System Adequacy, Investment and Risk

1

CERTS, Cornell University, 10/8/06

• System Adequacy, Investment and Risk
• Tim Mount
• Applied Economics and Management
• Cornell University
• tdm2_at_cornell.edu

2
OUTLINE

• Conclusions from Testing Market Designs
• Co-Optimization in Markets for Energy and
  Reserve Capacity.
• Joint Markets for Energy and Reactive Power
  (VArs).
• Power Transfers and the Cost of Meeting Native
  Load.
• Paying for Reliability Theory and Practice
• Energy-only Markets versus Energy plus Capacity
  Markets.
• Locational Capacity Markets in New York State.
• Forward Capacity Markets in New England.
• Paying for Reliability Proposal for Future
  Research
• Who should be Responsible for Investment
  Decisions?
• Limitations of the Current Structure of the
  Electric Utility System.
• Challenges for future research.

3
PART 1
Testing Market Designs

• References on ltwww.e3rg.pserc.cornell.comgt
• Mount, Timothy D. and Surin Maneevitjit
• Paying for Reliability in Deregulated
  Electricity Markets
• 25th Annual Eastern Conference, Rutgers CRRI,
  May, 2006
• Mount, Timothy D., Steen Videbaek and Ray D.
  Zimmerman
• Testing Alternative Market Designs for Energy
  and VArs in a Deregulated Electricity Market
• 25th Annual Eastern Conference, Rutgers CRRI,
  May, 2006
• Mount, Timothy D. and Robert J. Thomas
• Testing the Effects of Power Transfers on
  Market Performance and the Implications for
  Transmission Planning,
• Proceedings of the IEEE PES Conference, June
  2006.

4
Testing Markets for Energy and Reserves1.
PowerWeb AC Network
Region A Competitive
Region B Load Pocket
Reserve req40MW
17
18
Total Reserve req 60 MW
5
Testing Markets for Energy and Reserves 2. The
Markets Tested

• PowerWeb Network has two Regions
• Region A Competitive
• 4 firms --- marginal cost offers submitted by
  software agents
• Region B Load Pocket caused by limited
  transmission capacity
• 2 firms --- price/quantity offers submitted by
  students
• Three markets were tested
• Test I Joint Market with Fixed Locational
  Reserves (JMwFR),
• The current market structure used in New York
  State Test II Joint Market with Responsive
  Reserves (JMwRR),
• Co-Optimization for an explicit set of
  Contingencies
• Test III Integrated Market with Responsive
  Reserves (IMwRR)
• Co-Optimization and pay the Opportunity Cost
  for Reserves plus a Make-Whole Startup Cost

6
Testing Markets for Energy and Reserves 3.
Average Price Paid to Meet System Load
Test I (JMwFR), Test II (JMwRR) and Test III
(IMwRR)
7
Testing Markets for Energy and Reserves 4.
Analysis of Earnings/Firm

• Average Earnings for Two Firms in the Load Pocket
• Test I Joint Market with Fixed Locational
  Reserves
• Inelastic demand for both energy and reserve
  capacity
• More than 50 of earnings come from providing
  reserve capacity, particularly from peaking units
  --- reserve market is easy to exploit
• Test II Joint Market with Responsive Reserves
• Effects of market power are partially mitigated
  by substitution between energy and reserve
  capacity
• More capacity withheld from the market due to the
  downward pressure on earnings/firm
• Earnings from reserve capacity drop substantially
  to 15 for base load units and 45 for peaking
  units
• Test III Integrated Market with Responsive
  Reserves
• Opportunity cost payments for reserve capacity
  are zero, implying that reserve capacity is
  free
• Startup costs are the dominant source of
  earnings, 80 for base load units and 95 for
  peaking units
• Withholding capacity is no longer a major problem

8
Testing Markets for Energy and Reserves 5.
Conclusions

• Using Responsive Reserves (Co-Optimization) is an
  effective way to make the market more competitive
  and reduce the average price paid to meet system
  load compared to Fixed Locational Reserves.
• Paying the the Opportunity Cost for reserves
  using co-optimization is even more effective
  because speculating in the energy auction is
  punished by lower opportunity costs for
  reserves.
• BUT there is an underlying incompatibility
  between getting low prices and maintaining system
  reliability because capacity is withheld in
  competitive auctions (generating units with high
  operating costs are dispatched at minimum levels
  most of the time because they are only needed as
  reserves to cover specific contingencies, and as
  a result, expected profits are generally low for
  these units).
• In competitive markets,supplementary payments are
  needed to ensure that all generating units
  required to maintain system reliability are
  financially viable. Using Make-Whole Payments is
  an effective way to allocate money where it is
  needed and reduce withholding (similar to the
  Forward Capacity Market proposed for New England)
  

9
Pilot Tests of Markets for Energy and VArs 1.
The Markets Tested

• PowerWeb Network is relatively uncongested
• 6 firms represented by students
• Three markets were tested
• Test I Pay nodal prices for real energy with a
  contract for VArs.
• Firms submit price/quantity offers for energy
  only
• 30 periods normal and 30 periods with limited
  VAr capability
• Test II Pay nodal prices for real energy and
  nodal prices for VArs
• Same offers as Test 1 plus price offers for
  VArs
• 30 periods normal and 30 periods with limited
  VAr capability
• Test III Pay nodal prices for real energy and
  nodal prices for VArs
• 1) 30 periods with Interruptible Load
• 2) 30 periods with Local Dispatchable VAr
  Capacity
• Same offers as Test 2
• All periods with limited VAr capability

10
Pilot Tests of Markets for Energy and VArs 2.
Nodal Prices for VArs in Tests 1 and 2
TEST 1 Energy-Only Auction
TEST 2 Offers for VArs
Periods 1-30 normal VAr capability Periods
31-60 limited VAr capability
11
Pilot Tests of Markets for Energy and VArs 3.
Nodal Prices for VArs in Test 3
TEST 3 Offers for VArs (2 sessions) Periods
1-30 Interruptible Load Periods 31-60 Local
Dispatchable VAr Capacity
Periods 1-60 limited VAr capability
12
Pilot Tests of Markets for Energy and VArs 4.
Conclusions

• The nodal prices of VArs are close to zero most
  of the time (capacitors etc. are used to
  compensate under normal conditions), but nodal
  VAr prices can be very high when contingencies
  occur (need dynamic VArs).
• The production cost of dynamic VArs for
  generators is zero unless the dispatch is on the
  capability curve (less real energy generated).
• In a VAr market, firms find that it pays to
  speculate for both units. (unlikely that
  short-run competitive prices will ever be
  realized).
• Since VArs dont travel, distributed sources of
  VArs provide an effective way to mitigate
  speculative behavior by generators to a limited
  extent.
• It is unlikely that market signals can be trusted
  to determine optimum levels of investment in VAr
  capacity needed to maintain system reliability.

13
Power Transfers and Market Performance 1. The
Rationale for the Tests

• Merchant Transactions on the TVA system have
  grown over 1,000 since 1996
• Many of these transactions occur when system
  resources are stressed --- e.g. voltage limits
  are reached on some lines

Source Tennessee Valley Authority
14
Power Transfers and Market Performance2. The
Markets Tested

• PowerWeb Network is relatively uncongested
• 6 firms represented by students
• System load varies from period to period
• Three markets were tested
• Test I No transfers.
• Firms submit price offers for energy for five
  blocks of capacity
• Periods 1-25
• Test II Fixed 40 MW transfer from bus 28 to
  bus 14 (NW to SE).
• Firms submit price offers for energy for five
  blocks of capacity
• Periods 26-50
• Test III Fixed 40 MW transfer from bus 14 to
  bus 28 (SE to NW).
• Firms submit price offers for energy for five
  blocks of capacity
• Periods 51-75

15
Power Transfers and Market Performance3.
Conclusions

• Even though the quantity of real energy flowing
  though the network stayed the same for a number
  of periods in the experiments, the flows on
  individual lines were quite erratic. Identifying
  fixed accounting pathways for physical bilateral
  contracts defies the laws of physics and may be
  highly misleading on a congested network.
• Although maintaining reliability is a socially
  efficient decision, the cost of supporting
  generating units with low capacity factors is
  very high. Increasing the capacity of
  transmission, and lowering the spot prices in a
  congested region, like New York City, may
  undermine the financial viability of generators
  needed to maintain system reliability.
• The pipeline analogy for paying transmission
  owners is appropriate for a DC intertie but not
  for an AC network. The AC transmission system
  should be fully regulated using performance-based
  rates of return that penalize unscheduled outages
  and losses and reward low operating costs (e.g.
  like the UK).
• The speculative behavior of generators in a
  deregulated market is affected by the level of
  congestion on a network. Planning models for
  system expansion should take into account how
  this behavior is likely to change in order to
  estimate the net benefits of making a
  transmission upgrade.

16
Power Transfers in New York State1. New York
Control Area (NYCA)
Zones Analyzed (NYMEx trading) A - Niagara G -
Hudson Valley J - New York City K - Long Island
17
Power Transfers in New York State2. Ranked Nodal
Prices in NYCA for 2000
18
Power Transfers in New York State3. Ranked Nodal
Prices in NYCA for 2005
19
Power Transfers in New York State4. Zonal Price
Differences in NYCA
G - A
J - G
K - J
2 0 0 0
2 0 0 5
20
Power Transfers in New York State5. Sub-Zonal
Price Differences within NYC
G - J1
J2 - J1
J3 - J2
2 0 0 0
2 0 0 5
21
PART 2
Paying for Reliability Theory and Practice
22
Textbook Analysis of Generation Adequacy1. Total
Cost of Generation/Year
Specified Costs Variable Capital (/MWh)
(k/MW/Year) Peak 60 80 Shoulder
30 159 Baseload 15 238
Capacity Factors for Least-Cost Choices Peak
lt 30 Shoulder 30-60
Baseload gt 60
23
Textbook Analysis of Generation Adequacy2.
Net-Revenue by Type of Generator
Specified Costs Variable Capital (/MWh)
(k/MW/Year) Peak 60 80 Shoulder
30 159 Baseload 15 238
Additional Revenue Needed to Cover the Capital
Costs (k/MW/Year) Peak
80 Shoulder 80 159 - 79 Baseload 80
238 - 158
24
Textbook Analysis of Generation Adequacy3. The
Solution --- Allow Scarcity Pricing
Specified Costs Variable Capital (/MWh)
(k/MW/Year) Peak 60 80 Shoulder
30 159 Baseload 15 238
Capacity Factors for Least-Cost Choices Shed Load
lt10 Peak 10-30 Shoulder
30-60 Baseload gt60
Shed Load (10 36.5 Days/Year)
152/MWh NERC Reliability Standard (2.4
Hours/Year) 33,393/MWh
25
Textbook Analysis of Generation Adequacy4.
Net-Revenue with Scarcity Pricing
Specified Costs Variable Capital (/MWh)
(k/MW/Year) Peak 60 80 Shoulder
30 159 Baseload 15 238
Additional Revenue Needed to Cover the Capital
Costs (k/MW/Year) Peak 0 80 -
80 Shoulder 0 159 - 159 Baseload 0
238 - 238 Problem Solved!
26
Generation Adequacy in Reality1. Net-Additions
to U.S. Generating Capacity
Source 2005 NERC Long-Term Reliability
Assessment, Fig. 4
27
Generation Adequacy in Reality3. Spot Price
Behavior in New York City (NYC)
Price /MWh
Regulatory response to the Californian Energy
Crisis ---gt Automatic Mitigation Procedures and
regulatory threat have suppressed high prices
28
Generation Adequacy in Reality3. Average Price
Duration Curves for NYC
Average Price /MWh
lt--- 2000/01
2002/03 2004/05 ---gt
Hours/Year (1000 Hours 11.4 Capacity Factor)
29
Generation Adequacy in Reality4. Average Prices
in NYC versus LRAC
Av. Price gt Long-Run Average Cost (LRAC) is
RED Max. value for each row is BOLD
30
Generation Adequacy in Reality5. Annual
Earnings/Generator in NYC
The Capacity Market is the dominant source of
income for peaking units in NYC and LI BUT this
income is still not enough to attract
new merchant generators
Capital is the upper Hudson valley, and LI is
Long Island Source Figure 16 on p. 23 of the
NYISO 2004 State of the Market Report
ltwww.nyiso.comgt
31
Generation Adequacy in Reality6. Projected
Reserve Margins for New York
NYISO standard --- A reserve margin of 18 is
needed to meet the proposed NERC reliability
standard (Fail lt1 day in 10 years) Reserve
Margin is the amount of Installed Capacity above
the Forecasted PEAK Load ()
Source NYISO PowerTrends
32
Generation Adequacy in Reality7. Australia, New
York and New England

• Generation Adequacy is a minimal requirement for
  maintaining the operating reliability because
  blackouts are very expensive. Since the electric
  supply system is unforgiving, policies for
  maintaining Generation Adequacy must be
  sufficient.
• The Australian energy market works because
  allowing price spikes results in an average price
  duration curve that approximates the long-run
  average costs of different types of capacity.
  However, it is financially risky for operations
  and investment and is NOT sufficient.
• The Capacity Market in New York is expensive
  (1billion/year in NYC) but there is no
  obligation to invest this money in new generating
  capacity. This market is NOT sufficient. The
  Demand Curve clears only one month before real
  time --- too little, too late for merchant
  generators.
• The proposed Forward Capacity Market in New
  England clears 3 years ahead, and merchant
  generators can lock in a price for up to 5 years.
  There are alternative procedures if insufficient
  capacity is entered into the auction, and this
  market is likely to be sufficient. Regulators
  take the primary responsibility for ensuring that
  generation adequacy is maintained far enough in
  advance to correct deficiencies.

33
PART 3
Paying for Reliability A Proposal for Future
Research
34
Current Reliability Standards1. Two Different
NERC Criteria

• DEFINITIONS OF RELIABILITY
• North-American Electric Reliability Council
  (NERC), 2005
• System Adequacy The ability of the electric
  system to supply the aggregate electrical demand
  and energy requirements of customers at all
  times, taking into account scheduled and
  reasonably expected unscheduled outages of system
  elements.
• Ensuring there is enough generation and
  transmission capacity --- the investors problem
• Operating Reliability The ability of the
  electric system to withstand sudden disturbances
  such as electric short circuits or unanticipated
  failure of system elements.
• Determining the dispatch of installed capacity
  and levels of reserves --- the system
  operators problem
• NERC has little expertise on markets???

35
Current Reliability Standards 2. FERC is now
Responsible for Reliability
The Energy Policy Act of 2005 (EPAct05) was
signed into law in August 2005, and it gives
greater authority to the Federal Energy
Regulatory Commission (FERC) to enforce
reliability standards by imposing penalties on
end-users if the standards are violated. In
addition, a new organization, the Electric
Reliability Organization (ERO), will be given the
authority to establish these reliability
standards. Prior to EPAct05, FERC was primarily
an economic regulator of the wholesale
transactions and tariffs on the bulk power
system. At this time, it is not clear exactly
how FERC will implement the new responsibilities
for enforcing reliability. FERC will enforce
standards for Operating Reliability
but not for System Adequacy???
36
Proposed Structure for a Deregulated Electric
Utility Industry
Operating Reliability O M Payment System Adequacy Ownership
GENCO Energy ISO Owner Spot Market Central Authority Private or Public
GENCO Reserves ISO Owner Spot Market Co-Optimization Central Authority Private or Public
GENCO VArs ISO Owner Contingent Claim Market Central Authority Private or Public
TRANSCO ISO Owner Performance Based Regulation Central Authority Private or Public
DISCO DERALCO ??? Owner Performance Based Regulation DERALCO ??? Private or Public
Micro Grid DERALCO ??? Owner Spot Market DERALCO ??? Private or Public
Load Response DERALCO ??? Owner Spot Market DERALCO ??? Private or Public
Distributed Energy Resources and Active Load
Company (DERALCO)
37
Current Challenges for the Electric Utility
Industry

• Tension between Federal and State regulators
• FERC is responsible for Operating Reliability
  only
• Fragmented control of a single AC Network
• Markets are expected by many to deliver too much
• Efficient pricing for a reliable system is very
  risky
• Inadequate investment for System Adequacy and
  Security
• Little innovation, particularly for load response
  and DER
• Environmental and Energy Security
• Fuel diversity
• Clean coal and carbon sequestration
• New nuclear plants
• Renewables
• Transmission upgrades and NIMBY
• Efficiency such as Combined Heat and Power
• Load response and DER
• Weather extremes