Intro to electric power

Prof. Richard Sweeney

Second half of the course is about electricity. Why?

• Electricity currently account for 43% of US energy consumption.
• But eventually we want to electrify as much as possible.
  • And switch to low-carbon electricity….
• Also growing demand from AI will be predominantly met with electricity.
• As such we will spend most of the rest of course talking about electricity.

Overview of next few lectures

• Today:
  • Overview of technology and core engineering constraints.
  • What does economic efficiency look like in this context?
• Next class: How electricity markets work
  • Start of electricity market game
• Lecture 3: Market power.
• Lecture 4: Long run investment

Lecture 1: Electric Power Background

• How we generate electricity
• What’s special about electricity?
• Goals for power markets and how they work

Some terminology: Watts vs watt-hours

• A watt (W) is an instantaneous measure of the rate of power flow (equal to 1 joule per second)

• A watt-hour (Wh) refers to one watt of power flowing for an entire hour. This is what you see on your electricity bill.
  • Avg US household consumes 10,000,000 Wh per year.
• Common units of aggregation are kilowatt-hours (kWh = 1000 Wh) and megawatt-hours (MWh = 1,000,000 Wh)

• The capacity of a power plant in is denoted in megawatts (MW). This is the amount of power it can produce when operating optimally at any moment.

• If a 1 megawatt (MW) plant operates optimally for 1 hour, it creates 1 megawatt-hour of electricity (MWh).

  • This would power approximately 800 homes.

Generation Technology

Most electric power generators driven by a turbine

Source: EIA

Coal fired power plant

Source and more info here.

• Average plant is around 600 MW
• C02 emissions of 2.25 lbs per kWh

Natural gas power plant

Source and more info here

• Average plant is around 300-500 MW
• C02 emissions of 0.91 lbs per kWh. (~1/2 coal)

Nuclear power plant

Source and more info here.

• Average plant is around 2,000 MW
• Can still impose massive social costs (Fukushima)
• Waste needs to be stored / processed

Hydroelectric power plant

Source and more info here

• Sizes varies a lot.
• Large plants can be very ecologically disruptive.

Wind turbines

• Average turbine around 3 MW.
• Offshore turbines of 20 MW on the horizon.

Solar thermal plant

Solar photovoltaic (PV) is the exception

Average residential system is 0.005 MW

Some questions

• Which technologies should we use more / less of in the short run?
• What plants should we build more / less of in the long run?
• How does the market make these decisions?
• Is the market mix economically efficient?

Grid

Overview of electric power systems

Source: Wikipedia

• Power plants often located far a way
  
• Transmitting lots of power requires high voltage (dangerous)
• Needs to be “stepped down” to lower voltage to get to customers.

Electricity is different from other goods

• Generators don’t create electrons, they send waves across existing electrons on the grid
  • Flow reverberates to balance differences in potential energy across nodes (Kirchoff’s law)
• Analogy: Electricity like a pond where demanders draw water, suppliers put more in
  • Cannot track “electricity” across the grid
  • From consumer’s perspective, electricity generation is totally undifferentiated (ie it does not matter how it was produced)

Challenge: Supply and demand have to balance perfectly at every point in space and time

• Too much supply/ too little demand causes the frequency to increase.
  • Our appliances are designed to expect a very specific voltage. Even minor deviations can damage them.
  • Large deviations can cause explosions and fires on grid equipment.
  • Implication: Can’t just bring more power unless you know someone wants it.
• Too little supply/ too much demand and the frequency drops.
  • This can lead to a blackout, which precludes anyone from consuming electricity.
  • Contrast that with a typical market (Avocado example)
  • Implications: In order for someone to increase electricity use, a power plant has to concurrently increase supply.

Goal #1 for power grid: Keep the lights on

Demand

Balancing supply and demand is nontrivial because demand varies considerably across time

• When do you think we use more electricity over the course of a day? Year?

Example from New England

Challenge: Electricity not (cheaply) storable

• Many markets have seasonal / fluctuating demand (fireworks)
• Solution is often to produce at a nearly constant rate, and store the good
• Historically, energy storage has been inefficient or prohibitively expensive
  • This is changing! We will talk about batteries in a few weeks

Varying demand and limited storage means we need more plants at some times than others

• Supply and demand must balance at every point in time.
• Implication: There is a plant somewhere that only produces power during the highest demand hour of the year!
• Which plant should that be?

Efficient Dispatch: Use lowest MC plants first

Example dispatch curve, Source: EIA

From short run to the long run

So far we have been talking about efficient dispatch holding the set of plants available fixed.

Questions

1. Why do we have this mix of technologies?
2. What is the “cheapest” technology?

Power production involves both fixed and variable costs

Which technology is cheapest?

If a plant has capital cost \(K\) and variable cost \(v\) (per hour), and it runs for \(q\) hours, it’s total lifetime cost \(C\) is

\[C = K + v*q\]

Dividing \(C\) by lifetime production \(q\), we get the average cost per unity of electricity over the life of the asset, called the Levelized Cost of Energy (LCOE)

\[LCOE = K/q + v\]

[note typically we want to discount]

This is the minimum average price the plant must be paid in order to cover costs.

Which energy tech is “cheapest”?

The answer depends on how often the asset will be used.

Lecture 1 summary

• Electricity is an undifferentiated good, produced by very differentiated technologies
• Supplied in an environment with important engineering constraints
• In the short run, we want to ensure that the plants with the lowest marginal costs produce power every hour.
• In the long run, we want to build plants with the lowest average costs, inclusive of capital costs, that can meet demand.

Electric Power Markets

Goals for the electric power system

1. Keep the lights on! (avoid blackouts)
2. At the lowest possible cost (or price?)

Challenge:

• Demand varies a lot over time.
• Supply must exactly match it.
• Electricity not storable.

Goals for the electric power system (cont.)

Short run:

• Capacity fixed
• Make sure the cheapest options produce power every hour.

Long run goal

• Build capacity mix that matches demand patterns as closely as possible.

How deregulated power markets work

• Private investors put up (a lot of) capital to build a power plants, at their own risk.
  • They sell power to earn revenue, but return is not guaranteed.
• An Independent System Operator (ISO) is given full control over balancing the grid.
  • Have to approve generation requests (solves balancing problem)

ISO-NE control room

Every hour the the ISO holds an auction

• Generators tell the ISO the minimum price they’d need to be to produce power for next hour.
• ISO “dispatches” cheapest generators first until demand is satisfied.
• Every generator that produces gets paid the marginal (ie highest clearing) bid.
  • This is called a uniform price auction.

Example

• Consider a market with following tech:

Fuel	Capacity (MW)	Marginal Cost
Hydro	10	0
Coal	30	40
Gas	20	60

Off-peak equilibrium

Demand = 35 Market clearing price P = 40 (coal’s MC).

Fuel	Cap.	MC	Q	P	Revenue	Cost	Profit
Hydro	10	0	10	40	400	0	400
Coal	30	40	25	40	1000	1000	0
Gas	20	60	0	0	0	0	0

Notes: Hydro earns inframarginal rents; Coal is marginal and breaks even.

Peak equilibrium

Demand = 55 MW Market clearing price P = 60 (gas’s MC).

Fuel	Cap.	MC	Q	P	Revenue	Cost	Profit
Hydro	10	0	10	60	600	0	600
Coal	30	40	30	60	1800	1200	600
Gas	20	60	15	60	900	900	0

Notes: Gas is marginal and breaks even; Hydro and Coal earn inframarginal rents.

Takeaways

• Market price equals the marginal unit’s marginal cost.
• Inframarginal profit: \((P - MC_j) \times \text{MWh}_j\).
• Moving from off-peak to peak raises market price from 40 to 60, expanding inframarginal rents.

Energy Market Game

Based on California electricity markets

• Goals:
  • Understand how bidding and profits work in deregulated markets
  • Explore possibilities for market power
  • Study how a carbon tax works in this market (next week)
• More info:
  • Player guide
  • Game Concepts

7 Generation companies (full portfolio)

A day consists of 4 demand periods

Each period teams submit their bids

What the game does with your bids

• ISO arranges bids into a supply curve for each period.
• Intersects with demand curve to determine market clearing price.
  • Every plant that bids below the market clearing price is dispatched.
• If a plant is dispatched:
  • it earns revenue = price * MWh produced.
  • it incurs cost = variable cost * MWh produced.
• If a plant is not dispatched, it earns no revenue and incurs no variable cost.
  • It may still have fixed O&M costs.

Later we will introduce congestion

This splits CA into two markets during high demand.

Getting started

• Take a minute to figure out the best way to keep track of things with your team
  • It may be useful to start a new google sheet
• Look at the “Plant List” tab in the google sheet I sent you. This has all the information about your genco’s plants.
  • Look at your genco portfolio and make a table of your plants, their capacities, and their marginal costs.
  • MC = varom + fuelcost

Example game (In Class - Baseline)

• Go to the Class Bidding Page I emailed.
  • Click on “Game Links”.
  • Go to the “In Class” section.
  • Click the dropdown for “Baseline”.
  • From there click on the “Open Dashboard” link.
• Go to the “place bids” tab, and confirm the default bids match these marginal costs.

Example game - Market Clearing

• Leave bids as is (= MC)!
• I am going to solve the game, and let’s walk through each period in turn.

For each period, make note of:

• Which of your plants produced power? (were dispatched)
• What price did they get?
• What profits did they earn?

Example game 2 – Teams set bids

• Now go to the “New Demand” dropdown, and click your team’s “Open Dashboard” link.
• Take 5 minutes to discuss with your team what bids you want to set for each of your plants in each of the 4 periods.
• Goal: Earn as much profit as possible over the day.

Playing for real

• Review the Energy Market Game assignment

At Home - Assigned Gencos

For this round, each team will be in charge of the Genco they were assigned. Bids will be submitted by 9 pm on the following days:

• Thursday 10/22: Baseline
• Friday 10/23: New Demand
• Sunday 10/25: Congestion

You should be able to review the results of that game on the website by 10 pm (first thing the next morning at the latest).

Group Deliverable 1 Due in Class Next Tuesday

Write a short memo (1 page) or slides (3-4) summarizing the first three days of the game. Must include:

• What was your strategy for bidding in each round?
• Did things work out as you expected? Why or why not?
• Which of your plants was most profitable? Which was least profitable? If you could get rid of one plant, which would it be?

Come prepared to discuss your strategy and results in class.

Lecture 3: Long run investment

From short run to the long run

So far we have been studying outcomes conditional on the set of plants available.

Questions for today

1. Why do we have this mix of technologies?
2. What is the “cheapest” technology?

Review your plant performance so far

• Which of your plants produced the most power?
• Which plants made the most money?
• Did any plants lose money?
• If you could get rid of one of your plants, which would it be?

What does LCOE miss?

• If we want to know what technology is “best”, can we just look at LCOE?
• What important factors are not included in LCOE?

• Externalities (eg pollution)
• Grid integration costs (eg need for backup, transmission)
• Risk / Volatility
• Other factors (eg land use, water use, social acceptance)

In many studies, renewables look cheapest

Why don’t renewables just take over the grid? To increase the share of renewables on the grid, do they just need to be cheaper than fossil LCOE?

Module 1 wrapup

• Electricity is an undifferentiated good, produced by very differentiated technologies
• Supplied in an environment with important engineering constraints
• In the short run, markets allocate production to plants that bid the lowest.
• Firms earn money by being “inframarginal”
• If markets are competitive, this will keep lights on at lowest private cost
• In long-run, plants need to be paid at least their LCOE to make money.
• Next up: social cost

Backup