Economies of Scope and Value of
Coordination in the Evolving
Electric Power
Marija Ilic, Marcelo Elizondo, Michael Patnik,
Zayra Romo and Zhyiong (Richard) Wu
Carnegie Mellon University
milic@ece.cmu.edu
Presentation at the CMU T&D
Conference, December 16,04
Paper outline
•
Future quality of system performance is likely to be primarily
determined by the information technology (IT) supported justin-time (JIT) and just-in-place (JIP) decision making.
•
Performance metrics need to account for
various
economies of scope underlying management of an electric
power interconnection whose characterization is very complex
over space and time.
Examples are given to illustrate the inter-dependencies
between the technical characteristics and the economic
outcome (social welfare and outcomes of suppliers and
consumers) and the magnitude of their effects in some detail.
A conjecture that in this environment it becomes possible to
define private goods such as delivery service, reactive
power/voltage support, reliability service, all of which are
currently viewed as public goods; this would, in turn, enable
compatible incentives for a variety of services and products
and lead to genuine choice.
IT-tools essential for doing this. R&D needs.
•
•
•
Presentation at the CMU T&D
Conference, December 16,04
Basic ideas
• A dichotomy between decentralization (technological /
organizational), and coordination for system-wide
reliability/efficiency. Industry at standstill.
• The basic problem is lack of systematic ways to quantify
the industry performance for various types of
decentralization.
• We need to answer questions concerning : (1) complete
product/service design as a function of industry rules; (2)
relations between products and industry organization in
place; (3) mechanisms to facilitate electricity service
unbundling without loosing benefits of coordination.
Presentation at the CMU T&D
Conference, December 16,04
Open R&D problems
• Future solutions are multi-fold with
technical and economic outcomes never
studied before.
• Paper offers examples illustrating the
criticality of capturing inter-dependencies
between the physical operations, industry
rules in place and the IT support.
• Synergies between the physical and
economic networks previously unimaginable are plausible by careful
design of a supporting IT network.
Presentation at the CMU T&D
Conference, December 16,04
Product differentiation and
economies of scope
• Careful product differentiation and
decomposition (aggregation) of users and
suppliers are the key determinants of capturing
these synergies.
• One way of quantifying the implications of
service unbundling is by defining economies
of scope present in today’s bundled electricity
service, and by analyzing potential loss of
economies of scope resulting from the service
unbundling.
Presentation at the CMU T&D
Conference, December 16,04
Economies of scope in the electric
power industry
• The existing literature concerning economies of
scope in the electric power industry falls short of
recognizing unique temporal and spatial
characteristics of this industry.
• Electric power delivery, reactive power support
and capacity provision are identified as three
major components of the unbundled electricity
service exhibiting economies of scope.
• N.B: Extracting economies of scope of this
kind has NOT been optimally done in the
regulated industry of today.
Presentation at the CMU T&D
Conference, December 16,04
Economies of Scope in Electric
Power Systems: Basic Concepts
• Corporate unbundling-related (T,G and D)
• Industry structure-dependent (real power
the only product; everything else an “uplift”
public good; versus near-complete set of
products/services)
• One possible set of near-complete set of
services (real power; delivery; reactive
power/voltage; load following; capacity to
manage failures)
Presentation at the CMU T&D
Conference, December 16,04
Economies of Scope in Electric
Power Systems: Basic Concepts
• A. Unbundling real power and its delivery;
• Examples of usage-based transmission
charges and nodal price based
C (G, T ) < C (G,0) + C (0, T )
Presentation at the CMU T&D
Conference, December 16,04
16%
400
13.7%
Total cost [$]
350
14%
300
12%
250
10%
8.3%
200
7.3%
150
100
8%
7.6%
6.1%
6%
4%
3.2%
1.8%
50
0.0%
0
4
7
11
15
18
22
25
29
1.0%
0.0% 0.0% 0.0% 0%
33
36
40
44
2%
0%
47
Growing total system demand [MW]
(Gen + Tran)
(Gen) +(Tran)
Presentation at the CMU T&D
Conference, December 16,04
Ec. Of Scope achieved
nodal price [$/MW]
Nodal price for bundled service ( G+ T )
20
15
10
5
0
-5
-10
-15
-20
-25
-30
-35
4
7
11
15
18
22
25
29
33
36
40
44
Growing total system demand
λ1
λ2
Presentation at the CMU T&D
Conference, December 16,04
λ3
47
Effects of reactive power/voltage
support
16
14
Load (1000MW)
12
10
8
6
4
2
0
1
2
3
4
5
6
7
8
9
10 11 12
13 14 15
16 17
hour
Presentation at the CMU T&D
Conference, December 16,04
18 19 20
21 22
23 24
8
7
real power output (1000MW)
6
5
4
3
2
1
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
hour
gen1 Presentation
gen2 at the
gen3
CMU T&Dgen4
Conference, December 16,04
120
LMP($/MWh)
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
hour
bus 1
bus 2
bus 3
Presentation at the CMU T&D
Conference, December 16,04
bus 4
bus 5
LMPs difference as percentage of DC case
30%
20%
10%
0%
-10%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-20%
-30%
-40%
-50%
-60%
hour
bus1
bus2
bus3
Presentation at the CMU T&D
Conference, December 16,04
bus4
bus5
difference as percentage of DC case
30%
20%
10%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-10%
-20%
-30%
-40%
-50%
hour
load payment
generator revenue
Presentation at the CMU T&D
Conference, December 16,04
congestion revenue
2
1.8
1.6
Million $
1.4
1.2
1
0.8
0.6
0.4
0.2
0
9
10
11
12
13
14
load (1000MW)
load payment
generator revenue
Presentation at the CMU T&D
Conference, December 16,04
congestion revenue
15
difference as percentage of DC case
60%
50%
40%
30%
20%
10%
0%
-10% 9
10
11
12
13
14
-20%
-30%
-40%
-50%
load (1000MW)
load payment
generator revenue
Presentation at the CMU T&D
Conference, December 16,04
congestion revenue
15
Inter-temporal effects—capacity
markets
• Examples of spot market and stratum
market
2.2
x 10
4
Total Load
yearly
monthly
hourly
2
1.8
MWh
1.6
1.4
1.2
1
0.8
0.6
0
0.5
1
1.5
Hour
Presentation at the CMU T&D
Conference, December 16,04
2
2.5
3
x 10
4
Prices - stratum
110
yearly
monthly
hourly
100
90
80
$
70
60
50
40
30
20
10
0
0.5
1
1.5
Hour
Presentation at the CMU T&D
Conference, December 16,04
2
2.5
3
x 10
4
Prices - spot
110
100
90
80
$
70
60
50
40
30
20
10
0
0.5
1
1.5
Hour
2
Presentation at the CMU T&D
Conference, December 16,04
2.5
3
x 10
4
Effects of types of controllers on
transmission capacity limits
Hz
Gov. and AVR
Frequency
Gov. only
60
59
No control
(f continues decreasing)
58
57
AVR only
56
55
Time
0
5
10
15
20
25
Presentation at the CMU T&D
Conference, December 16,04
30
35
40
s
PU
Voltage B3
Gov. and AVR
Gov. only
0.95
No control
0.90
AVR only
0.85
Time
0.80
0
5
10
15
20
25
Presentation at the CMU T&D
Conference, December 16,04
30
35
40
s
Hz
Frequency
Gov. only
60
59
Gov. and AVR
No control
58
57
AVR only
56
55
Time
0
5
10
15
20
25
Presentation at the CMU T&D
Conference, December 16,04
30
35
40
s
Possibilities to design nearcomplete set of products/services
• These sub-services are generally interdependent; need IT-supported protocols which
- Define suppliers and producers of these subservices and ways of settling for quantity and
price (charge);
-Must be done with respect to space and time (in
the regulated industry using a hierarchical
approach; strong assumptions that may no
longer hold under open access);
-It is possible to do dynamic clustering for
transforming data into relevant information (at
present open problem).
Presentation at the CMU T&D
Conference, December 16,04
Conclusions--Moving forward
• Need to clearly define “ancillary” services as real
products/services; not negligible effects which must be
managed reliably and efficiently; compatible incentives
needed for giving incentives to provide these services (in
particular investment in this).
• MUST develop IT tools for product differentiations,
aggregation of users and providers of these services.
• Inherent part of future markets. It is potentially simple
and doable.
• AN EXCITING R&D MULIT-DISCIPLINARY AREA.
Presentation at the CMU T&D
Conference, December 16,04