This document discusses dynamic forward pricing as a way to provide price signals for the smart grid. It argues that centralized optimization by system operators is too complex and won't scale for billions of devices. Dynamic forward pricing publishes binding prices for 5-minute intervals ahead of time that smart devices and generators can use to decide when to consume or supply power. As conditions change, prices are adjusted to balance supply and demand, allowing the grid to integrate intermittent resources like solar and wind at large scale.
1 of 11
Downloaded 17 times
More Related Content
Smart Price Signals For The Smart Grid UCGEC
1. Smart Price Signals for the Smart Grid:
Dynamic Forward Pricing
1
US-CHINA GREEN ENERGY COUNCIL (UCGEC)
SMART GRID SEMINAR
PA L O A LT O , C A L I F O R N I A
APRIL 29, 2009
E D W A R D G . C A Z A L E T, P H D
THE CAZALET GROUP
AND
M E G A W AT T S T O R A G E F A R M S , I N C .
W W W . C A Z A L E T. C O M
E D @ C A Z A L E T. C O M
? 2009 Edward G. Cazalet
2. The Internet and the Electric Grid
2
The electric grid connects hundreds of
g
millions of customers and meters and
billions of devices.
Many of these devices will soon be
inter-connected on the Internet.
By about 2012 most of California will
have communicating smart meters.
The Smart Grid is the merger of the
Internet and th electric net.
It t d the l t i t
? 2009 Edward G. Cazalet
3. Smart Grid Price Signals
3
Everybody¡¯s vision of the Smart
Grid uses ¡°price signals¡± to
devices.
But what are these price
signals?
Typically, they are so undefined
that they might as well be
smoke signals.
Proposing practical price
signals to operate the Smart
Grid is the purpose of this talk.
? 2009 Edward G. Cazalet
4. Power Price Volatility
4
Dec 2006
CAISO
Retail l t i
R t il electric prices t d
i today 5-min
5 min
prices
are mostly static (either
flat or time-of-use prices.)
But, five-minute Solar
w olesale ea t e power
wholesale real-time po e
prices are highly volatile.
With large amounts of
wind and solar, price
volatility will increase. Wind
? 2009 Edward G. Cazalet
5. Long-Term and Real-Time Pricing
5
The Smart Grid needs real-time retail and wholesale prices.
Customers and suppliers want long-term contracts.
Solution :
? Blocks of power at long-term prices for each time of day and season.
?PPower above or b l
b below bl k amounts i transacted at d
block is d dynamic real-time
i li
prices.
Full incentives to respond to real-time prices and
participation in real time transactions is voluntary.
voluntary
Both restructured and vertically integrated power systems
can implement retail real-time pricing.
p p g
? 2009 Edward G. Cazalet
6. System Operator Pricing
6
System Operators (ISOs
and RTOs) use complex
centralized optimization
and bids to compute prices.
? Real-time prices are published
several minutes to h
li hours or d
days
after each 5-minute interval
passes.
? T act on the real-time price you
To t th l ti i
have to bid into the ISO¡¯s real-time
markets and have the ISO control
your device
device.
? 2009 Edward G. Cazalet
7. Complexity and the Smart Grid
7
ISO optimization systems are straining
under the complexity of their current
mission.
? New ISO systems in California and Texas cost
hundreds of millions each and ten years to build.
? Even the most powerful computers are not
enough.
Centralized control of billions of
customer devices by an ISO is
impossible.
? The Smart Grid will fall short from the cost of its
own complexity if we attempt central control.
? 2009 Edward G. Cazalet
8. Dynamic Forward Pricing
8
Forward real-time pricing provides binding offer prices for each 5-
minute interval published by an operator or virtual operator before
each 5-minute interval.
? Smart devices and generators, buy or sell at these p
g y prices and the transaction is binding.
g
? Customers and devices choose whether and how to respond to prices.
As grid conditions change the forward prices increase or decrease to
balance supply and demand for the 5-minute interval.
The ISO or virtual operators manage the price adjustment process
in response to changes in device schedules to balance supply and
demand within grid constraints using g
g g grid-wide data and models.
? Instantaneous loads only need the current forward 5-min price.
? Inertial loads and generation such as heating, cooling and storage need a set of forward
prices.
? 2009 Edward G. Cazalet
9. Smart Responses to Dynamic Forward Pricing
9
120 $12.00
Air Conditioner Operation
$10.00
$10 00
Smart air conditioner
S i di i 100
thermostats frequently receive
$8.00
80
Deg F
the current set of forward prices.
$6.00
60
$
$4.00
? Smart thermostats efficiently cool the 40
$2.00
building taking into account, weather, 20 $-
building thermal inertia, customer
comfort preferences and more
more. 0 $(2.00)
( )
15 15 16 16 17 24 6 12 18 24
As dynamic forward real time Outside Temp F Temp Deg F
prices change, the timing of Hour Beginning Price Per kWh
AC kW
power use changes.
h
? Customers pay or get paid for the
scheduled power changes at the new
prices.
? 2009 Edward G. Cazalet
10. Storage Needs Dynamic Forward Pricing
10
Grid scale
Grid-scale storage is needed for integration of
large amounts of wind and solar.
? Central optimization by the ISO of storage is impossible.
? Dynamic forward prices allow smart storage to decide how
much to store or deliver in each interval and to change
instantly upon receipt of new prices.
Plug-in vehicles with storage need dynamic
forward prices.
? Based on forward prices, a vehicle is charged at the lowest
cost perhaps with excess renewable generation.
? When prices are high the vehicle may sell power to the
grid.
? If vehicles attempt to charge at the same time, prices
hi l h h i i
increase causing some vehicles to be charged at a different
time.
? This is the only practical way to manage millions of
vehicles.
? 2009 Edward G. Cazalet
11. Smart Take Away
11
Dynamic forward pricing
provides is the only practical
coordination signals for the
Smart Grid.
Any other approach will
cause the Smart Grid to fail to
fully achieve its goals and
cost more because of its own
complexity.
? 2009 Edward G. Cazalet