Download pdf - Technical Working Group #3: Demand Curve Shape Analysis · • Several shapes of downward sloping demand curves were tested to reveal their merits and shortfalls • Vertical , Deeply

Technical Working Group #3:

Demand Curve Shape Analysis

May 04, 2018

Intent

• Develop a shared understanding of the rationale for the

proposed demand curve shape

• Provide further explanation as to why the demand curve is

offset/right shifted at 1.0 x Net CONE

2

Outline

• Resource Adequacy Considerations

• Demand Curve Performance

– Desired characteristics

– Simulation results

• Test specific shapes

– Shortcomings of vertical curve, deeply convex curve

– Reasoning for selecting proposed curve

– Reasoning for right shifting the curve

3

4

Resource Adequacy Study Draft Results

• Expected Unserved Energy (EUE) is measured in MWh at various

Reserve Margin levels

• EUE increases exponentially as Reserve Margin declines

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

15.0% 20.0% 25.0% 30.0% 35.0%

Exp

ecte

d U

nse

rved

En

egy,

MW

h

Reserve Margin

Resource Adequacy Study - Draft Results

964 MWh EUE (0.0011% EUE) 400 MWh EUE

100 MWh EUE 1 MWh EUE

5

Resource Adequacy - Historical

• The historical range for resource adequacy has been 17.5% to

38.7% reserve margin between 2008 and 2017

• The range of the proposed demand curve is between 18.7% and

30.5%, within the historical reliability range over the past decade

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

15.0% 20.0% 25.0% 30.0% 35.0% 40.0%

Exp

ecte

d U

nse

rved

En

egy,

MW

h

Reserve Margin

Resource Adequacy Study - Draft Results

964MWh EUE (.0011% EUE) 1 MWh EUE 2009 2017

6

Supply & Demand Shocks

• Supply shocks are based on historical retirements and additions

• Demand shocks are based on load forecast errors

• Demand and supply shocks follow normal distributions

• These shocks are used in the Monte Carlo simulation to evaluate

average performance of demand curves

-

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 20% 40% 60% 80% 100% 120% 140%

Net

CO

NE

Mu

ltip

le

% of Reliability Requirement

Indicative Range of Demand and Supply Shocks

Low Demand Low Supply High Demand High Supply

0

20

40

60

80

100

120

# o

f O

bse

rvat

ion

s

MW

Distribution of Supply & Demand Shocks

Supply Shocks Demand Shocks

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

15.0% 20.0% 25.0% 30.0% 35.0% 40.0%

Exp

ecte

d U

nse

rved

En

egy,

MW

h

Reserve Margin

Resource Adequacy Study and Distribution of Shocks

Distribution of Shocks Resource Adequacy Curve

964MWh EUE (.0011% EUE) 400 MWh EUE

1 MWh EUE

7

Resource Adequacy – Supply Shocks and

Target Offset

• Expected unserved energy grows exponential when supply is scarce

• The distribution of shocks is approximately normal

• As a result, a demand curve with the target set at the target reserve

margin of 400MWh EUE and 1X Net CONE will not achieve 400MWh

EUE on average– A target set at 400 MWh EUE would achieve 2400 MWh EUE on average

• The asymmetry of the Resource Adequacy curve requires an offset of

approximately 5% to achieve the resource adequacy target on average

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

15.0% 20.0% 25.0% 30.0% 35.0% 40.0%

Exp

ecte

d U

nse

rved

En

egy,

MW

h

Reserve Margin

Resource Adequacy Study Target Offset

Distribution of Shocks Resource Adequacy Curve

964MWh EUE (.0011% EUE) 400 MWh EUE

1 MWh EUE

8

Resource Adequacy Target

• The Government has provided a minimum reliability standard of 0.0011%

unserved energy– This equates to a minimum of approximately 964MWh EUE in the reference

year

• A target level has not yet been defined, but a range of 100MWh EUE to

400 MWh EUE has been tested

Select Demand Curve Evaluations

10

Demand Curve Performance

• In line with the Demand Curve Principles developed in SAM 3.0, a

demand curve can be considered to perform well if it:

Can accommodate shocks in supply and demand

Performs consistently to a reliability target, and typically above

the minimum reliability standard

Accommodates a wide range of possible outcomes, with

changes in supply, allowing for a well functioning and

competitive market

Minimizes the scope of market power exercise

Provides adequate revenue to incent entry and exit when

required i.e. leads to the ‘right’ level of capacity

Demand Curve Evaluation

• Market Conditions are simulated using a Monte Carlo

simulation model

– Model solves to an average price of net cone

– Tests the impacts of supply/demand shocks on price and

reliability outcomes under various demand curves

• Key ‘characteristics’ are evaluated to assess if the curve is

meeting the performance requirements

– Price Volatility

– Cleared capacity

– Average reliability, and frequency

11

12

Selected Curves for Review

• Several shapes of downward sloping demand curves were tested to reveal their merits and

shortfalls

• Vertical , Deeply Convex, and the CMD1.0 Proposed Curve are evaluated

• Each of these curves performs to the 400 MWh EUE reliability level on average over 1,000

simulations

– Proposed Curve with 100 MWh EUE also tested

• Each curve maintains a price cap of 1.75X Net CONE

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

95% 100% 105% 110% 115% 120%

Net

CO

NE

Mu

ltip

le


400 MWh EUE Demand Curves with Various Shapes

Deeply Convex, Cap at 1.75x Net CONE (400 MWh EUE)

Tuned Convex, Cap at 1.75x Net CONE

Vertical Curve, Cap at 1.75x Net CONE (400 MWh EUE)

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 50% 100% 150%

Net

CO

NE

Mu

ltip

le


400 MWh EUE Demand Curves with Various Shapes




13

Vertical Curve

• A vertical curve produces significant price volatility, with no change in procurement level– Vertical curve is perfectly inelastic

• Vertical curve is not robust to load forecast errors in resource adequacy modeling

• Price outcomes are heavily concentrated in the low % of Net CONE range & price cap

• Highly susceptible to market power exercise, since any change in volume can materially

change price

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 20% 40% 60% 80% 100%120%140%160%

Pri

ce

as

% o

f N

et

CO

NE

Quantity as % of Reliability Requirement


0

100

200

300

400

< 2

0%

32

%

48

%

64

%

80

%

96

%

11

2%

12

8%

14

4%

16

0%

Freq

uen

cy o

f P

rice

Price (% of Net CONE)

Foot Cap

Net CONE

Price Reliability

Average Standard

Deviation

Frequency

at Cap

Average

EUE

Average

LOLH

Average Quantity

as % of

Rel. Req.

Average Excess

(Deficit) Above

Rel. Req.

Average

Uncleared

Supply

Frequency

Below

Rel. Req.

Demand Curve ($/MW-d) ($/MW-d) (%) (MWh) (Hours) (%) (MW) (MW) (%)

Vertical Curve, Cap at 1.75x Net CONE (400 MWh EUE) $381 $236 35% 400 1.3 103% 353 446 11%

Tuned Convex, Cap at 1.6x Net CONE $381 $128 7% 402 1.3 105% 578 237 13%



0

200

400

600

800

< 8

5%

88

%

92

%

96

%

10

0%

10

4%

10

8%

11

2%

11

6%

12

0%

Freq

uen

cy o

f Q

uan

tity

Cleared Quantity (% of RR)

Foot

964 MWH EUE

400 MWH EUE

14

Demand Curve with Inflection Point Set at 1.0X

Net CONE and 100% of the Reliability Target

• With the inflection point at 1.0X Net CONE and 100% of the reliability target requires a deeply convex

curve to meet reliability outcome– The foot is meaningfully wider and longer than the middle section, skews price outcomes to the foot

• Provides a very small buffer to errors in net CONE estimation and resource adequacy modeling

• Tends to procure more supply than a less convex curve

• Has a high frequency of occurrences below the reliability requirement

• The steepness of the curve above Net CONE is highly susceptible to market power exercise, like the

vertical curve

0

20

40

60

80

100

< 8

5%

88

%

92

%

96

%

10

0%

10

4%

10

8%

11

2%

11

6%

12

0%

Freq

uen

cy o

f Q

uan

tity


FootCap

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 20% 40% 60% 80% 100%120%140%160%

Pri

ce

as

% o

f N

et

CO

NE



Price Reliability

Average Standard

Deviation

Frequency

at Cap

Average

EUE

Average

LOLH

Average Quantity

as % of

Rel. Req.

Average Excess

(Deficit) Above

Rel. Req.

Average

Uncleared

Supply

Frequency

Below

Rel. Req.


Deeply Convex, Cap at 1.75x Net CONE (400 MWh EUE) $382 $94 6% 399 1.2 105% 599 212 16%




0

50

100

150

200

250

< 2

0%

32

%

48

%

64

%

80

%

96

%

11

2%

12

8%

14

4%

16

0%

Freq

uen

cy o

f P

rice


Foot Cap

Net CONE

15

Proposed 1.75X Net CONE Convex Curve:

400 MWh EUE

• A slightly convex curve provides a robust range of price outcomes– Inflection point 5-6% above reliability requirement target level

• Reflects increasing value of capacity in times of scarcity

• Width helps mitigates market power exercise

• Amount of time market clears below minimum reliability requirement a concern

• Provides a meaningful buffer to errors in net CONE and resource adequacy modeling

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 20% 40% 60% 80% 100% 120%

Pri

ce

as

% o

f N

et

CO

NE



0

20

40

60

80

100

120

< 8

5%

88

%

92

%

96

%

10

0%

10

4%

10

8%

11

2%

11

6%

12

0%

Freq

uen

cy o

f Q

uan

tity


FootCap

Price Reliability

Average Standard

Deviation

Frequency

at Cap

Average

EUE

Average

LOLH

Average Quantity

as % of

Rel. Req.

Average Excess

(Deficit) Above

Rel. Req.

Average

Uncleared

Supply

Frequency

Below

Rel. Req.






0

20

40

60

80

100

< 2

0%

32

%

48

%

64

%

80

%

96

%

11

2%

12

8%

14

4%

16

0%

Freq

uen

cy o

f P

rice


Foot Cap

Net CONE

16

Proposed 1.75X Net CONE Convex Curve:

100 MWh EUE

• A 100 MWh EUE convex curve reduces the frequency below the minimum reliability

requirement

• The foot of the 100 MWh EUE curve increases from 116% to 126% of reliability target

• Provides a relatively normal distribution of prices an quantities

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

0% 20% 40% 60% 80% 100% 120% 140%

Pri

ce

as

% o

f N

et

CO

NE


Tuned Convex, Cap at 1.75x Net CONE (100 MWh EUE)

0

10

20

30

40

50

60

70

< 2

0%

32

%

48

%

64

%

80

%

96

%

11

2%

12

8%

14

4%

16

0%

Freq

uen

cy o

f P

rice


Foot Cap

Net CONE

Price Reliability

Average Standard

Deviation

Frequency

at Cap

Average

EUE

Average

LOLH

Average Quantity

as % of

Rel. Req.

Average Excess

(Deficit) Above

Rel. Req.

Average

Uncleared

Supply

Frequency

Below

Rel. Req.


Tuned Convex, Cap at 1.75x Net CONE (100 MWh EUE) $382 $122 2% 100 0.2 108% 874 229 4%




0

20

40

60

80

100

120

< 8

5%

88

%

92

%

96

%

10

0%

10

4%

10

8%

11

2%

11

6%

12

0%

Freq

uen

cy o

f Q

uan

tity


Foot

964 MWH EUE

100 MWH EUE

17

Conclusions

• A downward-sloping, convex, demand curve is preferred to vertical

curve

• Too much convexity can cause concentration of the price

outcomes below Net CONE

• Very steep curves enable market power exercise

• Stakeholder concerns regarding over procurement in CMD 1.0

proposed curve – Proposed curve clears below minimum resource adequacy standard

6.5% of iterations

• AESO continues to evaluate resource adequacy target and

demand curve shapes– Continued feedback welcomed