1 Introduction 2 The power system in context 3 Model - - PowerPoint PPT Presentation

Chaire European Electricity Markets (CEEM) Université Paris-Dauphine

OPTIAL CAPACITY INVESTMENTS AND FLEXIBILITY RESOURCES:

AN INVESTMENT MODEL INTEGRATING THE SHORT-TERM REQUIREMENTS

WITH THE LONG-RUN DYNAMICS

PhD student, Manuel Villavicencio

Lulea 24/08/2016

Agenda

Introduction

1

The power system in context

2

Model presentation

3

Simulation results: optimal mix with increasing I-RES shares

4

• 1. Introduction

3

DIFLEXO model: Dispatch, investments and flexibility optimization model A self-developed tool for market design benchmarking

System cost and “multiple services” approach: investment and operational costs Hydrothermal optimization: when and how to use available hydro resources Operational constraints: Ramping limits, min/max capacities, part-load efficiencies, etc. Reliability issues: reserve requirements as a function of I-RES penetration (non-event control)

Source: Strbac. Imperial College London, 2012.

Minutes (5 – 30 min)

LT ST RT

• 2. The power system in context

4

The shock of increasing I-RES shares on electricity markets

Long-term (years)

• Scissor effect
• Capacity adequacy problems
• Energy security issues

Short-term (h)

• Merit Order Effect
• Less flexible dispatch but higher flexibility is needed
• Higher operational cost for committed units due to cycling
• Missing money problem

Real time (sec-min)

• Increased need for ancillary services
• Balancing issues
• Need for an enhanced congestion management

• 2. The power system in context

5

System dependent and interrelated issues

Long-term (years)

• Scissor effect
• Capacity adequacy problems (scissor effect)
• Energy security issues

Short-term (h)

• Merit Order Effect
• Less flexible dispatch but higher flexibility is needed
• Higher operational cost for committed units
• Missing money problem

Real time (sec-min)

• Increased need for ancillary services
• Balancing issues
• Congestion management

Multiple services, values and revenue sources

Capacity and flexibility adequacy (long-term)

• Available capacity
• Investment savings

(generation and network)

Flexibility for power and energy supply (short-term)

• « Peak Shaving » (Intraday)
• I-RES integration
• Weekday/weekend arbitrage

Reliability and flexibility for security supply (real time)

• Balancing and load following
• Congestion management
• System stability
• Other ancillary system services
• 2. The power system in context

6

The cheapest power generation technologies might not deliver the greatest value to the system

• 3. Model presentation: Research questions

7

• 1. Do considering flexibility and reliability requirements matter while optimizing

capacity investments? – What is their impact? – Do new flexibility resources have a role to play ? (e.i. electric energy storage (EES) and demand side management (DSM) capabilities) – Are they complementary or in competition?

• 2. What is the real value of generation technology’s capacity (conventional, I-

RES)? – Is that value dependent on the power system representation adopted? – How much flexibility is accounted from conventional units ? At what cost ?

What electricity mix to minimize the total system cost and comply with

• perability and reliability requirements?

• 3. Model presentation: Objective function

8

𝒁 = 𝐽𝑑𝑝𝑜 +

𝑑𝑝𝑜

𝑃&𝑁𝑑𝑝𝑜,𝑢 + 𝐺

𝑑𝑝𝑜,𝑢 + 𝐷𝑃2𝑑𝑝𝑜,𝑢 +∆𝐻𝑑𝑝𝑜,𝑢 𝑢 𝑑𝑝𝑜

+ 𝐽𝑆𝐹𝑇

𝑠𝑓𝑡

+ 𝑃&𝑁𝑠𝑓𝑡,𝑢 + 𝑆𝐹𝐷𝑠𝑓𝑡,𝑢

𝑢 𝑠𝑓𝑡

+ 𝐽𝑓𝑓𝑡

𝑓𝑓𝑡

+ 𝑃&𝑁𝑓𝑓𝑡,𝑢

𝑢 𝑓𝑓𝑡

+ 𝐽𝑒𝑡𝑛

𝑒𝑡𝑛

+ 𝐸𝑇𝑁𝑚𝑑,𝑢

𝑢 𝑚𝑑

+ 𝐸𝑇𝑁𝑚𝑡,𝑢

𝑢 𝑚𝑡

S.T. operational constraints:

• Ramping constraints
• Min/max generation level
• Part load efficiencies
• Spinning and non-

spinning reserve supply capabilities

• DSM and EES operation

related constraints

• Clean energy policies…

min

System cost represented as Y :

LT ST RT

Set Element Description T 𝑢 ∈ T Time slice I I ∈ I Generation technologies I ⊇ Con 𝑑𝑝𝑜 ∈ I Conventional technologies I ⊇ RES 𝑠𝑓𝑡 ∈ I Renewable energy sources I ⊇ EES 𝑓𝑓𝑡 ∈ I Electric energy storage technologies DSM ⊇ LC 𝑚𝑑 ∈ DSM Demand side management able to supply load curtailment DSM ⊇ LS 𝑚𝑡 ∈ DSM Demand side management able to supply load shifting

• 3. Model presentation: The energy only market (EOM)

9

Clearing the power market including new flexibility resources (i.e. DSM and EES):

Lt

base 1 + 𝜀 = 𝐻𝑚𝑠𝑓𝑡,𝑢− 𝐻𝑑𝑣𝑠𝑓𝑡,𝑢 𝑠𝑓𝑡

+ 𝐻𝑑𝑝𝑜,𝑢

𝑡𝑧𝑜𝑑 𝑑𝑝𝑜

+ 𝑇𝑓𝑓𝑡,𝑢

𝑒𝑑ℎ − 𝑇𝑓𝑓𝑡,𝑢 𝑑ℎ

+ 𝐸𝑇𝑁𝑚𝑚𝑑,𝑢

𝑚𝑑 𝑓𝑓𝑡

+ 𝐸𝑇𝑁𝑚𝑡,𝑢𝑢,𝑢

𝑒𝑝 𝑢𝑢=𝑢+𝑀𝑚𝑡 𝑢𝑢=𝑢−𝑀𝑚𝑡 𝑚𝑡

− 𝐸𝑇𝑁𝑚𝑡,𝑢

𝑣𝑞 𝑚𝑡

∀ 𝑢

Supply side flexibility Demand side flexibility

Variable Unit Description 𝐻𝑑𝑝𝑜,𝑢

𝑡𝑧𝑜𝑑

[GW] Synchronized power level of technology con

• n time t

𝐻𝑚

𝑠𝑓𝑡,𝑢

[GW] Power level of RES unit res 𝐻𝑠𝑓𝑡,𝑢

𝑑𝑣

[GW] Power curtailed by res on hour t 𝑇𝑓𝑓𝑡,𝑢

𝑑ℎ

[GW] Power demanded by storage unit ees on time t 𝑇𝑓𝑓𝑡,𝑢

𝑒𝑑ℎ

[GW] Power supplied by storage unit ees on time t 𝐸𝑇𝑁𝑚

𝑚𝑑,𝑢

[GW] DSM curtailment of load lc on time t 𝐸𝑇𝑁𝑚𝑡,𝑢

𝑣𝑞

[GW] DSM shifting up ls on time t 𝐸𝑇𝑁𝑚𝑡,𝑢,𝑢𝑢

𝑒𝑝

[GW] DSM shifting up ls on time tt from t

EOM

10

• 3. Model presentation: The balancing market

“Operating Reserves and Variable Generation”

Balancing markets: tackling variability and uncertainty of net load with the FRR

Source: NREL 2011, “Operating Reserves and Variable Generation”

Total forecast error Residual imbalance Forecast error covered by BRPs

𝜁𝑚

𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑀𝑢 𝑐𝑏𝑡𝑓 1 + 𝜀 + 𝜁𝑠𝑓𝑡 𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑄 𝑠𝑓𝑡 = 𝐻𝑑𝑝𝑜,𝑢 𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑞

𝑡𝑞

+ 𝐻𝑑𝑝𝑜,𝑢

𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜

𝑜𝑡𝑞

+ 𝑇𝑓𝑓𝑡,𝑢

𝑑ℎ,𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 + 𝑇𝑓𝑓𝑡,𝑢 𝑒𝑑ℎ,𝑛𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑓𝑓𝑡 𝑑𝑝𝑜 𝑠𝑓𝑡

𝜁𝑚

𝑏𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑀𝑢 𝑐𝑏𝑡𝑓 1 + 𝜀 + 𝜁𝑠𝑓𝑡 𝑏𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑄 𝑠𝑓𝑡 = 𝐻𝑑𝑝𝑜,𝑢 𝑏𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 + 𝑇𝑓𝑓𝑡,𝑢 𝑑ℎ,𝑏𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 + 𝑇𝑓𝑓𝑡,𝑢 𝑒𝑑ℎ,𝑏𝐺𝑆𝑆𝑣𝑞/𝑒𝑝𝑥𝑜 𝑓𝑓𝑡 𝑑𝑝𝑜 𝑠𝑓𝑡

Residual forecast error due to uncertainty Residual imbalance due to variability

Residual forecast error

• 3. Model presentation: The balancing market

11

Not-event secondary control driven by variability and uncertainty of RE generation

aFRR up mFRR up aFRR down mFRR down

Bidders: All spinning and non-spinning units + EES units Bidders: Synchronized units only + EES units

Assets competing to supply multiple services at least cost

• 3. Model presentation: Solving algorithm

12

Capturing flexibility needs Optimal dispatch for unit scheduling to minimize

• perational cost

Verifying reliability compliance

Min Total Cost

Optimal investments Optimal dispatch Optimal reserve supply

Capacity adequacy level Defining investments (capacity and cost) given operational constraints Defining annual net load Defining reserve requirements

aFRR up aFRR down mFRR up mFRR down EOM

LT RT Optimal dispatch

Optimal scheduling of available capacity (generation, EES and DSM) regarding

• perational constraints and variable cost

Optimal FRR allocation

Verifying reliability compliance given dispatched unit regulating capabilities

Optimal investments

Investments: set installed capacity (I-RES, conventionals, EES) and fixed cost Defines annual net load Defines reserve requirements

ST

• 4. Simulation: Valuating flexibility resources by the difference

13

Experimental setup: optimal mix increasing I-RES shares progressively allowing for investments on flexibility resources (e.i. EES and dsm)

• Greenfield system without interconnections
• Perimeter and dataset: Load (Lt) and I-RES capacity factors of France on 2013
• Hourly time slice and 8760 hours simulated
• Considered portfolio of technologies: endogenous investment on

– Power generation technologies: Nuclear, reservoir hydro, hard coal, lignite, CCGT, CT (high peak), wind and solar (including curtailment decisions). – Bulk storage technologies (EES) : PHS, CAES, VRFB, NaS, Li-ion – DSM: Load curtailment and load shifting less than 1% and 2% of Lt respectively – Other RES: Fatal hydro, Biomass and other RE accounted on the residual load.

• Cost and parameters compiled from reports of DIW, Black and Veatch, IEA, EPRI, NREL and
• ther scientific publications.

Solved on GAMS under CPLEX 12.5 using the barrier algorithm

• 4. Simulation results: Total system cost

14

Subject to hypothesis:

CO2 tax: 20€/ton No interconnection considered DSM < 2% Lt Greenfield optimization

• System

cost increases “exponentially” with RE shares.

• Investing on new flexibility

resources allows to integrate RES at least cost.

40,7 42,5 45,6 48,9 53,2 59,2 66,9 76,0 86,7 99,7 150,7 25000 50000 75000 100000 125000 150000 175000 200000 225000 250000 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [M€] RE share

Total system cost with new flexibility resources

Load following cost CO2 emissions cost Total O&M cost Total fuel cost Total overnight cost

42,0 43,9 47,2 51,6 58,0 66,1 76,0 88,9 103,3 125,9 234,8 25000 50000 75000 100000 125000 150000 175000 200000 225000 250000 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [M€] RE share

Total system cost w/o new flexibility resources

• 4. Simulation results: valuation flexibility by the difference

15

• The value of flexibility is always positive and smoothly increases with RE penetration according to the cost

assumptions adopted (e.i. CO2 and fuel prices) . a. At low RE shares, investing on flexibility assets causes additional overnight and O&M cost, but they are compensated with cost savings on fuel, CO2 and load following. b. At mid to high RE penetration levels, the value of flexibility is mainly driven by fuel savings. The divergence between overnight cost and O&M cost becomes negligible. c. At very high RE penetration levels, the value of flexibility depends mainly on overnight and O&M savings.

• 400
• 200

200 400 600 800 1000 1200 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [€/Kw yr] RE share

Value of flexibility investments on the optimal electrcity mix (by the difference)

Load following cost CO2 emissions cost Total fuel cost Total O&M cost Total overnight cost System value of flexible resources

• 4. Simulation results: optimal power generation mix

16

Subject to assumptions:

• No interconnections
• DSM is not allowed to supply

reserve

• 20€/ton of CO2
• W/o flexibility resources capacity

investments increases exponentially when increasing RE shares because of the low capacity value of RES.

• Investing on EES and DSM assets

highly increases capacity value of RES.

…

15 36 59 83 110 140 172 203 240 379 100 200 300 400 500 600 700 800 0,1 0,2 0,3 0,4 0,5 0,6 1 Capacity [GW] RE share

Optimal mix with new flexibility resources

PV Wind Reservoir CT CCGT Hard coal Lig Nuclear

15 36 61 91 123 159 200 239 313 735 1 5 30 62 100 200 300 400 500 600 700 800 0,1 0,2 0,3 0,4 0,5 0,6 1 Capacity [GW] RE share

Optimal mix w/o new flexibility resources

…

20 40 60 80 100 120 140 160 180 200 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [GW] RE share

Share of flexibility resources

CT DSM capacity CAES PHS VRFB NaS Li-ion

• 4. Simulation results: optimal investments on flexibility resources

17

• DSM is the most cost competitive flexibility source.
• DSM alone is not enough to cover up the entire flexibility need (capacity and reliability).
• Assuming 2020 cost forecast for EES technologies, PHS and CAES becomes cost competitive

and optimal from low RE shares. Investments on VRFB technologies are cost effective for mid to high RE penetrations as well as CT.

Current PHS capacity in the EU Current PHS capacity in the world

• 4. Simulation results: CO2 emissions

18

50 100 150 200 250 300 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

t CO2/yr

RE share

CO2 emissions

Full - RMSE=1 With EES & DSM Full - RMSE=1 w/o EES or DSM

• When investing on flexibility resources are not allowed there are no CO2 emission reductions due to the

low capacity value of I-RES and higher operating reserve required supplied by more peaking units.

• Only investments on low emission flexibility resources like EES and DSM can palliate this effect.

Subject to:

Approach: System cost minimization study (Greenfield case) CO2 tax: 20 €/ton No interconnection considered Only con and EES can supply reserve

Because of the non-dispatchable nature of RE resources developed in France. It would be a better case for dispatchable based RE goals such as biomass or hydro (Sweden..).

• 4. Simulation results: optimal mix with increasing I-RES shares

19

Flexibility value and opportunities for EES technologies:

100 200 300 400 500 600 700 800 900 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [€/kW yr] RE share

Technology forecasted cost for 2013 - PNNL, 2012

System value of flexible resources Li-ion NaS VRFB PHS ACAES

100 200 300 400 500 600 700 800 900 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [€/kW yr] RE share

Technology forecasted cost for 2020 - PNNL, 2012

100 200 300 400 500 600 700 800 900 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 [€/kW yr] RE share

Technology forecasted cost for 2050 - Pape et Al. 2014

Chaire European Electricity Markets (CEEM) Université Paris-Dauphine

Thank you for your attention. Any questions?

manuel.villavicenio@dauphine.com

APPENDIX

• 1. The power system in context

22

The impact of increasing shares of I-RES on power systems and electricity markets: Amplified uncertainty and variability of net load in the short-term

Short-term (sec-min)

• Balancing: augmented need for non-event operating reserve (Power control and

load following): Need for improved forecast

• Higher need for other ancillary services: need for enhanced BRP
• Congestion management: LMP, market splitting, market coupling.

Source: EURELECTRIC, 2010

Uncertainty Variability

• 1. The power system in context

23

The impact of increasing shares of I-RES on power systems and electricity markets: Low short-run marginal cost I-RES enter first in the merit order

Exceeding offers

• n the balancing

market Source: IEA, 2014 10GW of wind 10GW of peaking units

DV DP

S-D equilibrium on hour t

Mid-term (h)

• Merit Order Effect: reduced volumes and prices => reduced revenues
• There are more constringent ramping restrictions binding the dispatch, but

there is a higher need for availability and flexibility. Capacity mechanisms.

• Additional cost are incurred due to load following, wear and tear costs and

part load efficiencies => higher operational cost of individual units

• Net load duration curve decreases and becomes stepper

=> missing money problem

• 1. The power system in context

24

The impact of increasing shares of I-RES on power systems and electricity markets: Low short-run marginal cost I-RES enter first in the merit order

Mid-term (h)

• Merit Order Effect: reduced volumes and prices => reduced revenues
• There are more constringent ramping restrictions binding the dispatch, but

there is a higher need for flexibility

• Additional cost are incurred due to load following, wear and tear costs and

part load efficiencies => higher operational cost of individual units

• Net load duration curve decreases and becomes stepper => missing money

problem (“Missing money or missing markets”, Newbery 2015)

Net load duration curve

Exceeding offers

• n the balancing

market

DCF

• 1. The power system in context

25

The impact of increasing shares of I-RES on power systems and electricity markets: Depreciated profits: peaking plants mothballing and no investment incentives

Long-term (years)

• Cumulated losses of profits causes a SCISSOR EFFECT in the long-run

=> Retirement of peaking plants. E.x: Mothballing of 20GW CCGT capacity from EU markets of which 8,8GW were “recently” installed units

• Capacity adequacy problems: depreciated prices cause no inframarginal rent

threatening incentives for new investments.

• Energy security issues: not enough capacity when needed => blackout risk

Source: Robinson, 2015

• 4. Model presentation

26

Source: Palmintier, 2014.

• Palmintier. “Flexibility in Generation Planning : Identifying Key Operating Constraints”. PSCC 2014.

FULL = Complete MILP formulation with unit clustering 8760h: MIP gap = 0.1% => solution time > 60h

Variations

Relative accuracy

• 50 combinations of UC+Maintenance+Planning
• MILP relaxations are considered for each variation

• 4. Model presentation

27

Source: Palmintier, 2014.

• Palmintier. “Flexibility in Generation Planning : Identifying Key Operating Constraints”. PSCC 2014.

FULL = Complete MILP formulation with unit clustering 8760h: MIP gap = 0.1% => solution time > 60h

variations

Relative accuracy

• 50 combinations of UC+Maintenance+Planning
• Relaxations are considered for each variation

Palmintier’s formulation:

• No EES or DSM considered
• No endogenous investments on RES
• Deterministic reserve dimensioning

𝑍 = 𝐽𝑑𝑝𝑜 +

𝑑𝑝𝑜

𝑃&𝑁𝑑𝑝𝑜,𝑢 + 𝐺𝑑𝑝𝑜,𝑢 + 𝐷𝑃2𝑑𝑝𝑜,𝑢 +∆𝐻𝑑𝑝𝑜,𝑢

𝑢 𝑑𝑝𝑜

+ 𝐽𝑆𝐹𝑇

𝑠𝑓𝑡

+ 𝑃&𝑁𝑠𝑓𝑡,𝑢 + 𝑆𝐹𝐷𝑠𝑓𝑡,𝑢

𝑢 𝑠𝑓𝑡

+ 𝐽𝑓𝑓𝑡

𝑓𝑓𝑡

+ 𝑃&𝑁𝑓𝑓𝑡,𝑢

𝑢 𝑓𝑓𝑡

+ 𝐸𝑇𝑁𝑚𝑑,𝑢

𝑢 𝑚𝑑

+ 𝐸𝑇𝑁𝑚𝑡,𝑢

𝑢 𝑚𝑡

• 4. Model presentation

28

Source: Palmintier, 2014.

• Palmintier. “Flexibility in Generation Planning : Identifying Key Operating Constraints”. PSCC 2014.

Complete MILP formulation with unit clustering 8760h: MIP gap = 0.1% => > 60h variations

Relative accuracy

• 4. Model presentation

29

Source: Palmintier, PSCC 2014.

A LP formulation seems to be the best compromise between accuracy and resolution time

Pareto front

variations

Relative accuracy

• 4. Model presentation

30

• Min power limits: when using a technology based dispatch and

Pmin > 0, it implicitly contains must-run obligations which are not convenient to schedule peak and extreme peak units.

• Ramping constraints issues: technology ramping in MW/min

can overestimate real ramping capabilities on hourly scheduling.

• Part load efficiencies: non-linear by nature they use to be

step-wise linearized

• r

linearly approximated, thus,

• verestimating fuel consumption and CO2 emissions.

Modeling issues when adopting LP formulations:

Base unit

• Flex. asset

(peak unit?)

• 4. Model presentation

31

Modeling issues for representing flexibility assets:

• EES technologies:
• Investments: energy and capacity should be separately optimized.
• Operation: Constrained by installed capacity but also by energy stock (path dependence)
• DSM operation: Using the “virtual stock analogy” to model load shifting (LS) is insufficient

=> the “debit/credit moving window” formulation (Zerranh and Schill, 2015) was adopted

EES

Source: Zerranh and Schill, 2015.

e.x: stock(t=0) = 0

LS

e.x: stock(t=0) = 0

32

Accounting for variability and uncertainty of net demand

• 4. Model presentation

Source: NREL 2011, “Operating Reserves and Variable Generation”

Forecast error Residual imbalance

• 4. Model presentation

33

Probabilistic vs. deterministic methodologies for dimensioning FRR

Deterministic method: Probabilistic method: based on the recursive convolution method of residual system

imbalances

𝐺𝑆𝑆 = 10 𝑀𝑛𝑏𝑦 + 1502 - 150 But how much FRR is required?

Source: Stiphout, 2014. FRR dimensioning based on ELIA methodology, 2012 Source: Hirth & Ziegenhagen, 2013. Control Power and Variable Renewables: A Glimpse at German Data

• 1. Normalisation:

1/Cap

• 2. Convolution:

• 4. Model presentation

34

ENTSO-E secondary control

Source: NREL 2011, “Operating Reserves and Variable Generation”

• 4. Model presentation

35

How will be covered this the gap? By who?

Source: ELIA, 2012.

• 4. Model presentation

36

aFRR up/down mFRR up/down

Source: NREL 2011, “Operating Reserves and Variable Generation”

ENTSO-E NERC

• 4. Model presentation

37

LT MT ST

Energy balancing: J-1 market Secondary control aFRR up mFRR up aFRR down mFRR down

Multiple services over the entire time horizon

• 4. Model presentation

38

Source: ELIA, 2012.

Net demand including I-RES curtailment: Variability pattern

RE_share = 0% RE_share = 20% RE_share = 40% RE_share = 60%

• 4. Model results: optimal mix with increasing I-RES shares

39

Effect of I-RES curtailment

• 4. Model results: optimal mix with increasing I-RES shares

40

Net demand including I-RES curtailment: duration pattern RE_share = 0%

RE_share = 20% RE_share = 40% RE_share = 60%

Effect of I-RES curtailment

• 4. Model results: optimal mix with increasing I-RES shares

41

There is a high level of flexibility supplied by DSM, subject to hypothesis: Load curtailment: LC < 4% Lt Load shifting: LS < 3% Lt

? …

DSM_CURT DSM load curtailment h1 h170 h411 h652 h894 h1161 h1454 h1747 h2040 h2333 h2627 h2920 h3213 h3506 h3799 h4092 h4385 h4678 h4972 h5265 h5558 h5851 h6144 h6437 h6730 h7023 h7316 h7610 h7903 h8196 h8489 3,5 3,4 3,3 3,2 3,1 3 2,9 2,8 2,7 2,6 2,5 2,4 2,3 2,2 2,1 2 1,9 1,8 1,7 1,6 1,5 1,4 1,3 1,2 1,1 1 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 DSM_UP DSM load shifting up h1 h167 h404 h642 h879 h1142 h1430 h1719 h2007 h2295 h2583 h2872 h3160 h3448 h3737 h4025 h4313 h4601 h4890 h5178 h5466 h5755 h6043 h6331 h6620 h6908 h7196 h7484 h7773 h8061 h8349 h8638 3,5 3,4 3,3 3,2 3,1 3 2,9 2,8 2,7 2,6 2,5 2,4 2,3 2,2 2,1 2 1,9 1,8 1,7 1,6 1,5 1,4 1,3 1,2 1,1 1 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1

Load shifting Load curtailment

Lpeak = 82,83 GW => Total_DSMpeak = 5,79 GW

Operational Benefits Monetizing the Value of Energy Storage. Source: EPRI 2010.

Positioning of Energy Storage Technologies. Source : EPRI 2010.

EES technologies: Which ones and what for?