© OECD/IEA 2015

Renewable Energy Training for Latin America Day 1

IEA Training and Capacity Building - Latin America, Santiago de Chile, 19-23 Oct 2015

Simon Müller

Analyst – System Integration of Renewables

© OECD/IEA 2014

Day 1 – The Integration Challenge and International Experience: Session 1: Integration of renewable energy into power grids

Session 2 & 3: International context - Successes and failures (EU and USA)

Country presentations: Mexico (Markets & Policies), Uruguay (System operation)

Course overview

© OECD/IEA 2014 2

© OECD/IEA 2014

Session 1: Integration of renewable energy systems into power grids

1. Basic concepts of power systems and integration

2. Detailed discussion of the technical properties of VRE and what integration effects they cause

3. Relevant properties of power systems that shape degree of flexibility

Exercise 1: Constructing net load for different shares of wind and PV

Integration effects depending on penetration rate and country context

Session agenda

© OECD/IEA 2014 3

© OECD/IEA 2014

Grid based power systems have historically been organised into generation, transmission and distribution

This is still main paradigm for most systems

Grid based power systems

© OECD/IEA 2014 4

© OECD/IEA 2014

Power demand varies by region, season and daytime

© OECD/IEA 2014 5

© OECD/IEA 2014

Why is it challenging to balance supply and demand of electricity? What are relevant time scales?

© OECD/IEA 2014 6

© OECD/IEA 2014

Grid integration of renewables in one slide

Properties of variable renewable energy (VRE)

Flexibility of other power system components

© OECD/IEA 2014 7

Grid integration of VRE is about understanding, managing and systematically improving the interaction between VRE and other system components

Systems are all different, but there are common issues across systems – learning from each other is possible and valuable

© OECD/IEA 2014

Properties of VRE

© OECD/IEA 2014 8

© OECD/IEA 2014

The 6 VRE properties that matter

Variable Maximum output varies depending on wind and sunlight

Uncertain No perfect forecast for wind and sunlight available

Non-synchronous technologies VRE connect to grid via power electronics, have little or no

rotating mass

Location constrained resource is not equally good in all locations and cannot be

transported

Modularity Wide range of sizes and may be much smaller than other options

Low short-run cost Once built, VRE generate power almost for free

sec

yrs

1km

100s km

© OECD/IEA 2014 9

© OECD/IEA 2014

Netload = Load - VRE

Note: Load data and wind data from Germany 10 to 16 November 2010, wind generation scaled, actual share 7.3%. Scaling may overestimate the impact of variability; combined effect of wind and solar may be lower, illustration only.

© OECD/IEA 2014 10

0

10

20

30

40

50

60

70

80

1 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Hours

Net

load

(G

W)

0.0% 2.5% 5.0% 10.0% 20.0%

Illustration of netload at different VRE shares

© OECD/IEA 2014

Variability: Examples of solar PV and wind

© OECD/IEA 2014 11

© OECD/IEA 2014

Variability: Examples of solar PV and wind

© OECD/IEA 2014 12

© OECD/IEA 2014

A typical week of solar (top) and wind (bottom) generation in summer 2011 in Spain.

© OECD/IEA 2014 13

© OECD/IEA 2014

Geographic smoothing within a wind plant with 15 turbines (red) and 200 turbines (blue)

© OECD/IEA 2014 14 Source: NREL ,2009

© OECD/IEA 2014

Seasonal complimentary between wind and PV in Germany in 2012

© OECD/IEA 2014 15

Source: Fraunhofer ISE 2013

© OECD/IEA 2014

Uncertainty reduces dramatically with shorter horizon

Real-time generation data key for short-term accuracy

Forecasts generally more mature for wind than for PV

Uncertainty

© OECD/IEA 2014 16

0%

5%

10%

15%

20%

25%

1 5 9 13 17 21 25 29 33 37 41 45

Mea

n a

bso

lute

err

or

/ av

erag

e p

rod

uct

ion

Forecast horizon (Hours before real-time)

2008 2009

2010 2011

2012

Accuracy of wind forecasts in Spain

Source: REE

© OECD/IEA 2014

Uncertainty: Problems with extreme weather

Source: SHORT-TERM PREDICTION OF SOLAR,PHOTOVOLTAIC POWER, Detlev Heinemann, Elke Lorenz, University of Oldenburg, Germany,Institute of Physics,Energy Meteorology Group, International Conference Energy & Meteorology (ICEM), 10 November 2011, Gold Coast, Australia

© OECD/IEA 2014

Location constraints

© OECD/IEA 2014 18

Brazil: Wind Power Density (W/m2) and population density (persons per km2)

Source: GPW, 2002

© OECD/IEA 2014

Location constraints

© OECD/IEA 2014 19

Japan: Wind Power Density (W/m2) and population density (persons per km2)

Sources: http://www.stat.go.jp/english/dat

a/handbook/pdf/c02cont.pdf

Sources: 2012-12-5 RE2012展示(縮小版).pdf

© OECD/IEA 2014

Location constraints

© OECD/IEA 2014 20

2030 maximum excess PV capacity (MW) for resource-driven and load-driven deployment scenarios

Source: EPIA, 2012

© OECD/IEA 2014

Modularity: German case

© OECD/IEA 2014 21

© OECD/IEA 2014

Modularity: Example of high solar PV capacities in Wolkshausen, Germany

© OECD/IEA 2014 22

© OECD/IEA 2014

Modularity: Danish case

© OECD/IEA 2014 23

© OECD/IEA 2014

Non-synchronous generation

Source: Integrating Wind in Ireland: Experience and Studies Mark O’Malley Director, MIT Wind Week January 21th 2011

PV: non-synchronous, non-mechanical generation

Wind: non-synchronous, mechanical generation

© OECD/IEA 2014

Synchronous vs non sychronous

SOURCE: IEC-White paper Grid integration of large-capacity Renewable Energy sources ,October 2013

© OECD/IEA 2014

Relevant integration effects

© OECD/IEA 2014 26

© OECD/IEA 2014

Germany, 1993, 0.1% wind power in total generation

“Renewable energies such as sun, hydro or wind cannot cover more than 4% of our electricity consumption – even in the long run” Joint statement by German power utilities, published in Die Zeit, 30 July 1993, page 10

Ireland, 2003, 2% wind power in annual generation

“This amount of wind generation does, however, pose an increased risk to the security and stability of the power system which the transmission system operator feels exceeds the level normally likely to be accepted by a prudent system operator.“

Kieran O'Brien, Managing Director of ESB National Grid, Ireland, 1 December 2003

2014, 20% wind power in annual generation

2020, target, 37% wind power in annual generation

VRE and initial concerns

© OECD/IEA 2015 27

© OECD/IEA 2014

VRE break an old ‘golden rule’ of system operation: “We cannot control load, so we must control generation to keep the lights on. VRE are not controllable, so we cannot use them.” Think of VRE as negative load at low shares. This will solve many issues.

VRE challenge operational patterns of existing assets: “The majority of our generation assets are technically incapable of changing their output to follow wind and solar power.” Critically assess technical performance characteristics, targeted upgrades

including better monitoring equipment and training of operating crews.

VRE may be incompatible with existing contract frameworks: Obligations to feed-in flat generation profile

Obligations to report schedules far ahead of time with no way of changing

Guaranteed operating hours of technically flexible plants constrain operations

Review and possibly change contractual framework.

Main barriers at low shares

© OECD/IEA 2015 28

© OECD/IEA 2014

When VRE are added to a system with adequate capacity:

Situations of low load and high VRE generation VRE curtailment if flexibility insufficient

Negative market prices due to inflexible generation and VRE support mechanisms

Grid bottlenecks in regions with high VRE density Limitations feeding production from the distribution to the

transmission grid

Insufficient evacuation capacity in regions with rapid build out of new VRE capacity

Main short-term challenges

© OECD/IEA 2014 29

© OECD/IEA 2014

Power systems already deal with a vast demand variability Can use existing flexibility for VRE integration

No problem at low shares, because …

© OECD/IEA 2014 30

Exceptionally high variability in Brazil, 28 June 2010

© OECD/IEA 2014

Main market impacts

Reduced market prices (merit order effect)

Reduced operating hours (utilisation effect)

Displacement effect mainly due to

low short-run cost of VRE and

reinforced by support policies

influenced by variability, in particular PV

Shift in German spot market price structure

© OECD/IEA 2014 31

© OECD/IEA 2014

Seconds Years 100km 1km

Low short run cost

Non synchro-

nous

Stability

Uncertainty

Reserves

Variability

Short term changes

Asset

utilisation

Abundance Scarcity

Location constrained

Trans-mission

grid

Modularity

Distribution grid

Properties of variable renewables and impact groups

Systems are different – impacts will vary too

But common groups of effects

© OECD/IEA 2014 32

Balancing Profile / Utilisation Location Stability

© OECD/IEA 2014

Main persistent challenge: Balancing

Note: Load data and wind data from Germany 10 to 16 November 2010, wind generation scaled, actual share 7.3%. Scaling may overestimate the impact of variability; combined effect of wind and solar may be lower, illustration only.

© OECD/IEA 2014 33

0

10

20

30

40

50

60

70

80

1 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Hours

Net

load

(G

W)

0.0% 2.5% 5.0% 10.0% 20.0%

Larger ramps at high shares

Higher uncertainty

Larger and more pronounced changes

Illustration of Residual power demand at different VRE shares

Lower minimum

© OECD/IEA 2014

Netload implies different utilisation for non-VRE system

Main persistent challenge: Utilisation

Note: Load data and wind data from Germany 10 to 16 November 2010, wind generation scaled, actual share 7.3%. Scaling may overestimate the impact of variability; combined effect of wind and solar may be lower, illustration only.

© OECD/IEA 2014 36

0

10

20

30

40

50

60

70

80

90

1 2 000 4 000 6 000 8 000

Net

load

(G

W)

Hours

0.0% 2.5% 5.0% 10.0% 20.0% Maximum

remains high: Scarcity

Lower minimum:

Abundance

Changed utilisation pattern

Base - load

Mid - merit

Peak

Mid - merit

Peak

Mid - merit

Base - load

-

-

© OECD/IEA 2014

Competitive Renewable Energy Zones (CREZ), Texas

Wind built before completing all transmission lines

Curtailment peaked at 17% in 2009

Curtailment reduced to 1.6% in 2013 after implementing locational pricing and expanding the grid

Wind in Texas – Operational issues while waiting on transmission

© OECD/IEA 2014 37

345 kV double-circuit upgrades identified in

CREZ transmission plan

CREZ, Texas

Source: NREL

© OECD/IEA 2014

Voltage rise common issue

Smart inverters and transformers help

Distributed PV and the grid

© OECD/IEA 2014 38

Vo

ltag

e le

vel d

evia

tio

n

Transformer

Transformer

O%

+9%

Voltage rise in a rural distribution system in Germany

© OECD/IEA 2014

There is no single factor that would put a maximum technical limit on the long-term amount of variable generation

However, more and more measures are needed to achieve high shares

In the short term, many institutional and some technical issues can be a constraint.

Question rather is: How far can I go before it get‘s expensive? And what do I need to do for that?

No principal technical limit on VRE share

© OECD/IEA 2014 42

What is the maximum speed at which you can driver a car?*

*Apart from the speed of light

© OECD/IEA 2014

Relevant properties for system flexibility

© OECD/IEA 2014 43

© OECD/IEA 2014

Four sources of flexibility …

Grid infrastructure

Dispatchable generation Storage

Demand side integration

© OECD/IEA 2014 44

© OECD/IEA 2014

What is flexibility?

Definition “Extent to which a power system can adjust the balance of electricity

production and consumption in response to variability, expected or otherwise.”

Measurement of

Flexibility supply

Flexibility demand

Comparison of supply and demand

Flexibility options may Increase supply

Reduce demand

© OECD/IEA 2014

Capable of: “Achieving and sustaining any consumption or generation level at arbitrarily small response times at no cost.”

Relevant dimensions?

Possible levels: “Adjustability”

Max. duration of output: ”Durability”

Possible changes: “Ramping”

“Lead time” for change

“State dependency”

Real power sources, loads & storage approximate ideal source on these dimensions

Looking at the ideal power device

Handful of relevant time scales

© OECD/IEA 2014

Flexible generation: Conventional plants and firm RE adjustable in wide range, very durable,

more or less rampable, different lead time and state dependency (nuclear to OCGT)

Adjustability and durability of VRE limited by resource, but within these bounds they are very rampable, little response time, almost no state dependency

DSM Adjustable (consumption), varying durability, varying rampable, lead

time, possibly high state dependency

Storage Perfect adjustability, determined durability …

Interconnection Can reduce demand and increase supply along all dimensions

The 4 flexible resources

© OECD/IEA 2014

Integration vs. transformation

Classical view: VRE are integrated into the rest

Integration costs: balancing, adequacy, grid

More accurate view: entire system is re-optimised

Total system costs

Integration is actually about transformation

Remaining system

VRE

FLEXIBLE Power system

• Generation • Grids • Storage • Demand Side Integration

© OECD/IEA 2014 49

© OECD/IEA 2014

2. Make better use of

what you have

Op

eratio

ns

1. Let wind and solar play their

part

3. Take a system wide-strategic

approach to investments!

System friendly

VRE

Technology spread

Geographic spread

Design of power

plants

Three pillars of system transformation

© OECD/IEA 2014 50

Investm

ents

© OECD/IEA 2014

Exercise 1: Netload & Integration effects

© OECD/IEA 2014 51

© OECD/IEA 2014

Session 2: International context

1. Co-ordination and consolidation of balancing areas

2. Benefit of sub-hourly scheduling and rapid dispatch

3. Centralised vs non-centralised power markets

4. Definition of market products and timing of electricity trade

Exercise 2: Benefit of rapid system operation

Session agenda

© OECD/IEA 2014 52

© OECD/IEA 2015

Policy and regulatory framework

Border = seam

Policy and regulatory framework Policies 2

3 Operational rules / Markets

Interconnectors 1

Country/ Area A

Country/ Area B

© OECD/IEA 2015

Impact of wind and solar power

Comparison of net power exchanges on the French-German border with wind and solar power generation in Germany, sept. 2011 (MW)

MW

MW

Production Exchanges

Imports from France

Exports to France

Net exchanges Wind production in Germany Wind and solar production in Germany

Source: RTE

© OECD/IEA 2015

PJM footprint

Note: this map is without prejudice to the status of a sovereignty over any territory, to the delimitation of international frontiers and boundaries, and to the name of any territory, city or area.

Source: Monitoring Analytics, 2013.

© OECD/IEA 2015

Price patterns before and after integration

Note: this map is without prejudice to the status of a sovereignty over any territory, to the delimitation of international frontiers and boundaries, and to the name of any territory, city or area.

Source: PJM, 2013d.

Pre-integration price pattern (PJM) Post-integration price pattern (PJM)

140

110

40

Marginal Cost (USD/MWh)

© OECD/IEA 2015

Indicative regulating requirements for a balancing authority as a function of peak demand (%)

Indicative regulating requirements for a balancing authority as a function of peak demand (%)

0.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.7%

0.8%

0.9%

1.0%

0 20 40 60 80 100

Peak demand (GW)

Perc

enta

ge o

f pea

k de

man

d

Source: NREL, 2011.

© OECD/IEA 2014

Germany has four balancing areas (historic reasons)

Reserve sharing mechanism across four areas

Reduced requirements despite rapid increase of VRE

Co-operation with neighbours

© OECD/IEA 2014 58

0

10

20

30

40

50

60

0

1

2

3

4

5

6

7

8

2008 2009 2010 2011 2012

Inst

alle

d V

RE

cap

acit

y (G

W)

Res

erve

req

uir

emen

t (G

W) VRE

capacity

Upward reserves

Downward reserves

+100%

+0%

-40%

Required frequency restoration reserves in Germany

© OECD/IEA 2014

Short scheduling intervals (5min best practice)

Adjust schedules up to real time (5min best practice)

Generation and transmission schedules

© OECD/IEA 2014 59

6 7 8 9

Cap

acit

y (M

W)

Time (hours)

Actual load curve

Load schedule -15 minutes

Load schedule -60 minutes

Balancing need 15 min schedule

Balancing need60 min schedule

Impact of scheduling interval on reserve requirements, illustration

© OECD/IEA 2014

… but you also need the tools to do this Example: The Control Centre for Renewable Energy - CECRE

© OECD/IEA 2014 61

© OECD/IEA 2014

Portfolio vs unit specific bidding may lead to different performance for integrating high shares of VRE

Energy only market ≠ energy only market Centralised vs Decentralised Markets

© OECD/IEA 2015 62

Day aheadFuture and

forward / OTCIntra-day

18:00

13:50

6:00 10:00 13:30 14:0013:00

13:1512:00

Unit schedule adjustment period

Intra-day price Balancing price(ex post)

Real timeprice

Day aheadprice

Day aheadprice

Portfolio bid Portfolio bid

Unit bid

Gateclosure

Gateclosure

ERCOT

GERMANY

BalancingHour of

operation

15:00

15:00

Schematic representation of ERCOT and German (EPEX Spot) market design

© OECD/IEA 2014

Are your operating reserve and system service definitions VRE ready?

Example Ireland DS3 programme

System service definitions Prepared for a variable future?

© OECD/IEA 2014 64

• Synchronous Inertial Response• Fast Frequency Response• Fast Post-Fault Active Power

Recovery

• Ramping Margin

0 – 5s 5 – 90s 90s – 20min 20min – 12hr

Inertial

Response

Reserve

Ramping

POR

SOR

TOR1

TOR2

RR

Ramping

SIR

FFR

time

Source: EirGrid • Dynamic Reactive Power

ms – s

Transient Voltage Response

Voltage Regulation

Network

Dynamic

Reactive

Power

Network

Adequacy

Grid 25

s – min min – hr

Steady-state

Reactive

Power

• Steady-state Reactive Power

Frequency Services

Voltage Services

© OECD/IEA 2014

Session 3: International context

1. Transmission grid planning and RE deployment

2. Renewable support policies and grid integration

Session agenda

© OECD/IEA 2014 65

© OECD/IEA 2014

Grid planning

© OECD/IEA 2014 66

© OECD/IEA 2014

Importance of coordinated development of grid and generation well understood

Chicken and egg problem for first-off, distant VRE projects Competitive

Renewable Energy Zones (CREZ), Texas

Irish gate system

Appropriate cost recovery is key!

Planning the grid - transmission

© OECD/IEA 2014 67

345 kV double-circuit upgrades identified in

CREZ transmission plan

CREZ, Texas

Source: NREL

© OECD/IEA 2014

About ENTSO-E, an EU institution with legal mandates

41 TSOs

from 34 countries

525 million

citizens served

1000 GW generation

capacity

310 thousands km

of transmission lines

Ten-Year

Network

Development

Plans

Adequacy

forecasts R&D plans

Tools for

Market

Integration

Network

Codes

Source: ENTSO-E

© OECD/IEA 2014

Ten Year Network Development Plan

© OECD/IEA 2014 69

Source: ENTSO-E

© OECD/IEA 2014

Ten Year Network Development Plan

© OECD/IEA 2014 70

50000 km of new or refurbished investments 21000 km of new HVDC lines 15% of all investments are upgrade of existing assets

Source: ENTSO-E

© OECD/IEA 2014

Renewable energy policy and integration

© OECD/IEA 2014 71

© OECD/IEA 2014

As policies provide sufficient remuneration, deployment picks up and costs decrease

Policies have stimulated deployment at falling costs in some circumstances

© OECD/IEA 2015 72

Development of LCOE, remuneration levels and installed capacity for utility scale PV, Germany

LCOE Revenues Capacity

© OECD/IEA 2014

Different sub-technologies show different response patterns to market conditions

Residential customers show more ‘inertia’ in response to profitability of investments

Adjusting policies to sub-technology

© OECD/IEA 2015 73

- Capacity

Residential Commercial Utility

LCOE Revenues

Differences between solar PV sub-segments in Italy

0

10

20

30

40

50

60

70

80

90

100

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

c$/kWh

3.51 GW

0.00

1.00

2.00

3.00

4.00

5.00

6.00

GW

0

10

20

30

40

50

60

70

80

90

100

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

c$/kWh

11.88 GW

0.00

1.00

2.00

3.00

4.00

5.00

6.00

GW

0

10

20

30

40

50

60

70

80

90

100

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

c$/kWh

3.73 GW

0.00

1.00

2.00

3.00

4.00

5.00

6.00

GW

© OECD/IEA 2014

Adjusting policies to sub-technology

© OECD/IEA 2015 74

0

1

2

3

4

5

6

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

GW

PVRES PVCOMM PVUT

0

1

2

3

4

5

6

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014G

W

PVRES PVCOMM PVUT

Italian PV sub – segments additional capacity German PV sub – segments additional capacity

Rapid spiked in deployment occurred in the commercial and utility solar PV market segments

FITs did not lead to out-of-control deployment in residential segment

© OECD/IEA 2014

Adjusting policies to sub-technology

© OECD/IEA 2015 75

Japan PV sub – segments additional capacity USA PV sub – segments additional capacity

More gradual evolution of residential PV market also in Japan and United States

0

1

2

3

4

5

6

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014G

W

PVRES PVCOMM PVUT

0

1

2

3

4

5

6

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

GW

PVRES PVCOMM PVUT

© OECD/IEA 2014

Reaping the benefit of competition …

Revenues

Competitive procurement can reduce policy cost and stimulate further cost reductions

Land-based wind >25 MW in Italy

Auction introduced - Capacity

0

2

4

6

8

10

12

14

16

18

20

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

c$/kWh

8.54 GW

0.00

1.00

2.00

3.00

4.00

5.00

6.00

GW

Land-based wind >25 MW in Brazil

0

2

4

6

8

10

12

14

16

18

20

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

c$/kWh

5.46 GW

0.00

1.00

2.00

3.00

4.00

5.00

6.00

GW

© OECD/IEA 2014

Some form of capacity control usually advisable

But must not be overly restrictive

Differentiate by sub-sector and technology

Put in place clear connection standards / grid codes from the beginning

Find a way to include information about the grid in the support mechanism

Preferential development zones

Limited pass through of grid costs

Maximise technological and geographic diversity

Safeguards as markets expand

© OECD/IEA 2014

1. Excessive geographic, technical concentration

Key lessons: analysis based diversification of location and technology

2. Ill-adapted technical performance standards

Key lessons: focus on grid codes! (fault ride through, 50.2 Hertz); visibility and controllability

3. No or ineffective VRE production forecasts

Key lessons: use forecasts in unit commitment and dispatch of other generation

Three typical mistakes to avoid when deployment begins

© OECD/IEA 2014 78