system analysis advisory committee - a new metric - michael schilmoeller tuesday, september 27, 2011
TRANSCRIPT
System AnalysisAdvisory Committee
- A New Metric -
Michael SchilmoellerTuesday, September 27, 2011
2
Why Consider a New Metric?
• We have observed that decision makers tend to emphasize least-risk plans
• The choice of least-risk plans is strongly (exclusively?) influenced by a handful of futures. About 70 of the 75 “worst” (highest-cost) futures are common among all the plans on the efficient frontier
• The costs in these futures is largely determined by the higher loads in these futures
• Is it appropriate that least-risk plans are strongly influenced by high-load futures?
3
What’s the Difference?
• If there is only a single, fixed load forecast, there isn’t any difference: a plan that provides for minimum cost for the Region provides minimum unit energy cost.
• However, if loads vary from future to future, there may be a significant difference
4
A Few Key Points
• Loads are “frozen efficiency” loads– Loads changes therefore do not correspond
to any energy efficiency improvements or changes
• A unit energy cost (¢/kWh) is NOT a utility rate (¢/kWh)
5
Frozen Efficiency Loads• The loads in the RPM are what we refer to as “frozen
efficiency loads”• Because we want to consider all sources of conservation
as resources, we use a load forecast that assumes:– Discontinuation of utility conservation programs– No new appliance standards– No new building codes– No electricity price-induced conservation
• They do include transmission and distribution losses• Candidate conservation shows up on the resource side
of the ledger: we treat its cost and energy like any other future resource
6
Load Futures• Changes in loads among futures have nothing to do
changes in conservation policy• One way to think of load futures is in terms of
households and businesses moving into or out of the region, rather than changes in usage by individual customers
• It is also true that individual customers may buy more or fewer gadgets and require more electricity
• What is true in any of these situations, however, is that individual regional customer satisfaction will track their unit energy cost, irrespective of whether their requirements are met through energy efficiency or purchases of electricity from their utility
7
Unit Energy Cost ≠ Rates
• Costs in the rate numerator include embedded costs; costs in unit energy cost do not
• Energy in the rate denominator is sales; energy in the unit energy cost is frozen efficiency load, including system losses
8
Start with the Cash Flows
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Period
Costs
9
Calculating NPV for One Future(discount rate = 4%)
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Period
Costs
1)0.041(
$12.3
2)0.041(
$12.8
3)0.041(
$13.4
20)0.041(
$23.5
20
1
004.0
)0.041(TT
TT
CNPV
NPV = $227.1 B
10
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
MW
a
Period
Costs and Loads
Calculating NPV Cost per MWh for One Future
8760)(22,429MWa1)0.041(
B $12.3
8760)(22,805MWa2
)0.041(
B $12.8
8760)(23,180MWa3
)0.041(
B $13.4
8760)(29,571MWa20
)0.041(
B $23.5
20
10
04.0
)04.01(tt ttL
tC
VNP
NPV/MWh = $1,017.03
11
Second Future: Lower Loads
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
MW
a
Period
Costs and Loads
12
Thinking of Load Growth inTerms of More Customers
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
MW
a
Period
Loads
13
Reduced Load in Terms of Fewer Customers …
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
MW
a
Period
Costs and Loads
14
Costs are Also Lower….
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
MW
a
Period
Costs and Loads
15
21,000
22,000
23,000
24,000
25,000
26,000
27,000
28,000
29,000
30,000
31,000
10.0
12.0
14.0
16.0
18.0
20.0
22.0
24.0
26.0
28.0
30.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
MW
a
Period
Costs and Loads
… And Costs Are Distributed Over Fewer Units of Energy
Higher load future: 29,571 MWa Lower load future: 25,428 MWaDifference: 4,143 MWa (-14 %)
Higher load future cost: $23.5 BLower load future cost: $21.5 BDifference: $2.0 B (-8.5 %)
While cost in the last year goes down 8.5 percent, cost per kWh increases by 6.4 percent !
064.01254285.23
295715.21
29571/5.23
29571/5.2325428/5.21
16
Is This Future Better or Worse Than the First Future?
• Is it a bad outcome if all customers see higher bills because regional loads have fallen more than regional cost?
• Is it a bad outcome for all customers to see lower bills although total regional costs are higher?
• As we currently measure cost and risk, the first is a good outcome and the second is a bad outcome.
17
While You Consider The Previous Questions …
• How would the new calculation be incorporated into RPM resource choice?
• What are the likely consequences of the new metric?
• What are the differences between this and a rate calculation?
18
The Existing Distribution
0
10
20
30
40
50
60
70
80
9030
40
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
Freq
uenc
y
Billions of 2006 Constant Dollars
NPV 20-Year Study Costs
C:\Documents and Settings\Michael Schilmoeller\Desktop\NWPCC - Council\SAAC\Presentation materials\L813 NPV Costs.xlsm
19
1. Use of the New Observations
• We get a distribution with a little different shape and a much different scale
• We would calculate the expected cost and risk measures exactly as before but would apply them to the new “observations”
20
The New Distribution
0
10
20
30
40
50
60
70
80
9050
0
667
833
1,00
0
1,16
7
1,33
3
1,50
0
1,66
7
1,83
3
2,00
0
2,16
7
2,33
3
2,50
0
2,66
7
2,83
3
3,00
0
3,16
7
3,33
3
3,50
0
3,66
7
3,83
3
Freq
uenc
y
2006 Constant NPV Dollars per MWh
NPV 20-Year Study Normalized Costs
The first future we just examined is probably counted in this bin
21
2. There is a Difference in Plans on the Efficient Frontier
• We ran the same model twice, but optimized the plan selection and creation of the efficient frontier using the two metrics
• Both model runs produce both metrics for cost and risk for each plan
• We identified the plans on efficient frontier with respect to the new load-normalized NPV metric
• Using the traditional NPV metric values for these plans, we plotted them on the feasibility space with traditional coordinates
22
What do Plans Selected Using the Modified Metric Look Like?
23
Differences between theLeast-Risk Plans
• Because high-load futures play a less prominent role in the selection of resources, we would expect to see less resource capacity optioned
Cns
rvn_
Lost
Opp
ortu
nity
Cns
rvn_
Dis
patc
habl
e
CC
CT
_CY
_Dec
19
CC
CT
_CY
_Dec
21
CC
CT
_CY
_Dec
23
SC
CT
_CY
_Dec
17
SC
CT
_CY
_Dec
19
SC
CT
_CY
_Dec
21
SC
CT
_CY
_Dec
23
New metric 50 60 1512 1512 1512 324 810 810 810Existing metric 60 100 1890 1890 1890 648 1458 1620 1620
24
Differences between theLeast-Risk Plans
Cns
v_M
Wa
CO
2Avg
2025
woT
CO
2Avg
2030
woT
Cns
v_LO
_en
Cns
v_LO
cst
Cns
v_N
LOen
Cns
v_N
LOcs
t
Cnv
t_cs
t
CO
2Avg
2030
wT
CO
2Avg
2025
wT
Rat
eStD
evIn
cr
Rat
eMax
Incr
New metric 5890.4 34.1 35.3 3087.0 33.6 2803.4 35.3 34.4 26.2 25.3 0.1 0.3Existing metric 6074.0 33.9 34.7 3157.0 35.2 2917.1 40.0 37.5 25.3 25.0 0.1 0.3
• The effect on conservation targets, CO2 produced, and rate variation is minimal, however (e.g., conservation drops 184MWa)
• With fewer new, cleaner turbines, CO2 production increases slightly
25
Conclusions• This metric provides an alternative concept of
“bad” – or risky – and “good” futures.• The new metric
– Is simply the application of the existing statistical measures to new values for the observations
– Will result in different least-risk plans, probably with less capacity and conservation optioned
• This metric would be added to the existing metrics, not replace them
26
Final Observation
• The cost per unit energy metric is not a surrogate for utility rates
• Rates are calculated differently!– The energy value appearing in the denominator of a rate
calculation is sales; it is reduced by conservation energy and T&D losses
– Cost of conservation often is expensed
• We know that using rates as an objective does not guarantee lowest total resource cost; TRC is in fact likely to be higher
27
End