Renewables Grid Integration and Variability
Session 1: Introduction, variable generation, flexibility and netload
IEA Training and Capacity Building - Latin America, Santiago de Chile, 10-14 Nov 2014
© OECD/IEA 2014
Course overview
 Session 1:
Introduction, variable generation, flexibility, netload
Excurse: Flexible Generation from coal?
 Excercise: Net load analysis (after coffee in different room)
 Session 2:
System impacts and adapting operations
 Excercise: Reserve requirements and VRE forecasts
 Session 3:
Economics of grid integration, system transformation
 Excercise: The value of variable generation
© OECD/IEA 2014
© OECD/IEA 2014
2
Session agenda
 Session 1:
Introduction, variable generation, flexibility, netload
1.
2.
3.
4.
5.
Identifying relevant questions and issues
Grid based power systems: generation, transmission, distribution, loads
Relevance of balancing supply and demand across different time scales
6 properties of variable renewable energy (VRE)
4 sources of power system flexibility

Flexible Generation from coal?
6. Variable renewables as negative loads – the concept of net load
 Exercise 1:
 Constructing net load for different shares of wind and PV
 Analysing net load ramps
 Average vs. Instantaneous penetration
© OECD/IEA 2014
© OECD/IEA 2014
3
Relevant questions and issues?
© OECD/IEA 2014
© OECD/IEA 2014
4
Some basics
© OECD/IEA 2014
© OECD/IEA 2014
5
Grid based power systems
 Grid based power systems are historically been
organised into generation, transmission and
distribution
 This is still main paradigm for most systems
© OECD/IEA 2014
© OECD/IEA 2014
6
Power demand varies by region, season
and daytime
© OECD/IEA 2014
© OECD/IEA 2014
7
Why is it challenging to balance supply and
demand of electricity?
What are relevant time scales?
© OECD/IEA 2014
© OECD/IEA 2014
8
Properties of VRE
© OECD/IEA 2014
© OECD/IEA 2014
9
The 6 VRE properties that matter
 Variable
yrs
 Maximum output varies depending on wind and sunlight
 Uncertain
 No perfect forecast for wind and sunlight available
 Non-synchronous technologies
sec
100s
km
1km
 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
© OECD/IEA 2014
 Once built, VRE generate power almost for free
© OECD/IEA 2014
10
Variability:
Examples of solar PV and wind
© OECD/IEA 2014
© OECD/IEA 2014
11
Examples of wind and solar PV
© OECD/IEA 2014
© OECD/IEA 2014
12
Uncertainty
 Uncertainty reduces
25%
Mean absolute error / average production
dramatically with
shorter horizon
 Real-time generation
data key for shortterm accuracy
 Forecasts generally
more mature for wind
than for PV
Accuracy of wind forecasts in Spain
20%
15%
10%
5%
2009
2010
2011
2012
0%
1
© OECD/IEA 2014
2008
5 9 13 17 21 25 29 33 37 41 45
Forecast horizon (Hours before real-time)
Source: REE
© OECD/IEA 2014
13
Excurse: Non-synchronous generation
Source:Characteristics of Wind Turbine Generators for
Wind Power PlantsIEEE PES Wind Plant Collecto
System Design Working Group
© OECD/IEA 2014
© OECD/IEA 2014
14
Netload and flexibility
© OECD/IEA 2014
© OECD/IEA 2014
15
Netload = Load - VRE
Illustration of netload at different VRE shares
0.0%
2.5%
5.0%
10.0%
20.0%
80
70
Net load (GW)
60
50
40
30
20
10
0
1
10 20 30 40 50 60 70 80 90 100 110 120 130 140
Hours
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
of wind and solar may be lower, illustration only.
© OECD/IEA effect
2014
© OECD/IEA 2014 16
Four sources of flexibility …
Grid
infrastructure
© OECD/IEA 2014
Dispatchable
generation
Storage
Demand side
integration
© OECD/IEA 2014
17
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
Operationalization of
technical flexibility
 Extent of adjustment
 Maximum change of supply/demand balance over a certain
time horizon
 Two scales [+/-MW] per time horizon
 But capability will depend on:
 Past system state
 Current system states
 Future system state
 VERY many dimensions
 This gives rise to complex effects that make mapping of
a system on above scales challenging
© OECD/IEA 2014
Looking at the
ideal power device

 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”
Handful of
relevant
time scales
 Real power sources, loads & storage approximate ideal
source on these dimensions
© OECD/IEA 2014
The 4 flexible resources
 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 varRE 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
© OECD/IEA 2014
Focus on flexible generation
Or: can a coal plant really ramp?
© OECD/IEA 2014
© OECD/IEA 2014
22
Flexibility: ask for it…. and it appears
A sunny 1st May 2013 in Germany – actual production
Source: Fraunhofer ISE
 German hard coal plants carry most of ramping duty in Germany
 Lignite and nuclear ramp as well, even nuclear at some times
 Ramping costs can be minimised at low cost; retrofits are
possible e.g. Flexible Coal: Evolution from Baseload to Peaking Plant (NREL, 2013) 23
© OECD/IEA 2014
© OECD/IEA 2014
Experiences from coal in Denmark
© OECD/IEA 2014
© OECD/IEA 2014
24
It‘s not the fuel –
it‘s the design and the operation
© OECD/IEA 2014
© OECD/IEA 2014
25
No problem at low shares, because …
 Power systems already deal with a vast demand variability
 Can use existing flexibility for VRE integration
Exceptionally high variability in Brazil, 28 June 2010
© OECD/IEA 2014
© OECD/IEA 2014
26
Exercise 1: Netload
© OECD/IEA 2014
© OECD/IEA 2014
27
Session agenda
 Session 2:
System impacts and adapting operations
1.
2.
3.
4.
5.
6.
7.
Overview of impact groups
Avoiding typical mistakes when beginning deployment
Focus on balancing
Focus on utilisation
Focus on grids
Specific challenges and opportunities of distributed, small-scale PV
Strategies for improving system operations
 Exercise 2:
 Calculate impact of scheduling intervals on reserve requirements
 Assess benefit of using close to real-time forecast data
© OECD/IEA 2014
© OECD/IEA 2014
28
Different impact groups
© OECD/IEA 2014
© OECD/IEA 2014
29
Interaction is key
Properties of variable
renewable energy (VRE)
yrs
 Variable
Flexibility of other power
system components
Grids
Generation
 Uncertain
sec
 Non-synchronous
100s
km
 Location constrained
1 km
 Modularity
Storage
Demand Side
 Low short-run cost
© OECD/IEA 2014
© OECD/IEA 2014
30
Properties of variable renewables
and impact groups
 Systems are different – impacts will vary too
 But common groups of effects
Low short
run cost
Non
synchronous
Stability
Uncertainty
Reserves
Variability
Short term
changes
Abundance
Seconds
Stability
© OECD/IEA 2014
Asset
utilisation
Location
constrained
Modularity
Transmission
grid
Distribution
grid
Scarcity
Years 100km
Balancing
Profile / Utilisation
1km
Location
© OECD/IEA 2014
31
(Unit Commitment
& Replacement
Reserve &
Network Control
Reserve)
• Australia (NEM)
• Iberia (REE & REN)
• Texas (ERCOT)
• Hydro Québec
• Ireland (All Island)
• Germany (DENA)
• New Zealand (Transpower)
Electromechanical Stability
(Rotor Angle Stability,
Small Signal Stability,
Sub-Synchronous
Resonance)
Grid
Adequacy
Congestion
Management
(Transmission
Efficiency)
years
1 month
Security
Dispatch
Balancing
Operation
Planning
1 day
1hr
1min
1s
Time-constant/Period time
Source: Dr Thomas Ackermann, Energynautics
© OECD/IEA 2014
regional
CO2
& Fuel
Prices
MOST RELEVANT
SYSTEM STUDIES
Secondary
Reserve
(AGC
Regulation)
Transient
Voltage
Stability
(LVRT)
Power Quality
(Harmonics,
Electromagnetic
Transients)
Fault
Current
Level &
Protection
100ms
10ms
local
Generation
Adequacy
Tertiary
Reserve
system wide
Issues investigated in integration studies
1ms
© OECD/IEA 2014
32
Grid integration studies – best practice
 IEA Wind Implementing
Agreement Task 25:
Design and Operation of
Power Systems with
Large Amounts of Wind
Power

http://www.ieawind.org/index_page_postings/1
00313/RP%2016%20Wind%20Integration%20Stud
ies_Approved%20091213.pdf
© OECD/IEA 2014
© OECD/IEA 2014
33
Persistent impacts
© OECD/IEA 2014
© OECD/IEA 2014
34
Three typical mistakes to avoid
when deployment begins
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
© OECD/IEA 2014
© OECD/IEA 2014
35
No principal technical limit on VRE share
What is the maximum speed at
which you can driver a car?*
 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?
*Apart from the speed of light
© OECD/IEA 2014
© OECD/IEA 2014
36
Main persistent challenge: Balancing
Illustration of Residual power demand at different VRE shares
 Higher
2.5%
5.0%
10.0%
20.0%
80
70
60
Net load (GW)
uncertainty
 Larger and
more
pronounced
changes
0.0%
50
Larger
ramps
at high
shares
40
30
20
Lower
minimum
10
0
1
10 20 30 40 50 60 70 80 90 100 110 120 130 140
Hours
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
of wind and solar may be lower, illustration only.
© OECD/IEA effect
2014
© OECD/IEA 2014 37
VRE impact – hourly values (MW)
© OECD/IEA 2014
Load
PV
Wind
Netload
Hourly values in MW for different times of a day and days of a year in
Spain and Portugal
Note: figures refer to year 2011
VRE impact – hourly differences (MW)
© OECD/IEA 2014
Load
PV
Wind
Netload
Hourly differences in MW for different times of a day and days of a year
in Spain and Portugal
Note: figures refer to year 2011
Main persistent challenge: Utilisation
 Netload implies different utilisation for non-VRE system
90
0.0%
2.5%
5.0%
10.0%
Maximum 80
remains high:
Scarcity 70
20.0%
Peak
Net load (GW)
Peak
60
Midmerit
50
Midmerit
40
Changed utilisation
pattern
30
20
Baseload
Lower
minimum:
Abundance
10
Baseload
0
1
2 000
4 000 Hours 6 000
8 000
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
of wind and solar may be lower, illustration only.
© OECD/IEA effect
2014
© OECD/IEA 2014 40
Distributed PV and the grid 1/2
 Voltage rise common
issue
 Smart inverters and
transformers help
© OECD/IEA 2014
Transformer
Transformer
Voltage level deviation
Voltage rise in a rural distribution system in Germany
O%
+9%
© OECD/IEA 2014
41
Distributed PV and the grid 2/2
Power flows across HV-MV transformer
example form Bavaria, Germany
Source: Bayernwerk, Fraunhofer IWES
From Transmission to Distribution
From Distribution to Transmission
 Reverse flows no big technical issue
 But too high concentration can imply costly retrofits
© OECD/IEA 2014
Wind in Texas – Operational issues
while waiting on transmission
 Competitive
CREZ, Texas
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
© OECD/IEA 2014
345 kV double-circuit
upgrades identified in
CREZ transmission plan
Source: NREL
© OECD/IEA 2014
43
Excurse: Maximum penetration levels
 26/27 January 2013
© OECD/IEA 2014
© OECD/IEA 2014
44
Maximum non-synchronous penetration in
Ireland
W0
W25
W50
W75
W100
WMAX
SMAX
WMIN
SMIN
http://www.eirgrid.com/media/Renewable%20Studies%20V3.pdf
© OECD/IEA 2014
© OECD/IEA 2014
45
Adapting operations
© OECD/IEA 2014
© OECD/IEA 2014
46
VRE production forecasts
Accuracy of wind forecasts in Spain
 Forecasting of VRE
 Adjust schedules to
incorporate newest
information
 Allows to extract all
available flexibility
25%
Mean absolute error / average production
production key strategy
for cost-effective
operation
 But operations need to
make use of them:
20%
15%
10%
5%
2009
2010
2011
2012
0%
1
© OECD/IEA 2014
2008
5 9 13 17 21 25 29 33 37 41 45
Forecast horizon (Hours before real-time)
Source: REE
© OECD/IEA 2014
47
Generation and transmission schedules
Impact of scheduling interval on reserve requirements, illustration
Actual load curve
Capacity (MW)
Load schedule 15 minutes
Load schedule 60 minutes
Balancing need
15 min schedule
6
7
Time (hours)
8
9
Balancing need
60 min schedule
 Short scheduling intervals (5min best practice)
 Adjust schedules up to real time (5min best practice)
© OECD/IEA 2014
© OECD/IEA 2014
48
Co-operation with neighbours
Required frequency restoration reserves in Germany
60
7
+100%
50
+0%
40
6
5
4
30
3
-40%
20
2
10
1
0
0
2008
2009
2010
2011
Installed VRE capacity (GW)
Reserve requirement (GW)
8
VRE
capacity
Upward
reserves
Downward
reserves
2012
 Germany has four balancing areas (historic reasons)
 Reserve sharing mechanism across four areas
 Reduced requirements despite rapid increase of VRE
© OECD/IEA 2014
© OECD/IEA 2014
49
System service definitions
Prepared for a variable future?
 Are your operating reserve and system service
definitions VRE ready?
 Example Ireland DS3 programme
Frequency Services
Inertial
Response
Reserve
Ramping
Voltage Services
SIR
POR
TOR1
RR
FFR
SOR
TOR2
Ramping
0 – 5s
5 – 90s
90s – 20min
20min – 12hr
time
Transient Voltage Response
Voltage Regulation
Network
Dynamic
Reactive
Power
Steady-state
Reactive
Power
Network
Adequacy
• Synchronous Inertial Response
• Fast Frequency Response
• Fast Post-Fault Active Power
Recovery
• Ramping Margin
Grid 25
ms – s
s – min
• Dynamic Reactive Power
© OECD/IEA 2014
min – hr
• Steady-state Reactive Power
Source: EirGrid
© OECD/IEA 2014
50
Exercise 2:
Reserve requirements and VRE forecasts
© OECD/IEA 2014
© OECD/IEA 2014
51
Session agenda
 Session 3:
System friendly VRE, economics of grid integration, system
transformation
1.
2.
3.
4.
5.
6.
Options for system friendly VRE deployment
Market fundamentals: stable and dynamic power systems,
Merit-order effect and utilisation effect
System costs and system value of variable generation
Economic analysis of the four flexible resources
System transformation and total system costs
 Exercise 3:
 Calculate total system costs in a system where VRE are just added
 Calculate total system costs in a system where VRE and other plants can be
matched to each other
© OECD/IEA 2014
© OECD/IEA 2014
52
System friendly VRE
© OECD/IEA 2014
© OECD/IEA 2014
53
Economics of VRE integration
© OECD/IEA 2014
© OECD/IEA 2014
58
Transformation depends on context
• Sluggish demand growth
• Little general investment
needed short-term
Dynamic Power Systems
• Dynamic demand growth
• Large general investment
needed short-term
Example: Europe
Example: Emerging economies
Stable Power Systems
Slow demand growth*
Dynamic demand growth*
* Compound annual average growth rate 2012-20 , slow <2%, dynamic ≥2%; region average used where country data unavailable
© OECD/IEA
2014
This map
is without
prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.
© OECD/IEA 2014
59
Main market impacts in stable systems
Shift in German spot market price structure
 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
 Economic impact on gas generation result of several factors
© OECD/IEA 2014
© OECD/IEA 2014
60
The ‘merit order effect’
© OECD/IEA 2014
© OECD/IEA 2014
61
Main persistent challenge: Utilisation
 Netload implies different utilisation for non-VRE system
90
0.0%
2.5%
5.0%
10.0%
Maximum 80
remains high:
Scarcity 70
20.0%
Peak
Net load (GW)
Peak
60
Midmerit
50
Midmerit
40
Changed utilisation
pattern
30
20
Baseload
Lower
minimum:
Abundance
10
Baseload
0
1
2 000
4 000 Hours 6 000
8 000
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
of wind and solar may be lower, illustration only.
© OECD/IEA effect
2014
© OECD/IEA 2014 62
Utilisation: short term impact
90
Peak
Mid-merit
 Displacement
Base
of units with
highest
operating
costs
80
Net load (GW)
70
60
50
40
30
20
10
0
1
© OECD/IEA 2014
2 000
4 000
6 000
Hours
8 000
© OECD/IEA 2014
63
Utilisation: long term impact
90
Peak
Mid-merit
 Re-adaptation
Base
leads to
displacement
of technologies
that are costly
at low
utilisation
rates
80
Net load (GW)
70
60
50
40
30
20
10
0
1
© OECD/IEA 2014
2 000
4 000
6 000
Hours
8 000
© OECD/IEA 2014
64
Impacts of utilisation on total costs of
residual system
35
Total cost (billion USD/year)
Solar PV only
30
Wind only
Mix
25
Baseload
20
Benchmark
15
100%
90%
80%
Remaining net load
70%
60%
 At growing shares of variable generation, cost savings in the residual
system tend to slow
 (Most) relevant economic effect, exact numbers system specific
 Not captures by standard ‘back-up costs’
© OECD/IEA 2014
© OECD/IEA 2014
66
Examples of calculating balancing costs
Ireland (SEAI)
UK, 2002
UK, 2007
Colorado US
Minnesota 2004
Minnesota 2006
California US
EWITS, US
Pacificorp US
Greennet Sweden
Greennet Norway
Greennet Denmark
Greennet Finland
Greennet Germany
8.0
7.0
USD/MWh of wind
6.0
5.0
4.0
3.0
2.0
1.0
0.0
0%
5%
10%
15%
20%
Wind penetration as share of gross demand
25%
30%
 Modelled balancing costs are usually in the order of a few USD/MWh, highly
system dependent, typically rise with VRE share
 Empirical data from German market: approx. 3 EUR/MWh to balance the
FIT-portfolio
© OECD/IEA 2014
© OECD/IEA 2014
67
Investment in additional flexibility
Four sources of flexibility …
Grid
infrastructure
© OECD/IEA 2014
Dispatchable
generation
Storage
Demand side
integration
© OECD/IEA 2014
68
Flexibility options - Transmission
Where do you get inexpensive flexibility?
 Interconnection can offer low-cost flexibility
© OECD/IEA 2014
© OECD/IEA 2014
69
Flexibility options – Electricity Storage
Where do you get inexpensive flexibility?
 Storage shows relatively high LCOF compared to other
options
 But not only costs count – also the value matters
© OECD/IEA 2014
© OECD/IEA 2014
70
Flexibility options – DSI
Where do you get inexpensive flexibility?
 DSI potentially very favourable benefit/cost ratio
 Moving demand to when wind blows and sun shines
increases firm capacity credit of variable sources
 But what is the real potential?
© OECD/IEA 2014
© OECD/IEA 2014
71
Flexibility options – Generation
Where do you get inexpensive flexibility?
© OECD/IEA 2014
© OECD/IEA 2014
72
Benefit/cost of flexibility options
North West Europe - DSI
1 500
DSI
1 000
Million USD per year
500
128
899
0
- 500
Fixed Opex
-29 -437 -140
• 71% made of heat and
other schedulable
demand (110 TWh)
• 29% EV demand (44 TWh)
Start up costs
-1 000
-1 500
Capex CCGT
DSI assumed to be 8% of
annual power demand:
-1 539
Variable cost
incl. fuel
-2 000
CO2 price USD 30 per tonne
Coal price USD 2.7/Mmbtu
Gas price USD 8/Mmbtu
CO2 costs
-2 500
System benefit
DSI costs
 Overall system savings of 2.0 bln $/year
 DSI costs of 0.9 bln $/year
Benefit/cost ratio: 2.2
Note:
graph
© OECD/IEA
2014 represents the differences between DSI scenario (DSI 8% of overall demand) and baseline scenario
© OECD/IEA 2014
73
Benefit/cost of flexibility options
North West Europe
Benefit/cost ratio
2.3
2.2
0.6 - 1.5
1.2
1.1
1.1
IC
Storage + IC
Storage
1
DSI + IC
DSI
Hydro capacity
boost
 DSI has large benefits at comparably low costs
 Interconnection allows a more efficient use of distributed flexibility options
and generates synergies with storage and DSI
 Cost effectiveness of hydro plant retrofit depend on project specific
measures and associated investment needs
Notes:
1) CAPEX assumed for selected flexibility options: interconnection 1,300$/MW/km onshore and 2,600$/MW/km offshore, pumped hydro storage 1,170$/kW, reservoir
hydro 750 $/kW -1,300$/kW (repowering of existing reservoir hydro increasing available capacity). Cost of DSI is assumed equal to 4.7 $/MWh of overall power
demand (adjustment of NEWSIS results)
2) Fuel prices and CAPEX ($/kW) for VRE and flexibility options are assumed constant across all scenarios
© OECD/IEA
2014
© OECD/IEA 2014
Source:
IEA/PÖYRY
74
Investments in system flexibility –
Need for a suite of solutions
 Abundance
 Flexible generation  
 DSI 
 Storage 
 Curtailment 
 Scarcity
 Flexible generation  
 DSI o
 Storage 
 Curtailment  
  : very suitable,  :suitable, o : neutral,   : unsuitable
Data:
Germany 2011, 3x actual wind and solar PV capacity
© OECD/IEA 2014
80
60
40
GW
it all!
 Example:
Solar and wind can be abundant …
20
0
-20
33
Wind
34
… or scarce.
35
36
37
Day of the year
Solar
Demand
38
39
Net load
80
60
40
GW
 No single resource does
20
0
-20
321 322 323 324 325 326 327
Day of the year
© OECD/IEA 2014
75
System transformation and total system
costs
© OECD/IEA 2014
© OECD/IEA 2014
76
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
© OECD/IEA 2014
Remaining
system
VRE
Power system
• Generation
• Grids
• Storage
• Demand Side Integration
© OECD/IEA 2014
77
Three pillars of system transformation
of power
plants
System
friendly
VRE
© OECD/IEA 2014
2.
Make better use of
what you have
Operations
1.
Let wind
Geographic
spread
and
solar
play their
part
Design
Investments
Technology
spread
3.
Take a system wide-strategic
approach
to investments!
© OECD/IEA 2014
78
Total system costs (USD/MWh)
Cost-effective integration means
transformation of power system
Grid cost
140
+40%
120
+10-15%
100
DSI
80
Fixed VRE
60
Emissions
40
Fuel
20
Startup
0
Legacy
low grid costs
0% VRE
Legacy
high grid costs
Transformed generation &
8% DSI, low grid costs
45% VRE penetration
Fixed
non-VRE
Test System / IMRES Model
 Large shares of VRE can be integrated cost-effectively
 Significant optimization on both fixed and variable costs
© OECD/IEA 2014
© OECD/IEA 2014
79
Exercise 3:
Effects of VRE on total system costs
© OECD/IEA 2014
© OECD/IEA 2014
80
Add-on: integration costs
© OECD/IEA 2014
© OECD/IEA 2014
81
Frequent classification
 Decomposition always challenging:
 Balancing, adequacy, grids
 Uncertainty, profile, location
Unclear that all effects are covered with these practices;
categories not independent
 Calculation requires reference technology:
 Common benchmark for all technologies can skew the
calculation (flat block benchmark) or miss important
adaptation options (load correlated benchmark)
Portfolio of technologies (not just generation) needed to
minimise total system costs.
© OECD/IEA 2014
© OECD/IEA 2014
82
Problems with
calculating integration costs
 Decomposition always challenging:
 Balancing, adequacy, grids
 Uncertainty, profile, location
Unclear that all effects are covered with these practices;
categories not independent
 Calculation requires reference technology:
 Common benchmark for all technologies can skew the
calculation (flat block benchmark) or miss important
adaptation options (load correlated benchmark)
Portfolio of technologies (not just generation) needed to
minimise total system costs.
© OECD/IEA 2014
© OECD/IEA 2014
83
Alternative: System value
Savings
Costs
Generation costs per
MWh of VRE
Net savings in
residual system per
MWh of VRE
generation
System value
LCOE
 Benchmark VRE against net savings in residual power
system – more useful than trying to calculate
integration costs
 System value depends on transformation of the system 84
© OECD/IEA 2014
© OECD/IEA 2014