Development of Demand Response Price Thresholds - PowerPoint PPT Presentation

Loading...

PPT – Development of Demand Response Price Thresholds PowerPoint presentation | free to download - id: 67f1b8-ZmZmY



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Development of Demand Response Price Thresholds

Description:

Development of Demand Response Price Thresholds NEPOOL Markets Committee June 2, 2011 – PowerPoint PPT presentation

Number of Views:29
Avg rating:3.0/5.0
Slides: 29
Provided by: England8
Learn more at: http://www.iso-ne.com
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Development of Demand Response Price Thresholds


1
Development of Demand ResponsePrice Thresholds
  • NEPOOL Markets Committee
  • June 2, 2011

2
Development of Demand ResponsePrice Thresholds
  • Background
  • Study objectives and analytical approach
  • Findings

3
Background
  • FERCs Order No. 745 prescribes a simple method
    to determine the price above which dispatch of DR
    offers is cost-effective (Net Benefit Test
    Threshold, NBTT)
  • i.e., savings to consumers exceed payments to DR
    providers
  • The method uses a smoothed supply curve to find
    the lowest point on the curve above which the Net
    Benefit Test is always satisfied
  • The method ignores numerous complicating factors
    such as congestion, imports and exports, pumped
    storage, outages, startup costs, min. gen., unit
    commitment, load etc.
  • ISO-NE engaged Charles River Associates to test
    the method and compare its results to those of
    more sophisticated analysis

4
Study objectives
  1. Calculate monthly NBTTs using hourly
    security-constrained dispatch model (GE MAPS) for
    2010
  2. Calculate monthly NBTTs using a simplified
    smoothed supply curve approach based on the same
    supply curves as in (1)
  3. Compare results from these two methods
  4. If the results of the two approaches are similar,
    apply simplified approach to RT supply offer data
    for 2010

5
Hourly dispatch analysis Analytical approach
  • Using 500 GE MAPS runs, find heat rate thresholds
    (and corresponding LMPs) yielding net benefit
    maxima for each month and DR quantity
  • Use implied heat rate rather than LMP, then
    translate results to corresponding LMPs using gas
    prices
  • to improve convergence in the simulations and for
    flexibility in application, given fuel price
    volatility
  • Test over a range of DR quantities to determine
    sensitivity of results
  • GE MAPS base case calibrated to 2010 real-time
    prices, using actual zonal loads, generator
    outages, and weekly fuel prices
  • GE MAPS runs with demand response performed for
    range of heat rate thresholds and DR quantities
  • Heat rate thresholds 100 27,700 BTU/kWh in
    increments of 300 BTU/kWh
  • DR quantities of 300, 500, 750, 1000, and 1500 MW
  • DR allocated to zones using 2010 asset
    distribution
  • All DR in a zone dispatched when zonal implied
    heat rate threshold

6
2010 gas prices used in analysis
7
Hourly dispatch analysis Principal finding
  • Results are robust
  • Net benefit maximizing (NBTT) heat rate is
    relatively consistent across months
  • NBTT heat rate is relatively insensitive to DR
    quantity
  • NBTT heat rates are in the 8,200 to 8,800 BTU/kWh
    range
  • Corresponding LMPs vary more widely due to gas
    price volatility 33 - 69/MWh
  • The following graphs illustrate the results

8
(No Transcript)
9
(No Transcript)
10
Hourly dispatch analysis results NBTT Heat rate
DR Quantity (MW) 300 500 750 1000 1500
Month Heat Rate Threshold (Btu/kWh) Heat Rate Threshold (Btu/kWh) Heat Rate Threshold (Btu/kWh) Heat Rate Threshold (Btu/kWh) Heat Rate Threshold (Btu/kWh)
1 8,200 8,200 8,500 8,500 8,500
2 8,200 8,200 8,200 8,200 8,500
3 8,800 8,800 8,800 8,800 8,800
4 8,800 8,800 8,800 8,800 8,800
5 8,500 8,500 8,500 8,500 8,800
6 8,500 8,500 8,500 8,500 8,500
7 8,500 8,500 8,500 8,500 8,500
8 8,500 8,800 8,800 8,800 8,800
9 8,500 8,500 8,500 8,500 8,500
10 8,500 8,800 8,800 8,800 8,800
11 8,500 8,500 8,500 8,500 8,500
12 8,500 8,500 8,500 8,500 8,500
11
Hourly dispatch analysis results NBTT LMPs (2010)
DR Quantity (MW) 300 500 750 1000 1500
Month LMP Threshold (/MWh) LMP Threshold (/MWh) LMP Threshold (/MWh) LMP Threshold (/MWh) LMP Threshold (/MWh)
1 63.2 63.2 65.6 65.6 65.6
2 53.9 53.9 53.9 53.9 55.9
3 42.0 42.0 42.0 42.0 42.0
4 39.1 39.1 39.1 39.1 39.1
5 39.1 39.1 39.1 39.1 40.5
6 44.7 44.7 44.7 44.7 44.7
7 44.0 44.0 44.0 44.0 44.0
8 40.8 42.3 42.3 42.3 42.3
9 36.6 36.6 36.6 36.6 36.6
10 32.8 33.9 33.9 33.9 33.9
11 39.8 39.8 39.8 39.8 39.8
12 69.2 69.2 69.2 69.2 69.2
12
Supply curve analysis
  • The method uses a smoothed supply curve to find
    the highest point on the curve above which the
    curve is always inelastic
  • The inelastic condition is (Delta Price / Price )
    gt (Delta Supply / Supply)
  • Rearranged, this means that the local slope
    (Delta Price / Delta Supply) is greater than the
    average slope from the origin (0,0) to that point
    (i.e., Price / Supply)
  • The NBT is satisfied in the inelastic portion of
    the supply curve
  • A real supply curve consists of a staircase of
    flat increments for each generating unit or group
    of units with the same offer price
  • It is necessary to fit a smooth curve to the
    portion of interest in order to apply this test,
    because each flat segment (slope 0) is locally
    elastic

13
Supply curve analysis (continued)
  •  

14
June 2010 GE MAPS Supply Curve Fit
15
Observations on the function used to smooth curves
  • Exponential portion of the curve allows
    reasonably good fit to the steep tail of the
    supply curve, where expensive and small peaking
    units dominate the supply stack
  • Cubic portion allows better fit to the broader,
    more gently rising portion of the curve going
    from large steam units to intermediate gas-fired
    combined cycle units
  • Lowest portion of the supply stack, consisting of
    low cost hydro, wind, nuclear, and some base load
    coal units, is not included in the fit
  • It is clear from casual inspection that the
    supply stack has large elastic portions (flat
    portions) above these units

16
Observations (continued)
  • The raw curve has many short flat portions above
    the threshold price
  • near 43.58 for June 2010, corresponding to a
    heat rate of 8,294 (using the average natural gas
    price in that month of 5.25/MMBTU)
  • At a fine scale, each of these flat portions is
    elastic (zero slope)
  • The threshold metric is sensitive to the details
    and granularity of the fitted curve choice A
    smooth global fit will generally lead to lower
    threshold values than a more detailed fit with
    more local inflection points (bumps)
  • The selected six-parameter expression yields a
    smooth curve with good overall fit to the supply
    curves examined, providing unambiguous thresholds
    at the highest elasticto-inelastic transition
    points

17
Results MAPS supply curve analysis agrees well
with simulations
  • For most months, supply curve thresholds were
    similar to those from MAPS analysis
  • NBTT Heat rates in the same range 8,200 8,800
    MBTU/kWh
  • Thresholds were somewhat higher for one month
    October (9,500)
  • Explained in large part by low gas prices,
    causing larger contribution from variable OM
    costs to the implied heat rate
  • Implied HR (Gas Price HR VOM)/Gas Price
    HR VOM/Gas Price

18
MAPS Supply Curve NBTT Heat Rate Results
19
NBTT Heat Rates and LMPs, MAPS Supply Curve
Month HR Threshold LMP Threshold
1 8,830 68.1
2 8,590 56.5
3 8,800 42.0
4 8,810 39.1
5 8,480 39.1
6 8,290 43.6
7 8,250 42.7
8 8,260 39.7
9 8,680 37.4
10 9,520 36.7
11 8,960 41.9
12 8,460 68.9
rounded to nearest 10 BTU/kWh
20
Supply curves from ISO-NE RT generator offer data
yield similar results
  • Having validated the simpler approach, it then
    made sense to apply it to actual offer data for
    2010
  • We assembled monthly average supply curves from
    daily RT generator offers (masked data posted on
    ISO-NE website)
  • We fit curves using the same cubic exponential
    function, for the 25 - 300 portion of each
    curve, the portion of interest
  • Initial fits to full range showed we could ignore
    offers above 300
  • Smaller range allowed closer fits and more
    uniform results
  • Results are somewhat lower, but in the same
    general vicinity, except for January and
    December, which are in the 6,500 range
  • Given the convergence of the simpler approach
    with the more sophisticated approach, we
    recommend that NBTT prices be based on a supply
    curve analysis using RT generator offer data

21
June 2010 Average Offer Curve
22
NBTT Heat Rates and LMPs, 2010 RT Generator
Offer Curves
Month HR Threshold LMP Threshold
1 6,630 51.2
2 7,020 46.2
3 8,100 38.7
4 8,740 38.8
5 8,430 38.8
6 8,310 43.7
7 7,790 40.3
8 8,010 38.5
9 8,460 36.5
10 9,200 35.5
11 8,480 39.6
12 6,480 52.8
23
How often would demand response clear, given the
NBTTs determined using 2010 offer curves?
  • More generally, how does the number of
    consecutive business days with demand response
    vary with HR threshold?
  • 2010 RT hourly prices were analyzed over the
    range of implied heat rates to determine maximum
    number of consecutive business days at or above a
    range of LMP / implied heat rate thresholds
  • Approach
  • Calculate monthly LMP thresholds over range of
    implied heat rates using monthly weighted average
    gas prices
  • Compare hourly RT zonal prices to LMP thresholds
  • Tabulate maximum number of consecutive business
    days in any zone in each month at or above
    threshold LMPs

24
Consecutive business days with DR vary with HR
threshold
  • There were many consecutive business days
    spanning several consecutive months when prices
    met or exceeded NBTT levels

25
Baseline integrity thresholds
  • Development of Demand Response Price Thresholds

26
Determining Baseline Integrity (BI) thresholds
  • CRA was asked by ISO-NE to determine implied heat
    rates and LMP thresholds needed to satisfy the
    13-day criterion, using actual hourly RT prices
  • Defined as the minimum heat rate for which the
    corresponding average LMP was not equaled or
    exceeded in any zone more than 13 consecutive
    business days in a month
  • Results
  • Results vary greatly by month, and range from
    7,900 to 14,200 BTU/kWh
  • For all months, BI thresholds are higher than the
    NBTT levels

27
Heat Rate Thresholds Meeting BI Criterion
28
Results of BI analysis using 2010 RT prices
Month 13-day BIPs 13-day BIHRs
1 70.2 9,100
2 55.9 8,500
3 43.5 9,100
4 39.1 8,800
5 55.7 12,100
6 51.0 9,700
7 67.2 13,000
8 59.5 12,400
9 61.2 14,200
10 43.2 11,200
11 52.4 11,200
12 64.3 7,900
About PowerShow.com