Connecting Solar to the Grid is Harder Than You Think

On June 4, 2022, a small piece of equipment  (called a lightning arrestor) at a power  

plant in Odessa, Texas failed, causing  part of the plant to trip offline. It  

was a fairly typical fault that happens  from time to time on the grid. There’s a  

lot of equipment involved in producing and  delivering electricity over vast distances,  

and every once in a while, things  break. Breakers isolate the problem,  

and we have reserves that can pick up the slack.  But this fault was a little bit different.

Within seconds of that one little short  circuit at a power plant in Odessa,  

the entire Texas grid unexpectedly lost 2,500  megawatts of generation capacity (roughly 5%  

of the total demand), mainly from solar plants  spread throughout the state. For some reason,  

a single 300-megawatt fault at a single power  plant magnified into a loss of two-and-a-half  

gigawatts, dropping the system frequency to  59.7 hertz. The event nearly exceeded Texas’s  

“Resource Loss Protection Criteria,” which  is minimum loss of power that requires having  

redundancy measures in place. Another  fault in the system could have required  

disconnecting customers to reduce demand.  In other words, it was almost an emergency.

If you lived in Texas at the time, you  probably didn’t notice anything unusual,  

but this relatively innocuous event sent alarm  bells ringing through the power industry.  

Solar plants, large-scale batteries, and wind  turbines don’t produce power like conventional  

thermal power plants that make up such a big  part of the grid. The investigation into the  

2022 Odessa disturbance found that it wasn’t  equipment failures that caused all the solar  

plants to drop so much production all at once,  at least not in the traditional sense. Instead,  

a wide variety of algorithms and configuration  settings in the power conversion equipment  

reacted in unexpected ways when they  sensed that initial disturbance.

The failure happened just before noon on  a sunny summer day, so solar plants around  

the state were at peak output, representing  about 16% of the total power generation on the  

grid. That might seem high, but there have  already been times when solar was powering  

more than a third of Texas’s grid, and that  number is only going up. The portion of the  

grid comprised of solar power is climbing  rapidly every year, and not just in Texas,  

but worldwide. So the engineering challenges in  getting these new sources of power to play nicely  

with the grid that wasn’t really built for them  are only going to become more important. And,  

of course, I have some demos set up in the garage  to help explain. I’m Grady and this is Practical  

Engineering. In today’s episode, we’re talking  about inverter-based resources on the grid.

Solar panels and batteries work on direct current,  DC. If you measure the voltage coming out,  

it’s a relatively constant number. This  is actually kind of true for wind turbines  

as well. Of course, they are large spinning  machines, similar to the generators in coal  

or natural gas plants. But unlike in thermal  power plants that can provide a smooth and  

consistent source of power through a  steam boiler, winds vary a lot. So,  

it’s usually more efficient to let the turbine  speed vary to optimize the transfer of energy from  

the wind into the blades. There are quite  a few ways to do this, but in most cases,  

you get a variable-speed alternating current from  the turbine. Since this AC doesn’t match the grid,  

it’s easier to first convert it to DC.  So you have this class of energy sources,  

mostly renewables, that output DC, but the grid  doesn’t work on DC (at least not most of it).

Nearly all bulk power infrastructure, including  the power that makes it into your house,  

uses an alternating current. I won’t go  into the Tesla versus Edison debate here,  

but the biggest benefit of an AC grid is that  we can use relatively simple and inexpensive  

equipment (transformers) to change the voltage  along the way. That provides flexibility between  

insulation requirements and the efficiency of  long-distance transmission. So we have to convert,  

or more specifically invert, the DC power from  renewable sources onto the AC grid. In fact,  

batteries, solar panels, and most wind turbines  are collectively known to power professionals as  

“inverter-based resources” because they are so  different from their counterparts. Here’s why.

The oldest inverters were mechanical devices: a  motor connected to a generator. This is pretty  

simple to show. I have a battery-powered  drill coupled to a synchronous motor. When  

I pull the trigger, the drill motor spins the  synchronous motor, generating a nice sine wave  

we can see on the oscilloscope. Maybe you  can see the disadvantages here. For one,  

this is not very efficient. There are losses in  each step of converting electricity to mechanical  

energy and then back into electrical energy on  the other side. Also, the frequency depends on  

the speed of the motor, which is not always  a simple matter to control. So these days,  

most inverters use solid-state electronic  circuits, and look what I found in my garage.

These are practically ubiquitous these  days, partly because cars use a DC system,  

and it’s convenient to power AC devices from  them. I just hook it up to the battery, and  

get nice clean power from the other end…

haha just  kidding. These cheap inverters definitely output  

alternating current, but often in a way that  barely resembles a sine wave. Connecting a load  

to the device smooths it out a bit, but not much.  That’s because of what’s happening under the hood.  

In essence, switches in the inverter turn on and  off, creating pulses of power. By controlling the  

timing of the pulses, you can adjust the average  current flowing out of the inverter to swing up  

and down into an approximate sine wave. Cheaper  inverters just use a few switches to create a  

roughly wave-like signal. More sophisticated  inverters can flip the switches much more quickly,  

smoothing the curve into something closer to a  sine wave. This is called pulse width modulation.  

Boost the voltage on the way in or the way out,  add some filters to smooth out the choppiness of  

the pulses, and that’s how you get a battery  to run an AC device… but it’s not quite how  

you get a solar panel to send power into the  grid. There is a lot more to this equipment.

For one, look at the waveform of my inverter and  the one from the grid. They’re similar enough,  

but they’re definitely not a match. Even the  frequency is a little bit off. I will not be  

making an interconnection here, since I don’t have  a permit from the utility, but even if I did, this  

inverter would let out the magic smoke. A grid-tie  inverter has to be able to both synchronize with  

the phase and frequency of the grid and be able  to vary the voltage of the waveform to control how  

much current is flowing into or out of the device.  The synchronization part often involves a circuit  

called a phase-locked loop. The inverter senses  the voltage of the grid and sets the timing of all  

those little switches accordingly to match what  it sees. So, these are often called grid-following  

inverters. They synchronize to the grid frequency  and phase and only vary the voltage to control the  

flow of power. And that hints at one of their  challenges: they only work when the grid is up.

I’ve done a video all about black starts,  so check that out after this if you want  

to learn more, but (in general),  inverter-based resources like solar,  

wind, and batteries can only follow what’s  already on the grid. When the system’s down,  

they are too, regardless of whether the sun’s  shining or the wind’s blowing. That’s why  

most grid-tied solar systems on houses  can’t give you power during an outage.

There’s another interesting thing that inverters  do for solar panels, and I can show you how it  

works in my driveway.  

I have a solar panel  hooked up to a variable resistor, and I’m 

measuring the voltage and current produced by  the panel. You can see as I lower the resistance,  

the output voltage of the panel goes down and  the current it supplies goes up. But this isn’t a  

linear effect. I recorded the voltage and current  over the full range, and multiplied them together  

to get the power output. If you graph the power as  a function of voltage, you get this shape. And you  

can see there’s an optimum resistance that gets  you the most power out of the panel. It’s called  

the maximum power point. If you deviate on either  side of it, you get less power out. In other  

words, you’re leaving power on the table. You’re  not taking full advantage of the panel’s capacity.

What’s even more challenging is that point  changes depending on the temperature of the  

panel and the amount of sun hitting it. I  ran this test again with a few more clouds,  

and you can see how the graph changes. So nearly  all large solar photovoltaic installations use  

what’s called a Maximum Power Point Tracker (or  MPPT) that essentially adjusts the resistance to  

follow that point as it changes with sunniness  and temperature. It’s really a separate device  

from the inverter, but often they’re located  right next to each other or inside the same  

housing. Even this panel came with a charge  controller that has this MPPT function,  

and you can see it adjusting the flow of  current to constantly try and stay at the  

peak of the curve while it charges this battery.  These can be used for an entire installation,  

but in many cases, each panel or group  of panels gets its own MPPT because that  

curve is just a little bit different  for each one. Tracking the peak power  

output individually can often squeeze a  little more capacity out of the system.

Squeezing out capacity is essential to address  another challenge associated with inverter-based  

resources on the grid: frequency.   

The rate at  which the voltage and current on the grid swing

back and forth is an important measure of how  well generation and demand are balanced. If demand  

outstrips the generation capacity, the frequency  of the grid slows down. Lots of equipment, both on  

the generation side and the stuff we plug in, is  designed to rely on a stable grid frequency, so if  

it deviates too far, stuff goes wrong: Devices  malfunction, motors can overheat, generators  

get out of sync, and more. It’s so important  that rather than let the frequency get too far  

out of whack, grid operators will disconnect  customers to get electrical demands back in  

balance with the available supply of power, called  an under-frequency load shed. Things go wrong on  

the grid all the time, so generators have to be  able to make up for contingencies to keep the  

frequency stable. Here’s the quintessential  example: an unexpected loss of generation.

Say a generator trips offline, maybe because of a  failed lighting arrestor like the Odessa example.  

The system frequency immediately starts dropping,  since power demand now exceeds the generation. And  

the frequency will keep dropping unless we inject  more power into the system. The first part of  

that, called Primary Frequency Response, usually  comes from automatic governors in power plants.  

If we do it fast enough, the frequency will reach  a low point, called the nadir (NAY-dur), and then  

recover to the nominal value. The nadir is a  critical point, because if it gets too low,  

the grid will have to shed load in order to  recover. The other important value is called  

the rate-of-change-of-frequency, basically  the slope of this line. It determines how much  

time is available to get more power into the  system before the frequency gets too low,  

and there are several factors that play into it:  How much generation was lost in the first place,  

how quickly we can respond, and how much inertia  there is on the grid. Thermal power plants that  

traditionally make up the bulk of generating  capacity are gigantic spinning machines. They’re  

basically a bunch of synchronized flywheels.  That kinetic energy helps keep them spinning  

during a disturbance, reducing the slope  of the frequency during an unexpected loss.

Maybe you can see the problem with a simple  grid-following inverter. It’s locked in phase  

with the frequency, even if that frequency  is wrong. And it has no physical inertia to  

help arrest a deviation in frequency. If we  keep everything the same and just increase  

the share of inverter-based resources, any  loss of generation will mean a steeper slope,  

reducing the time available to get backup  supplies onto the grid before it’s forced  

to shed load. Larger renewable plants, like  solar and wind farms, are increasingly required  

to participate in primary frequency response,  injecting power into the grid immediately when  

the frequency drops. And some inverters can even  create synthetic inertia that mimics a turbine’s  

physical response to changes in frequency.  But there’s another challenge to this.

Dealing with an over-frequency event is relatively  straightforward: just reduce the amount of energy  

you’re sending into the grid. But, response  to an under-frequency event requires you to  

have more energy to inject. In other words, you  have to run the plant below its maximum capacity,  

just in case it gets called on during  an unexpected loss somewhere else in  

the system. For a power company, that means  leaving money on the table, so in most places,  

the energy markets are set up to pay power plants  to maintain a certain level of reserve capacity,  

either through operating below maximum output  or including battery storage in the plant.

The last big thing that inverter-based resources  have to manage is faults. Of course, you need  

protective systems that can de-energize solar or  wind resources when conditions on the grid could  

lead to damage. These are expensive projects, and  there’s almost no limit to the things that can go  

wrong, requiring costly repairs or replacement.  But, for the stability of the grid, you can’t  

have those protective systems being so sensitive  that they trip at the hint of something unusual,  

like what happened in Odessa. This concept  is usually referred to as “ride-through.”  

Especially for under-frequency events,  you need inverters to continue supplying  

power to the grid to provide support. If  they trip offline, or even reduce power,  

in response to a disturbance, it can lead to  a cascading outage. This is kind of a tug of  

war between owners trying to protect their  equipment and grid operators saying, “Hey,  

faults happen, and we need you not to shut  the whole system down when they do.” And  

reliability requirements are getting  more specific as the equipment evolves,  

because every manufacturer has their own  flavor of protective settings and algorithms.

As inverter-based resources continue to grow  rapidly in proportion to the overall generation  

portfolio, their engineering challenges are only  becoming more important. We talked about a few  

of the big ones: lack of black start ability, low  inertia, and performance during disturbances. And  

there are a lot more. But inverters also provide  a lot of opportunities. They’re really powerful  

devices, and the technology is improving quickly.  They can chop up power basically however you want,  

and they aren’t constrained by the physical  limitations of large generating plants. So  

they can respond more quickly, and, unlike  physical inertia that will eventually peter out,  

inverters can provide a sustained response.  There are even grid-forming inverters that,  

unlike their grid-following brethren, can  black start or support an isolated island  

without the need for a functioning grid to rely  on. We’re in the growing pains stage right now,  

working out the bugs that these new  types of energy generation create,  

but if you pay attention to what’s happening in  the industry, it’s mostly good news. A lot of  

people from all sides of the industry are working  really hard on these engineering challenges so  

that we’ll soon come out with a more reliable,  sustainable, and resilient grid on the other end.

I build a lot of homemade demonstrations  for videos like this one, and I hope it  

comes across how much joy it gives me. I love  the challenge of making something useful with  

constraints on budget and tools. But I’ve  never built a hot air balloon! One of my  

fellow creators who runs the Neo channel  just released a video on this incredible  

story of two families escaping East Germany  in maybe the most creative way possible.

I don’t know about you, but I have to say that  almost everything I watch these days is produced  

by independent creators. There’s just something  really authentic and original about content that  

hasn’t had to go through 5 levels of studio  executives before it gets made. Neo’s episode  

on The Balloon Escape is a perfect example. Just  a fascinating story about homemade engineering,  

including an interview with one of the  men who made the attempt, all set to the  

beautiful animations they’re known for. And, if  you want to see it, it’s only available on Nebula.

You probably know about Nebula now, even if  you’re not subscribed. It’s a streaming service  

built by and for independent creators. No studio  executives deciding what gets the green light,  

no algorithm driving the content into  a single style, and no ads getting in  

the way. We just released a huge update  that completely redesigned the home page,  

making it easier to find new stuff in addition  to your favorites. There's tons of originals, and we’re always adding creators,  

so the new categories can help you  discover content related to your interests.

My videos go live on Nebula before they  come out here, and my Practical Construction  

series, where I embedded on a construction  site for a year, was specifically produced  

for Nebula viewers who want to see deeper dives  into specific topics. I know there are a lot of  

streaming platforms out there right now, and no  one wants another monthly cost to keep track of,  

but I also know that if you’re watching a  show like this to end, there is a ton of  

other stuff on Nebula that you’re going to  enjoy as well. So I’ve made it dead simple:  

click the link below and you’ll get 40% off an  annual plan. That means you pay just one time,  

30 dollars, for an entire year’s access at  nebula.tv/practical-engineering. Or if you have  

subscription fatigue, but still want to support  what I’m doing, you can get a lifetime membership.  

Pay once and have access for as long as you  and Nebula last. Hopefully that’s a long  

time! If you’re with me that independent  creators are the future of great video,  

I hope you’ll consider subscribing. Thank you  for watching, and let me know what you think!