Integrating electricity balancing markets in Europe

Report
Integrating electricity
balancing markets in
Europe:
opportunities and
constraints
Gianluigi Migliavacca
Utility Week, Amsterdam
4th November 2014
Contents
•
Why is reserve needed
•
Different kinds of reserve
•
Balancing and reserve procurement within the “target model”
•
Advantages and criticalities of trans-national balancing
•
Regulatory context
•
Presentation eBADGE simulator
•
Presentation of some scenario simulation results
Why is
reserve
needed?
RES (wind,
solar)
Market (s)
and operation
•
Lacking still nowadays a reliable technology for bulk storage,
energy has to be consumed as soon as it is produced: power
demand and supply have to remain balanced in real time in
order to maintain frequency stable
•
In order to maintain real time balancing, an amount of reserve
generation has to stay available to move its program for
compensating real time deviations from the scheduled
generation and load programs.
•
Directive 96/92/EC has produced the unbundling of generation
and retail from the transmission system operation. GenCos
need economic signals to provide reserve services in an
economic way: balancing market were created with this aim.
•
The explosion of variable RES puts the network under
increasing stress: (EC 2050 Roadmap: between 31.6% and
48.7% of electricity production from wind by 2050). Need to
compensate variability (real time balancing) while maintaining
an adequate reserve level from conventional (dispatchable)
generation.
•
Necessity to reduce costs by sharing reserve among the EU
Countries. Need to coordinate relevant markets and TSO
operation. However, there is a clear competition with CBT in
interconnector allocation.
•
Prompt from EC to complete the IEM progressively coupling
the whole market chain from forward till reserve and
balancing. In this way, it is possible to increase markets liquidity
and optimize the usage of complementary generation
resources in different Countries (gas, wind, PV, nuclear…)
•
VPPs increased penetration (distribution side): can they
prove a viable resource within a trans-national
balancing/reserve market?
•
Interaction of reserve/balancing with storage
Load
Dispatchable
generation
VPPs
Storage
Different kinds of reserve
Source EWEA
•
•
•
ACER definitions:
•
(Balancing) Reserves – power
capacities (MW) available for TSOs
to balance the system in real time.
These capacities can be contracted
by the TSO with an associated
payment for their availability and/or
be made available without
payment. Reserves can be either
automatically or manually
activated.
•
Balancing Energy – energy (MWh)
activated by TSOs to maintain the
balance between injections and
withdrawals.
Frequency containment reserves – operating reserves for constant containment of frequency
deviations from nominal value in the whole synchronously interconnected system. Activation of these
reserves results in a restored power balance at a frequency deviating from nominal value. Operating
reserves have activation time up to 30 seconds and are activated automatically and locally.
Frequency restoration reserves – operating reserves to restore frequency to nominal value after system
imbalance. Activation up to 15 minutes, typically managed by an automatic controller.
Replacement Reserves – operating reserves used to restore the required level of operating reserves to
be prepared for a further system imbalance. This category includes operating reserves with activation
time from 15 minutes up to hours. They may be contracted or subject to market.
A “target model” for the EU market design
In December 2009 the
Electricity Regulatory Forum
("Florence Forum"), chaired
by the European
Commission and composed
of all member states and
relevant stakeholders,
endorsed the establishment
of the European Target
Model for congestion
management in the electricity
market. This work, continued
in 2010 through the Ad-Hoc
Advisory Group to ERGEG,
has constituted one the main
basis for the Framework
Guidelines on Capacity
Calculation and Congestion
Management.
Source ENTSO-E
Reserves procurement [MW]:
ahead of real time (before gate
closure of last market). Reserve
can be contractualized or
procured with market
mechanisms.
Balancing energy [MWh]:
Close to real time,
available reserve can be
activated (automatically
or manually by the TSO)
Advantages and
criticalities of a transnational reserve
procurement
CONS
•
•
•
More complicated and requires a
better synchronization of the markets
(gate closure, harmonization of
standard products)
Higher ICT costs
Ensuring cross-border dispatchability
of balancing orders can be nonobvious
PROS
• Greater dispatching efficiency,
costs reduction, less generators
in stand-by
• Better usage of generation park
with “complementary”
characteristics
Barriers to trans-national
balancing markets due to
non-harmonized design
variables of national
regulation
(source: eBADGE D2.2)
Optimal allocation
has to be promoted
between cross-border
day-ahead trade and
balancing
Source
ENTSO-E
The ENTSO-E Network Code
on Electricity Balancing
Submitted:
23.12.2013
•
ENTSO-E, in fulfillment to the Third Package and EC Regulation 714/2009, is working to 15
network codes upon relevant Framework Guidelines from ACER, among which the Network
Code on Electricity Balancing that was submitted to ACER for opinion on 23rd December
2013.
•
The NC EB provides for a phased approach to foster cooperation amongst TSOs in various areas
of Balancing. The key concept of Coordinated Balancing Areas. As time passes, the level of
cooperation within a Coordinated Balancing Area and between neighboring ones will increase;
neighboring Coordinated Balancing Areas will merge; and finally all Coordinated Balancing Area s
will merge to reach the final target of a single pan-European Common Merit Order list.
•
The NC EB creates a level playing field for all potential providers of Balancing Services, including
demand side response, energy storage and intermittent sources.
•
The harmonized processes and the use of Standard Products form a framework for providers to
offer Balancing Services to regional or pan-European Balancing Markets based on TSO-TSO
cooperation.
The eBADGE trans-national balancing market simulator
… models only the balancing energy
market as a real-time power dispatch, in
which:
•
imbalance values are known before
the activation for each modelled zone
•
(Secondary and tertiary) energy bids
are called on the basis of a price
merit order up to eliminate all
system imbalances.
•
The amount of cross-border capacity
still not allocated after the Intraday
trading is assumed as available for the
exchange of balancing energy.
Features of the simulator:
(1) based on historical bids
After an in-depth analysis, it was decided to assume both
historical balancing bids and historical imbalance figures as
known data to be used as an input for the simulation.
Pros: historical data naturally incorporate real system
characteristics.
Cons: difficulty of finding all data and national market
characteristics. The presence in our consortium of TSO helped
in solving a lot of problems (data incongruences, etc.).
Features of the simulator:
(2) netting is considered
what is the netting?
a procedure that allows the compensation of imbalances with different signs
between two neighbouring zones.
Austria
MWh
Slovenia
Imbalances
MWh
+50
-20
MWh
MWh
+30
Imbalances
after netting
0
Without netting, the solution is:
1. to accept an upward bid of +50
MWh in order to face the
positive imbalance
2. To accept a downward bid of 20 MWh in order to face the
negative imbalance
With netting, the solution is:
to accept only one bid in order to
solve the positive imbalance still
present in Austria
Features of the simulator:
(3) cost allocated to the zones generating imbalances
Costs should be allocated to the zones creating the imbalances, that
could be different from those where generation is activated
Austria
Upward
bid
accepted

Quantity= 5 [MWh]
Price= + 10 [€/MWh]
Slovenia
MWh
In this case the
imbalance occurs in
Slovenia, so this is a
cost for the Slovenian
TSO…
+5
COST= 5*10=50€
IMBALANCE
In a complex network, calculating cost
allocation may be very difficult.
The routine SCORGE by RSE, implementing
the Average Participation algorithm was
connected to the simulator
Features of the simulator:
(4) Secondary and tertiary bids separately considered
Simplified rules for selecting between secondary and tertiary bids are needed.
They were agreed with the Austrian and Slovenian TSOs.
For Austria and Slovenia, if the value of the imbalance is lower than a given
threshold, secondary bids are called for activation, otherwise tertiary bids are
activated.
MWh
MWh
THRESHOLD
THRESHOLD
IMBALANCE
secondary bids used.
IMBALANCE
Tertiary bids used.
For Italy the approach is the same, but the threshold is based on a percentage of
the zonal load.
Simulator model
 
min    cˆ g - up  p g  up 
  g  up
cˆ g  up  c g  up  K

p g - up 
g - up
EIE
Tl 


 VOLL
z
lz
0  p g - up  p g  up
 ENP z 
z
 ( EIE
z
 VOEE
z
 
EIE z  
 
K  max( abs ( c g  down ))  1
 ENP z ) 
 Imbalances
z

z
 p g  down 
g  down
p g - down 
 0
  PTDF
g - down
cˆ g  down  c g  down  K
g - down
z
 cˆ
To prevent the
simulator to
perform arbitrage
z
z
ENP z  0

  p g  up 

 g  up  z

z
p g  down  Imbalances
g  down  z
0  p g - down  p
g  down
z

 ENP z  EIE z  


T l  Tl  T l
Zones and network
• NTC values
T A-  B
CH
A
B
FR
T B-  A
• PTDF (power Transfer Distribution
Factors)
AU SL IT-NO
NO->AU -0.06 0.014 0.073
NO->SL -0.04 -0.3 0.24
AU->SL 0.14 -0.22 -0.07
0
0
NO->CN 0
0
0
0
CN->CS
0
0
0
CS->SU
0
0
0
SU->SI
0
0
0
CS->SA
IT-CN
0.073
0.24
-0.07
-1
0
0
0
0
IT-CS
0.073
0.24
-0.07
-1
-1
0
0
0
IT-SU
0.073
0.24
-0.07
-1
-1
-1
0
0
IT-SI
0.073
0.24
-0.07
-1
-1
-1
-1
0
IT-SA
0.073
0.24
-0.07
-1
-1
-1
0
-1
Scenarios
• Timeframe: 15 minutes
• Independent simulations
• … from 1 July 2012 to March 2013
(9 months32000 quarter of hours)
• Using historical values for imbalances
and bids
SCENARIO 1
Base Case
each zone solves the imbalances
with its own resources (no NTC
available between zones)
SCENARIO 2
Common Balancing Market
each zone solves the imbalances
with all the resources available
(NTC available between zones)
RESULTS:
Transits
(GWh)
Base Case
AU
Common Balancing Market
IT-NO
SL
145
264
IT-CN
IT-CS
ITSA
226
99
IT-SI
IT-SU
182
202
RESULTS: Transits (GWh)
Common Balancing Market
145
264
If we consider that:
• Total imbalances = 6286 GWh
• Total exchanges = 3894 GWh
..total exchanges are 60 % of the
total imbalances that occur in the
system!!
226
99
182
202
RESULTS:
costs (M€)
Base Case
Common Balancing Market
AU
3,5
5,2
IT-NO
94,8
48,4
SL
20,2
37,3
9,9
IT-CN
Total:
177,7
M€
Total:
79,7
M€
-1,3
IT-CS
17,9
IT- 4,4
SA
IT-SI
-3,2
5,3
3,3
IT-SU
13,2
-4,7
3,2
Conclusions
• Simulation results highlight great benefits for the integrated system of a
common management of the balancing energy market
• The system cost is more that halved (55%), passing from 177,7 M€ (no
balancing market) to 79,7 M€ in case a transnational balancing market is
implemented; for Italy the cost reduction is of the 60%.
• A greater and better usage of interconnections: with this approach an
implicit auction is implemented, and this allow an optimal usage of
available interconnection capacity.
Thank you for your attention…
Gianluigi Migliavacca
Head of Transmission Planning Research Group
RSE S.p.A.
via Rubattino,54
20134 Milano
E-mail: [email protected]

similar documents