Priority Pricing-A Proposal for an Economic Demand

Download Report

Transcript Priority Pricing-A Proposal for an Economic Demand

Priority Pricing:
A Proposal for an Economic Demand-Side
Program in ERCOT
Jay Zarnikau, PhD
Frontier Associates
Shmuel Oren, PhD
UC-Berkeley
ERCOT Demand Side Working Group
June 8, 2007
Impetus for This Proposal
• Continuing concerns over resource adequacy.
• Additional steps are needed to promote the response of loads to
wholesale prices, which will in turn promote economic efficiency,
avoid or reduce some price spikes, and constrain the market power
of generators.
• While the BUL program is theoretically attractive, it hasn’t proven
popular and will be terminated when the market structure changes.
This proposed program could provide a substitute for the BUL.
• This proposed program could complement the EILS program (or, if
EILS is terminated, a modified priority pricing program could provide
an alternative to EILS).
•
Disclaimer: This proposal has not be sponsored by, and does not necessarily represent the views
of, any of our consulting clients.
Background
• I proposed a “Long Term BUL Contracts” program to the Demand
Side Working Group in October 2005.
• Professor Oren suggested a call option program (which would
involve both supply-side and demand-side resources) at a Resource
Adequacy Rulemaking workshop in April 2005 (“Market Friendly
Generation Adequacy Assurance”).
• Our two proposals had some common features and are both based
upon the same pricing theory that Prof. Oren helped to develop in
the 1980s.
• Consequently, we decided to collaborate on the development of a
more refined demand-side program.
Program Overview
• Program participants would contractually commit to curtail load at
three preset strike prices.
• Commitments would be for a period of one year.
• The program participant would agree to curtail whenever the MCPE
(balancing energy price) in the zone or LMPZ (zonal average of
LMPs) exceeds the strike price at any time during the contract
period.
• Balancing energy settlement (or its equivalent in the nodal market)
to participants during curtailment period based on strike price.
• A reservation payment based on committed capacity and strike price
would be paid to the program participant.
• Participants can enroll for either:
– One-Day Notice option (where curtailments are triggered by high prices in the
Day-Ahead Market), or
– Real-time Market option (where curtailments are triggered by a projection of a
high real-time energy price).
Program Overview (cont.)
• A higher capacity payment will be provided to program participants
that offer their curtailments at a lower energy price.
• Participants are insulated from the future RUC Capacity Short
charges.
• A bonus payment and priority status will be provided to any program
participant that agrees to interrupt with no notice under ERCOT’s
control.
• A bonus payment will be provided to any program participant that
would additionally agree to interrupt in the event of an ERCOTdeclared emergency. (If market prices are raised to the wholesale
price cap during an EECP, then this would not be necessary.)
• Of course, this is a voluntary program.
Theoretical Foundation
• This is consistent with the notion of “priority pricing,” whereby the
price paid by consumers for energy is related to the level of reliability
that they select. Pay more if you want higher reliability. Pay less (or
get an payment or credit) if you are willing to accept more outages).
• This is also consistent with financial options theory. The ISO is
granted an option to curtail the usage of participating customers at
the customer’s strike price (the MCPE at which the customer
provides an offer to curtail).
• This is really an “economic program” with some reliability features,
rather than a reliability program. So, it is quite different from EILS.
Similarities to Other Programs
• Prior to restructuring, HL&P offered a large menu of interruptible rate
options (e.g., firm service with LOS; instantaneous interruptible
service with IS-B; and interruptible with notice service through IS-10,
IS-30, and IS-60). Higher prices were paid for greater reliability or
firmness. However, system reliability conditions tended to trigger
curtailments, rather than economic interruptions.
• This is also analogous to a situation where a Real-Time Pricing
program participant or a consumer that purchases through an MCPE
Product always reduced usage at a predetermined price point.
• If a BUL (if we had any) had a standing fixed offer into the balancing
energy market in every interval over a year, it would look something
like this program.
• It is equivalent to a bilateral contract for load response since the
fixed premium locks in the payment to committed load eliminating
the risk of uncertain balancing energy settlement.
Benefits of This Approach
• The necessary system changes should be minimal (but some work
will be required to set baselines for some loads and forecast realtime prices).
• Knowing that demand will be reduced at various market prices will
have value to system operators.
• Since the program involves a commitment to curtail in an
emergency, demand reduction from these program participants
could be recognized in a reserve margin calculation. Further, the
committed load should receive a “gross up” that accounts for the
reduction in needed reserves.
• While demand-side offers to curtail into the day-ahead market
(DAM) moves us in this direction, a priority pricing program would
have more value since the curtailments under this program can be
relied upon. DAM participation is voluntary and there is no longterm commitment by a load to curtail when at a given price level.
Likely Participants
• Residential and small commercial direct load control program
participants with advanced metering systems.
– Good candidates for the Real-Time Market Option
– Water heaters, air conditioners, and pool pumps could be controlled and
cycled. Alternatively, the load at a facility could be limited with a fuse
during a high price, requiring the homeowner or facility manager to
decide which equipment to operate (like Demand Subscription Service
at SCE in CA during the 80’s).
– Advanced metering infrastructure facilitates performance monitoring
• Industrial loads that can tolerate some interruptions.
– Many of these would be good candidates for the One Day Notice
Option.
• LaaRs would not be good candidates.
Baseline for Measuring Curtailment
Performance
• We have the same problems that we encountered with EILS,
namely:
– It is difficult to predict load levels and the amounts that can be curtailed
at a facility up to a year in advance.
– Some loads are weather sensitive or have maintenance outages.
• Consequently, the formulas and procedures used by the ERCOT
Staff for EILS should be followed (treat as intermittent resources).
• In addition, we’ll have some special complications involving
residential direct load control (e.g., air conditions are not in use
during winter). Fortunately, many ISOs and utilities have addressed
these issues in the design of other programs and we can learn from
their experiences.
• Perhaps the NAESB effort (to be discussed later on today’s agenda)
will also provide guidance.
Quantities Provided by Participants
• For loads with “random” (i.e., difficult to predict) temporal load
patterns, an average quantity could be offered.
• For loads with a predictable pattern (e.g., a typical weekly
production schedule with typical maintenance schedules), the
pattern could be offered.
• For weather-sensitive residential air conditioning loads, a formula
expressing potential demand reduction as a function of weather
could be offered.
• Curtailment performance is considered to be adequate if the load
level during a curtailment is at or below the baseline minus the
quantity offered.
What the Participant Receives
• Reservation or priority payment.
• Avoids the high energy charges (as the customer would if it was
engaging in voluntary load response).
• There may be some mark-up of incentives for avoided losses.
• Protection from RUC Capacity Short Charges
• Receives price forecasts (if it subscribes to the Real Time Market
Option).
• If the load is controlled by ERCOT or a third-party, relieves the
participant of the need to take any manual actions to curtail.
One Day Notice Option
• Participant makes a “standing offer” to provide a curtailment into the
DAM over the one-year duration of its contract.
• Since this is an hourly market, the deployment/curtailment period is
one hour when its offer is struck.
• The Participant has roughly one day’s notice of the needed
curtailment.
Real-Time Market Option
• We need the Demand Side WG to first complete Item 2 on its Goals
for 2007, so that we can provide a price forecast to loads at least a
few minutes prior to each settlement interval when the nodal market
is introduced:
– There presently is a problem because the “real prices” will be calculated
every 5 minutes (roughly). So, at the start of a 15-minute interval, the
price for the first 5 minutes will be determined (with no advance notice).
But the price for the entire 15 minute settlement interval will not be
known.
– Until we resolve this problem, there is no way of knowing whether the
LMPZ for the 15 minutes will exceed the trigger price.
• Regarding minimum duration of a curtailment:
– Residential direct load control programs will want short durations.
– If an industrial energy consumer is curtailed, it will want a guaranteed
minimum curtailment period, and be paid an acceptable price for all
intervals of that period. Let’s handle this in the same manner as we
addressed it for the BUL program.
Price Triggers
• We suggest:
– $750 per MWh
– $1000 per MWh
– $1500 per MWh
These are the price levels at which you could subscribe to curtail.
As offer caps increase, these should change.
Calculation of Program Incentive:
Risk Premium Approach
• One option is to provide the participant with an upfront payment
based on the expected value of the high balancing energy or LMPZ
costs that the participant will avoid as a result of the customer’s
participation in the program.
• For example, the shaded area in the following graph might reflect
the expected value of the costs avoided by the customer. Let’s say
that (when multiplied by the customer’s expected load level) that the
shaded area equals $1000. The program participant would be paid
$1000.
• In an actual year, the price spikes might be more or less frequent
and higher or lower than the expected value. But, the program
participant is paid $1000, and is protected from any risk.
Calculation of Program Incentive:
Risk Premium Approach
• Additionally, there would be a “scaling up” of the value of the shaded
area of the curve to reflect the reduced line loss benefits associated
with this demand-side program.
• There is no distortion to the market (beyond all the distortions that
we already have). Risks are just shifted from the participant to the
market.
Calculation of Program Incentive:
Risk Premium Approach (cont.)
• Wouldn’t an energy consumer on an MCPE Product that curtails
(outside of any formal program) achieve the same result?
– Yes, on an expected value basis. But the customer removes some risk
by participating in the program. Most energy consumers are risk averse
when it comes to energy costs, so this is a benefit.
– The participant also gets better price information and protection from
RUC Capacity Short charges.
• This approach would require ERCOT to calculate price duration
curves (which might get a little contentious).
• But, it is an open question whether this removal of risk will be
sufficient to entice an energy consumer to participate in the
program. Further, when the consumer commits to a year-long
contract it surrenders some latitude to make operational decisions
based on real-time economic conditions.
Calculation of the Program Incentive:
Proxy Avoided Capacity Cost Approach
• Yes, yes. We have adopted an energy-only approach to resource
adequacy and generators do not receive capacity payments (unless
they are providing an ancillary service or RMR service).
• But if the risk premium approach does not provide sufficient
incentives for load participation, then additional incentives should be
considered.
• We could either:
– Administratively set a capacity payment, or
– Administratively set a capacity price ceiling and a target quantity (for
each price level) and then let the market determine the price.
Calculation of the Reservation Payment
(cont.)
• A price or price cap could be established based on:
– Annualized capital cost of constructing a combustion turbine
– Tied to the annual average price of non-spinning reserves or responsive
reserves in the previous year.
– A fraction of the PUCT’s Avoided Capacity Cost in its Energy Efficiency
Rule. It is presently $78.50/kW. PUCT Subst R §25.181(e).
– The PUCT-approved annual capacity payment for a load management
program operated by a TDU could also be used. This number is
roughly $15/kW/year for a summer-only program. See:
http://www.txuelectricdelivery.com/electricity/teem/services/elmsop/incentives.asp
– The estimated economic value of the program in reducing overall
market prices to the entire market.
– Or we could just pay make an additional payment to the participant
based on the price of responsive reserves at the time of the curtailment.
Bonus for “No Notice” Service
Participants that are willing to accept “no notice” can be treated as
Regulation or Responsive Reserves and get a premium
comparable to the average AS payment.
The usual concerns over having “too much” load providing such
services will need to be addressed.
And any declining marginal value of additional load providing such
services will need to be considered.
Market Prices During an EECP
• The price should be automatically raised to the price cap during
an EECP condition, if market prices as otherwise unlikely to
reach that level.
• If this occurs, then program participants (at least those who sign
up for the Real Time Market Option) will be curtailed during an
EECP (and will avoid those high prices).
• However, we need a mechanism to notify loads of this price with
sufficient notice time to accommodate their response.
Quantities to Procure
• The procured quantity could be based on reserve requirements.
• Alternatively the quantities can stay the same but the prices
changed to reflect the expected price forecasts.
• Or, perhaps there may be no need to limit this.
Performance for Payment Purposes
• QSEs submit offers to participate or enroll program participants
(consistent with how EILS program customers are enrolled).
• The performance of industrial participants is evaluated on an
individual customer basis. This is designed to avoid the problem
with the current BUL where multiple BULs under the same QSE
must be settled as a group, and the performance of a large BUL
within the group could affect the payments to the whole group.
• Residential and small commercial participants (e.g., a direct load
control program) may be settled on an aggregate basis.
Who Pays for the Program?
• The cost of the incentive payment could be assigned to load-serving
QSEs based on their load ratio share.
• A more sophisticated approach might be to assign the cost
associated with each strike price as a surcharge on load that
persists above that price.
• Self provision can be done through passive load response or
through participation in the program (depending upon the approach
used to calculate the incentives made to program participants, the
payment will offset the charge).
Penalties for Inadequate Performance
• Nonperformance can take several forms.
– If available load for curtailment is persistent below the original rating the
rating can be adjusted going forward.
– If, on the other hand the load does not curtail when asked, it should be
liable for the difference between the MCPE and the strike price and pay
a pro rata share of the program cost (including the scale-up portion for
RUC charges and distribution losses).