Cheap: If You Ignore $3Trillion

Lessons From The Texas Grid: The Levelized Costs of Electricity (LCOE) is a widely quoted, yet often abused 'statistic.' Failing to understand its pitfalls can lead to absurd conclusions.

“THE MACHINE KNOWS!” - Michael Scott yelled as he drove his car into the lake.

It’s a classic scene from the popular TV mockumentary, The Office (the US version, not the UK version: we can argue later which one is best). For those not familiar, Michael Scott is the dramatic, often well-meaning manger who lacks all self-awareness and common sense.

In that scene, Michael Scott and Dwight Schrute (his bumbling side-kick) are driving to a sales call. It’s the first time Michael has ever navigated with a GPS (aka The Machine). As they approach a fork in the road, the voice on the machine - GPS - says “make a right turn.”

Dutifully, Michael Scott turns hard to the right, drives off of the road and down a boat ramp towards the nearby lake. You know it’s bad when Dwight is the voice of reason. Dwight panics and yells “No! No! No!” trying to convince his boss that “make a right turn” means stay on the road, go up to the fork, then turn right.

“There’s no road!”- Dwight yells as they roll down the boat ramp.

“THE MACHINE KNOWS,” Michael shouts as he drives straight into the lake, flooding the car. In dramatic fashion, both characters escape the car that is now up to its windows in water.

A GPS is, in a nutshell, a model of the world. If a person doesn’t understand or use that model appropriately, it can lead you straight into the lake, so to speak.

Photo by Anthony A on Unsplash

---

An Absurd Assumption

Ever hear someone say that wind is cheap? There’s a really good chance they’re quoting the basic Levelized Costs of Electricity (LCOE). Because the LCOE provides a very simple, easy to use talking-point, it is heavily quoted by the media and by politicians.

But what exactly is it? It’s a model that divides the costs (capital costs, maintenance costs, etc.) by the electricity produced to get a simple average cost.

How can that be misused or misunderstood?

The average costs ignores the real world seasonality and intermittency of renewables. For example, a simple LCOE, using average costs, assumes an average wind output looks something like this:

Let’s check that assumption with a good dose of reality.

We’ll turn now to ERCOT (Electric Reliability Council of Texas). The ERCOT grid covers 75% of Texas’ land mass and supplies 90% of Texas’ power, servicing approximately 26 million customers. It has the most installed wind capacity of any state, totally a whopping 39.4 GW of wind. On an annual average, wind generates 25% of ERCOT’s power. But that average ignores the steep up and downs of wind output. Here’s what ERCOT wind output looks like, in 15-minute intervals.

No surprise, the basic LCOE average looks nothing like the real world.

If you unplug your TV from the wall, how long does it stay on? Not very long (assuming it doesn’t have an internal battery).

It’s not much different with a gird. If a power station drops off below demand, another generator has to instantly pick up the load or the grid will cutoff demand (blackouts, etc). Electricity isn’t magically held in inventory on a grid: electricity has a shelf life of basically zero. The grid is a big, complex circuit and requires electrons to flow in real time. That power to run your TV is generated instantly, even if it is from the chemical storage in a battery.

To keep electricity flowing in real time, the grid needs reliable generators.

There are two broad types of power generators:

• Dispatchable Power Stations: They’re a lot like your car; they have a ‘gas pedal’ and a ‘brake pedal.’ You can make them go ‘faster’ or ‘slower’ to meet changes in demand. Examples include coal and natural gas plants (if they are weatherized and have proper fuel supply).

• ‘Roller-coaster’ Power Stations that are intermittent and weather dependent. Beggars can’t be choosy: you get what you get, when you get it, and there’s not much you can do about it. It’s a lot like a roller coaster. Once you get going, you have a lot of ups and downs you can’t control. That pretty much defines wind and solar. One day the wind drops off and it’s cloudy. Yet the next day, the weather might be ideal and wind and solar might produce near max capacity. At best, you can temporarily take wind or solar offline if there’s too much. However, if you need more wind, there’s no one to call to make the wind blow faster: you just get what you can get.

Photo by Matt Bowden on Unsplash

---

(Wind Speed)³

But isn’t the wind always blowing in a place like Texas? Most of the time; but not all wind is created equal.

The power in wind is highly sensitive to wind speed. It takes a lot more force of nature to move wind faster across the surface of the earth. There’s 8x as much power in 16 mph wind as in 8 mph wind. That’s because the power of wind is function of wind speed cubed. That is why wind’s output is very ‘finicky’ and is highly sensitive to small changes in wind speed.

---

Level 2 Emergency Alert

During the evening of September 6, 2023, ERCOT issued a Level 2 Grid Emergency Alert, warning that Texas was teetering near edge of controlled outages.

That morning, Texas electricity prices floated in the $20-50 range per MW/hr. Yet by 8pm, prices skyrocketed to around $3,800 - $ 5400 her MW/hr (depending on the region).

What happened?

Around 3-4pm, wind power dropped low but solar held fairly steady. Wind started to slowly recover and prices declined. Yet wind couldn’t recover fast enough. By sunset, wind failed to fully recover by the time solar dropped off. Around 8pm, solar was producing at ~ 1% of its installed capacity and wind was struggling to produce around 14% of its installed capacity.

Remember that the grid is a complex circuit? If you start ‘unplugging’ generators, the grid scrambles to find someone to fill the gap.

The customer ends up paying for two ‘grids.’ The first ‘roller-coaster’ grid. It gets massive tax credits allowing renewables to temporarily flood the market, undercutting the competition- but only as long as the weather is good.

When wind and solar drop off, the second ‘dispatchable’ grid strains to pick up the slack. When the weather is good, the dispatchable grid makes room for renewables, forcing many dispatchable stations to run well below capacity. When they have to pick up the slack, they charge more to make up for the lost time.

This is compounded by the fact that the ‘dispatchable’ gird is smaller than it would be without wind and solar since so much money is diverted to renewables. When wind and solar drop off, the dispatchable grid doesn’t have the capacity to easily meet demand, sending prices through the roof. When the grid is that tight, if small amount of dispatchable generation has an emergency and unexpectedly goes offline, the grid can fall into rolling blackouts. And that can create perverse incentives. If just one or two stations go offline…(We don’t want to speculate on that).

Speaking on the September 6th emergency alert, ERCOT’s CEO Pablo Vegas stated:

“High demand, lower wind generation, and the declining solar generation during sunset led to lower operating reserves on the grid and eventually contributed to lower frequency, which precipitated the emergency level 2 declaration.”

---

Exercise In Absurdity.

Remember that the basic LCOE assumes wind output is flat? Unfortunately, a basic LCOE doesn’t consider the costs of battery storage or backup generation needed to smooth out the real world variations. It just assumes you get what you get, and doesn’t consider what to do when the wind drops off.

One of the biggest problems for ERCOT isn’t the day-to-day variations in wind: it’s that seasonal wind production doesn’t line up with seasonal demand. Beggars can’t be choosy. It gets hot in Texas, really hot in the summer with power demand peaking in June, July, and August. Unfortunately, that’s around the same time that wind drops off to the lowest seasonal output.

Last year, wind peaked at 11.7GWhrs in April for an average of 38% of ERCOT demand. By summer, wind dropped off to just ~15 % of demand in August and September, with a September low output of just 5.76 GWhrs. Note: while wind drops in half, demand climbs. This creates a big mismatch between the seasonal demand and seasonal output.

How Much Would It Cost to Smooth Out The Current Seasonal Variation?

Let’s just run with the LCOE assumption that wind is flat and figure out how much battery storage Texas would need to smooth out that seasonal variation. And to keep it simple, we’ll ignore that Texas demand goes up in the summer. Plus, we’ll keep wind’s output at 2022 levels, for an average of 25% of ERCOT demand.

Just to keep ERCOT’s wind flat, Texas needs approx 10,000 GWhrs of storage (with a 20% round trip loss in energy). And that doesn’t begin to address the massive spike in demand in the summer.

So how much would 10,000 GWhrs of storage cost?

California’s Moss Landing is the world’s largest battery storage facility. This July, it commissioned phase 3 of the project, bringing Moss Landing’s combined storage to 3 GWHrs. In the chart below, the green bar represents 10,000% of Moss Landing’s storage. When this chart was first made, 3 GWhrs didn’t show up. So, it was multiplied by 100x to make it visible.

The total price tag for Moss Landing’s 3GWhrs? - well over US$1 Billion. That works out to well north of $300 Million per GWhr of storage. But we’ll round down to a nice even $300 Million. (Someone needs to make wind look good, OK?)

So, here’s the math: $300 Million per GWhr of storage x 10,000 GWhrs of storage = a measly $3 trillion. That’s how much storage Texas would need to meet the LCOE assumption of flat average output! And that just with the current wind output.

For Texas, the basic LCOE of wind power ignores $3 trillion in storage costs!

Now, let’s be crystal clear: That $3 trillion is only for Texas’ current yearly wind output. This scenario only takes current wind output (that peaks ~ 38% in the spring and drops to ~ 15% in the summer) and simply smooths it out, just like the LCOE assumes, to the flat average.

Of course, it’s almost certain that Texas (with a GDP of $2.36 trillion) won’t spend $3 trillion backing up wind to match the LCOE assumptions of ‘average’ output. Batteries are expensive: they need to be used a lot to help cover their high costs. Using them for long term storage is a very UN-economic since they would only be used a few times a year, at most.

But something has to back up the grid when wind drops off. And that something is assumed to be free, under the basic LCOE.

Now, in all fairness, the people who create the LCOE often warn users against comparing intermittent resources with dispatchable ones without at least adjusting for the cost of backup or storage. But who reads warnings? The LCOE provides a soundbite that claims wind is cheap. So, politicians and environmental groups run with it.

This article is the first in a series of articles breaking down how the LCOE is often abused. Failing to understand that the basic LCOE ignores storage can lead to one to miss just a few $trillion in ‘hidden’ storage costs.

Here’s what the LCOE does: To claim wind is cheap, simply ignore the expensive parts.

Coming up, we’ll look at how LCOE can be abused by cherry picking the best in class locations for wind. This overestimates wind’s average output, lowering the ‘average’ costs. In addition, we’ll look at out how cheap wind also relies on low inputs.

The next time you hear someone say that renewables are cheap, here are a few things to keep in mind:

• Does that include in the cost of seasonal storage or backup?

• Does that cherry pick the best in class locations for wind or solar and then assume the rest of the country will get the same output? (Unfortunately, we see a lot of that these days.)

• Is the backup big enough to handle the slack without sending prices through the roof?

• Does ‘cheap’ assume never-ending low input costs for renewables? (Low interest rates, cheap manufacturing costs, etc.)

• How long does it assume the plants will last? (Some LCOE models assume a wind turbine and a nuclear power plant have the same lifespan which is another absurd assumption).

There’s a lot more for us to cover on the LCOE.

But for now, we’ll leave you with a formula on how to claim that wind is cheap:

‘Til next Time!