Grid Vibes

Vibe-nets and Mad Maxian parades: Our electrical infrastructure explained

Before we start…

Have you ever wondered why electrical transmission lines have alien-looking connectors?

Because of rain. Huh? Here’s a quick graphic I made to show you why:

They’re called high-voltage standoffs, and they are made out of porcelain insulators to prevent the high-voltage electricity in the cables from traveling down the metallic or wooden poles they are mounted on. Because water can be conductive (so can sand or dust), their weird shape makes it difficult for continuous conductive paths to form and zap you. It turns out engineers don’t make fun-looking things for no reason. The question is: was this thought of before or after someone got zapped in the rain? 🤔 Fun fact: each “disc” usually helps insulate ~15 kV (15,000 V), so the example on the left can guard against ~90 kV. Now you can go nerd out and estimate the voltage on your local power lines 🤓 .

Oh, and also:

Because that would be tragic. Also, because that’s not how electricity works. Voltage is the electrical analogy to altitude—it would be meaningless unless it’s measured with respect to something such as sea level. These birds are similar to you standing on top of the Empire State. The floor beneath you is your reference point and you won’t feel any different (apart from fear). But if someone were to kick you off (your fear), your new reference point will be W 34th Street, and you’d likely be dead. Similarly, if one of the birds standing on the high-voltage cable also touched a tree that isn’t magically floating, then its new electrical reference point would be “earth ground” and it would get zapped. Whether the bird will die or not depends on how resistive to electrical current the tree is (e.g. how juicy it is). But if a really large bird was able to put one foot on one cable and the other foot on another, it would almost certainly die, unless the two cables are seeing the same electrical phase (e.g. in parallel). For those of you that are wondering, no, the cables are not insulated. That would cost too much for the millions of miles of electrical conductors we have in just the U.S.

Now that we’ve answered the important questions, let’s talk about the electric grid.

---

The Electric Grid

In the last post, we talked about where we get our energy from, much of which comes from fossil fuels, to do useful things. Useful things can be transporting people, providing heat for manufacturing, or doing a whole bunch of stuff with electricity. Today, we’ll talk about the electricity part, and how that electricity actually makes it from the thermal plants (coal, natural gas, nuclear), solar farms, or wind farms to your home or factory. This is called the electric grid, and it will play one of the most critical parts in our decarbonization. The grid is the foundation to the first of the three buckets we talked about last time—the one called “electrify-that-shit”—to address our energy crisis. It’s also in dire need of upgrades, and the recent Texas power outages and California wildfires serve as painful reminders. Our goal is to fundamentally understand how it works, and we’ll do this in two parts: (1) the technical part and (2) the logistical part. Afterwards, we’ll do a quick deep dive into what happened in Texas. Given the length of this post, we’ll reserve deep dives into next generation grid/micro-grid technologies for a later date. But this should help us first establish a foundational understanding to build upon. In the meantime, you can subscribe to David Roberts’ “Volts” Substack, where he just did a month-long series on Transmission.

Warning: This is a long post as there is a lot to understand with our electrical grids—they’re technically the world’s largest machines so they kinda deserve this special attention. But bear with me because it really isn’t that complex, and you can read Part I and II separately. I find oversimplifications more confusing, so we’ll walk through it without using analogies that don’t work.

---

Part I: The Technical Stuff

Our massive vibe net

It’s in our common parlance to speak of electricity “flowing” and to think of it like water or gas molecules in pipelines. That’s not actually how it works, especially when we’re talking about the grid. You may have heard there was once a showdown between two dudes named Edison and Westinghouse in the late 1800s, and they were battling between alternate (pun intended) visions of how best to electrify America. Edison was a proponent of direct current (DC) electricity, whereas Westinghouse, along with Tesla, was a proponent of alternating current (AC) electricity. This was known as the War of the Currents, and it turned into a media-hyped nerd-battle. Ultimately, AC electricity proved to be the technologically superior method of electrifying cities and homes, despite shockingly (another pun!) low-blows from Edison’s crew to promote the use of AC for the first electric chair, electrocute stray animals with AC electricity as demonstrations, and fear-monger on safety concerns. Not cool, Tommy.

Ultimately, electricity is just the movement of electrons in conductors, and we can make use of it so long as the electrons are moving. The key question is just how we induce them to move.

Electrons in cables, wires, and other conductors are just like an unbroken chain of electron-buddies linking arms. In DC systems, the long chain of electrons march together in one direction with the electron at the front of the line pulling the rest along with it (as opposed to being pushed—key distinction) to wherever potential energy is lowered (e.g. to a lower voltage). Here, there is actually a net “flow” of electrons. You can find DC in battery-powered stuff (your phone, laptop, electric vehicle), solar panels that generate a voltage with sunlight, or in lightning ⚡️ .

In AC systems, the long chain of electron-buddies stay in relatively the same place but sway back-and-forth in harmony, either 50 or 60 times per second (50 or 60 Hertz) depending on what country you’re in. In a trance. Just vibin’. There is a tiny bit of local back-and-forth movement but there is no net “flow” of electrons (actual electron travel distance depends on conductor material, diameter, voltage, and current, but they oscillate no more than ~500 nanometers in a 16 gauge copper wire in a 240V 100W light bulb circuit¹). So when you flip a switch or hook up a new distribution line to an AC grid network, you’re not getting electrons from somewhere else, say your power plant. Instead, you’re just causing the electrons inside your conductors to vibe-in, lock arms, and sway at the same frequency. This means that the entire electric grid is just a massive vibe net—an everlasting party managed by the grid operators—where we tap into the rhythm whenever we need. Such unity 🥺 . This unity is actually how we used to make sure all of our clocks ran the same when we still used AC synchronous motors to turn their hands. But as we’ll see below, the electric grid is never perfectly at 60 Hz (50 Hz in Europe) despite our best efforts to keep within very tight tolerances. This means drift over time will cause old AC clocks to not run true.

Given more recent technology advancements in power electronics, there is now a resurgence of interest in DC transmission for grid-scale applications. What’s different is just the mega-high voltages used that were previously unattainable, aptly called high-voltage DC (HVDC) or ultra-high-voltage DC (UHVDC). These are transmission lines for direct one-way pipelining (the analogy works here since it’s DC) of electrons over long distances, which was generally the strong-suit of AC but only up to a voltage threshold (approximately 750,000 V²). HVDC makes sense for transporting electrical energy directly from, say, a solar farm in a remote desert area to an urban metropolis, where it must be converted into AC before being introduced into the grid. We’ll touch upon this lightly later, but we’ll do a deep dive one day.

What actually happens when you flip a switch

You may have heard that the electric grid is always perfectly balanced between supply and demand—where any new demand for electric power on the grid (such as when you flip that switch) is simultaneously matched with an equal supply of electrons being pumped in by a generator somewhere. If this explanation bothers you because it sounds too good to be true, that’s because you’re right, and the oversimplification can make things confusing. First of all, we already know that no electrons are being pumped over long distances in an AC grid. Second, it’s impossible for things to be balanced that perfectly—let’s be real…we’re humans and we’re error-prone. Let’s figure out how it really works.

First, let’s talk about supply

As we discussed earlier, electrons in an AC grid are just in a state of lock-armed, swaying trance at a particular frequency. But of course, nothing is frictionless and electrons are not just free to move within a conductor without an energy input. This is the role of generating plants—to provide electrical work, not electrons. Taking a nuclear thermal plant as an example, let’s walk through the process. A nuclear fission reactor is just a glorified water boiler, wherein steam is generated by the crapload of heat energy released by breaking nuclear bonds. When the hot, pressurized steam is allowed to expand through the various stages of a steam turbine, it pushes against the turbine blades to create rotary mechanical motion. All of this is known as the Rankine cycle, and is the same physics behind coal-fired power plants where steam is the working gas. This rotary shaft is connected to a three-phase electric generator, which is the same as a three-phase induction motor (thanks to Nikola Tesla) in reverse operation. Here, the generator “rotor” (the spinning part) spins at a constant 1800 rpm (3600 rpm in coal plants)³ and induces voltages and currents to form in three separate windings, called phases, in the outer shell, or “stator”. These phase windings are physically spaced apart by 120 degrees around the rotor, which corresponds to them being 120 “electrical degrees” apart. All of this sounds fancy, but it just means that three sinusoidal waves are generated in three separate cables, each one lagging behind another. Here’s a visual⁴:

The sinusoidal voltage output of each of the three phases is what gives the electrons in those conductors the energy to move back and forth, and the 1800 rpm of the rotor is what generates a constant frequency of 60 Hz (1800 rpm = 30 rotations/sec, and there are two sinusoids per revolution in nuclear power plant generators). These three phases are creatively labeled A-B-C. To be clear, one single-phase is enough to be useful (and to zap you). The 120 or 240V you get out of your wall is single-phase. But transporting three-phase power just gives us more bang-for-the-buck based on total conductor material required, and is more efficient for larger motors and generators in general. So while the actual voltages differ in different parts of the world, three-phase power is used everywhere for transmission, industrial plants, and commercial buildings, and they’re only later split into single-phases when powering homes or small businesses.

In essence, electric power generation for our grid simply comprises big (electrical) wave generating machines that sway the electrons in our massive vibe net (technically three vibe nets, one for each phase) in unison. There is no overall “supply” or injection of electrons into the system, only a supply of electrical energy to push and pull electrons inside conductors. The bigger our vibe net is to power more homes and industries, the more energy is required to sway our electrons. Adding more power plants is like enlisting more people to help push a kid on a swing.

In case you were curious, any power plant that wants to jump in and “contribute” electrical work to the grid must first be synced-up. Not only must it sway at the same frequency, the voltage waveforms must also rise and dive at the same time in order for it to “vibe-in”. Not doing so will be catastrophic, and would be similar to a swing-pusher pushing when they’re not supposed to. Nobody likes that. This syncing is done with a sci-fi-sounding thing called a synchroscope.

What about demand?

What happens when you flip a switch? You are essentially adding a circuit to the massive vibe net. It doesn’t matter if it is a light bulb, a toaster, or a time machine, you are tapping into the grid with your “electrical load”. The lock-armed electrons in your newly connected conductors sap some energy away from the vibe net as they participate in the sway. Since more electrons have joined the party, the entire system is a tiny bit (or a lot of bit) more sluggish, and more electrical work is required. What happens next? The large generator rotor in the power plant we met earlier actually slows down, causing the frequency on the entire grid to drop below 60 Hz. This frequency “sag” will not lift back up until steam valves are opened wider for more turbine power, which will require more nuclear fission to be allowed in the reactor, thereby generating more heat, boiling more water, and pressurizing more steam.

The opposite happens when you turn off your switch. The entire grid temporarily goes into overdrive as the generator rotor spins faster, the frequency rises above 60 Hz, and nothing will slow down until power plants are throttled. Of course, the degree to which these sags and overdrives happen depend on how big of an electrical load you are turning on or off.

One thing to note is that the bigger and heavier the generator rotors are in the power plants, the more rotational inertia they have, and the less willing they are to change their speed. They are so heavy that they can continue spinning for a brief time even if there is a plant failure. This heftiness is a good thing to maintain constant grid frequency, and is a disadvantage for smaller, distributed, and more variable renewable energy sources like wind and solar (which must rely on small inverters). This is a key reason why batteries are important for frequency response, which we’ll deep dive into in the next post. But first, here’s what a modern-day generator looks like⁵ 😍 .

Now it becomes clear why AC-powered clocks won’t be super accurate. It’s also clear why a larger grid with more electricity generators and users makes for a much more stable system. Each flick of a switch becomes an increasingly negligible ripple on the entire system, and we become more fault tolerant to plant failures.

Maintaining balance

What happens when frequencies dip or rise too far below or above nominal? Stuff breaks. A lot of things, especially power plants, are designed and optimized to operate at 60 Hz, and deviating too far from this can be catastrophic. Remember it’s the grid operators’ job to keep the party going, and so they can’t allow things to slow down or get too rowdy before the cops show up and shut things down. Below is a chart showing the various triggers for frequency adjustments on a 60 Hz grid, with the left and right axes showing coarse and fine deviations, respectively⁶. In case you were wondering, “Governor Response” here refers to automatic equipment that help keep speed in check. Don’t worry, your state governor is not getting pings all the time to change rotor speeds. As you can see, deviation of just ±1% is already a big deal.

What this means is that grid operators must accurately predict how much electrical work will be demanded by all the consumers on the vibe net on a real-time basis, and coordinate with generating plants on moment’s notice to ramp up, throttle down, connect, or disconnect their power production to match accordingly. While starting your car engine may take only a couple seconds, starting certain types of power plants can take hours or days. Therefore, grid operators plan in advance, with day-ahead and hour-ahead forecasts of what is likely needed. We’ll talk more about this in the logistical stuff of Part II. There will inevitably be error in the frequency, but the grid operators have gotten pretty good at this, and as we’ve seen above, there is slight tolerance for deviation. As always, the bigger the grid, the smaller the error.

Plans B, C, and D

To anticipate for generator failures, sudden increases in demand, or other flukes, grid operators have “operating reserves” to rely on in a cascading manner. The first line of defense are the power plants already on-line, which are often asked to have the capacity to ramp up power output typically within 10 seconds. This means power plants generally don’t run at full, pedal-to-the-metal “nameplate” capacity in case more output is needed from them. The second line of defense are the “spinning reserves”. These are generating plants that are asked in advance to get ready, spin-up their rotors, sync-up using their synchroscopes, and wait for the command to vibe-in. These can typically be called upon within 10 minutes. Finally, there are supplemental and backup “non-spinning” supplies that can be dispatched within an hour like “peaking plants” that can kickstart on a dime. The amount of operating reserves maintained at any given time is usually at least enough to compensate for the failure of the largest generator on the grid, should it suddenly fail. Since people are good at finding various ways to make money, there are often entire markets associated with each one of the above, especially in deregulated grid systems which we’ll talk about in Part II. These are often known as “Ancillary Services” to the conventional energy spot markets in the U.S., or “frequency control markets” in Europe. Below is a graph showing the day-ahead, hour-ahead, and actual power demand on the grid managed by the California Independent System Operator (CAISO) on the first day of 2021. You can see that their estimates are pretty close to actual demand⁷.

As you may already know (and shown above), our electricity usage throughout the day, week, season, or year is hardly close to constant. On the logistics side, it makes things interesting. On the technology side, it’s a nightmare. It requires us to split our electricity generation into various categories—baseload, intermediate load, and peak load—where the baseload plants are reserved for those that can produce electricity at the lowest cost and often the hardest ones to turn off (e.g. nuclear). The opposite is true for the peak load plants. With the proliferation of rooftop solar being installed in California which allows homes to power themselves emissions-free around noon (a good thing), it does however put downward pressure on the electricity demand and “cut-in” to baseload supply, shown below. Since baseload plants cannot simply be turned off or throttled, this can result in inefficient operation (a bad thing) or even negative electricity pricing where consumers are actually paid to consume electricity. But shortly after this mid-day depression, the after-work demand for electricity peaks as people return home and turn on their A/Cs and TVs—requiring a steep ramp-up in electricity production, which can be stressful on generating equipment. This non-flat consumption profile (the notorious “duck curve”) is the crux of the energy storage challenge we have today, and part of our goal in Honest Energy is to deep dive into the various technologies to address it.

Here is a look at the utilization rate (also known as capacity factor) of the different types of power plants in the U.S.⁸ You can see that the baseload plants are close to maximum utilization and are only turned off occasionally for maintenance reasons, whereas peaking plants don’t operate for very much at all. We’ll understand how they can still afford to exist when we talk about the markets in Part II.

We’ve now talked about the bookends—generation and consumption. But obviously there’s still a huge part in between the two that often forms what we typically think of physically as “the grid”.

Transmission & Distribution

The primary reason AC electricity won the War of the Currents is that power can be transferred efficiently across medium-to-long distances more easily (not ultra-long distances, though, that’s where HVDC comes in). It comes down to a simple physics problem involving just three (really two) equations:

It’s really two equations since plugging (1) into (2) gives you (3).

Since work is required to move electrons in a conductor, it must also mean that there are losses somewhere. As with most things, this loss is in the form of heat, and the inherent non-zero resistance of conductors cause them to heat up. This power loss can be easily calculated with equation (3), and we don’t like it. We can reduce this power loss by either reducing electrical resistance, R (hard to do), or reducing overall current, I. Since the current is squared, you get more bang-for-your-buck by reducing the latter. Halving resistance will reduce your power loss by a factor of 2, but halving current will reduce your power loss by a factor of 4. Of course, there’s no free lunch here, and we need to make up for this reduced current somehow. Equation (2) says that electric power is just the magnitude of the current multiplied by the voltage. So if we agree we need to halve our current, we then need to double our voltage to transmit the same amount of power. Simple enough.

The primary reason AC won the battle is that it is stupid-easy to increase (step up) and later decrease (step down) voltages highly efficiently with cheap devices known as transformers, which allows long-distance, high-voltage transmission with low currents, minimal power loss, and cheap small-diameter cables. Stepping DC voltages up or down is/was non-trivial, and therefore power loss would have resulted in terrible efficiencies or very costly transmission lines. Touché Tesla.

Our electric grid therefore comprises different sections having different voltage levels depending on generation, transmission, distribution, or consumption. But while voltage (amplitude) is intentionally changed, the nominal frequency does not change at any point in the process. So all electrons in our grid still vibe to the same beat—some just vibe harder than others. Here’s a schematic of the entire grid system⁹. Note the different voltages:

Also note the three lines representing the three phases. You may have noticed (or you will now) all transmission and distribution pylons/poles have cables in multiples of three, with an extra cable or two at the very top used to protect against lightning strikes. Like this¹⁰:

Finally, let’s nerd out for just a sec on the transmission cables. What are they made of? Definitely not copper—that would be too heavy and expensive, and they’d probably get cut and stolen a bunch for scrap money (for real though). Instead, they are called aluminum-conductor steel-reinforced (ACSR) cables, where aluminum strands are wrapped around steel strands. Here’s what they look like:

When you apply an alternating current to any conductor, electrons tend to vacate the core and move closer to the surfaces. This is called the “skin effect”, and it reduces the useable cross sectional area of conductor available, thereby increasing resistance and heat loss. To avoid too much loss, smaller strands are used to bundle into a larger cable (the higher the frequency, the more dramatic the effect, which is why radio frequency wires have super fine strands). Steel is strong but a terrible conductor, whereas aluminum is weak but a great conductor. Both are cheap. By wrapping aluminum strands around a steel core, you get decent strength and conductivity. Combo meal. Here, the skin effect on the overall cable makes it so that almost all of the electricity is carried in the outer aluminum strands anyway, and where each aluminum strand also has its own skin effect—it’s a…skinception 👀 .

Technical Challenges on the Grid

Now that we understand all the components to the largest machines in the world, we can start to understand what some of the key technical challenges are (logistical challenges are in Part II). These are almost always rooted in the physical constraints of the system that lead to something called “congestion”.

Congestion…ew

You may have heard of the term “grid congestion”, where electrons in transmission lines are compared to traffic in Los Angeles. Again, it’s not a good analogy because electrons aren’t really flowing (but I guess neither is LA traffic). Regardless, grid congestion is a real thing, and it happens when the cables we just talked about aren’t beefy enough to carry enough current to downstream demand. Resistance in cables is a direct function of cross-sectional area, as you can see in Equation (4):

The fatter cables are, the lower the resistance. And we know from earlier that power loss in the form of heat increases with more resistance. So when cables are too thin, they heat up, and become less efficient in power transmission. If unchecked, this heat can cause cables to expand and sag, touching trees, lighting them on fire, and causing blackouts. You may wonder why we can’t just bury all of the high-voltage transmission lines underground to avoid zapping trees. The answer is that hanging cables in open air helps to cool them down, whereas burying them would force them to roast in their own sauna and make them even more inefficient—that is unless we made them even beefier or built in cooling, all of which is expensive.

So even though our grid is a massive vibe net swaying in unison powered by thousands of generating plants, our grid’s ability to connect generators and end-users is only as strong as its weakest link. Grid operators, therefore, must take into account the current-carrying capacity of cables (technical term: ampacity) and throttle power generation (curtailment) upstream even when it is needed downstream. This is technically called “transmission constraint”, but the economic impact (and frustration) it inflicts on the end-users is called “transmission congestion”¹¹. Unfortunately, curtailment happens most with renewable power since they are generally located further away (such as a wind farm) than conventional thermal plants, and they are naturally intermittent. Let’s look at three example scenarios below:

In the first scenario, we have conventional, dirty thermal plants powering the grid to service end-users—industry, commercial, and residential. There’s no congestion here because the power plants are supplying what the consumers demand.

In the second scenario, we now have cheap, clean, renewable power that can be brought in from a wind farm far away from the city. Hooray! Since it’s cheaper than say coal, users want to use that instead. Unfortunately, our cables are anorexic, and servicing all consumers with the wind power will cause transmission lines to overheat, blow fuses, or zap trees. This is transmission constraint. In this case, cheap wind power must be curtailed and expensive coal plants must be reintroduced in order to service end-users, making them pay more for electricity. This negative economic impact is congestion.

In the third scenario, we invest in beefing up our transmission lines for 21st-century needs and increase the size of conductors so we can use more clean, renewable power. Here, we can finally eliminate our reliance on dirty thermal plants. Note that this doesn’t solve the intermittency problem of renewables, but is key to enabling more deployment to increase their overall power-generating capacity.

The above scenarios are for a made-up cartoon city. Here’s what it actually looks like in real life, with an example from the Midcontinent Independent System Operator (MISO) in 2018¹². All that red is no good.

Whadda we do?

We can of course invest in rebuilding large parts of the grid (duh!). But as you’ve probably guessed, that takes a lot of people coming together and agreeing to do something big…and we’re not very good at that. But that’s not to say there aren’t lower-hanging fruits we can start with now, and here I’m highlighting one example from a recent post from “Volts” regarding grid enhancing technologies.

For one, we may be imagining problems that are really not that bad (yet). We talked earlier about how we like to hang our cables in open air so they can be cooled. It turns out that even a gentle breeze can cool these lines substantially, and increase their current carrying capacity by a lot. Who would’ve thunk! On top of the fact that we already operate lines with a decent safety margin (most of the time…ahem PG&E), the “fixed ratings” on line ampacities are usually very conservative. By installing sensors on pylons that use LIDAR to see how much transmission lines sag and sway in the wind, we can calculate how much wind speed they see and generate real-time “dynamic line ratings” that grid operators can use to alleviate grid congestion. In some cases, more than twice the amount of rated power can be transmitted, even if for shorter periods of time. So some of the dark red we see in the map above could actually just be self-imagined, and perhaps even vanish without upgrading our lines at all. 🤯

What about HVDC?

As we mentioned earlier, high-voltage DC, or HVDC, is a transmission approach enabled by improved power electronics. This is suited for long-distance, one-way pipelining of electricity into a grid network, which is perfectly suited for large wind or solar farms that generate power in often remote areas where no one wants to live. Taking our earlier cartoon city as an example, this is what HVDC could look like:

Note that the DC must first be converted into AC and synced-up with the grid before vibing-in. As you can imagine, using long-distance HVDC could enable solar power generated in early afternoon of Arizona to be consumed by the evening demand of Kentucky. Or the night-time wind power from the Midwest to power the after-work demand in California¹³.

We now have a decent first-principles technical understanding of how the grid works, spanning from generation to consumption, electrons to infrastructure. But not only is the electric grid a ginormous machine, it is also the platform for entire markets and requires the intricate coordination of countless people to make function. Armed with a technical understanding, we can now turn to look at the logistics of the grid, how it’s managed, and who the key players are. As we’ve seen already, the technologies that comprise the grid are mostly 20th-century Physics 101 stuff that we’ve kinda already mastered. While there are certainly more opportunities for 21st-century tech deployment (LIDAR, AI/machine learning, improved power electronics), some key challenges ahead with respect to the grid are more related to economics, politics, and willpower.

---

Part II: The Logistical Stuff

To make such a massive system work, there are lots of entities that must coordinate with each other across different geographical and time scales. Who exactly does what is rather complex because of the way the grid has evolved into existence over time. You may have heard terms like ISOs, RTOs, TSOs, IOUs, and recently a whole lot about ERCOT. Instead of memorizing all of the who’s who and acronyms, it’s better to first understand the roles that must be played in any grid system, and later see how different grid systems delegate these roles across various entities.

The Power Players

First, an analogy that kinda works: let’s think of the grid system as a massive parade of floats being towed by different vehicles, where each float also has a bunch of smaller wagons in-tow. Our massive parade needs to move at a constant speed of 60 mph (our analogy to 60 Hz), no matter what the terrain and slope. It’s a big one-way journey. Here’s what that looks like. It’s got some Mad Max vibes:

On the far left, we have the tow vehicles. These are the generating plants. These tow vehicles come in various shapes and sizes, with several viable powertrain technologies available (fossil fuel, nuclear, wind, and solar). There are many more tow vehicles available but only the ones towing the floats are doing real work. Other vehicles may be doing pit stops, running maintenance, changing oil, or being decommissioned. We also have some vehicles catching up to speed ready to hitch their tow and contribute if more horsepower is needed to move our parade, and these are the spinning reserves we discussed in Part I. Some tow vehicles are gas guzzlers and aren’t very efficient. Others cost a lot to build but once they’re running, they are reliable and cheap to operate. Some vehicles stop and go based on the wind or the sun. Some are very agile, while others, once started, are almost impossible to stop on a dime.

On the far right, we have the multitude of wagons being towed by the floats which are in-turn pulled by the tow vehicles. These wagons are the end-users or consumers, which includes you and me in residential homes, commercial buildings, and industry. If we have the ability to generate our own power (say with a solar panel on the roof, or a generator within a manufacturing facility), we can power our own wagon and reduce our reliance on the tow line, or in some cases even be completely “off-grid”. But otherwise, we are the primary consumers of the work that is performed by the tow vehicles, and we pay for this service. When we turn on all the lights and appliances in our home, our wagon gets heavier and harder to tow.

In the center, we have the floats, which have busy people onboard managing the horsepower supply of the tow vehicles and the demand from the wagons. These are the balancing authorities, and they do their best to maintain the speed of the parade at a constant 60 mph. Each float has its own group of tow vehicles and trailing set of wagons, but the floats are also tied together with tie-lines (which also happens to be their technical term in the grid). If the tow vehicles pulling one float are short on horsepower, the floats around it can help maintain its speed. Similar to our analogy, the balancing authorities are solely responsible for managing the demand and supply of end-users and power plants within their given jurisdiction, but can import and export power to neighboring balancing authorities.

The busy people onboard the floats have several jobs. Some are looking ahead with binoculars to predict what kind of oomph is likely to be needed soon. Others are yelling at the current tow vehicles through megaphones to rev harder, let off, or calling on new vehicles to catch up and hitch their tow if there are sudden changes in demand. And some other folks will be calling neighboring floats to coordinate on sharing horsepower. One may know in advance that many of its vehicles are due for scheduled maintenance soon, and will need the other floats to rev up their vehicles instead.

Finally, the tow lines connecting the generating vehicles to the floats are the high-voltage transmission lines, and the tow lines connecting the end-user wagons to the floats are the low-voltage distribution lines. Apart from the physical wires, transmission and distribution also includes the poles, pylons, transformers, and other equipment involved in the transport of electricity.

In the U.S., we have more than 7,300 tow vehicles (generating plants) pulling a total of 66 floats (balancing authorities, with 8 more tied-in from Mexico and Canada) via more than 160,000 miles of transmission tow-lines and millions of miles of distribution tow-lines, ultimately pulling along 145 million customer wagons¹⁴. There is also not just one big parade in the U.S., but three. These are the three independent grids, formally known as “interconnections” because they evolved from small local grids being interconnected to form the large vibe nets they are today. They are known as the Western Interconnection, the Eastern Interconnection, and (Texas being Texas) the Electric Reliability Council of Texas (ERCOT). The three parades are almost completely independent from each other. Here’s what they look like in different colors, and the circles represent separate balancing authorities, or floats, within each parade:

Finally, there are rules, regulations, and standards to the parade, and they are managed by people off on the side. They include FERC, NERC, and other regional coordinating councils, who we’ll touch upon in a bit.

Who owns the parade?

This is where things start to differ across different parts of the U.S. and the world. It’s not always clearcut who owns what portion, and while you always own your home or business, who you pay is not always the same. Even within the same parade (grid or interconnection), you can have very different types of operating structures. Given the historical evolution, it’s easier to understand the journey of how things came to be.

The electric grid was not created by one large coordinated national effort…we’re obviously not that organized. Instead, it was created by a patchwork of local utility entities that served only their local areas. These utilities were the OGs, and were entities that ran the entire parade, including generation, transmission, distribution, and balancing. They were super vertically-integrated. Since they hooked up homes and businesses all the way back to their generating plants, they were natural monopolies (you only have one set of wires feeding your home so you don’t really have a choice), and you could either pay them for electricity or live in the dark. Up to you. As the local grids grew to become much more interconnected for increased reliability (discussed above), it evolved into the large interconnections that are physically tied to each other via tie-lines. Despite the tie-lines, you are still connected to the grid with only one set of wires, so the utilities still operated as monopolies in their respective territories.

If the word “monopoly” rings alarm bells in your head, that’s normal because we are trained to dislike them starting from a very young age. But exceptions are made if there’s no other way, and we allow them to exist only if they promise to give something in return. Just like how patents grant you a monopoly on your idea for 20(ish) years if you promise to teach the world your invention, electric utilities are granted monopolies if they promise to serve all customers on their distribution lines and to be regulated by state-owned public utilities commissions to keep them in check. Up until the mid-1990s, this is how basically all of the U.S. operated. Since these were “regulated monopolies”, this became known as regulated electricity markets. Here’s a simplified version of what that looked like¹⁵:

Order 888

Then on one spring day in 1996, the Federal Energy Regulatory Commission, or FERC, issued a monopoly-killing order called Order 888, not to be confused with the Jedi-killing Order 66 from Star Wars: Revenge of the Sith. OK, it didn’t really “kill” monopolies, but it focused their monopoly power on just the transmission and distribution part that are always going to be natural monopolies. This was the first part of Order 888. We’ll talk about the second part later.

Since the utilities were the ones who originally built the transmission and distribution lines connecting generation to consumers, you can imagine they weren’t too eager to let third-party generating plants use their power lines to service customers, even if third-party plants could generate electricity at lower costs. Typical monopoly attitude. So the heart of the Order was to create open access on the transmission and distribution lines to other electricity generating plants as a way to create more competition. This would enable a bunch of independent power producers (IPPs) to sell their electricity into the market, and compete with the big electric utility monopolies, driving costs down and spurring innovation. The utilities then either had to divest and sell their generation plants to third-parties, or perform “functional unbundling”, where they setup walls within their own companies to make sure their own generation-guys had no unfair strategic advantage in transmission and dispatch¹⁶. This is what that looks like (with functional unbundling):

This just means the entire parade is still owned and operated by the same utility monopolies, but they are now required to contract third-party tow vehicles for their horsepower if they could tow at lower costs. These are purchased by the utility at “wholesale” rates, but you as a paying customer in your wagon would never know. You just pay your “retail” rates to the one utility in your area. These retail rates are based upon “fair” rates-of-return on the utilities’ investments, which are approved by the public utility commissions (PUCs). Upgrades to generation and transmission, say with new power lines or clean generating plants, are also approved through long-term planning processes with the PUCs. Here, the retail customers bear the risk of investments because utilities can recover costs through electricity rates, no matter how good or bad the power plants perform. This is how electricity customers in South Carolina ended up paying for nuclear power plants that never got built¹⁷. Utilities also sometimes trade power with neighboring utilities on the wholesale markets if it is more cost effective to do so, such as purchasing excess hydroelectric power and importing via tie-lines¹⁸.

Regulated Electricity Markets

About one-third of all electricity served in the U.S. still operate under this regulated market construct. If you live in any part of the white, unshaded area in the next figure, you are in a traditionally regulated market. In this model, the utility can be an investor-owned utility (IOU), or a state- or municipally-owned utility. Most people in the entire U.S. are served by IOUs.

Deregulated Electricity Markets

The second, more interesting part of Order 888 was to encourage (not mandate) the formation of independent system operators (ISOs) to perform the balancing and dispatching role in an unbiased fashion, hence the “independent” part to the name. They were to do this fairly by establishing a market-based system for electricity generation. This focuses the original utilities to be mostly Transmission & Distribution Service Providers (TDSPs), e.g. just owning and maintaining the tow lines. Even though that sounds much less sexy, there’s a lot involved in transmission and distribution and it’s still got that natural monopoly flex. The formation of several ISOs happened shortly after the Order, followed by the maturation of some ISOs into regional transmission organizations (RTOs) that are just more grown-up versions of ISOs. We now have seven ISO/RTOs in the U.S. that also span into Canada and Mexico. See the map we showed above. Here is a breakdown of the new roles:

It’s starting to look complex, but let’s break it down. The towing service of our parade is now an openly competitive market of tow vehicles. Vehicles that are managed better, run more efficient, spend less on fuel, and therefore operate at lower-costs will make more money and beat out the less competitive, less efficient vehicles. Hooray capitalism. The utilities still have a regulated monopoly on the transmission and distribution tow-lines, without which there will be no parade. And finally, the floats are now the ISO/RTOs that are calling the shots on which vehicles get dispatched. How do they do this fairly? The energy spot markets.

Recall that we have a guy on each float with binoculars looking ahead to predict what horsepower is soon to be needed. For example, they may see an uphill climb on the horizon (say a heatwave or a cold snap…ahem, Texas), and they will prepare day-ahead and hour-ahead forecasts of the horsepower needed to keep the parade going at our constant 60 mph. Based on these forecasts, the owners and operators of the tow vehicles will begin bidding the float guys to use their towing service in the coming day. The float guys generally want to find the cheapest tow service so that end-customers don’t pay more than necessary. So ISO/RTOs will start by awarding contracts to the lowest-cost towing services, and work their way up to more costly towing services until they have satisfied their anticipated horsepower demand. At this point, the market “clears”, and all the contracted tow vehicles get paid the same $/horsepower equivalent to the last bid ($/MW in real life, but both are power units). In this model, the lowest-cost towing services make lots of cash, whereas the medium-cost towing services barely break even, and the most expensive towing services are priced out of the market. This system rewards efficient, low-cost generation, such as renewables (wind/solar) where the marginal cost of electricity is zero (or close to it). Once the contracts are awarded, the vehicle operators know exactly when they need to be in the driver’s seat to hitch in, and approximately how hard they need to tow in their agreed time slot. If vehicle operators who previously promised they’d show up become no-shows, are not very good drivers, or have unreliable vehicles, the float guys will penalize them and make it harder for them to win future jobs. This is why wind and solar plants need to have accurate forecasts of wind patterns and cloud cover. The vast majority of the energy supply is accounted for via the day-ahead bidding process. But of course, no one has a crystal ball and we will never be exactly on-target with our predictions. So the same bidding process happens at the hour-ahead mark and sometimes even the 15-minute-ahead mark prior to the actual generation and consumption. While there are only fine adjustments to horsepower at these last-minute time intervals, the contracts are more lucrative because the demand is close to immediate, and only the more agile tow vehicles can play in these markets, such as natural gas peaking plants. This entire cycle happens in a rolling, staggered fashion constantly in the background 24/7. In your time reading this post thus far, some of these auctions have likely already been cleared.

Similar to the spot markets, there is also the ancillary services market. This operates basically the same but instead of selling actual towing services, what is sold is the potential horsepower available to jump in when needed—kinda like an insurance. This is how the spinning reserves we discussed in Part I are monetized, and are used to ensure the frequency does not rise or fall to dangerous levels.

Zooming out to look at longer term stuff, another critical market to increase grid reliability and reduce price volatility are the capacity markets, which are auctions that happen three-years in advance. Here, the generators lock-in the price of power they will make available to the grid operators, whenever they demand and regardless of whether they are produced or not¹⁹. This is kinda like paying AAA road-service in advance for potential future on-demand tow service when you’re suddenly stranded on the road. These markets help generators guarantee revenue while ensuring grid reliability and reduced price volatility. While most grid systems operate capacity markets, ERCOT famously lacks one, and it was one of the primary reasons for the recent power outages in Texas.

Finally, there may also be long-term bilateral contracts between commercial and industrial (C&I) end-consumers with independent power producers (IPPs) using the transmission and distribution service providers (TDSPs). Acronyms aplenty! This is done to reduce investment risk for the generating plants while also ensuring predictable costs for large consumers.

Like I said, people are good at finding ways to make money, but these various markets are efficient and actually play a critical role in the reliability of our grid systems.

Moar dereg!

Finally, in a completely deregulated model, there are “retailers” of residential electricity that are separate from utilities and don’t own any poles or wires. They simply sell the towing service provided by the generators straight to the consumer wagons, and only use the utilities as a provider of transmission and distribution tow lines. The U.K., Norway, Chile, and Australia operate close to this model, and Texas is the only state in the U.S. with this degree of “retail choice”.

So, how unprepared are we?

As we discussed in our last post, a key pillar to making our clean energy transition will be to “electrify-that-shit”—basically transitioning as many things from fossil fuel-burning to being electric, while simultaneously decarbonizing our electricity generation. If we are successful in executing that plan, then our current grid infrastructure is massively under-equipped to handle the inevitable increased electricity demand. Here’s a plot showing historical and projected electricity demand in the U.S. assuming various degrees of electrification²⁰:

As we discussed last time, the transportation sector consumes the majority of its energy in the form of petroleum, but battery electrification is well within our means and already happening. In scenarios where we have deep electrification of vehicles (what we want) by 2050, this would translate to dramatic increases to electricity demand that our current grid is not ready for. Even the grid enhancing technologies we discussed earlier to improve existing transmission lines would be overwhelmed. What will undoubtedly be needed is significantly more construction of transmission lines. Not only does this include upgraded AC transmission lines to relieve worsening congestion, it also includes significantly more HVDC lines that criss-cross the entire country to provide clean power where needed.

Unfortunately, just like how the electric grid originally evolved (locally), the key barriers impeding new construction of transmission lines are local resistance. Unlike natural gas pipelines where the Federal government can permit pipelines and seize lands via eminent domain (for better or worse), siting transmission lines falls largely under state-level jurisdiction²¹. While I find transmission pylons rather sexy, it’s understandable that folks generally don’t want to live next to them given the potential hazards, necessary brush clearing, and constant humming noise that I guess could be annoying, to name a few. This “not-in-my-back-yard” (NIMBY) attitude all-around then makes any long-distance, interstate transmission project (key for a lot of HVDC projects) a nightmare to site and acquire proper permits for, especially given the politics of short-term election cycles.

However, one idea that has been gaining traction is to build HVDC transmission lines along (or buried underneath) interstate highways and railways that already have established Federal “rights of way”, meaning less red tape and no contentious new land seizures required²². Additionally, having high-power lines installed alongside interstate highways will enable the rapid deployment of fast-charging stations for electric vehicles and trucks, which will need significantly higher power draw that would undoubtedly encounter transmission constraints on our existing outdated grid.

As we’ve discussed, in the case of the electric grid—specifically the transmission and distribution of energy—we largely already have the technologies mature and ready for implementation, so long as we can muster the willpower to execute. This is not true for many of the energy storage technologies that we will deep dive into next, which still have many technical hurdles before they can be implemented at wide scale.

Voilà!

And there you have it. You now understand how the grid works—technically and logistically. It’s a beast of a machine, and also a beauty of capitalism. Most of us certainly take it for granted as we flick our switches on and off everyday, not realizing the tremendous amount of engineering and economics behind each watt-hour we consume. While it is a modern marvel, it is of course imperfect, and there’s a lot more work to do especially as we begin experiencing more frequent extreme weather events due to climate change. The recent Texas power outages in February 2021 is a lesson to learn from, and we will talk about that next. Before that though, here’s a glimpse into what that future grid could look like in our parade analogy. We’ll certainly touch upon these things in the future.

---

Uri messed with Texas

Since we’re talking about the grid, it would be irresponsible not to talk about the recent Texas power outages due to Winter Storm Uri. At the end, there were more than 58 lives lost due to the avoidable tragedy, and is therefore not a light subject. It serves as a relevant case study that touches upon a lot of what we have discussed here, while also providing a clear warning of what happens when we under-invest in this critical infrastructure, not just in the Lone Star State. There had also been unsubstantiated knocks on clean energy, and so it’s important to set the record straight.

While trucks, hair, and chicken fried steaks are bigger in Texas, not everything is. As we learned earlier, the independent Texas grid is substantially smaller than the Western or Eastern Interconnection, and is managed by ERCOT. Texas has resisted joining either of the larger interconnections in order to avoid being regulated by FERC, which steps in when electricity is transmitted across state borders as interstate commerce (there are only two links between ERCOT and the Eastern Interconnection, but these are HVDC lines which are not regulated due to a FERC loophole²³). As we learned above, the larger the grid, the more reliable the vibe net. But given the severity of the outages across the entire “ERCOT Island”, in addition to the winter storms also hitting neighboring states, it is unlikely that being more intimately tied-in with the Western or Eastern Interconnections would have significantly alleviated the crisis. Transmission constraints and congestion at the hypothetical tie-lines also would have limited the usefulness for the majority of folks deep in the heart of Texas (longer-range HVDC lines could have helped though). So what went wrong?

In this case, the primary root causes are (1) the lack of weatherization, specifically winterization, of power plants that led to widespread outages, compounded by (2) competing demand of natural gas between electricity generation and residential heating, and (3) the lack of capacity markets in ERCOT.

1. Weatherization: In our parade analogy, the independent owners and operators of the tow vehicles failed to invest in snow chains or anti-freeze. With the typically balmy Texas climate, independent power producers did not see economic justification to make these improvements. When the cold snap hit, power plants across the board, especially thermal plants (natural gas, coal, nuclear), failed. Frozen sense lines, frozen steam-related pipes and valves, and frozen condensates in compressed air lines can cause an entire thermal plant to trip²⁴. Lack of blade-deicing on wind turbines can cause them to freeze-in-place. Cloud cover and snow cover on solar panels will obviously reduce power output. While Winter Storm Uri was particularly bad, this was not the first time Texans saw power plant failures due to lack of winterization. In 2011, a similar cold weather event caused state-wide power outages, after which FERC and the North American Electric Reliability Corporation (NERC) recommended that Texas power plants be properly weatherized for future events. Of course, ERCOT did not mandate such upgrades, and instead made them voluntary guidelines only²⁵. NERC is now developing mandatory standards for winterizing power plants. As we experience more extreme-weather events due to climate change, climate resilience will be key to ensuring security in our critical infrastructure.

2. Natural Gas Supply: Natural gas is one of the few fuels used in both electricity generation and heating for residential homes (and industry, but that’s less relevant in this example). Given that almost half of Texas homes use natural gas for heating²⁶, there was strong competing demand for natural gas in an event such as Winter Storm Uri, especially when natural gas dominates the majority of the Texas power generation mix. Additionally, natural gas must be pumped by compressors through pipelines, many of which lost power in the rolling blackouts, further constraining supply.

3. Lack of Capacity Markets: As mentioned in Part I and II, many grid operating systems have contingency plans in case there are generation failures or unanticipated spikes in consumption demand—both of which happened in Texas. While most ISO/RTOs across the country operate capacity markets, ERCOT never bothered to create one because it had always enjoyed excess electricity supply—barring occasional catastrophes. After the 2011 power outages 10 years ago, NERC and FERC jointly concluded:

   “the massive amount of generator failures that were experienced raises the question whether it would have been helpful to increase reserve levels going into the event. This action would have brought more units online earlier, might have prevented some of the freezing problems the generators experienced, and could have exposed operational problems in time to implement corrections before the units were needed to meet customer demand.”²⁷

   These comments issued a decade earlier are disappointingly applicable to February 2021.

Chain of Events

When the winter storm started, the concurrence of power plant failures and the increased electricity demand caused the ERCOT grid frequency to dip well below 60 Hz. No operating reserves were available to come to the rescue, especially due to a lack of capacity markets. Other power plants were incapable of starting due to improper winterization, despite how much they wanted to…wholesale spot market prices skyrocketed 10,000% to record values >$9,000/MW (pre-storm values were < $50/MW)²⁸ due to the supply glut. With the inability to satisfy demand and experiencing 30,000 MW of simultaneous outages, ERCOT balancing authorities were forced to implement “rolling blackouts”, by disconnecting various sections of the grid to reduce demand. Not doing so would otherwise have resulted in much wider-spread damage to equipment throughout the entire grid, and extended the power outage to several months. The president of ERCOT said they were only “seconds to minutes” away from catastrophic equipment damage²⁹. While some were quick to blame freezing wind turbines, the fact is that power plant failures were overwhelmingly due to thermal plants, particularly natural gas plants. Wind and solar combined constituted only 8% of the total anticipated electricity surplus in the winter months, as published by ERCOT in November 2020³⁰ and were not significant contributors to the power outage crisis. The following figure shows the actual power generation in Texas during Winter Storm Uri, which shows the degree to which natural gas plants dominated the Texas power generation mix, and their significant dip after the storm started³¹. One nuclear power plant also tripped, along with several other coal plants.

Finally, the the crazy high spot prices on the wholesale markets translated into mind-blowing electricity bills for residential end-customers of certain electricity retailers. While Texas’ “retail choice” should in theory reduce overall rates for consumers due to increased free-market competition, retailers are free to be creative with how they price electricity and leave those in the fine print. Most retailers sell electricity at a flat rate, but some choose to base their rates on the fluctuating spot prices of wholesale electricity. When Texas homeowners are signing up for utilities upon move-in (often hurriedly), they may miss these details, not to mention wholesale electricity spot markets aren’t really something normal people talk about over brunch or are well-versed in. So when wholesale spot prices rose to >$9,000/MW, many average home-owners who were lucky to not have lost power were then surprised with electricity bills totaling more than $15,000 in some cases. This is market failure.

---

W(way)TL;DR

• Our “electrify-that-shit” bucket relies on a strong electric grid foundation that needs upgrades.

• The grid is a ginormous machine that is just a massive vibe net of electrons. It’s really not that complex and can be understood with stuff from Physics 101.

• The grid also forms the foundation of critical markets that are running 24/7 in the background. We use a giant parade analogy to learn about the various players.

• The recent Texas power outages were avoidable and tragic due to three primary root causes that does not include frozen wind turbines. We should learn from this to invest in grid infrastructure.

• Energy storage is important, and we’ll deep dive into batteries in the next post.

---

Thanks for sticking with me to the end, and I look forward to continuing our jam-sesh in the next deep dive. If you want to show some big pic energy, please share with your nerd friends to subscribe:)

Thanks and much love,

Adrian

---

Corrections:

1. In the original post, I miswrote that ERCOT lacked ancillary service markets instead of capacity markets. Texas does actually have the former, but does not have the latter, which was a key reason for reduced grid reliability. My B. Thanks Joyce for pointing it out :)

1. https://www.quora.com/In-a-typical-AC-circuit-how-far-does-each-electron-travel-along-the-wire-in-one-direction-before-the-polarity-is-reversed-in-a-single-alternating-cycle

2. https://energycentral.com/c/ec/ac-vs-dc-powerlines-and-electrical-grid

3. https://www.brown.edu/Departments/Engineering/Courses/ENGN1931F/AC_Electrical_Generators_ASOPE.pdf

4. https://instrumentationtools.com/three-phase-ac-generators/

5. https://www.ge.com/power/gas/generators/water-cooled-gigatop-4-pole

6. https://info.ornl.gov/sites/publications/Files/Pub57419.pdf

7. http://www.caiso.com/TodaysOutlook/Pages/index.html

8. https://www.eia.gov/todayinenergy/detail.php?id=1710

9. https://www.energy.gov/sites/prod/files/2017/01/f34/Maintaining%20Reliability%20in%20the%20Modern%20Power%20System.pdf

10. https://www.electranet.com.au/our-approach/safety/transmission-lines/

11. https://www.energy.gov/sites/prod/files/2020/10/f79/2020%20Congestion%20Study%20FINAL%2022Sept2020.pdf

12. https://sustainableferc.org/new-analysis-midwest-and-southern-leaders-are-letting-crucial-clean-energy-projects-slip-away/

13. https://www.volts.wtf/p/transmission-week-why-we-need-more

14. https://www.eia.gov/todayinenergy/detail.php?id=27152

15. http://sites.utexas.edu/energyinstitute/files/2016/09/UTAustin_FCe_History_2016.pdf

16. https://www.nap.edu/read/5482/chapter/3#9

17. https://www.postandcourier.com/business/sce-g-customers-to-pay-another-billion-for-failed-nuclear/article_5e8a6fb2-fe1b-11e8-8737-3783cd65f48f.html

18. https://www.rff.org/publications/explainers/us-electricity-markets-101/

19. https://www.reuters.com/article/us-usa-weather-power-prices-explainer-idINKBN2AG2KD

20. https://www.nrel.gov/docs/fy18osti/71500.pdf#:~:text=The%20Reference%20scenario%20has%20the,of%20total%20consumption%20by%202050.

21. https://www.volts.wtf/p/transmission-week-why-we-need-more

22. https://thehill.com/opinion/energy-environment/531793-a-transportation-infrastructure-and-climate-priority

23. https://www.tdworld.com/disaster-response/article/21156772/potential-policy-outcomes-from-texas-power-outages

24. https://www.powermag.com/prepare-your-gas-plant-for-cold-weather-operations/

25. https://www.texastribune.org/2021/02/17/texas-power-grid-failures/

26. https://www.reliant.com/en/residential/ac-and-heat/gas-vs-electric-heating-systems-in-houston.jsp

27. https://crsreports.congress.gov/product/pdf/IN/IN11608

28. https://www.reuters.com/article/us-electricity-texas-prices/texas-wholesale-electric-prices-spike-more-than-10000-amid-outages-idUSKBN2AF19A

29. https://thehill.com/homenews/state-watch/539501-texas-power-grid-was-seconds-and-minutes-from-total-failure-officials#:~:text=Bill%20Magness%2C%20president%20of%20the,uncontrolled%20blackouts%20lasting%20for%20months.

30. https://abcnews.go.com/Politics/republicans-texas-power-outages-spread-false-claims-green/story?id=75947664

31. https://www.nytimes.com/interactive/2021/02/19/climate/texas-storm-power-generation-charts.html