LMP The Very Basics

1
LMP -- The Very Basics
Committee on Regional Electric Power
CooperationApril 30, 2002
Arne OlsonEnergy and Environmental Economics,
Inc.
2
Outline

• Economic Dispatch
• Security Constrained Economic Dispatch
• LMP with Transmission Rights
• LMP in a Hydro-based System
• Some Observations

3
What is LMP?

• Locational Marginal Pricing
• Method of determining the value of energy at
  various locations on the grid at specific times
• Method of pricing energy and transmission
• Value of energy is a function of
• Demand for energy at various locations
• Capacity to supply energy at various locations,
  and the marginal cost of doing so
• Ability of transmission system to deliver energy
  from resources to loads

4
The Marginal Cost of Energy

• The marginal cost of energy at a particular
  location is determined by the cost of the last
  unit required to deliver energy to that location

25
20
P 20
15
D
10
5
5
3 Bus Transmission System
Unconstrained System
B
400 MW Gen.Capacity
200 MW Gen.Capacity
Deer Creek
A
Wild Run
Thermal Limit 300 MW
(10/MW)
(20/MW)
Key
Load Flow Generation
C
6
Unconstrained Example

• Total load 400 MW
• Bus B 75 MW
• Bus C 325 MW
• Constraint on line A-C (300 MW thermal limit)
• Deer Creek dispatched up to full 400 MW capacity
• Wild Run idle
• Same LMPs at each bus

7
What About Transmission?

• Congestion occurs when a transmission system
  limitation prevents the lowest-cost set of
  generators from being dispatched
• Must dispatch higher-cost resource that does not
  violate the constraint
• Result is that different units become marginal
  at different locations on the grid

8
Effect of Congestion

• Transmission limitations can be
• thermal limit
• voltage limit
• stability limit
• LMPs
• are calculated as the cost of serving an
  additional MW of load at each location
• are equal when transmission system is
  unconstrained
• vary by location when transmission system is
  constrained

9
3 Bus Transmission System
Constrained System
B
400 MW Gen.Capacity
200 MW Gen.Capacity
Deer Creek
Wild Run
(10/MW)
A
Thermal Limit 200 MW
(20/MW)
Key
Load Flow Generation
C
10
Constrained Example

• Constraint on line A-C reduced to 200 MW
• Wild Run dispatched to 62 MW
• Deer Creek ramped down to 338 MW
• Next MW of load at A can be served by Deer Creek
  at 10 per MWh
• Next MW of load at C must be served by Wild Run
  at 20 per MWh

11
Effect of Real Power Flows

• Next MW of load in Area 2 is served by a
  combination of Deer Creek and Wild Run
• Depends on the contribution of each generator to
  congestion on the constrained element
• In this example, a 1 MW increase from A to B must
  be matched with 1 MW from C to B to stay below
  the limitation
• Calculation of LMP at B
• LMPB 0.5 10/MWh 0.5 20/MWh 15/MWh

12
For the Math Geeks

• Economic dispatch is a linear optimization
  problem
• Minimize a1G1 a2G2 anGn
• such that
• Sum(aiGi) Sum (biDi) Demand Supply
• Sum(fiGi) lt TLi Transmission Limitations
• Locational Marginal Prices represent cost of
  serving additional MW of load at each bus
• Shadow prices of each binding constraint indicate
  value of additional MW of transmission capacity

13
LMP with Transmission Rights

• FTRs Firm Transmission Rights, Financial
  Transmission Rights, Fixed Transmission Rights,
  Transmission Congestion Contracts, etc. etc. etc.
• FTRs are financial contracts that entitle holder
  to a stream of revenues (or charges) based on the
  hourly energy price differences between an
  injection point and a withdrawal point
• Financial hedges that provide price certainty to
  Market Participants when delivering energy
• A source of revenue to offset fixed-cost
  contributions by long-term transmission customers

14
FTR Value

• Economic value determined by hourly Day Ahead
  LMPs
• MW schedule (Sink LMP - Source LMP)
• Benefit (Credit)
• same direction as congested flow
• Liability (Charge)
• opposite direction as congested flow
• No strategic blocking value in real-time

15
Energy Delivery Consistent with FTR
Bus B
Bus A
Source (Sending End)
Sink (Receiving End)
LMP 30
LMP 15
Congestion Charge 100 MWh (30-15)
1500 FTR Credit 100 MW (30-15) 1500
16
Energy Delivery Not Consistent with FTR (I)
Bus A
LMP 10
Bus C
LMP 15
Congestion Charge 100 MWh (30-15) 1500
FTR Credit 100 MW (30-10) 2000
17
Energy Delivery Not Consistent with FTR (II)
Bus A
LMP 20
Congestion Charge 100 MWh (30-15) 1500
FTR Credit 100 MW (30-20) 1000
18
Characteristics of PJM-style FTRs

• Defined from source to sink
• MW level based on transmission reservation
• Rights are obligations -- receive revenue whether
  your generator is dispatched or not
• Must pay grid operator if price differential is
  negative
• Financial entitlement, not physical right
• Independent of energy delivery

19
LMP with Hydro

• Hydro has several unique characteristics
• Zero marginal cost
• Ability to store fuel for a day and/or a season
• Enormous dispatch flexibility
• Occasionally binding non-power operational
  constraints
• NW hydro-thermal system runs thermal units at
  baseload and uses hydro to meet peaks

20
Locational Price Differences in a Hydro System

• Two reasons why locational marginal prices are
  relevant even in a hydro-thermal system
• Hydro dispatch has impact on which thermal units
  must run, affecting total system dispatch cost
• Not much of an issue if thermal used only for
  baseload and peaks met exclusively by hydro
• Important in systems where the marginal unit is
  always a dispatchable thermal unit, regardless of
  what other resources exist
• Operators of hydro systems that arent strictly
  run-of-river will have differing views on future
  value of energy during period in which storage is
  feasible

21
Bilateral Schedules with LMP

• Schedule 100 MW from A to B from unit with
  marginal cost of 10
• Pay LMP at B
• Receive LMP at A
• If LMP at A lt 10, generator ramped down and
  energy purchased from a different unit
• Must-run units can be assured of operating by
  bidding zero (thereby acting as price taker)
• RTO West system of balanced schedules has same
  effect

22
Some Observations

• No need for market to know how the power flows
• System operator monitors constraints, dispatches
  and redispatches generators as necessary to avoid
  overloads
• Pricing mechanism black box too complicated for
  market participants to get much use out of
• However, price determination can be treated as
  stochastic process, rather than a deterministic
  process

23
More Observations

• Limits the effect of marginal unit to localized
  area
• In single price model, price for all is set by
  highest unit
• With more price points, effect of high unit is
  limited
• Nodal pricing confines effect to smallest
  possible area

24
But is it a Market?

• In real markets, prices arent calculated with a
  computer, theyre arrived at through trading
• Optimal solution defined by system operator,
  rather than aggregate of individual optima
• Subject of vigorous debate in academics, RTO
  formation discussions
• The Wild Wild West model bilateral trading of
  physical transmission rights at defined
  constraint points, to work with bilateral energy
  market
• Or, is this the only way power markets can work?

25
More Information

• This presentation based on training material
  available from PJM website http//www.pjm.com
  -- click on Training
• Feel free to contact me
• Arne OlsonEnergy and Environmental Economics,
  Inc.arne_at_ethree.com(415) 391-5100