on the characteristics of the load. A major part of these characteristics is covered by its power factor (PF), which describes the effect of the load on the AC phase. If the PF is 1, the load is purely resistive with no phase shift. If the PF is 0, it’s a purely reactive load and no net current flows. Most AC-powered devices have a power factor that’s somewhere between 0.5 to 0.99, meaning that they appear to be a mixed reactive and resistive load.
PF can be understood in terms of the two components that define AC power, being:
- Apparent Power (S, in volt-amperes or VA) and
- Real Power (P, in watts).
The PF is defined as the ratio of P to S (i.e. `PF = P / S). Reactive Power (Q, in var) is easily visualized as the angle theta (Θ) between P and S if we put them as respectively the leg and hypotenuse of a right triangle. Here Θ is the phase shift by which the current waveform lags the voltage. We can observe that as the phase shift increases, the apparent power increases along with reactive power. Rather than being consumed by the load, reactive power flows back to the generator, which hints at why it’s such a problematic phenomenon for grid-management.
From the above we can deduce that the PF is 1.0 if S and P are the same magnitude. Although P = I × V gets us the real power in watts, it is the apparent power that is being supplied by the generators on the grid, meaning that reactive power is effectively ‘wasted’ power. How concerning this is to you as a consumer mostly depends on whether you are being billed for watts or VAs consumed, but from a grid perspective this is the motivation behind power factor correction (PFC).
This is where capacitors are useful, as they can correct the low PF on inductive loads like electric motors, and vice versa with inductance on capacitive loads. As a rule of thumb, capacitors create reactive power, while inductors consume reactive power, meaning that for PFC the right capacitance or inductance has to be added to get the PF as close to 1.0 as possible. Since an inductor absorbs the excess (reactive) power and a capacitor supplies reactive power, if both are balanced 1:1, the PF would be 1.0.
In the case of modern switching-mode power supplies, automatic power factor correction (APFC) is applied, which switch in capacitance as needed by the current load. This is, in miniature, pretty much what the full-scale grid does throughout the network.
Traditional Grids
Based on this essential knowledge, local electrical networks were expanded from a few streets to entire cities. From there it was only a matter of time before transmission lines turned many into few, with soon transmission networks spanning entire continents. Even so, the basic principles remain the same, and thus the methods available to manage a power grid.
Spinning generators provide the AC power, along with either the creation or absorption of reactive power on account of being inductors with their large wound coils, depending on their excitation level. Since transformers are passive devices, they will always absorb reactive power, while both overhead and underground transmission lines start off providing reactive power, overhead lines start absorbing reactive power if overloaded.
In order to keep reactive power in the grid to a healthy minimum, capacitive and inductive loads are switched in or out at locations like transmission lines and switchyards. The inductive loads often taken the form of shunt reactors – basically single winding transformers – and shunt capacitors, along with active devices like synchronous condensers that are effectively simplified synchronous generators. In locations like substations the use of tap changers enables fine-grained voltage control to ease the load on nearby transmission lines. Meanwhile the synchronous generators at thermal plants can be kept idle and online to provide significant reactive power absorption capacity when not used to actively generate power.
Regardless of the exact technologies employed, these traditional grids are characterized by significant amounts of reactive power creation and absorption capacity. As loads join or leave the grid every time that consumer devices are turned off and on, the grid manager (transmission system operator, or TSO) adjusts the state of these control methods. This keeps the grid frequency and voltage within their respective narrowly defined windows.
Variable Generators
Over the past few years, most newly added generating capacity has come in the form of weather-dependent variable generators that use grid-following converters. These devices take the DC power from generally PV solar and wind turbine farms and convert them into AC. They use a phase-locked loop (PLL) to synchronize with the grid frequency, to match this AC frequency and the current voltage.
Unfortunately, these devices do not have the ability to absorb or generate reactive power, and instead blindly follow the current grid frequency and voltage, even if said grid was going through reactive power-induced oscillations. Thus instead of damping these oscillations and any voltage swings, these converters serve to amplify these issues. During the 2025 Iberian Peninsula blackout, this was identified as one of the primary causes by the Spanish TSO.
Ultimately AC power grids depend on solid reactive power management, which is why the European group of TSOs (ENTSO-E) already recommended in 2020 that grid-following converters should get replaced with grid-forming converters. These feature the ability absorb and generate reactive power through the addition of features like energy storage and are overall significantly more useful and robust when it comes to AC grid management.
Although AC doesn’t rule the roost any more in transmission networks, with high-voltage DC now the more economical option for long distances, the overwhelming part of today’s power grids still use AC. This means that reactive power management will remain one of the most essential parts of keeping power grids stable and people happy, until the day comes when we will all be switching back to DC grids, year after the switch to AC was finally completed back in 2007.
The power factor can be zero when the load is not consuming any real power i(t) x v(t) = 0 and that doesn’t mean the current has to be zero – just that it always applies when no voltage across the device exists, which could be a dead short.
Or, if i(t) is always zero then that’s a cut wire – open circuit.
In both cases the load seen by the power system is actually the transmission line itself.
If the pf is zero, then the current can not be zero, nor the potential.
If pf is zero, there is no net power delivered. Energy is sloshing back and forth from source to load, the only dissipation being delivery losses (I squared R loss, transformer and parasitic losses, etc)
So, renewables (solar, wind) are not so good in providing reactive power for stability, even if that’s mostly driven by cost. In the absence of stability provided by spinning reserves (gas, generally, but also hydro), other mechanisms are available. Inverters and batteries are good for hour-scale corrections. But plain old flywheels are a viable means of grid support too. https://beaconpower.com/technology/ has been doing it for a couple of decades.
Homes even as backup.
I hope there are some ways to “multi-purpose” wind turbines to
A) use the existing generator with a flywheel during off-time (no wind / already too much energy in the grid)
B) and/or maybe even use the generator(=motor) to lift a lead weight inside the high towers and use it as energy storage
Could even be possible to do both at the same time: Use generator + flywheel as grid support and simultaneously use the wind-power to lift the weight.
I assume neither option is really feasible – both from a physical/engineering and the economical one – but one can dream… :-)
Another interesting driver of the adoption of HVDC for long transmission lines is that long lines start approaching a quarter wavelength at 60 Hz, and start acting like quarter-wave impedance transformers themselves: Load current changes manifest as voltage excursions at the generator end, making voltage control difficult and causing generator and line trips. HVDC makes 1000 km lines possible by trading those problems for a different set of problems.
“trading those problems for a different set of problems” – so many engineering decisions are exactly this.
That picture up top, it just screams “robot overlord army.”
Easily disabled (shorted) by drone-dropped piece of bicycle brake cable. Don’t ask me how I know (; (; (;
I can’t imagine a puny little bicycle cable is going to be much challenge to a kiloamp supply like that. Did it really trip the line, or just make a nice spectacular but inconsequential little pzzzzt?
Question: do HVDC lines have electrolysis corrosion problems? I would think that in the 500kV realm even air would cause erosion of the anode cable.
HVDC is usually made with two wires. One with a positive voltage, and the other with a negative voltage. Only in emergencies and during maintenance the earth is used as GND. There probably are limitations on the impedance of the GND electrode, but I never looked into details. For the majority of the time, the problem is simply avoided.
The Wikipedia article I linked to below does node that electrode design for the anode and cathode are different, but again, I don’t know further details.
My Grandfather was a Catholic Protection Pioneer for Union Oil Company in California. California did a HVDC Transmission Line from Northern to Southern Ca sometime in the Late 60’s Early 70’s. When they started up the HVDC Line his Cathodic Testing to verify Corrosion Stability went haywire.
Ends up the HVDC line was only ONE wire, with a Earth Ground Return. The Return Current of the HVDC Line was going back all the North South Pipe Lines in California, screwing up the Cathodic Protection..
He retired shortly after that and I never got a resolution to that Problem..
Cap
I don’t agree with the introduction here. Back then AC was chosen because it was simpler. Not because it was more complicated. For example, distribution of DC power is complicated, while for AC you only need a bunch of wire and a stack of steel plates on both ends. And by switching taps at a distribution transformer, you can stabilize the voltage of local sections.
DC distribution is a real thing, but because of the extra complexity, it is only used when it has clear benefits. For more info:
https://en.wikipedia.org/wiki/High-voltage_direct_current
Synchronizing AC generators is easy. The simplest method is to put two light bulbs in between the local generator and the net. From the blink rate of the lamps, you can see the frequency and phase difference, and when the lights are off, then they are in sync and you can short circuit the light bulbs. And from that moment on, the “generator” either delivers or consumes energy, depending on whether you push it, brake it or freewheel it.
https://www.youtube.com/watch?v=xGQxSJmadm0
Here is a real world example of reactive power. The design was the exit signs you have seen in all the commercial buildings. This design used an elector-luminiescent sheet that lit up behind the word EXIT. The task was to measure the power consumed. All power meters were reading zero watts. Upon closer look the elector-luminescent sheet was a BIG capacitor and the current consumed was 90 degrees out of phase with the voltage. The power meters read zero watts as they were measuring volts times amps. I think our conclusion was that the device consumed zero power since power meters at the time could not measure what was going on.
Real power meters read real power. They don’t care about power factor.
Electroluminescent signs draw very little real power: much less than a watt each.
The reason a building-scale power meter registers exit lighting as “zero” power is because it is too small to measure, not because it is low power factor.
the aticle is just plain wrong on this part:
“Unfortunately, these devices do not have the ability to absorb or generate reactive power, and instead blindly follow the current grid frequency and voltage, even if said grid was going through reactive power-induced oscillations.”
grid following inverters can absolutely generate reactive power by injecting current out of phase with the AC voltage. and this is very well used and even mandatory in some countries now, because it’s the core function for voltage regulation. what they can’t do is contribute to frequency regulation, because that requires active power injection thus need energy storage for that.
What you said is true, but wrong in context. Read the surrounding text in the article, and look at the report on the Iberian blackout.
I have a new Trane heat pump. Trane was famous for having poor RF noise emissions on their equipment, so naturally I measured mine the day it went in. It’s dead silent. Amazing.
Then I measured the current draw: 0.78 amps continuous, even when OFF. 190 “watts” (actually VA…). So I called in Trane, escalated it all the way up to find out what’s up. While waiting for the tech to arrive I measured the power factor too. 0.04. Hmm. OK, so not real real power, but still: why so much current?
Tech arrived, replicated my measurements, was mystified as I was, and took a deeper dive into the thing.
There’s a massive line filter on these units, and all that current is actually the filter capacitors across the line: entirely reactive load, very effective at nuking the RFI, but still presenting a huge load on the line.
That 0.78 amps might be entirely capacitive, but the line feeding it still must supply the current, and I^2*R still applies.