Requirements for electrical energy storage · Generation Storage DLR.de • Felix Cebulla • Chart 7 40 GW (4 TWh) Compressed air Pumped hydro Lithium-ion Hydrogen Nuclear Lignite-fired

Requirements for electrical energy storage

Dipl. Wi.-Ing. Felix Cebulla

German Aerospace Center (DLR)

Systems Analysis and Technology Assessment

State of research and results from the energy system model REMix

2nd German-Japanese Workshop on Renewable Energies

Stuttgart, 07/05/2017

DLR.de • Felix Cebulla • Chart 2

Agenda

I. The need for system flexibility

II. Status quo: storage capacity requirements

III. Storage demand in highly renewable European energy scenarios

IV. Conclusions

* J. Haas, F. Cebulla, K. Cao, W. Nowak, R. Palma-Behnke, C. Rahmann, and P. Mancarella, “Challenges and trends of energy storage expansionplanning for flexibility provision in low-carbon power systems – a review,” Renewable and Sustainable Energy Reviews, vol. 80, pp. 603–619, 2017.


The need for flexibility*

StorageOperationstrategies

Transmission

Variabilityofload VariabilityofvRES

Uncertaintyofload UncertaintyofvRES

Forecasterrorsofload Uncorrelatedandspatiallydistributed

Outageofgenerators

Maintenanceofgenerators

Conventionalneedforflexibility Newneedforflexibility

Flexiblegeneration&

DSMOtherenergy

sectors

U.S.

25% 50% 75% 100%0

150

300

450

600

25% 50% 75% 100%

Bertsch et al.

Babrowski et al.

Frew

Inage

Adamek et al.

Zerrahn and Schill

Hartmann

Kühne

NREL Ref 2012

NREL Ref 2014

Steffen and Weber 2012

Ueckerdt et al.

Scholz et al.

Pape et al.

Steffen and Weber 2013

Pow

er c

apac

ity [G

W]

Variable renewable energy share [%]

Europe

* F. Cebulla, J. Haas, J. Eichman, W. Nowak, and P. Mancarella, “How much energy storage do we need? A review andsynthesis for the U.S., Europe, and Germany”.


Broad ranges of storage requirements* (I)Review of model-based assessments

Pow

er c

apac

ity [G

W]



Broad ranges of storage requirements (II)Review of model-based assessments

0

150

300

450

600

25% 50% 75% 100%0

150

300

450

600

25% 50% 75% 100%

Very PV-dominated

PV-dominated

Very wind-dominated

Wind-dominated

Balanced mix

U.S.Europe

… What other main drivers are there?

Energy system model REMix*

* H. C. Gils, Y. Scholz, T. Pregger, D. L. de Tena, and D. Heide, “Integrated modelling of variable renewable energy-based powersupply in Europe,” Energy, vol. 123, pp. 173–188, 2017.

Renewable Energy Mix Linear (mixed-integer) optimization model, written in GAMS, solved with CPLEX Minimize overall system costs Decision variables: capacity invest (single year, myopic, or path optimization) and

hourly dispatch of all assets

Sectors: power, heat, transportation, hydrogen infrastructure Renewables: wind (onshore, offshore), photovoltaic, hydro (pumped, run-of-river,

reservoir), biomass, geothermal, concentrated solar power (CSP) Fossil and nuclear thermal power plants (incl. CHP) Flexibility options: electricity storage, transmission grid expansion (AC,DC),

flexible CHP with thermal storage, demand response, controlled charging of BEV

Typical constraints: renewable shares or ratios, CO2 cap, minimum firm capacity, secondary or tertiary reserve, domestic generation shares


Storage requirements in Europe (I)Renewable and storage capacity

Generation Storage


40 GW (4 TWh)

Compressed air

Pumped hydro

Lithium-ion

Hydrogen

NuclearLignite-firedHard coal-firedCCGTGas turbineGeothermalHydro reservoirHydro run-of-riverPVOnshore windOffshore wind

Overall share of wind & PV [-]

50 GW50 GW

240 GW @ 95% RE share (EU)

Short-term storage

Storage requirements in Europe (II)Storage utilization


PV-dominated

Wind-dominated (onshore)

Wind-dominated (offshore)

Mid-term storage

Long-term storage

… But how robust are these results?

Robustness of storage capacity requirements*Influence of data assumptions and methodology


* F. Cebulla, “Storage demand in highly renewable energy scenarios for Europe: The influence of methodology and dataassumptions in model-based assessments,” University of Stuttgart, 2017, submitted.

Data Methodology

Investment costs for renewables, storage, and grid

Operational costs: fuel cost, CO2

certificates Variations of weather year for PV and

wind generation

Detailed unit-commitment vs. simple LP power plant modeling

Unlimited vs. restricted curtailments 20 node model vs. single node

representation (“copper plate”) Influence of temporal resolution

Further possible drivers Modeling approach: optimizations vs. simulation Model-inherent foresight, e.g. for investment and dispatch decisions (myopic, path, or

rolling horizon) Consideration of other technological details, e.g. DC approximation versus load-flow Multi criteria optimization (not solely system costs) and sector coupling

Influence of power plant modeling approach* (I)LP versus MIP


* F. Cebulla and T. Fichter, “Merit order or unit-commitment: How does thermal power plant modeling affect storage demand inenergy system models?,” Renewable Energy, vol. 105, pp. 117–132, 2017.

Over‐estimation of flexibility of fossil fired power plants …

… leads to an under‐estimation of storage utilization.

Realistic consideration of the flexibility leads to less ramping …

… and fosters an increase in storage utilization.

66%

Influence of power plant modeling approach (II)Effect on storage capacity expansion and utilization


0.00

0.25

0.50

0.75

1.00

LP MILPG

ener

atio

n sh

are

[-]LP MILP

33% 100%

a Before curtailments, storage- and transmission losses

Utilization

LP MILP0.00

0.25

0.50

0.75

1.00

LP MILP

Pow

er [r

el. t

o pe

ak lo

ad]

Li-ion Wind PV GT CCGT Coal Lignite Nuclear

LP MILP LP MILP

66%33% 100%

Storage expansion

Theoretical RE sharea Theoretical RE sharea

Influence of further assumptions (I)Renewable and storage capacities


0

50

100

150

200

250

300

350

Ref

Sto

r_In

v_lo

w

Stor

_Inv

_hig

h

VR

E_I

nv_l

ow

VR

E_I

nv_h

igh

FP_l

ow

FP_h

igh

CO

2_lo

w

CO

2_hi

gh

G++ G

+

G+_

Inv_

high

G+_

Inv_

very

high

Cur

.003

Cur

.010

VR

E_E

xp_C

O2_

med

VRE

_Exp

_CO

2_hi

gh

VRE

_Exp

_CO

2_ve

ryhi

gh

Wea

ther

200

7

Wea

ther

200

8

Wea

ther

200

9

Wea

ther

201

1

Wea

ther

201

2

Red

ox-fl

ow_I

nv_l

ow

PH

S_w

/o_o

ld_s

tock

0

250

500

750

1,000

1,250

1,500

1,750(I) Investment cost scen. (2) Operating cost scen. (3) Grid scen. (4) Cur scn. (5) VRE constr. (6) Weather scen. (7) Misc.

Redox-flow aCAES PHS new PHS stock Li-ion H2

Nuclear Lignite Coal CCGT GT CSP GeothermalBiomass Hydro reservoir Run-of-river PV Wind onshore Wind offshore

Restricted curtailments

Grid endog.

Sto

rage

[GW

]G

ener

atui

on[G

W]

Influence of further assumptions (II)Technology-specific ranges of storage capacity


0

20

40

60

80

100

120

140

H2 Li-ion aCAES

Pow

er c

apac

ity [G

W]

0

100

200

300

400

500

600

Li-ion aCAES

Ene

rgy

capa

city

[GW

h]

H20

10

20

30

40

50

60

Ene

rgy

capa

city

[TW

h]

Spreads for EU scenarios with a RE share ≥ 90%a

1st and 3rd quartile, median, min, and max Technology-specific capacities in large parts robust, however, some outliers exist

Power capacity Energy capacity

a Pumped hydro & redox-flow excluded due to marginal bandwidths and insignificant capacities

Conclusions (I)Can we narrow down the range of storage requirements?


0

150

300

450

600

25% 50% 75% 100%

Pow

er c

apac

ity[G

W]


10

100

1000

10000

100000

25% 50% 75% 100%

Ene

rgy

capa

city

[GW

h]

Europe

130−270 GW

16−54 TWh

Conclusions (II)


State of research Current studies result in broad ranges for storage requirements PV-dominated scenarios tend to foster higher storage capacities, compared to wind-

dominated or balanced mixes

Storage requirements in Europe Storage sensitive to scenario and methodological assumptions:

EU: 130–270 GW, 16–54TWh Power plant modeling affects storage requirements only in small systems with low shares

of variable renewable energies Large parts of storage capacity can be substituted by transmission grid expansion However, grid and storage are complement options and temporal decoupling of load and

supply is still necessary, even under perfect grid assumptions („copper plate”) Technology-diverse storage portfolio essential; each storage fills a certain niche Spatial storage capacity distribution mainly influenced by the shape of the net load

Dipl. Wi.-Ing. Felix Cebulla

German Aerospace Center

Systems Analysis and Technology Assessment

Thank you!Questions?

References


[1] J. Haas, F. Cebulla, K. Cao, W. Nowak, R. Palma-Behnke, C. Rahmann, and P. Mancarella,“Challenges and trends of energy storage expansion planning for flexibility provision in low-carbon power systems – a review,” Renewable and Sustainable Energy Reviews, vol. 80, pp.603–619, 2017.

[2] F. Cebulla, J. Haas, J. Eichman, W. Nowak, and P. Mancarella, “How much energy storage dowe need? A review and synthesis for the U.S., Europe, and Germany”.

[3] H. C. Gils, Y. Scholz, T. Pregger, D. L. de Tena, and D. Heide, “Integrated modelling of variablerenewable energy-based power supply in Europe,” Energy, vol. 123, pp. 173–188, 2017.

[4] F. Cebulla, T. Naegler, and M. Pohl, “Storage demand, spatial distribution, and storagedispatch in a European power supply system with 80% variable renewable energies,” 2017.

[5] F. Cebulla, “Storage demand in highly renewable energy scenarios for Europe: The influenceof methodology and data assumptions in model-based assessments,” University of Stuttgart,2017, submitted.

[6] F. Cebulla and T. Fichter, “Merit order or unit-commitment: How does thermal power plantmodeling affect storage demand in energy system models?,” Renewable Energy, vol. 105, pp.117–132, 2017.

Backup Transmission grid assumptions


Scenario Technology Invest land [k€/km]

Invest sea [k€/km]a

Interest rate [-]

Amor. time [a] O&Mfix

G+ AC 380kV 1,000 1,000 0.06 40 0.003G+ HVDC_2200_UC 913 1,815 0.06 40 0.010G+ HVDC_3200 384 2,640 0.06 40 0.010a For the modeling of the AC transmission grid no differentiation between land and sea investment costs is considered. The values of invest land and invest sea are therefore identical and should not be understood additively.

BackupCost assumptions


Technology Invest [k€] Unit Life time [a] O&Mfix [%/a]

AC 380 kV 1,000 Km 40 0.3

HVDC 2200 913a

1,815b Km 40 1.0

HVDC 3200 384a

2,640b Km 40 1.0

Photovoltaic 900 MW 20 1.0Wind onshore 900 MW 18 4.0Wind offshore 1,300 MW 18 5.5

Pumped hydro 45010

MWMWh

2060 1.0

Compressed air 57047

MWMWh

2040 1.0

Lithium-ion 50150

MWMWh

2525 0.5

Hydrogen 1,2001

MWMWh

1530 2.0

Redox-flow 630100

MWMWh

2020 3.2

a Land-basedb Sea-based

Documents

Requirements for electrical energy storage · Generation Storage DLR.de • Felix Cebulla • Chart 7 40 GW (4 TWh) Compressed air Pumped hydro Lithium-ion Hydrogen Nuclear Lignite-fired