integration of renewables in swiss energy system schaffen wissen – heute fÜr morgen integration...

WIR SCHAFFEN WISSEN – HEUTE FÜR MORGEN

Integration of Renewables in Swiss Energy System

Evangelos Panos :: Energy Economics Group:: Paul Scherrer Institute

Kannan Ramachandran :: Energy Economics Group:: Paul Scherrer Institute

72th Semi-annual ETSAP meeting, 11-12th December, Zurich

Overview of the Swiss Energy System, 2015

0

20

40

60

80 Solar

Wind

Other

Nuclear

Hydro

ELECTRICITY GENERATION (TWh)

0

5

10

15

20

25Solar

Wind

Other

Nuclear

Hydro

ELECTRICITY CAPACITY (GW)

0

200

400

600

800

1000 Heat

RES

Wood

Electricity

Other

Gas

Oil

Coal

FINAL ENERGY CONSUMPTION (PJ)

0

10

20

30

40Power generation

Services and agriculture

Industry

Residential

Transport

ENERGY RELATED CO2 EMISSIONS (Mt)

Swiss energy strategy 2050:

aims at gradually phasing out nuclear and promoting renewables and demand side efficiency

56%

38%

13.7

3.3

Solar

• We study integration measures for variable renewable generation (VRES) from wind

and solar PV in Switzerland for the horizon 2015 – 2050, such as:

Reinforcing and expanding the grid network

Deploying local storage, complementary to pump hydro, like batteries and ACAES

Deploying dispatchable loads such as P2G, water heaters and heat pumps

• The study was performed in the context of the ISCHESS project, which is a

collaboration between the Paul Scherrer Institute and the Swiss Federal Institute of

Technology (ETH Zurich), funded by the Swiss Competence Center Energy and

Mobility (CCEM) http://www.ccem.ch/ischess

Objectives of the research

http://www.ccem.ch/ischess

• High intra-annual resolution with 288 typical hours (3 typical days, 4 seasons, 24h/day)

• In this research, the model includes:

Higher detail in the electricity sector, variability in the RES generation, ancillary

services and power plant dispatching constraints

But, it excludes oil transport and has aggregate representation of industry

Methodology – The Swiss TIMES Energy Systems Model (STEM)

Swiss TIMES Energy system Model (STEM)

Fuel supply

module

Fuel

distribution

module

Demand modulesElectricity supply

module

Resource module

Electricity import

Uranium

Natural gas

Hydrogen

Electricity export

Electricity

Gasoline

Diesel

Renewable· Solar

· Wind

· Biomass

· Waste

Electricity

storage

Hydro resource· Run-of rivers

· Reservoirs

CO2

Demand

technologies

Residential- Boiler

- Heat pump

- Air conditioner

- Appliances

Services

Industires

Hydro plants

Nuclear plants

Natural gas

GTCC

Solar PV

Wind

Geothermal

Other

Taxes &

Subsidies

Fuel cell

Energy

service

demands

Person

transportat

ion

Lighting

Motors

Space

heating

Hot water

Oil

Transport

Car fleetICE

Hybrid vehicles

PHEV

BEV

Fuel cell

Bus

Rail

Mac

roec

onom

ic d

river

s (e

.g.,

popu

latio

n, G

DP

, flo

or a

rea,

vkm

)

Inte

rnat

iona

l ene

rgy

pric

es (o

il, n

atur

al g

as, e

lect

ricity

, ...)

Tec

hnol

ogy

char

acte

rizat

ion

(Effi

cien

cy, l

ifetim

e, c

osts

,…)

Res

ourc

e po

tent

ial (

win

d, s

olar

, bio

mas

s, …

.)

Biofuels

Biogas

vkm-Vehicle kilometre

tkm-tonne kilometre

LGV-Light goods vehicles

HGV-Heavy good vehicles

SMR-steam methane reformer

GTCC-gas turbine combined cycle plant

Oil refinery

Process

heat

FreightsTrucks

HGV

Rail

Natural gas

Heating oilOther

electric

• Different grid levels, characterised by transmission costs and losses

• A set of power plants can be connected to each level:

Power plants are differentiated by costs, efficiency, technical constraints & availability

• A linearised approximation of the Unit Commitment problem is formulated

Representation of the electricity sector in STEM

Very High Voltage Grid Level 1

NuclearHydro DamsImports Exports

Distributed Power Generation

Run-of-river hydroGas Turbines CCGas Turbines OCGeothermal

High Voltage Grid Level 3

Large Scale Power Generation

Medium Voltage Grid Level 5

Wind FarmsSolar ParksOil ICEWaste Incineration

Large scale CHP district heating

Wastes, BiomassOilGasBiogasH2

Low Voltage Grid Level 7

Large Industries & Commercial

CHP oilCHP biomassCHP gasCHP wastesCHP H2Solar PVWind turbines

Commercial/Residential Generation

CHP gasCHP biomassCHP H2Solar PVWind turbines

Lead-acid batteriesNaS batteriesVRF batteries

PEM electrolysis

Pump hydro

CAES

Lead-acid batteriesNaS batteriesVRF batteries

Li-Ion batteriesNiMH batteries

PEM electrolysis

Lead-acid batteriesLi-Ion batteriesNiMH batteries

+ 4 nodes for nuclear

power plants

• With a reduction algorithm from FEN/ETHZ we map the detailed transmission grid to an

aggregated grid with 𝑁 = 15 nodes and 𝐸 = 319 lines, based on a fixed disaggregation

of the reduced network injections to the detailed network injections

Representation of electricity transmission grid

MAPPING

−𝐛𝐝 ≤ 𝐇𝐝 × 𝐠𝐝 − 𝐥𝐝 ≤ 𝐛𝐝

𝐇𝐝 is the PTDF matrix of the detailed network, 𝐃 is the fixed dissagregation matrix,𝐠𝐝 , 𝐠𝐚 are vectors with elelectricity injections,

𝐥𝐝 , 𝐥𝐚 are vectors of electricity withdrawals, and 𝐛𝐝 , 𝐛𝐚 are vector of line capacities; superscripts d=detailed, a=aggregated network

The matrix 𝐃 is not unique, since there are infinite ways in which an aggregate injection can be distributed between multiple nodes;

here, it allocates power injections according to the original distribution of generation capacity in the detailed power flow model

−𝐛𝐚 ≤ 𝐇𝐝 × 𝐃 × 𝐠𝐚 − 𝐥𝐚 ≤ 𝐛𝐚

DETAILED NETWORK AGGREGATED NETWORK

• To capture variability, we need the variance of the mean availability of wind and solar

per typical hour in STEM

• By bootstrapping a 20-year meteorological sample, we calculate the distribution of

mean availability per typical hour in STEM

• We move ± 3 st.dev in this distribution to introduce variability of RES per typical hour

Representation of stochastic RES variability

Bootstrapped Distribution of Mean Photovoltaic

Capacity Factors (availability): Summer (left), Winter (right)

• Storage capacity must accommodate downward variation of the Residual Load Curve

(RLDC) and upward variation of non-dispatchable generation

• The dispatchable peak generation capacity (incl. storage) must accommodate upward

variation of the RLDC and downward variation of non-dispatchable generation

Source: Singh, A. PSI, 2016

• Power plants commit capacity to the reserve market based on their operational

constraints and shadow price of electricity generation and reserve provision

• In each of the typical hours the demand for reserve is endogenously calculated from

the joint probability distribution function (p.d.f.) of the individual p.d.f. of forecast

errors of supply and demand

The sizing is based on both probabilistic and deterministic assessment

We move ± 3 s.d. on the joint p.d.f to estimate the reserve requirements

Ancillary services markets – provision of reserve

The standard deviation of the joint probability

distribution of forecast errors is :

𝜎2𝑠𝑜𝑙𝑎𝑟

+ 𝜎2𝑤𝑖𝑛𝑑

+ 𝜎2𝑙𝑜𝑎𝑑

And the reserve requirement in hour 𝑡 is given as:

By assuming uncorrelated random variables:

𝑅𝑡 = 3 ∗ 𝜎2𝑠𝑜𝑙𝑎𝑟

∙ 𝐺𝑡2

𝑠𝑜𝑙𝑎𝑟+ 𝜎2

𝑤𝑖𝑛𝑑∙ 𝐺𝑡

2

𝑤𝑖𝑛𝑑+ 𝜎2

𝑙𝑜𝑎𝑑∙ 𝐿𝑡

2 + 𝑎 ∙ 𝑃𝑚𝑎𝑥

𝐺𝑖 generation from option 𝑖 , 𝐿 load, 𝑃𝑚𝑎𝑥 largest grid element

A range of “what-if” scenarios was assessed along three main dimensions:

1. Future energy policy and energy service demands, assuming nuclear phase out by 2034:

Energy service demand growth: low growth vs high growth

Electricity imports: allow net imports vs autarky

Climate change mitigation policy: mild policy vs 70% reduction in CO2 emissions

in 2050 from 2010 levels

2. Location of new gas power plants and installed capacity as % of the total national capacity

3. Grid expansion: allowing grid reinforcement beyond the “Grid 2025” plans of Swissgrid or not

in total about 100 scenarios were assessed with the STEM model based on the

Cartesian product of the above combinations across the three dimensions

Long term scenarios analysed with STEM

Corneux (NE) Chavalon (VS) Utzenstorf (BE) Perlen (LU)

Schweizerhalle

(BL)

Case 1 20.0 20.0 20.0 20.0 20.0

… … … … … …

Case 11 0.0 33.3 33.3 33.3 0.0

… … … … … …

Case 26 33.3 33.3 0.0 0.0 33.3

• Electricity consumption increases 4 – 30% from 2015 ( 0.1 – 0.8% p.a)

• New gas power plants replace existing nuclear capacity

• Under climate policy VRES could provide up to 28% of the domestic supply

Electricity consumption continues to increase and gas, VRES & imports replace nuclear by 2050

-100

1020304050607080

2015

2050 (max)

2050 (min)

ELECTRICITY GENERATION & CONSUMPTION IN 2050 (TWh)

The results correspond to ranges among the 100 scenarios assessed

• The requirements for secondary reserve almost double in 2050 from today’s level and

peak demand shifts from winter to summer

• Hydrostorage is the main contributor, followed by gas turbines (after 2040)

• Flexible CHPs and batteries could participate in the reserve provision by forming

“virtual plants”, especially in those scenarios with high reserve requirements

Electricity consumption continues to increase and gas, VRES & imports replace nuclear by 2050

SECONDARY RESERVE MAXIMUM DEMAND AND PROVISION IN 2050 (MW)

0100200300400500600700800

2015

2050 (max)

2050 (min)

Maximum contribution per technology


• High shares of VRES require electricity storage peak capacity of ca. 30 – 50% of the

installed capacity of wind and solar PV (together)

• Above 14 TWh of VRES generation there is a significant storage deployment

• About 13% of the excess summer VRES production is seasonally stored in P2G (~ 1 TWh)

Storage needs increase with VRES deployment

0.00.51.01.52.02.53.03.54.04.55.05.5

8 10 12 14 16 18 20 22 24

Inst

alle

d s

tora

ge c

apac

ity

(GW

)

TWh of wind and solar PV electricity production

Pump hydro(pumpingcapacity)

Batteries (4hmaxdischarge)

P2G(only as

seasonal

storage)

• Small scale batteries are driven by

distributed solar PV installations

• Medium scale batteries are driven by

wind and large scale PV and CHP

• Large scale batteries complement pump

hydro storage when it is unavailable

Max. Total battery capacity 5.5 GW

Large scale batteries 0.5 GW

Small scale

batteries 3 GW

Medium scale

batteries 2 GW

WIND AND SOLAR PV ELETRICITY VS STORAGE CAPACITY

Each data point in the graph corresponds to a different long term scenario and year

Contribution of each sector to the total electricity stored in water

heaters and heat pumps

Residential 85-97%

Industry <2%

Services 3-23%

• Electricity storage in water heaters and heat pumps could account for 8 – 24% of the

total electricity consumption for heating

• Accelerated deployment of electric load shifting when electricity consumption for heat

exceeds 13 TWh

• Large potential for load shifting is in water heating (resistance heating) followed by

space heating in buildings

Dispatchable loads help in easing electricity load peaks in the stationary end-use sectors

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

8 13 18

Ele

ctri

city

sto

red

in w

ate

r h

eat

ers

an

d h

eat

pu

mp

s (T

Wh

)

Electricity consumption for heating (TWh)

ELECTRICITY STORED IN WATER HEATERS AND HEAT PUMPS

VS ELECTRICITY CONSUMPTION IN HEATING IN 2050 (TWh)

Each data point in the graph corresponds to a different long term scenario

• Restrictions in grid expansion results, due to congestion, in:

Higher system costs, up to 90 BCHF (+10% when grid expansion is allowed) in 2020-50

Non-cost optimal deployment of electricity supply options

Less electrification of demand and reliance on fossil-based heating

The system-wide benefits from the electricity grid expansion outweigh the costs

1100

1200

1300

1400

1500

1600

1700

No gridexpansion

Gridexpansion

No gridexpansion

Gridexpansion

Reference Stingent climate policy

Max

Min

CUMULATIVE UNDISCOUNTED ELECTRICITY AND HEAT SYSTEM COST, BHCF/yr, 2020 – 2050


• Grid expansion results in cost savings, mainly in heating sectors, directly (e.g. technology

change via heat pumps) and indirectly (e.g. less costs for imported fuels)

Under a stringent climate change mitigation policy and high demand, these costs could

be up to 3 BCHF/yr

About 1/3 of the savings occur in the electricity supply and 2/3 in the stationary sectors

The system-wide benefits from the electricity grid expansion outweigh the costs

-3500-3000-2500-2000-1500-1000

-5000

5001000

Electricitypowerplants

ElectricityT&D

Netimports ofelectricityand fuels

Heatsupply

Total

Max

Min

ELECTRICITY AND HEAT SYSTEM COSTS SAVINGS DUE TO GRID EXPANSION, MCHF/yr, 2020-2050


• Electricity consumption continues to increase by 0.1 – 0.8% p.a., and especially

after 2030 due to new electricity uses; it could reach over 70 TWh/yr by 2050

• VRES can contribute up to 24 TWhe (or 28% of the domestic supply) after the

nuclear phase-out

• Storage peak capacity investments about 30 – 50% of the installed wind and solar

PV capacity; beyond 14 TWhe accelerated deployment of storage is inevitable

• About 13% of the excess electricity production from VRES in summer is seasonally

stored in P2G pathways, driven by the high seasonal electricity price differences

under stringent climate change mitigation scenarios

• Water heaters and heat pumps could contribute in easing electricity peaks and

they could shift 8 – 24% of the electricity consumption for heat

Conclusions

• Grid reinforcement results in net economic benefits for the whole electricity and

heat supply system of Switzerland on the order of 0.5 – 3.0 BCHF/yr. over the

period 2030 - 2050

• Rebound effects prevail in grid reinforcement, and a social planner shall prioritize

the expansion of those lines with high benefit/cost ratio for the energy system

(insights are provided form the dual values of the grid constraints in the model)

• Further work is needed to overcome some important limitations, such as:

Regional representation also for the heat supply and not only for electricity

Consideration of N-2 grid security constraints

More detail in technical representation of storage technologies (e.g. depth of

discharge, ramping rates, max discharging times, etc. )

Conclusions ctd.

Wir schaffen Wissen – heute für morgen

Thank you for your attention. Evangelos Panos

Energy Economics Group

Laboratory for Energy Systems Analysis

Paul Scherrer Institute

[email protected]

mailto:[email protected]

• Without batteries and grid, there is 30 – 50% less deployment of wind and solar electricity

compared to the case when both options are available

Batteries are important for the integration of VRES to cope with their variability

Grid expansion is important to integrate large amounts of VRES production ( >16 TWh)

• Total system costs can be 10 – 14% higher if both batteries and grid expansion are unavailable

In particular climate policy costs could increase by more than 50% (from 103 to 160 BCHF)

Storage and grid expansion are required to realise the VRES potential and lower climate policy costs

0

5

10

15

20

25

Reference Climate policy

TWh

of

win

d a

nd

so

lar

PV

NoBatteries,No gridexpansion

WithBatteries,Gridexpansion

IMPACT OF BATTERIES AND GRID EXPANSION IN

THE DEPLOYMENT OF WIND AND SOLAR PV POWER

1100

1150

1200

1250

1300

1350

1400

1450

Reference Climate policy

Un

dis

cou

nte

d c

um

ula

tive

co

st B

CH

F

NoBatteries,No gridexpansion

WithBatteries,Gridexpansion

IMPACT OF BATTERIES AND GRID EXPANSION IN

THE TOTAL SYSTEM COSTS

Results on the left graph corresponds to ranges; results on the right graph is for a scenario with high demand and electricit y imports

integration of renewables in swiss energy system schaffen wissen – heute fÜr morgen integration...

Documents