Market Power Evaluation in Power System with Congestion

Market Power Evaluation in Power System with Congestion

1. Market Power Evaluation in
Power Systems with Congestion
Tom Overbye, George Gross, Peter Sauer
Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
Urbana, IL
Mark Laufenberg, Jamie Weber
PowerWorld Corporation
Urbana, IL

2. Introduction
•Power industry is rapidly restructuring
•Key goal of restructuring is to reap benefits
of competitive marketplaces
•Significant concerns benefits could be lost
through development of horizontal market
power

3. Horizontal Market Power
•Market power is antithesis of competition
–ability of a particular group of sellers to
maintain prices above competitive levels
•An extreme case is a single supplier of a
product, i.e. a monopoly.
•In the short run, Price monopolistic
producer can charge depends upon price
elasticity of the demand.

4. Horizontal Market Power
•Market power can sometimes lead to decreased
prices in the long run
–Accompanying higher prices can result in a
quickening of the entry of new players and
technological innovation
•Some market power abuses are actually self-
inflicted by consumers by their reluctance to
respond to favorable prices offered by new
vendors in deregulated markets

5. Symptoms of Market Power
•Economic theory tells us that in a market with
perfect competition, prices should be equal to
the marginal cost to supply the product
•Therefore prices above marginal cost can
indicate market power

6. Market Power Analysis
•Market power analysis requires 3 steps
–identify relevant product/services
–identify relevant geographic market
–evaluate market concentration

7. Relevant Product
•FERC defines at least three distinct products
–non-firm energy
–short-term capacity (firm energy)
–long-term capacity
•Emphasis shifting to short-term energy
markets
•Presentation considers short-term
•Challenge in electricity markets is demand
varies over time

8. Relevant Geographic Market
•Most difficult step in electricity market due to
impact of transmission system
•Size of market is dependent on
–competitive prices of generators
–impacts of charges from transporting energy in
transmission network
–physical/operational characteristics of transmission
network

9. Herfindahl-Hirshman Index
(HHI)
•HHI is a commonly used methodology for
evaluating market concentration
•where N is number of participants
•q
iis percentage market share


N
i
iqHHI
1
2

10. HHI Examples
•For monopoly HHI = 10,000
•If N=4, q
1=40%, q
2=25%, q
3=25%, q
4=10%,
then HHI = 2950
•DOJ/FTC standards, adopted by FERC for
merger analysis
–HHI below 1000 is considered to represent an
unconcentrated market
–anything above 1800 is considered concentrated

11. Market Power Without
Transmission Considerations
•If transmission system is ignored, market
power depends only on concentration of
ownership relative to other producers in
interconnected system
•Without considering any constraints (using
NERC 1997 peak data)
–Eastern Interconnect HHI = 170
–ERCOT HHI = 2415

12. Market Power with Transmission
Charges
•In determining geographic market, FERC
requires that suppliers must be able to reach
market
–economically
•supplier must be able to deliver to customer at cost no
greater than 105% of competitive price to customer
•delivered cost is sum of variable generation cost and
transmission/ancillary service charges
–physically

13. Pricing Transmission Services
•Goal is to move energy from source to sink
•A number of different mechanisms exist;
examples include
–pancaking of transmission service charges along
contract path
–establishment of Independent System Operator
(ISO) with single ISO-wide tariff

14. Market Power with Transmission
Constraints
•Market size can be limited by physical ability to
delivery electricity
•Whenever physical or operational constraints
become active, system is said to be in state of
congestion
•Congestion arises through number of mechanisms
–transmission line/transformer thermal limits
–bus voltage limits
–voltage, transient or oscillatory stability

15. Radial System with Market PowerLine Limit = 100 MVA
Bus A
300.0 MW
100%
200.50 MW
99.5 MW
Rest of
Electric
System
100 MVA limit on line limits
bus A imports to 100 MVA
Models the
remainder
of the
electrical
system

16. Networked SystemLimit = 100 MVA
Limit = 100 MVABus A
300.0 MW
25%
175.00 MW
25.0 MW
Rest of
Electric
System
100%
100.0 MW
Analysis is
substantially
more complex.
Transfer
capability
into bus A
is NOT equal
to sum of
tie-line limits

17. Three Bus Networked Example
Imports = 74 MWBus B
Bus C
Bus A
300.0 MW
100%
226.00 MW
99.6 MW
26%
25.7 MW
23%
300.0 MW
50.0 MW
224.4 MW
324.0 MW
25 MWs of power is wheeling through bus A
In this
example the
allowable
interchange
is less than
limit either
line

18. Congestion in Networks
•Need to introduce several definitions
concerning network power transfers
–source: set of buses increasing their injection of
power into network
–sink: set of buses decreasing their injection of
power into network
–direction: source/sink pair
•Power transfer is then associated with a
particular direction

19. Congestion in Networks
•To understand impact of congestion in
networks, need to consider two interrelated
issues
–power transfer in a particular direction may impact
line flows in large portion of system
•this impact is commonly defined as the power transfer
distribution factor (PTDF)
–once a line is congested, any new power transfers
with a PTDF on the congested line above 5% can
not take place

20. Nine Bus, Nine Area ExampleA
G
B
C
D
E
I
F
H
400.0 MW 400.0 MW 300.0 MW
250.0 MW
250.0 MW
200.0 MW
250.0 MW
150.0 MW
50.0 MW
39%
Each area contains one bus/one 500 MVA generator.
Each line has 200 MVA limits. HHI = 1089
Pie charts
show
percentage
loading
on lines
Figure
shows
base case
flows

21. PTDF Values for A to I DirectionA
B
C
D
F
E
G
H
44%
56%
30%
13%
10%
20%
10%
2%
32%
35%
34% 34%
34%
PTDF show
the incremental
impact on
line flows, in
this case for
a transfer from
area A to area
I.
Pie charts now show the percentage PTDF
value; arrows show the direction.

22. PTDF Values for G to F Direction A
B
C
D
F
E
G
H
6%
6%
18%
12%
6%
12%
6%
19%
61%
20% 21%
21%
Note that
for both the
A to I and
the G to F
directions
almost all
PTDFs are
above 5%
Example: For 200 MW transfer from G to F, line
H to I MW flow will increase by 200*21%=42MW

23. AEC
DPL
JCP&L
AE
KACP
CILCO
DECO
DUKE
LGEE
IP
IPL
PECO
NI
STJO
SWEP
ALTE
NYPP
PSE&G
VP
SOUTHERN
AMRN
TVA
CPLW
CPLE
HE
WERE
SWPA
MIPU
SIPC
CIN
OVEC
DLCO
IMPA
PENELEC
PJM500
SCPSA
SCE&G
ENTR
EEI
DOE
NEPOOL
ONT HYDR
NIPS
CWLP
FE
AEP
PP&L
PEPCO
METED
LEPA
OMPA
PSOK
EMDE
SPRM
GRRD
ASEC
KACY
INDN
DPL
BG&E
BREC
SIGE
EKPC
CONS
NSP
MEC
SEPA-RBR
SEPA-JST
OPPD
HARTWELL
YADKIN
WEP
WPS
MGEDPC
MEC
ALTW
KAMO
NPPD
OTP
MPW
SMP
6%
24%
14%
5%
13%
17%
8%
19%
12%
9%
6%
38%
19%
6%
8%
14%
15%
7%
23%
7%
21%
22%
9%
9%
6%
6%
11%
5%
7%
24%
20%
15%
32%
5%
5%
5%
7%
16%
11%
9%
45% Large Case PTDF Example:
Direction Southern to NYPP
Figure shows the area to area interface PTDFs
Pie
charts
show
percentage
PTDF on
interface

24. Riverhead
Wildwood
Shoreham
Brookhaven
Port Jefferson
Holbrook
Holtsville
Northport
PilgrimSyosset
Bethpage
Ruland Rd.
Newbridge
Lcst. Grv.
07MEROM5
KEYSTONE
01YUKON
CONEM-GH
JUNIATA
SUNBURY
SUSQHANA
WESCOVLE
ALBURTIS
HOSENSAK
BRANCHBG
ELROY
WHITPAIN
LIMERICK
DEANS
SMITHBRG
3 MILE I
RAMAPO 5
HUNTERTN
CNASTONE
PEACHBTM
KEENEY
BRIGHTON
W CHAPEL
CLVT CLF
CHALK500
BURCHES
8POSSUM
8OX
8CLIFTON
8LOUDON
08MDWBRK
8MORRSVL
8MT STM
8VALLEY
8DOOMS
8BATH CO
8LEXNGTN
8NO ANNA
8LDYSMTH
8ELMONT
8MDLTHAN
8CHCKAHM
8CARSON
8SEPTA
8YADKIN
8FENTRES
8SURRY
8PERSON
8MAYO 1
8PARKWOD
8WAKE
8PL GRDN
8CUMBERL
8RICHMON
8MCGUIRE
8JOCASSE
8BAD CRK
8OCONEE
8NORCROS
8BULLSLU
8BIG SHA
8BOWEN
8KLONDIK
8UNIONCT
8VILLA R
8WANSLEY
8SNP
8WBNP 1
8ROANE
8BULL RU
8VOLUNTE
8SULLIVA
8PHIPP B
05NAGEL
8WILSON
8MONTGOM
8DAVIDSO
8MARSHAL
8SHAWNEE
8JVILLE
8WEAKLEY
8JACKSON
8SHELBY
8CORDOVA
8FREEPOR
WM-EHV 8
8UNION
8TRINITY
8BFNP
8LIMESTO
8BNP 2
8MADISON
8BNP 1
8WID CRK
8RACCOON
8FRANKLI
8MAURY
8MILLER
8LOWNDES
8W POINT
MCADAM 8
8S. BESS 8SCHERER
8ANTIOCH
8CLOVER
ROCK TAV
COOPC345
ROSETON
FISHKILL
PLTVLLEY
HURLEY 3
LEEDS 3
GILB 345
FRASR345
N.SCOT99
ALPS345
REYNLD3
EDIC
MARCY T1
MASS 765
OAKDL345
WATERC345
STOLE345
LAFAYTTE
DEWITT 3
ELBRIDGE
CLAY
VOLNEY
SCRIBA
JA PITZP
9MI PT1INDEPNDC
OSWEGO
PANNELL3
ROCH 345
KINTI345
NIAG 345
BECK A
BECK B
NANTICOK
MIDD8086
MILTON
TRAFALH1
TRAFALH2
CLAIRVIL
HAWTHORN
ESSA
BRUJB561
BRUJB569
BRUJB562
LONGWOOD
Barrett
E.G.C.
Valley Stream
Lake Success
Rainey
Jamaica
Greenwood
Fox Hills
Fresh Kills
Goethals
Cogen Tech
Gowanus
Farragut
E 15th St.
W 49th St.
Tremont
Shore Rd.
Dunwoodie
Sprain Brook
Eastview
Pleasantville
Millwood
Buchanan
Indian Point
Dvnpt. NK
Hmp. Harbor
Vernon
Corona
Greelawn
Elwood Southern to NYPP Line PTDFs
Color contour of PTDFs on 345 kV and up lines
PTDFs
key

25. PTDF Implications on Market
Power
•Once congestion is present on line, any power
transfer with PTDF above 5% on congested
line, in direction such that line loading would
be increased, is not allowed
•Congestion on a single line can constrain many
different directions

26. Nine bus example -Area I buying
•Table : Line G to F PTDF Values
•Seller to BuyerPTDF for Line G to F
•A to I 35%
•B to I 29%
•C to I 11%
•D to I 5%
•E to I -1%
•F to I -20%
•G to I 41%
•H to I 21%A
G
B
C
D
E
I
F
H
400.0 MW 400.0 MW 300.0 MW
250.0 MW
250.0 MW
200.0 MW
250.0 MW
150.0 MW
50.0 MW
39%

27. Nine Bus ExampleA
G
B
C
D
E
I
F
H
400.0 MW 400.0 MW 300.0 MW
250.0 MW
250.0 MW
200.0 MW
250.0 MW
150.0 MW
50.0 MW
39%
If the line from G to
F were congested,
then area I could
only buy from areas
E, F or I.
When congestion is present, area I load only has
possibility of buying from three suppliers. If we
assume each supplier has 1/3 of the potential
market, resultant HHI is 3333.

28. Strategic Market Power
•Characteristic that congestion can limit market
size allows possibility that generator portfolio
owner may unilaterally dispatch generator to
deliberately induce congestion
–this results in market power
–allows charging of higher prices
•Ability to induce congestion depends on
generator portfolio and transmission system
loading

29. Portfolio Flow Control
•A portfolio of N generators may be
redispatched to unilaterally control the flow on
a particular line, i, by an amount
•where S
ikis sensitivity of line i MW flow to
change in generation at bus k


N
k
gk
N
k
gkiki
PthatsuchPsP
11
0max

30. Portfolio Flow Control
•Once a line is congested, any generators with a
PTDF to a particular load pocket that would
increase loading on the congested line are
prevented from selling to that market.
•Likewise affected loads are prevented from
buying from the “blocked” generators.

31. Merged Areas F and G Blocking
LineA
G
B
C
D
E
I
F
H
400.0 MW 400.0 MW 300.0 MW
250.0 MW
430.0 MW
200.0 MW
70.0 MW
150.0 MW
50.0 MW
21%
Generators F and G are deliberately
dispatched to congest line G to F
With G-F
congestion
area I can
only buy
from FG,
or E

32. Cost to the Congestors
•Such a strategy of deliberate congestion could
certainly involve additional costs to congestors
(since they presumably would have to move
away from an economic dispatch)
•Congestors need to balance costs versus
benefits from higher prices

33. Integrating Economics into the
Analysis
Maximize “Social Welfare”
Include the
Power Flow Equations
Include Limits such as:
* transmission line limits
* bus voltage limits
0d)s,g(x,
0d)s,h(x,
sd
ds,x,

s.t.
-max CB
•The first step to doing this is developing an
optimal power flow
•Lagrange multipliers then used as spot-prices
BenefitsCosts

34. Market Simulation Setup: Get
away from “costs” and “benefits”
•Suppliers and Consumers will submit price-
dependent generation and load bids
–For given price, submit a generation or load levelPrice = p
[$/MWhr]
Demand Bid
[MW]
Price = p
[$/MWhr]
Supply Bid
[MW]
pmin
pmax
ms
md
=
d
B


=
s
C



35. Market Simulation Setup
•Consumers and suppliers submit bid curves.
•Using the bids, an OPF with the objective of
maximization of social welfare is solved
–This will determine the MW dispatch as well as
Lagrange multipliers which will determine the spot
price at each bus.
–The consumers and suppliers are paid a price
according to their bid, but their bid will effect the
amount at which they are dispatched.

36. Limit Possible Bids to Linear
Functions
•Each supplier chooses some ratio above or
below its true marginal cost function
Price = p
[$/MWhr]
Supply Bid
[MW]
pmin
ms
k*pmin
ms
k
“True” Marginal Bid
“k times” the “True”
Marginal Bid

37. What does an Individual Want to
do? Maximize its Welfare
•Maximize An Individual’s Welfare
–Individual may control multiple supplies and
multiple demands
–Note: An individual’s welfare is not explicitly a
function of its bid (implicitly through s,d,l) 

supplies
controlled
demands
controlledi
])([])([)λ(
iiiiiiii ssCddB,,f llds
+Benefits
-Expenses
-Costs
+Revenue

38. Determining a Best Response in
this Market Structure
•A “Nested Optimization Problem”












0d)s,g(x,
0d)s,h(x,
kskd
ds
ds
ds,x,
k
s.t.
,-,max
by determined are )λ(s.t.
)λ(max
CB
,,
,,f
The OPF Problem is
a “constraint” now
Individual’s Welfare
s,d,lare implicit
functions of k
“OPF Sub-Problem”

39. Economic Market Equilibriums:
The Nash Equilibrium
•Definition of a Nash Equilibrium
–An individual looks at what its opponents are
presently doing
–The individual’s best response to opponents
behavior is to continue its present behavior
–This is true for ALL individuals in the market
•This is a Nash Equilibrium
•Nash Equilibrium be found by iteratively
solving to individual welfare maximization

40. •C
i(s
i)=b
sis
i+c
sis
i
2
=suppliercost
•B
i(d
i)=b
did
i+c
did
i
2
=consumerbenefit
Example: Use 9-bus system and
Assign Cost and Benefit CurvesBus
Supplier bsi
Coefficient
Supplier csi
Coefficient
Consumer bdi
Coefficient
Consumer cdi
Coefficient
1 (A) 18 0.05 80 -0.10
2 (B) 18 0.05 80 -0.10
3 (C) 21 0.07 80 -0.10
4 (D) 21 0.07 80 -0.10
5 (E) 21 0.07 80 -0.10
6 (F) 21 0.07 80 -0.10
7 (G) 17 0.05 80 -0.10
8 (H) 0 0.10 440 -0.50
9 (I) 30 0.07 440 -0.50

41. Solution for All True Marginal
Cost Bids 14%
46%
32%
42%
49%
36%
22%
3%
95%
14%
63%
A
G
B
C
D
E
I
F
H
286.4 MW 286.4 MW
183.2 MW
183.2 MW
296.4 MW
233.2 MW
183.2 MW 183.2 MW
118.9 MW
60%
1 2
3
7 6
5
8 9
393.4 MW
166.8 MW
166.8 MW
393.4 MW
166.8 MW
166.8 MW
166.8 MW
166.8 MW
166.8 MW
46.64 $/MWh 46.64 $/MWh
46.64 $/MWh
46.64 $/MWh
46.64 $/MWh
46.64 $/MWh
46.64 $/MWh
46.64 $/MWh 46.64 $/MWh
4

42. Market Behavior
•Assume all consumers always submit bids
corresponding to true marginal benefit (k=1)
•Assume supplier A-F and I all act alone to
maximize their profit
•Assume suppliers G and H collude (or
merge) together
–G and H now make bid decisions together

43. What are General Strategies for G
and H?
•G and H could act to raise their prices hoping
to increase profit
•Also could act to take advantage of the
transmission constraint between them
–G lowers price hoping that overload on the line
between G-H will result in increased profit by H
•Nash Equilibria are found for each of these
two general strategies by iteratively solving
the individual welfare maximum

44. Nash Equilibrium Found When
Both G and H raise pricesBus Price
[$/MWhr]
Supplier
Output [MW]
Supplier
Profit [$/hr]
Consumer
Demand [MW]
Consumer
Welfare [$/hr]
A 48.51 275.8 4,612.36 157.4 2,478.55
B 48.51 275.8 4,612.36 157.4 2,478.55
C 48.51 183.0 2,690.69 157.4 2,478.55
D 48.51 183.0 2,690.69 157.4 2,478.55
E 48.51 183.0 2,690.69 157.4 2,478.55
F 48.51 183.0 2,690.69 157.4 2,478.55
G 48.51 262.1 4,824.89 157.4 2,478.55
H 48.51 216.1 5,813.56 391.5 76,630.97
I 48.51 123.1 1,218.26 391.5 76,630.97
Totals 1885.0 31,844.19 1885.0 170,611.81
•Combined profit for G and H of $10,638 $/hr

45. Nash Equilibrium Found G and H
try to Game the ConstraintBus Price
[$/MWhr]
Supplier
Output [MW]
Supplier
Profit [$/hr]
Consumer
Demand [MW]
Consumer
Welfare [$/hr]
A 47.08 241.9 4,108.89 164.6 2,709.01
B 47.80 257.5 4,357.63 161.0 2,592.32
C 49.95 192.4 2,978.58 150.3 2,257.62
D 50.67 196.1 3,125.79 146.7 2,151.16
E 51.38 198.3 3,272.70 143.1 2,047.09
F 50.67 196.1 3,126.40 146.7 2,150.68
G 46.36 295.9 4,310.76 168.2 2,828.57
H 60.73 183.3 7,771.83 379.3 71,921.82
I 54.29 84.0 1,546.03 385.7 74,387.47
Totals 1845.4 34,598.62 1845.4 163,045.74
•Combined profit for G and H of $12,082 $/hr

46. Contour Plot of Combined Profit
of G and H when A-F,I bid k = 1.0

47. 3-D Plot of Combined Profit
of G and H when A-F,I bid k = 1.0

48. Results
•G and H acting together can increase their
profit by gaming around the transmission
constraint
•Transmission Analysis MUST be included in
Market Power Analysis
•Engineering Analysis and Economic Analysis
can be integrated together

49. Conclusions
•Market power abuses in a large power system
need to be assessed.
•Regulators need to be cognizant of ability of
market participants to act strategically
•Portfolio owners need to be cognizant of their
own, and their competitors potential for
strategic behavior

50. Conclusions
•Rules of the game can make it more difficult to
act strategically, but it would be difficult to
eliminate possibility completely.
•Load’s ability to respond to market power is an
important consideration.
•Slides and free 12 bus version of the
PowerWorld Simulator software are available
at www.powerworld.com