Electricity Basics (and some advanced)

I received my second request for a post! This time the submitter asks for information about electricity, transmission, and how intermittent renewables like wind and solar fit in.

So, the first question:

1. Electricity, for the most part needs needs to be consumed the instant it is produced?

Yes. Storage of electricity can be done in batteries, or with pumped-water energy storage, but these are all just ways of being able to make electricity at some moment later in time. In short, electricity, once produced, is either used immediately or stored. Massive storage is not practical at the moment, so it's used.

2. Wind or Solar electricity is essentially in addition or parallel to the base load, and do little to lessen the use of coal, NG, or nuclear derived electricity!

This bring up an interesting point about electricity production. In the US, we have 60hz electricity. It's made 60hz by the generator design (in the US, Europe and other places use 50Hz power). Thermal power plants, those that burn things to produce power, rely on spinning a turbine in a magnetic field to produce power. The magnetic field is part of the turbine design, and is too complicated for this post to discuss in further detail. The turbine is spun because water, turned into steam by the heat from burning things or other reactions (coal, natural gas, or even heat from fission), expands rapidly from water to steam. It creates pressure, and then pushes through the turbines to spin them. The turbines spin at the exact rate they need to in order to produce 60hz electricity.

If we produce slightly too much electricity, the turbines start spinning slightly faster. To keep the grid at the right speed, electricity production is reduced at plants. If there is too little electricity, the turbines will slow down, and we'll fall below 60hz. There is a constant dance of the power plants and the electricity users to make everything balance. It's mostly automated, and happens very quickly.

What does this have to do with solar and wind? A lot. Solar and wind power output can be predicted, but not perfectly. If we want to maintain a perfect 60hz grid, we need to be able to adjust for wind and solar output. Because, again, electricity is used when it is made, and not stored. Coal and nuclear power plants aren't great at changing how much electricity they produce in a short timescale, so if we are going to have power plants to make the balance necessary, we need hydro and natural gas to account for the variability of the solar and wind. There isn't enough hydro to do that all over the country.

In  other words, if we want to maintain a 60hz grid, we are always going to have some amount of natural gas power plants.

But beyond that little wrinkle, solar and wind power absolutely offset coal-fired power plants. The more solar and wind we have, the less nuclear and fossil fuel power we need, in general.

In practice, do renewables offset much? See the chart below.

US primary energy consumption. Source: eia.gov info

Short version: Wind was about 1.2% of primary energy (primary energy counts burning oil for cars as well), and solar is 0.16%. So wind and solar can replace coal and nuclear, but it barely does currently.

Longer version: We can let the 60hz grid go from exactly 60hz to let it slide between 58 and 62. And then we can fairly easily do away with a lot of other power plants, as long as we have enough wind and solar. Note, however, that there aren't enough good wind sites in the US for this, and solar is currently too expensive and resource-demanding to replace fossil fuels.

3. Electricity is bought and sold just like a commodity?

In some ways, yes, but not exactly! There is a complicated daily bidding process, and several factors are brought into play.

This one is a bit confusing. I'll do my best. Power plants bid on the day-ahead market. They submit their bids to what is typically called an ISO, for Independent System Operator (some places call it differently, like RTO for Regional Transmission Organization. The ISO/RTO looks at the bids, looks at their best guess for power the next day, and then figures out how many of the power plants they need to hire for the day. Those that don't get hired don't actually burn anything or produce power. Those that do get hired, get hired at the rate of the highest bidder. Let's do an example to explain better.

Note that a MWh is one hour of one MW production. So a 600MW plant produces 600MWh in one our, and 1800MWh in 3 hours.

A plant says, "I can produce this many megawatts at this many dollars per megawatt." Power Plant 1 might say, "I can produce 600MW of coal power at $80/MWh." Power Plant 2, "I can produce 1000MW of natural gas power at $100/MWh." Power plant 3, a nuclear power plant, doesn't shut down. They just keep running. They say, "I can produce 1200MW at $0/MWh." Why? Cause they have to run anyways. They are delivering that power at any price. Power plant 4 is an old coal-fired power plant that has already paid for itself, so it's really cheap, and says, "I can provide 300MW at $50/MWh"

Let's assume it is determined that all of the less expensive power plants, along with Power Plant 2, need to run in order to satisfy electricity demand. They want $100/MWh. Power plant 1, despite bidding in at $80 per MWh, gets $100/MWh, nuclear plant 3 also gets $100/MWh, and coal plant 4 also gets $100/MWh.

On another day, it is determined that only enough electricity is needed for power plant 4 (and all the ones who bid below it). So Power plants 1 and 2 do not produce electricity, power plants 3 and 4 each get $50/MWh.

Should inputs become more expensive, then the power plant has to raise its price. Natural gas, for example, became a lot less expensive in the past 5 years. So they now produce electricity for less than a new coal fired power plant would. So they bid in for less.

A bit confusing, right? It gets more complicated than that. This is a great example to show that electricity is not exactly treated like a commodity.

Now what about solar and wind? Pretty much, if solar and wind is produced in the US, it is purchased, pretty much outside the normal bidding system. What happens to the bidding system if all power becomes solar and wind? There probably will still be some version of it, changed to fit the new system!

That's all for now, thanks for reading!

- Jason Munster

Solar 2

Photovoltaic solar cells. Solar PV.

This is not an easy thing to describe. For some, you may want to just skip past the technical section, cause it is pretty technical.

Solar PV: they used to take as much electricity to make as they produced in their lifetime. Now they they produce about 5x as much energy as they take to make, and the time to break-even on emissions compared to our cleanest fossil fuel stations is about 6 years (see Kannan, 2005, Lifecycle Assessment Study of Solar PV). Of all the clean technologies (nuclear power excluded), this is the only one with the potential to supply world energy needs (that is the subject of a later post).

In other words, when you hear some fool saying that solar panels take as much energy to manufacture as they ever produce, they are referring to a specific type of solar cells called thin film. A type that was made in the 70s and 80s and only goes into things like calculators. Feel free to ask them to stop being foolish.

Some Math

The light we see is not a homogenous single color. In fact, the light we see is not even all the light that is coming from the sun. Infrared and UV rays are also light, but we cannot see them at all. All this light is just an electromagnetic wave. The waves have different wavelengths, but the same speed, and so all the different wavelengths travel together. What we see is a blend of a tiny part of the electromagnetic spectrum.

This is the electromagnetic spectrum. Visible light, what we can see, is only a small part of it.

This is the electromagnetic spectrum. Visible light, what we can see, is only a small part of it.

The amount of energy contained in a photon is equal to

where is the wavelength. h and c are Planck's constant and the speed of light, respectively.

Smaller (shorter) wavelengths give more energy. This is easily shown just by plugging a smaller number into the denominator. Stuff in the infrared is long wavelength, and stuff in UV, X-ray, Gamma ray, etc, are really short wavelength.

Technical Stuff

There is no easy way to do this. I am going to use some terminology that most of you all are unfamiliar with.

A Photo Voltaic (PV) solar panel is a sandwich of two materials. The materials are largely the same, with a few key differences. Both are likely made of silicon (processed sand). But each one has very specific impurities put into them, in a process called doping (not the same type that Lance Armstrong does). This doping is incredibly technical, and very skilled chemists are paid a ton of money to figure out how to do it.

Doping

I won't get into specific materials. Some elements cause there to be a shortage of electrons, or a electron hole, in the whole material (p-type semiconductor). Other elements cause an excess of electrons (n-type semiconductor). So you have one material that can accept electrons, and another material that can give electrons. Putting them together (literally stacking them together) makes magic happen. And by magic, I mean quantum mechanics. Which to most people, including many who study it, is no different than magic.

Just because one has more electrons doesn't mean it wants to be nice and share them. The electron in the n-type literally needs to be excited to be shared. And in PVs, what turns the n-type material on is sunlight. More specifically, photons. Photons are particles of light. ("But Jason!" you say, "Isn't light, like, a wave?" to which I say, "It is both a wave and a particle! Please don't ask me why, just accept it.") Photons contain energy based on their wavelength. Shorter wavelength, more energy (see above).

Here's the fun part. It takes energy to make the n-type semiconductor want to party with the p-type semiconductor. There is a threshold level of energy that needs to be met to kick that electron up from the n-type to the p-type. Too little energy, and the photon doesn't get excited enough to go to the electricity-production party. If there is enough energy from the particular wavelength of light to make that electron jump, then is does jump.

But what if there is more than enough? This threshold level is pretty much determined. Any extra energy will be wasted. This is why PV cells are not particularly efficient. There is a huge amount of light that is too low-energy (all of infrared) for the cell to gather any energy from. There is a lot of light that has much higher energy than required to meet the threshold energy as well. The excess energy is wasted as heat. This can be solved by having a multiple junction cell (multiple junctions just means it has a bunch of different width absorption gaps, so it can harness tons of different energy levels in light). It is capable of absorbing more wavelengths of light, increasing efficiency. And since it has multiple junctions, it is also more expensive and complicated to produce.

Hokay, so, what happens next? You have an electron that has jumped the gap. Then you close the circuit by connecting them with a wire. The electron will go home, back to the n-type, and create an electric charge on its way down. That's about it.

New Methods 

Transistors are expensive. Normal glass optics are relatively cheap. The transistors absorb about 20% of the light that hits them. So it would make sense to use cheap optics to focus more light on the transistors and make the transistors small, yes?

One problem with these cells is how warm they get. If they warm up too much, they begin to lose efficiency. Here we see that we have a tradeoff. We want more optics and fewer cells, but if we do this, they get too warm. They stop being efficient. Some scientists and engineers are working on increasing the efficiency of the cells instead, to get more electricity from light. Otherwise, there are clever ways that some mechanical engineers are trying to get around these issues. One new design uses focusing mirrors and liquid cooling to get around this issue.

The pertinent stuff

Everything pertinent, like insolation and weather, was in the last solar article.

Next post: rounding up some of the stray power sources: tidal, geothermal, wave, and then I am pretty much done.

Thanks for reading.

-Jason Munster

Solar Power

Solar power. It comes in two primary flavors: photovoltaics (PV) and concentrated solar power (CSP). The latter is easy. I decided to do solar power this week, and go back to the dams next week. Big picture: CSP is a bridge technology at best; an investment in most places is little more than a show that the investor is serious about green tech. Moreover, not all places are created equal to invest in solar power. Many of the places that offer the best incentives to have solar power (NJ, MA, Germany) are far from the best places to have solar power.

So this time: insolation, what it means, where it happens. And CSP. PV comes later. Cause it involves quantum mechanics.

So, first, solar insolation map, AKA "Where is the sun shining all the time" map.

Solar power resources in the US. Darker colors indicate better regions for solar power.

Solar power resources in the US. Darker colors indicate better regions for solar power.

Who is not surprised that Alaska is awful for solar power? But check out MA and NJ. Why are they giving tax breaks to install solar cells? Easy answer. To drive the technology forward. Solar panels are really useful in places without any other power source. Like small villages in Africa and other depopulated places. California also has big incentives to build solar, and at least that makes sense, yes?

What determines how much insolation a place gets? Well, you need sun to have solar power. The sun doesn't come out to party at night, so no solar power. A huge one is how much atmosphere the sunlight has to pass through on the way to the the solar panel. More atmosphere means more absorption and dispersing of sunlight (the atmosphere reflects, absorbs, and spreads out sunlight). So higher elevation, like mountains, helps. Less atmosphere. On a related note, the latitude is also very important. Far northern places don't get as much sun annually (Canada, Alaska). Finally cloud and moisture make a huge difference. If there are clouds or moisture in general, sunlight is blocked. This explains most of the east coast of the US, as well as why Nevada, a giant desert, has great insolation. It has a high elevation, and no moisture to make clouds or block sunlight.

The equivalent amount of sunlight hitting the earth at a high latitude spills out over a larger area. In other words, there amount of energy per area is lower. link

The equivalent amount of sunlight hitting the earth at a high latitude spills out over a larger area. In other words, there amount of energy per area is lower. link

CSP is easy. There are a bunch of mirrors, either flat or parabolic (to focus the light even more intensely), and they reflect light to a single point. It produces heat and and then that heat is used to make steam and drive a turbine, just like the basic thermal power plants we have discussed. The heat is typically stored in molten salt, cause it can store a whole lot of energy before it rises a degree in temperature (kind of like water). The heat from this molten salt is slowly released to make that steam for the thermal part.

CSP in action. Lots of light reflected to a single point that then gets very hot. link.

Given that some places on Earth receive upwards of 500W/m directly to the surface (assuming no clouds, no pollution, and daytime), a CSP plant that is 500m*500m could produce 125MW of power. Sounds great, right? 'Cept we know from basic thermodynamics that a thermal power plant that this thing is likely going to be 30% efficient. So something with a quarter of a square kilometer footprint might produce 40MW of power.

So why don't we use this? First, the depiction above is too rosy a picture. CSP is not all that efficient, because if you look at the picture above, you see that not all the area is used for gathering light. There are plenty of empty spaces. Moreover, the transfer of heat from the salt to water is not very efficient. Cause the high temperature and low temperature of the Carnot cycle are closer together (review the thermal power plant post for a review of Carnot efficiencies for all heat engines). Finally, this stuff is expensive. It is easily 2x as expensive as almost any other power technology (other than PV). It requires water to clean the mirrors and has other maintenance costs, the mirrors themselves are quite expensive, and the entire design is expensive. And, if you want to harness the power of the sun, there are better alternatives. Like PV.

As you can tell, I don't have a very high opinion of CSP. Why is that? Take a look at this guy again:

U.S. average levelized costs (2011 $/megawatthour) for plants entering service in 2018
Plant type Capacity factor (%) Levelized capital cost Fixed O&M Variable O&M (including fuel) Transmission investment Total system levelized cost
Dispatchable Technologies
Conventional Coal 85 65.7 4.1 29.2 1.2 100.1
Advanced Coal 85 84.4 6.8 30.7 1.2 123.0
Advanced Coal with CCS 85 88.4 8.8 37.2 1.2 135.5
Natural Gas-fired
Conventional Combined Cycle 87 15.8 1.7 48.4 1.2 67.1
Advanced Combined Cycle 87 17.4 2.0 45.0 1.2 65.6
Advanced CC with CCS 87 34.0 4.1 54.1 1.2 93.4
Conventional Combustion Turbine 30 44.2 2.7 80.0 3.4 130.3
Advanced Combustion Turbine 30 30.4 2.6 68.2 3.4 104.6
Advanced Nuclear 90 83.4 11.6 12.3 1.1 108.4
Geothermal 92 76.2 12.0 0.0 1.4 89.6
Biomass 83 53.2 14.3 42.3 1.2 111.0
Non-Dispatchable Technologies
Wind 34 70.3 13.1 0.0 3.2 86.6
Wind-Offshore 37 193.4 22.4 0.0 5.7 221.5
Solar PV1 25 130.4 9.9 0.0 4.0 144.3
Solar Thermal 20 214.2 41.4 0.0 5.9 261.5
Hydro2 52 78.1 4.1 6.1 2.0 90.3
Solar thermal is expensive. And the capacity factor is junk. There are places for it, but those are so few that it is not worth further exploring this technology.
That's it for now. Thanks for reading!
-Jason Munster