7.
СОВЕТУВАЊЕ
Охрид, 2 − 4 октомври 2011
Guenter Bramboeck
EVN AG
BALANCING AS A KEY FUNCTION IN A LIBERALIZED MARKET
ABSTRACT
With the liberalization of electricity markets the dispatching of generation units is getting
decentralized and has to follow the contracts concluded in the open market. Whereas the quality of
supply also depends on the permanent balancing of generation according to actual consumption, the
System Operator has to provide reserve power capacity according to the international obligations. A
market has to be established for assigned products. The specifications have a remarkable influence on
the available liquidity of this specialized market and therefore on the costs. The mechanism, how these
costs are transferred to the market participants has an essential influence upon the stability of
balancing the system. The report shows, how to avoid abuse of the reserve capacity by a proper cost
allocation. The price formula has to be transparent and traceable and support the stability of the
system. The experience is positive and the monitoring by the market and the regulator didn’t show a
request for adaption.
Keywords: Liberalization, Balancing, Balancing mechanism, market monitoring, Electricity market.
1 INTRODUCTION
With the formerly integrated utilities and the closed grid and supply area, balancing was a
matter of planning and forecasting based on a few essential load meter data and actual energy flow in
these metering points. Although the operation of the generation units have followed a portfolio
management based on different cost structures of the different types of generation, the optimization
has not been that advanced as the costs have been covered by regulated tariffs. The liberalization of
the market aims to reduce the price of energy by competition and that includes the power reserve
capacity as well. With more and competing market participants the heritated system has to be adapted
and gets much more complex. The question was about how to organize the balancing of the system
with the highest priority still focused on the absolute stability of the electric energy system.
2 DEFINITION
Although the operation principles and rules for the transmission system operators (TSOs) are
laid down in the „UCTE Operation Handbook“, there is no explicate definition of system balancing.
According to the „Guidelines of Good Practice for Electricity Balancing Markets Integration“,
published by ERGEG,
„… the secure real time operation of a power system requires that TSOs ensure a
continuous balance between supply and demand. In competitive electricity markets, a balancing
market therefore generally exists such that TSOs can undertake balancing actions – that is, they
identify the need for, and procure, adjustments in generation or demand – in order to maintain
balance in their control area. Balancing markets differ from other market timeframes as TSOs
are a sole counterparty while in day-ahead and intraday timeframes market participants openly
trade between themselves to adjust their physical positions.
C5-139I 1/11
MAKO CIGRE 2011 C5-139I 2/11
Imbalance settlement can be used to encourage market players to maximise their efforts
to be in balance. Balancing markets therefore form an integral part of the overall wholesale
electricity trading arrangements and time schedules.“
To ensure the balance means the procurement of power reserve capacities, their operation by
the control area manager (CAM) and the provider of the capacity, the contribution of the market
participants to balance the system and therefore also the allocation of the costs for reserve capacity via
appropriate clearing prices.
3 BALANCING AS KEY FUNCTION
The technical challenge and the high costs for storage of electricity energy led to a system,
where the actual generation is permanently accommodated to the demand in the system. The load of a
control area and also of many distribution grids is permanently metered and with the data a forecast for
the following day or days is possible. The forecast relies also on wether forecast, type of day and may
include a lot of further parameters. All participants doing business with generation, supply or trading
have the responsibility to plan and forecast the balance in procurement and sales at any time.
A balanced system indicates that all trading deals are closed and the procurement by the
suppliers covers the actual needs of their customers. In the past decade we could see a vast deployment
of generation from renewable energy sources (RES). In order to maximise the harvesting of renewable
energy, conventional generation has to be adapted also in this respect. The capacity of such generation
underwent a tremendous development and forecasting for generation from RES has been developed
simultaneously.
The increasing number of actors on the market implies more complexity to ensure the
permanent balance in the multinational system. Each country is responsible for the market rules within
it’s territory but a failure may cause a blackout all over the continent.
4 IMAPCT OF LIBERALIZATION – MORE PERTICIPANTS, MORE DATA
Before the electricity market reform has started, in many European countries the electricity
system was a basically state owned monolith. It comprised generation as well as transmission and
distribution to supply and has been intransparent in many aspects. The, compared to actual status,
simple structure alleviated the balancing of the overall system by a load sequence process (Figure 1:
Starting position, CAM as dispatcher and balancing responsible). Although cost structures were
considered in dispatching of production, costs have been covered by tariffs approved by the
government. Generators with different ownership have been integrated under the regime of the main
dispatcher.
Figure 1: Starting position, CAM as dispatcher and balancing responsible
As soon as a market becomes more liquid, the economic performance of a generator depends
on his strategy towards the market and he wants to be totally independent from the central regime. He
no longer decides operation based on load forecasts, but based on market prices. It is the particular
demand for a specific period, normally an hour, which defines the price in the market. The suppliers
MAKO CIGRE 2011 C5-139I 3/11
have to know their future demand based on forecasts upon their sales to final customers. The way they
design their portfolio and their procurement strategy determine the economic result. The suppliers
have to provide the balance and take this responsibility as part of their commercial risk (Figure 2:
Sharing balancing responsibility). Failure and quality of planning will affect the system, but this risk is
distributed to more responsible parties than before.
Figure 2: Sharing balancing responsibility
5 ENERGY ACCOUNTING AS PREREQUISITE
The braking up of the single balancing responsibility requests a correspondent change in
planning and load sequence processes. As the grid and the supply area no longer cover the same
region, it is necessary to establish a clearing system to balance the energy flows actuated by a supplier
and his customers. The clearing is comparable to a financial account, but it has to be drawn up for each
single metering period of normally ¼ or one hour. The clearing has to be compiled within a reasonable
period of one week or one month. The clearing results in the calculation of the amount of balancing
energy for the metering periods and indicates the accuracy of the forecasting process (Figure 3: Tight
planning and forecasting process on accounts). Consequently, the market participants align their
market behavior to the rules for clearing and balancing energy price.
Figure 3: Tight planning and forecasting process on accounts
MAKO CIGRE 2011 C5-139I 4/11
The accounts are defined by metering points, physically represented by energy meters as well
as virtual covering the trades and short-term schedules (Figure 4: Accounts on physical and virtual
metering points). If the entire system is mapped in accounts, the total imbalances of all market
participants are equal to the imbalance of the control area.
Figure 4: Accounts on physical and virtual metering points
The suppliers respectively the balancing groups assume their responsibility to balance
procurement and sales by a forecasting based on the same principles of energy accounting. They have
to collect historic data, namely time series, from their customers connected to different grids within the
same control area as shown (Figure 5: Forecasting and scheduling by suppliers or balancing groups)
Figure 5: Forecasting and scheduling by suppliers or balancing groups
To facilitate the change of supplier and to get a sound data base, the provision and reading of
the meters should be carried out by the grid operator, also because of his neutral position in
competition (Figure 6: Collection and distribution of energy data by grid operators). As the grid
operator takes responsibility on grid losses, he has to ensure the metering of all energy flows. The
coordinator of the balancing groups established in a control area collects the aggregated data from the
grid operators and executes the clearing and settlement process.
MAKO CIGRE 2011 C5-139I 5/11
Figure 6: Collection and distribution of energy data by grid operators
The data are aggregated based on load profile metering or on standardised load profiles. A
comparison of the standardised profiles with the physical total load of the referring customers shows
differences of up to more than 50%. Therefore, the standardised load profiles may satisfy the starting
period of a competitive market, if the incumbent supplier takes the load risk for these customers. In a
market with distributed market shares, the physical load profiles get more and more important. The big
amount of data requests a system of Smart Metering and remote reading.
6 TRANSITIONAL SOLUTION FOR THE LOCAL PLAYER?
As long as not all meters are exchanged against load profile meters, a combined “bottom up”
and “top down” practice may be applied. Whereas new market players get their data based on
individual metering, the load of the incumbent supplier is calculated from grid data, according to the
calculation scheme, shown in (Figure 7: Solution for local player). It is necessary, that the change of
the meter doesn’t cause additional costs to the referring customers and that the grid losses can be
calculated on a reliable procedure.
Figure 7: Solution for local player
By this approach it is possible to get the load of the incumbent supplier and to enclose this big
part of national supply into the whole clearing and balancing process.
MAKO CIGRE 2011 C5-139I 6/11
The main challenge anyway is the change of meter reading and data collection. Whereas in the
old structure these data have been used for billing and sometimes for grid dimensioning, the liberalised
market requests more data with more attributes and much higher demand in timely response and
frequency.
7 POWER RESERVE CAPACITIES
With the balancing groups participating in balancing and economically responsible for
deviations, the System Operator on transmission grid level needs still reserve power capacity for the
final adjustment of frequency and load flows. The structure of providing the reserve capacity has not
been changed despite the change in roles and responsibilities after market opening. The capacity
options differ in the timely availability, the way how the control is facilitated and the range and speed
of reaction (Figure 8: Types of control reserve). The reserve means are controlled based on frequency
and/or load.
Figure 8: Types of control reserve
According to Policy in the relevant entso-e Handbook three main products are defined (Figure
9).
Primary reserve is used to maintain balance between generation and consumption within the
synchronous area. By the joint action of all interconnected parties the primary control aimes at the
operational reliability of the power system in the synchronous area by stabilizing the system frequency
at a stationary value after a disturbance or incident within seconds but without restoring the system
frequency or the power exchanges to their reference levels.
Secondary control maintains the a balance between generation and consumption within each
control area as well as the system frequency within the synchronous area, taking into account the
control program, without impairment on the primary control that is operated in the synchronous area in
parallel. Secondary control makes use of a centralised and continuous automatic generation control,
adjustments of controllable load in the time-frame of seconds up to typically 15 minutes after an
incident. Secondary control is based on generation resources made available by generation companies
to the TSOs
Tertiary control uses tertiary reserve that is usually activated manually by the TSOs in case of
observed or expected sustained activation of secondary control. It is primarily used to free up the
secondary reserves in a balanced system situation, but it is also activated as a supplement to secondary
reserve after larger incidents to restore the system frequency and consequently free the system wide
activated primary reserve.
MAKO CIGRE 2011 C5-139I 7/11
Figure 9: Structure of system power reserve capacities
8 PROCUREMENT IN AN OPEN MARKET
The products for the reserve capacities to provide system stability are not available in the usual
market and therefore tailored tenders have to be organised. Especially for primary and secondary
reserve capacities specific equipment is needed and monitoring of service has to be facilitated.
Providers have to undergo a pre-qualification procedure, where the technical and
organisational preconditions are fixed and checked. Generation units used for this service have to
provide minimal technical requests. Data communication and saving of operational data are important
elements to ensure service quality. If this service should be offered in the internal European market,
standardized parameters are obligatory.
To ease and speed up the frequent tendering, a master agreement between the tendering party
(TSO) and the providers is necessary. This standardised agreement defines the mutual rights and
obligations in the tendering process, invoicing, payment and default provisions, etc.
An electronic tender procedure attracts more potential providers by reducing administrative
efforts, reducing possible misinterpretation, providing a more transparent and standardised procedure
and faster operation of the tender procedure.
The liquidity of a market is a question of product design. Whereas in a supplier market the
product should attract demand, in the demand market the product has to be most attractive to possible
providers. Power reserve normally only can be provided from a generation unit in operation, especially
for primary and secondary regulation. Smaller minimum product size and shorter delivery periods ease
the offer for providers. Automatic control reduces the required minimum size, manually activated
capacity as tertiary reserve normally requests higher minimum capacity. The frequency should be
shorter than a month, with an optimum in a weekly tender. For unexpected additional short-term
capacity demand an electronic tender procedure for intraday should be provided for tertiary control.
Pricing for primary regulation is only for capacity, there is no specific energy demanded. Both,
secondary and tertiary regulations are offered in two components. A capacity price is paid for the
provided option and a energy charge according to the extent and duration of calls.
The selection of offers follows a merit order list. For products with two price components
different criteria might be applied. Either the selection of accepted offers is based on the capacity price
and the order of calls follows the merit order list of energy prices, or the selection is based on a typical
or historic shape and the resulting mixed price is bases for the merit order list.
Transparency in the whole tendering procedure and publishing prices in an appropriate way
should result in maximising offers and finally minimising costs. As reserve capacity often is marginal
MAKO CIGRE 2011 C5-139I 8/11
capacity at marginal costs, prices use to be very volatile. This makes it hard to transfer costs to fixed
tariffs and therefore other ways of financing have to be found.
Another aspect has to be considered. If a certain amount of costs is dedicated to balancing
energy price, it is not only about volatility of prices but also about vastly differing demand on
regulation energy from month to month. This results in the phenomen that higher demand of energy
from reserve capacities trends to result in lower specific costs for balancing energy.
9 BALANCING PRICE MODELS
The common used system in Europe is a two-price model based on the imbalance of the
balancing responsible party, the balancing group. The demand normally is fixed to a higher price than
average market price, whereas the price for surplus within a balancing group is significantly below
average market price. The price can be fixed for a day or longer periods or, in advanced markets, for
each clearing period. Two-price model however is in favour of big balancing groups and a
disadvantage for newcomers in the market. This system just reflects the situation within the balancing
group and normally not the imbalance of the control area. It is an incentive to permanently readjust
schedules to minimise costs for a balancing group but entails higher administrative effort.
A one-price model should be calculated on the amount of deviation of the control area. That
means that each party supporting the balance and therefore the stability of the system is rewarded
according to the need for such stabilisation. Although the price normally is calculated on the costs, it is
possible for market participants to forecast roughly the price. If the price also considers the market
price for the relevant clearing period, arbitrage isn’t possible. With a formula, where the price of
balancing energy is increasing with the shortening within a control area and reduced or even getting
negativ with increasing surplus, the participants get the right signal even in case of a failure of
generation or other incidents. Together with an adjusted information system, the imbalance of a
control area stays very low and by this the need for regulation capacity is minimised (Figure 10:
Clearing price dependent on system imbalance).
120,0
100,0
80,0 Clearingpreis
Börsepreis
60,0
EUR/MWh
40,0
20,0
avg. Control area imbalance
0,0
-300
-280
-260
-240
-220
-200
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
-20,0
Source: data 2010, published by Austrian Power Clearing & Settlement, own calculation
-40,0
Category MW
Figure 10: Clearing price dependent on system imbalance
The normal curve of distribution of imbalances should be very narrow to the stable system.
The corresponding normality to the price curve as in (Figure 10: Clearing price dependent on system
imbalance) is shown in (Figure 11: Normal distribution of imbalances in a system with imbalance
pricing) below.
MAKO CIGRE 2011 C5-139I 9/11
1600
1400
1200
1000
Frequency
800
600
400
200
0
00
80
60
40
20
00
80
60
40
20
00
0
0
0
0
0
20
40
60
80
0
0
0
0
0
0
0
0
0
0
0
-8
-6
-4
-2
10
12
14
16
18
20
22
24
26
28
30
-3
-2
-2
-2
-2
-2
-1
-1
-1
-1
-1
Category M W
Figure 11: Normal distribution of imbalances in a system with imbalance pricing
10 ALLOCATION OF COSTS
If a one-price model for balancing energy is applied, it should have the structure shown in
(Figure 12: Compilation of a clearing price dependent on imbalance).
Figure 12: Compilation of a clearing price dependent on imbalance
Using the market price for the relevant clearing period as a starting point for price calculation
is the best approach to an arbitrage-resistant system. If a reserve capacity is called and the relevant
energy price is higher than the market price because of further demand, the higher price should build
the basis. If there is a surplus and the energy deriving from called (negative) capacity is lower, the
lower price should be considered. Adding a dynamic price element depending on the degree of
imbalance of the control area automatically allocates the costs to those market participants being
responsible for the disturbance of stability. At the same time it offers an opportunity to other market
participants to activate accessable reserve capacities to support the reduction of imbalance.
The calculation may follow the scheme displayed in (Figure 13: Calculation scheme).
MAKO CIGRE 2011 C5-139I 10/11
Figure 13: Calculation scheme
The coincidence of volatility in price and changing volumes of balancing energy may
implicate very high cost risk to a small failure. It is obvious, that all market participants benefit of a
stable system and therefore should financially contribute to the costs as a kind of insurance rate. Such
a contribution may either be a specific grid tariff component, e.g. enclosed in a Tariff for ancillary
services or imposed to the suppliers as a separate clearing fee invoiced in a common process with
balancing energy. To make it a part of the clearing process would allow a more flexible model in
allocation of costs than in grid tariffs.
11 TRANSPARENT CALCULATIONOF CLEARING PRICE
The price calculation should be performed according to published formulas. The formula for
the basic price could be the following.
⎧min( Pt; PX , t ) Vt < 0
PB, t := ⎨
⎩max( Pt; PX , t ) Vt > 0 , (1)
= sgn(Vt ) ⋅ max (sgn(Vt ) ⋅ Pt; sgn(Vt ) ⋅ PX , t )
P B, t Basic price
PX , t PX price of relevant interval
Vt Imbalance of control area, positive if shortage to be compensated
The dynamic allocation of costs T may follow the shape of a “funnel”.
⎧ UMax − UMin 2
⎪UMin + 2
⋅ Vt Vt < VMax
T (Vt,UMax,UMin,VMax ) := ⎨ VMax , (2)
⎪UMax Vt ≥ VMax
⎩
UMax Maximal amount for dynamic element for relevant month
UMin Minimal amount for dynamic element for relevant month
VMax Imbalance value at which the dynamic allocation of costs has its maximum
MAKO CIGRE 2011 C5-139I 11/11
The Clearing price PC, t for the specific interval (15 min) results as follows:
PC, t := PB, t + sgn(Vt ) ⋅ T (Vt ,UMax,UMin,VMax ) . (3)
€/ M Wh
Clearingpreis
Grundpreis
De l t a RZ [ M Wh]
12 EXPERIENCES FROM THE PROPOSED MODEL IN AUSTRIA
The clearing price model displayed in this report is applied in APG control area in Austria
since July 2005 and has not been changed yet. The working group assigned in the national association
together with the regulatory body frequently monitors the market response and the impact on the
control area and makes and discusses analysis. There hasn’t been any negative experience reported by
the control area manager deriving from this kind of cost allocation. One positive effect could be
recognised in improving forecasting tools for RES.
In 2010 the control area’s imbalance energy (positive and negative together) amounted to
1,45% of the final consumption after 1,54% in 2009.
A comparison of specific balancing costs to other countries seems to be problematic according
to very different allocation via grid tariffs, fees, etc.
13 CONCLUSION
In designing procurement and pricing model for power reserve capacities and balancing energy
the focus should be on maintaining the stability of the system as all market participants benefit from
such an approach.
Distributed responsibility for balancing generation and consumption ensures the common
understanding of the main target. Market participants are open for contribution and more sensitive
against abuse of the system.
Only if a comprehensive system for data collection, data management and clearing process is
established, distributed responsibility may be applicated. Meter data have to be provided reliable,
within a short term, comprehensive and accurate. These data allow a generally accepted clearing and
sound forecasting process performed by each market participant. The well functioning and good
performance of the clearing and settlement has to be provided.
The clearing price structure has to incentivise support by those market participants having
additional power reserve capacities.
Transparency has to be maximised in regard to data and price building. Acceptance will be
improved by creating a trustful business atmosphere.
