Impact of bilateral contracts on the Italian electricity market by mifei


									Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy                                                                           189

                Impact of bilateral contracts on the Italian electricity market
                  A. Berizzi                          C. Bovo                       M. Delfanti                      M.S. Pasquadibisceglie
                                                              Politecnico di Milano - Italy

Abstract – The italian energy market adopted a hybrid structure:                configurations are a good representation of the unpredictable
generators are not mandated to submit energy bids in the aution                 strategies of the market operators.
mechanism, and are allowed to sign contracts with eligible customers and
traders. Such bilateral transactions are included in the day-ahead market,
receiving a scheduling priority. Since the operators strategies are difficult
                                                                                Day- ahead energy market model
to predict, a Monte Carlo approach is used to create different contract sets.
For each set a day ahead market is simulated to evaluate the possible
contract curtailment due to transmission constraints. In addition, a power      The Italian EHV transmission network is characterized by
flow calculation is carried out to analyze the EHV network performances         several bottlenecks that limit the total transfer capabilities
according to (n-1) security rules.                                              (TTC) among different areas. The PX takes into account these
Index terms – Bilateral Contracts, Electricity Market, Monte Carlo,             transmission constraints in a simplified way: the market is
Zonal Model                                                                     based on a zonal model.
                                                                                The process of the a priori identification of the zones is based on a
Introduction                                                                    specific procedure, described in the following [6]:
                                                                                1. no intra-zonal constraints are assumed to arise in the most
In Italy the Power Exchange (PX) started on April, 1st 2004. The                    frequent conditions;
cost-based fully centralized dispatch adopted in the last years of              2. power injections and withdrawals inside each zone do not
the vertically integrated operation has been abandoned towards a                    affect the transfer capacity among different zones;
bid-based dispatch with the presence of bilateral trades.                       3. an interface between neighbouring zones is identified as a set
The market rules have been discussed since 2001, starting from                      of transmission lines congested in the most frequent operating
the first issue of the PX regulatory framework [1] up to the most                   conditions.
recent documents published in late 2003 [2][3]. In the first                    According to the above procedure, the Italian ISO (Gestore della
proposal, bilateral trades were considered as an infrequent                     Rete di Trasmissione Nazionale - GRTN) identified 21 zones, 13
exception subject to the authorization of the regulator: all the                representing Italy, 8 representing Europe [6]. The national
energy had to be sold in the auction mechanisms. The last version               territory is divided in 7 geographical zones and 6 virtual zones.
of the PX rules changed the situation: now eligible customers and               The latter ones represent production areas with limits on export
generators can stipulate bilateral contracts without asking any                 capacity due to the presence of large power plants.
authorization. Moreover, all these contracts receive a dispatching              The zonal model is fully depicted in figure 1, where the virtual
priority in the day ahead market.                                               zones are shaded and the European ones are dotted.
The PX is managed by GME (Gestore del Mercato Elettrico) and
is organized in subsequent markets for energy, network services,                                 France      Switzerland              Austria       Slovenia

and reserve: this paper deals only with the day-ahead energy
market, with the presence of bilateral trades. The hourly clearing is                                UCTE N-W                              UCTE N-E

given by the point in the (€, MWh) space where demand and
supply curves meet, resulting in a single clearing price for the                         Turbigo/Ronco                      North                      Monfalcone
whole country, if no operating constraint is violated. A simple
mechanism for the solution of congestions is provided in the day-                                                      Middle North
ahead energy market, by a zonal approach, triggered by a violation
of the total real power flow between pre-defined zones. In this                                                            Piombino                      Corsica
case, the hourly clearing is found maximizing a total welfare
objective function, subject to energy balance and to simplified
                                                                                                                       Middle South                      Sardinia
transmission constraints. This algorithm provides zonal prices
which are directly applied to generators bids, while the consumers
are charged an unique price for the whole country, obtained                                                                 South

performing a mean value of the zonal prices weighted on the zonal
loads [3]. In order to receive a scheduling priority, bilateral                                           Rossano                        Brindisi

contracts are equivalent to supply bids at zero €/MWh and demand
bids without price upper limit in the PX.                                                                 Calabria                       Greece
The most common approach to investigate bilateral transactions in
a deregulated environment is a Monte Carlo simulation where                                                Sicily                         Priolo
many different scenarios are created by random extractions [4][5].
This method has been widely used so far, since its random                             Fig. 1 Zonal model for the Italian system (with UCTE equivalent).
190                                               Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy

The constrained market is cleared by performing a social                         na

welfare maximization, subject to the energy balance constraints                 ∑PD     i       i
and to the inter-zonal transmission limits [3]. The generated          PUN =    i =1                                            (5)
power is paid at the zonal prices, while the consumers pay a
single price (PUN, Prezzo Unico Nazionale) valid in the whole
                                                                                 i =1

country and computed as a mean of the zonal prices weighted on        where
the zonal loads. The only exception is given by the pumping           • Pi is the zonal price in area i;
storage plants that are charged the zonal price.                      • Di is the demand related to area i;
The presence of the PUN generally leads to a complex                  • na is the number of areas.
mathematical problem involving not only continuous variables,
but also logic ones, as detailed in [7]. This algorithm can be        The formula (5) implies that the total revenue from the final
simplified if the demand elasticity is set to zero (i.e the demand    customers is the same that would be achieved with zonal prices
bidding is considered fixed and independent from the market           applied both to supply and demand bids.
clearing price). This assumption is quite reasonable, since in        The introduction of bilateral contracts leads to very small changes
the day-ahead markets worldwide loads have usually shown so           in the market model described so far: in particular, the rules state
far small changes with respect to energy price. Moreover, with        that all the contracts stipulated outside the PX shall be notified to
respect to the Italian situation, during the first months of          the GRTN in order to receive a transmission right on the EHV
operation of the PX, the eligible customers cannot actively take      network [8][9]. This allocation is performed in the day-ahead
part in the market: all the demand bids are submitted by the          market by granting a scheduling priority: when a transmission
GRTN, according to its load forecast, and are independent             constraint arises, the dispatcher first satisfies the bilateral
from the clearing price [3].                                          transactions, then accepts the other supply bids. This result is
With elasticity set to zero, the welfare maximization becomes a       obtained considering the bilateral transactions as supply bids with
linear problem, that can be easily solved using traditional           zero price and as demand bids without price limit. In most
techniques:                                                           situations, all the contracts are scheduled, while the residual
                ⎛            ng
                                              ⎞                       demand is covered by the supply bids submitted by market
MAX             ⎜ Pnf ⋅ Dt − ∑ SQ j ⋅ prs , j ⎟        (1)            operators in the auction mechanism. On the contrary, in some
                ⎜                             ⎟
                ⎝            j =1             ⎠                       scenarios, it can be impossible to dispatch all the bilateral
s.t.                                                                  transactions, due to transmission constraints. This situation has not
                                                                      been explicitly considered so far in the regulatory framework: in
∑ SQ
 j =1
            j   = Dt                                   (2)
                                                                      particular, it is still unknown whether only the generation
                                                                      associated with the contract is affected by curtailment or the
PFr , s ≤ PF r , s                                     (3)            relevant load is involved too. Since in this paper a fixed demand is
                                                                      simulated (i.e. final customers will cover all their loads at any
SQ i ≤ SQi ≤ SQ i                                      (4)
                                                                      price) the following assumption is made: if a bilateral transaction
                                                                      can not be dispatched, only its supply side is curtailed in the day-
where                                                                 ahead market, while the relevant final customer is always
• Dt is the total demand;                                             completely satisfied; in the mean time, the curtailed energy is
• Pnf is a reference price assigned to the inelastic demand;          provided using the supply bids submitted in the PX by some other
• SQj is the quantity that generator j is willing to sell;            generators not involved in the shortened contract and located in
• prs ,j is the offer price of generator j;                           different zones.
• PFr,s is the flow on the interface between area r and area s        When the market clearing process is over, the GME pays the
  computed using fixed coefficients, given in advance;                GRTN the congestion rents (i.e. the difference from the total
                                                                      revenue from customers and the total expense for the suppliers);
• PF r , s is the flow limit on the interface between area r and
                                                                      the contracts holders are charged a congestion fee, equal to the
  area s;                                                             product of the contracted energy times the difference between the
• SQ , SQ i are the lower and upper bounds of SQi;                    PUN and the zonal price where the contract supplier is located [9].
• ng is the number of generators considered.                          This fee is non zero only when a congestion arises and represents
                                                                      the cost associated to the transmission rights granted in the day-
The social welfare (1) is given by two terms: one regarding the       ahead market. For this reason this fee is paid both if the contract is
customers and one regarding the suppliers. Since the first one is     curtailed and if it is fully dispatched. Obviously, only the
constant, the maximization of the social welfare is reduced to the    dispatched energy is charged, while the shortened one is not
minimization of the second one, that is equal to the total cost       subject to this fee.
associated to the supply bids.
The satisfaction of the total demand is forced by equation (2),       Bilateral transactions
while the transmission constraints are represented by relation (3)
(every interconnection between zones is represented with two          According to the recent documents [2] [3], for each generator two
different inequalities, one for each direction.).                     possibilities are given: stipulating a contract with a final customer
When the market gets to an equilibrium, the PUN can be easily         outside the PX structure (with a known fixed price stated a priori)
calculated as follows:                                                or submitting a bid on the day – ahead market (with an unknown
                                                                      and volatile price). Each operator uses these methods, according to
                                                                      his market strategy. In order to represent the unpredictability of
Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy                                                      191

these strategies, Monte Carlo simulations have been used to define           With reference to fig. 2, the total probability is divided into eight
the plants involved in bilateral transactions. In [4] and in [5] a           sections, corresponding to eight power plants. In A, all the sections
matrix is created to represent all the possible relationships among          have the same area and therefore the same probability (1/8) to be
the different operators involved. Concerning the real power, each            chosen. On the contrary, in B the power plants have different
contract is characterized by an upper bound (the lower bound is set          chances.
to zero) while there is no assumption on the production cost (each           In this paper, in order to have some control on the suitability of the
generator has the same probability to be chosen for a contract with          results and to give the most economic plants greater probabilities
a given consumer, independently of its marginal cost).                       to be involved in a bilateral transactions (according with the strong
In this paper, a different methodology is proposed: the Monte                competition on bilateral contracts, as detailed above), the
Carlo approach is adopted only to choose the set of generators               following weights are taken for the bilateral transactions:
involved in bilateral contracts and not the amount of energy                        1
                                                                              wi =
contracted. The generators chosen for bilateral transactions are                    ci
assumed to produce their maximum power; the contracts are                    where ci is the marginal cost of the i-th plant at its maximum real
cleared at a fixed price, while the market is characterized by high          power. In this way, the biggest weights are associated with the
price volatility. As a consequence, each operator is reasonably              most economic plants (i.e the coal plants), while oil and gas plants
willing to sign a contract for the whole plant power, instead of             are characterized by a lower probability.
submitting bids in the PX. Moreover, a supply bid submitted                  Regarding the final customers, no random extraction is considered
within the auction mechanism may be partially accepted (the plant            in this paper: since the demand is supposed to be inelastic and not
is the marginal one). If this situation appears, the generation              subject to any type of curtailment, the loads are completely
profile assigned to the plant may be unfeasible, i.e. in contrast with       supplied either when they are involved in a bilateral transaction, or
the turbine and generator dynamic limits. In order to face this              when they submit a demand bid in the PX. Therefore, the location
situation, the generator can buy or sell some energy in the                  of the contract withdrawal points does not affect the market
adjustment market or he can pay to the GRTN an unbalancing fee,              results, and consequently the load pattern on the EHV network.
related to the network services market. Both these solutions may
imply some risks due to the unpredictability of the energy and               Simulation tools
services price. On the contrary, a bilateral contract is a stable
agreement: the selling price is stated a priori between the two              This work follows the day-ahead market simulation, run at
operators and the generation profile is dispatched with priority             Politecnico di Milano during winter 2002-2003 [10], adopting the
(there is only a low curtailment risk associated to the transmission         same hypothesis. Realistic data for generation capacity and cost
constraints). Therefore it may be scheduled in advance according             curves are used together with typical load patterns, representative
to the plant dynamic performances, with no need to be corrected              of a high demand week in the year 2005. Hydro capacity is
by submitting bids in the adjustment market. For this reasons it is          excluded from the bidding process, assuming that the hydro plants
realistic to assume that the major part of energy is sold through            are always scheduled, according to their availability. This
bilateral transactions, while only a small fraction is exchanged             simplification is due to the high computational burden introduced
within the auction mechanism. This assumption results in a strong            by the explicit representation of a complete model of river
competition between the different operators for selling energy to            systems. Some particular power plants (typically renewable
final customers outside the PX: in order to reduce his production            resources) that are subject to specific rules (environmental
cost and to sell his energy at a competitive price, each supplier is         incentives) are dispatched with priority too.
reasonably induced to dispatch first the cheap plants, then the              For each operator, standard bids are defined according to the plant
expensive ones.                                                              generation cost: the full power is offered at the marginal cost plus
This situation is taken into account in this paper, by setting a             a 10% margin in order to provide each operator a certain profit.
different weight for each plant: this mechanism derives from the             Before solving the day-ahead market, Monte Carlo extractions are
widely used roulette wheel method and is represented in figure 2.            used to generate some bilateral transactions. If a plant is chosen by
                                                                             the Monte Carlo random extractions to be involved in a bilateral
                                                                             transaction, its bid is modified to grant a dispatching priority (the
                                                                             price is set to zero €/MWh); if the plant is not chosen, the original
                                               A) Equal probabilities        bid is kept unchanged and contributes to the supply of the residual
                                                                             In order to test the global performances of the Italian system,
                                                                             different amounts of bilateral contracts with respect to the total
                                                                             demand are simulated: in particular, 80%, 70% and 60% values
                                                                             are assumed. These values are related to the liquidity shown so far
                                                                             in the Italian PX. This parameter is usually defined as follows:
                                                                                  Dt − Db
                                               B) Different probabilities    L=
                                                                             where Dt is the total demand submitted in the market (including Db
                                                                             associated to bilateral contracts).
                                                                             In Italy, the GME adopts a slightly different formulation: the term
                                                                             Db represents not only the total demand satisfied by contracts
      Fig. 2 Monte Carlo simulation with equal and different probabilities
                                                                             stipulated outside the PX structure, but also the demand satisfied
                                                                             by other bids that are granted a dispatching priority (i.e. some
192                                                 Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy

renewable sources plants and the import). During the first two
months of operation (April and May 2004), Italian PX has shown                Index B

a mean value for liquidity equal to 30%. This means that only a                         0.99

limited amount of energy has been really exchanged in the PX,                           0.98

while most customers have preferred to cover their loads outside                        0.97

the PX structure. However the resulting liquidity was relatively                        0.96

high if compared with other European markets with a similar

organization. In particular, in the Nord Pool the bilateral contracts                   0.93
supply more than 95% of total demand [11], while, in the Italian                        0.92

PX, bilateral contracts are assumed to cover no more than 70-80%.                       0.91

                                                                                                                                          Termini Imerese

                                                                                                                                                            Brindisi Nord

                                                                                                                                                                            Turbigo Levante

                                                                                                      La Casella

This situation is due to the constraint imposed to AU (Acquirente

Unico, responsible of the supply to captive customers): this
operator has to buy in the PX at least 75% of the total captive
customers demand, and this limits the total amount of energy that
AU can purchase with bilateral contracts [2].                                                     Fig. 3 Index B for market liquidity equal to 20 %
For each liquidity level, 2000 different bilateral sets are generated
by the Monte Carlo random extractions, relevant to four different             Index B

peak load patterns. For each scenario, a day ahead market                               0.98
simulation is carried out to allocate the transmission rights, while a                  0.97
power flow calculation is performed on the EHV network to                               0.96

assess the power system security.                                                       0.95


Transmission rights allocation                                                          0.93


As stated above, a bilateral transaction shall get a transmission                       0.91

                                                                                                                                          Termini Imerese
                                                                                                      La Casella


                                                                                                                                                            Brindisi Nord

                                                                                                                                                                            Turbigo Levante


right in the day-ahead market; if a constraint arises, all the
generators will be re-dispatched in order to meet the power flow
limits among the zones (according to the day – ahead zonal market
model detailed above). Concerning the bilateral contracts,                                        Fig. 4 Index B for market liquidity equal to 30 %
although the generators are willing to sell their maximum real
power, the transmission constraints may allow them to inject only                        1

a fraction x.                                                                Index B

An index A is defined to measure the mean energy curtailed in the                      0.98

day ahead-market (in order to meet the flow limits on the                              0.97

interfaces), for each power plant involved in bilateral contracts:                     0.96

       ∑ (P             − Pki )

                k max                                                                  0.94

Ak =   i =1

              ns ⋅ Pk max


where                                                                                  0.91
                                                                                                                                          Termini Imerese

• ns is the number of scenarios where the plant kth is chosen for a

                                                                                                                                                                               Turbigo Levante
                                                                                                  La Casella

                                                                                                                                                            Brindisi Nord



  bilateral transaction.
• Pki is the real power the plant kth is allowed to inject on the EHV
  network for the scenario ith (only scenarios with the plant                                     Fig. 5 Index B for market liquidity equal to 40 %
  involved in bilateral contracts are considered to evaluate this
  index).                                                                All the plants located in continental Italy outside the virtual
                                                                         areas (for example La Casella and Ferrera Erbognone in North
• Pk max is the maximum real power of the plant k.
                                                                         Italy and Montalto in Middle South Italy) always have B=1,
If A ≈ 0, the plant has negligible probabilities to be curtailed in
                                                                         even if the market liquidity changes: these generators have
case a bilateral contract is signed (A is calculated by meaning on a
                                                                         negligible probabilities to be curtailed in the day-ahead market
ns randomly extracted scenarios); on the contrary, if A>0, the real
                                                                         and can freely choose their market strategy without suffering
power (1-A)*Pmax is very likely to get a transmission right,
                                                                         particular risks. On the contrary, the situation is quite different
while the remaining A*Pmax may be curtailed.
                                                                         for the plants located in the virtual areas (for example
Once defined B=1-A, B measures the mean power (in p.u.)
                                                                         Monfalcone, Turbigo Levante, Brindisi Nord) or in the islands
available for bilateral contracts, i.e the power that has great
                                                                         (for example Sulcis in Sardinia and Termini Imerese in Sicily):
chances to be fully dispatched.
                                                                         if a 20% liquidity is assumed, all these generators have
In this paper, the index B is computed for all the plants directly
                                                                         significant probabilities to be curtailed due to transmission
connected to the 400 kV and 230 kV transmission grids. Some
                                                                         constraints; with an higher liquidity, the shortening risk is still
simulations are carried out to study the impact of possible
                                                                         high for the plants located in Sardinia and in Sicily, while it
curtailments depending on the market liquidity. The graphs in
                                                                         becomes negligible for the plants located in the virtual areas.
figures 3, 4 and 5 show the values assumed by index B relevant
                                                                         These results are fairly realistic as the virtual areas are
to some thermal plants in Italy.
                                                                         identified by GRTN according to the limited TTC with the
Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy                                                    193

neighbouring zones: a curtailment has to be expected in most               transmission rights on the EHV network. Similar comments can
cases as the plants located in the area cannot export their full           be made for the generators.
power. The same considerations can be made for the islands                 This situation can be more critical in the actual market: in fact, in
(Sicily and Sardinia) where the interconnection to the                     this paper, the residual demand is covered using supply bids
continental Italy is given by two cables with limited capability.          submitted according to the generation cost (plus a 10% mark-up)
                                                                           and is independent from the total amount of bilateral contracts. On
Market prices                                                              the contrary, in the real market a strong competition between the
                                                                           operators has to be expected: the few plants not involved in
The mean zonal prices obtained in the market simulations are               bilateral contracts may raise the clearing price in order to get
reported in Table I.                                                       higher profits, resulting in interconnection and congestion fees
                                                                           greater than those shown above.
                                Prices with    Prices with   Prices with   Another critical aspect regards the price assumed by many
            Area                  L=20%          L=30%         L=40%       operators as a reference for bilateral transactions. In almost all the
                                 (€/MWh)        (€/MWh)       (€/MWh)      other markets worldwide, the PX price is assumed as a reference
       Turbigo - Ronco             41.47           51.18       55.03       for the bilateral contracts: the energy value defined by suppliers
         Monfalcone                14.60           17.33       26.18
                                                                           and customers in their agreements is often set slightly lower. This
            North                  42.27           51.70       55.08
        Middle North               43.86           52.09       55.19       position was not possible in Italy, since the PX started only in
          Piombino                 43.64           51.98       55.17       April, while many bilateral transactions had already been signed.
        Middle South               41.58           50.33       54.48       Lacking a market price, the only reference value available was the
            South                  41.20           50.13       54.45       captive customers rate defined by the regulator [12]: typical values
          Rossano                  39.02           48.46       53.65       are 70.00 €/MWh for peak load (F1 rate) and 35.00 €/MWh for
           Brindisi                39.02           48.46       53.65
          Calabria                 56.51           60.42       61.30
                                                                           off-peak load (F4 rate). Starting from these values, it is realistic to
            Sicily                 50.08           56.68       59.65       assume that many operators stipulated bilateral transactions with a
            Priolo                 44.99           52.06       55.61       energy value slightly lower than the captive rate. During the first
          Sardinia                 33.88           41.39       47.96       weeks of market operation, the energy price has been near the
                           Table I – Mean zonal prices                     captive customer rate, but the situation has strongly changed
                                                                           between the end of May and the beginning of June. In some days,
The zonal prices present significant differences. In particular            the liquidity level has resulted quite low, with high congestion fees
Middle North and Calabria are importing areas (they lack of                (about 1.50 €/MWh, against a mean value less than 1.00 €/MWh),
generation and their prices are higher than the neighbouring               according to the simulation presented in this paper. In other days,
ones) while Sardinia and Sicily are exporting zones. North area            however, at the beginning of June, the market price has got over
imports energy from Europe and the virtual areas Turbigo and               100 €/MWh, a price significantly greater than the captive customer
Monfalcone and exports energy to Middle North. South and                   rate. Moreover, together with the price increase, a similar increase
Middle South import energy from Rossano and Brindisi and                   of the congestion fees appeared, with a heavy impact on the
export towards Middle North and Piombino.                                  generators total revenues. In facts, the bilateral contract price is
If a lower liquidity value is assumed, the importing and                   fixed and cannot be modified, while the associated congestion fees
exporting areas remain the same, while the prices undergo a                change, according with the day-ahead market results: therefore,
significant diminution. This reduction results in an increase of           the operators always get the same income, while their expenses are
many interconnection fees (the differences between the prices              increased, especially when high clearing prices appear.
of two neighbouring areas), as shown in table II (only the most
important interconnections are reported)                                   The power flow analysis

                                    L=20%          L=30%       L=40%       Power flow calculations are performed in order to evaluate the
                                   (€/MWh)        (€/MWh)     (€/MWh)      impact of bilateral transactions on the EHV network operation.
                                                                           A detailed model of the Italian 400 kV and 230 kV network has
       Monfalcone - North           27.67          34.38        28.90
         Turbigo – North            0.80           0.52          0.05
                                                                           been developed according to [13] and [14]: it is made up of
      North – Middle North          1.60           0.38         0.11       more than 1400 busses and includes an equivalent
       Piombino - Sardinia          9.76           10.59         7.21      representation of the UCTE grid, taking into account the
    Piombino – Middle North         0.22            0.11         0.03      topology derived by the most recent plans issued by GRTN for
         Brindisi –South            2.17            1.68         0.80      year 2005. In particular the 380 kV line Matera – Santa Sofia,
        Sicily – Calabria            6.44           3.74         1.65
                                                                           considered crucial for the security and reliability of South Italy
                     Table II – Mean interconnection fees
                                                                           network, is simulated in operation.
The interconnection fee is strictly related to the congestion fee          For each liquidity level, the mean loading on all the
charged to contracts’ holders at the end of the day-ahead market:          transmission lines is studied, both in n security (i.e intact
thus, an increase of the differences between the zonal prices results      system) and in (n-1) security (grid affected by a single fault).
in an increase of the congestion fee which has a severe impact on          More than forty single different contingencies are taken into
the operators’ strategies. In particular, on one side, the customers       account: faults on all the 220 kV and 380 kV lines connecting
are willing to buy energy outside the PX structure, in order to be         different market areas are considered (with the exception of the
protected against spikes and price volatility; on the other side, this     two cables with Sardinia and Sicily), together with other faults
results in a lower liquidity level and in potentially high fees for the    on some other heavily loaded branches. The current flowing in
                                                                           branch r consequent to a fault on branch s is approximated
194                                                                          Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy

using the β r (s ) coefficients derived by applying the Inverse
Matrix Modification Lemma to the susceptance matrix of the                                          0.6
grid. A 20% overload is accepted in order to simulate the
corrective actions taken by GRTN after a contingency has                                            0.5
happened.                                                                                           0.4
While the market results are strongly influenced by the
assumed liquidity level (as shown in the last section), the                                         0.3                                                                       with Matera - S.
power flows on the transmission network seem to be fairly                                                                                                                     Sofia
independent from the total amount of bilateral contracts: the                                       0.2                                                                       without Matera -
load pattern does not change with the liquidity (the demand is                                                                                                                S. Sofia
supposed to have zero elasticity), while the generation profile                                     0.1
undergoes only negligible modifications, since both Monte
Carlo and the PX dispatch first the cheap plants, than the                                              0

                                                                                                                                        M a te r a S .

                                                                                                                                                          M a te r a
expensive ones.

                                                                                                                  F o g g ia
                                                                                                                  L a r in o

                                                                                                                                                           L a in o
                                                                                                                                           S o fia
All the lines are within their limits in n security cases, while in
(n-1) security cases there are congestions in some virtual areas,
particularly in the Brindisi one. The four plants located in this
area (two thermal, one with renewable sources, one actually                                                                Fig. 7 Sensitivity calculation in (n-1) security
under construction with full operation expected in late 2005)
can export their power mainly through four 400 kV lines:
Foggia – Larino on the Adriatic coast; Benevento – Foggia
                                                                                                  This paper presents a methodology to evaluate the impact of
connecting the Adriatic coast with Naples and the Tirreno sea;
                                                                                                  the amount of bilateral transactions on a real market
Matera – Santa Sofia, actually under construction and
                                                                                                  environment, like the Italian electricity market, where such
connecting the Brindisi area directly with the Naples; Matera –
                                                                                                  bilateral transactions are included in the PX structure with
Laino, along the Ionio coast. These lines can be overloaded if a
                                                                                                  dispatching priority.
single contingency occurs.
                                                                                                  As the Italian PX is at its beginning and the rules are still in
To analyze the impact of the Brindisi plants on these lines, the
                                                                                                  progress, the impact of such market structure is not fully
sensitivity calculations can be used: the distribution among the
                                                                                                  predictable: in facts, both the regulatory framework does not
different lines of the active power generated inside the Brindisi
                                                                                                  cover all the issues, and the percentage of bilateral contracts
area is shown in figure 6.
                                                                                                  over the total amount of exchanged energy depends on the
                                                                                                  market operators’ strategy. A detailed Monte Carlo simulation
                                                                                                  has been adopted for such goal. The market results show that
                                                                                                  some generators are likely to be curtailed in some scenarios,
  0.35                                                                                            depending also on their location.
   0.3                                                                                            The network studies confirm the suitability of the zonal
                                                                                                  structure in modelling the transmission constraints: no
                                                                               intact system      significant congestions arise and only some slight overloads
   0.2                                                                                            result in (n-1) security studies within the virtual areas. In
  0.15                                                                         without Matera -   particular, with respect to the situation of Brindisi area, the
                                                                               Santa Sofia        sensitivity analyses stress the importance of the new branch
                                                                                                  Matera – Santa Sofia in order to contain future lines
  0.05                                                                                            overloading.
                                               M a te r a S .

                                                                M a te r a
                                 F o g g ia
          B e n e v e n to

                                 L a r in o

                                                                 L a in o
             F o g g ia

                                                  S o fia

                                                                                                  [1]       GME: Disciplina del mercato elettrico (May 2001) – approved by
                                                                                                            Ministro dell’Industria del Commercio e dell’Artigianato. Available on
                             Fig. 6 Sensitivity calculation in n security                         [2]       MAP e AEEG: Sistema organizzato di offerte di vendita e di acquisto di
                                                                                                            energia Elettrica: indirizzi per il sistema Italia 2004 (July 2003)
                                                                                                            Available on e
The effect of the branch Matera – Santa Sofia is clearly                                          [3]       GME: Testo integrato della Disciplina del mercato elettrico (December
demonstrated: with this new branch in operation, all the                                                    2003)– approved by Ministro delle Attività produttive Available on
neighbouring lines are less loaded and this results in a more                                     
                                                                                                  [4]       J.W.M. Cheng, D.T. McGillis, F.D.Galiana: Probabilistic security
secure operating point.                                                                                     analysis of bilateral transactions in a deregulated environment - IEEE
Figure 7 shows the same sensitivities when a fault on the                                                   Transactions on Power Systems, Volume 14, Issue 3, Aug. 1999
branch Benevento – Foggia is simulated.                                                                     Pages:1153 – 1159
If the Matera – Santa Sofia is not working, almost all the                                        [5]       J.W.M. Cheng, D.T. McGillis, F.D.Galiana: Studies of bilateral
                                                                                                            contracts with respect to steady-state security in a deregulated
generated power is redistributed only on two lines, that may be                                             environment - IEEE Transactions on Power Systems, Volume: 13 ,
seriously overloaded.                                                                                       Issue: 3, Aug. 1998 Pages:1020 – 1025
Bulk Power System Dynamics and Control - VI, August 22-27, 2004, Cortina d’Ampezzo, Italy                                                                 195

[6]    GRTN: Individuazione zone della rete rilevante (January 2004)            Biographies
       Available on http://
[7]    GME: Analisi tecnica n°5/02. Determinazione dell’equilibrio del
                                                                                Alberto Berizzi (M'93) received his M.Sc. degree (1990) and his Ph.D. degree
       mercato del giorno prima dell’energia con prezzo uniforme per i
                                                                                (1994) both in Electrical Engineering from the Politecnico di Milano. He is
       consumatori e prezzi zonali per i generatori (October 2002). Available
                                                                                now Associate Professor at the same institution. His areas of research include
                                                                                power system analysis and voltage stability.
[8]    RTN: Regole di dispacciamento v.2.0 (December 2003). Available on
[9]    AEEG: Delibera 168/03 (December 2003) – Published on Gazzetta            Cristian Bovo (M'02) received his M.Sc. degree (1998) and his Ph.D. degree
       Ufficiale della Repubblica Italiana Supplemento Ordinario n°16           (2002) in Electrical Engineering from the Politecnico di Milano, where he is
       30/01/2004                                                               now Researcher at the Electrical Engineering Department. His areas of
[10]   Berizzi, C. Bovo, M. Delfanti, E. Fumagalli, M. Merlo: Simulation of a   research include power system analysis and optimization, and electricity
       bid-based dispatch subject to inter-zonal transmission constraints –     markets.
       2003 IEEE Powertech – Bologna, June 2003
[11]   Nord Pool Asa: The Nordic Power Market (January 2003) – Available        Maurizio Delfanti (M'99) received his M.Sc. degree (1994) and his Ph.D.
       on http.//                                                degree (1999) in Electrical Engineering from the University of Pavia. He is
[12]   AEEG: Delibera 238/00 (December 2000) – Published on Gazzetta            now Researcher at the Electrical Engineering Department of the Politecnico di
       Ufficiale della Repubblica Italiana Supplemento Ordinario n°4            Milano. His research interest is focused on the analysis of the optimal and
       05/01/2001                                                               secure operation of power systems in a market framework.
[13]   Ministero dell’Industria, del Commercio e dell’Artigianato: Decreto 25
       giugno 1999 Determinazione dell’ambito della rete elettrica di           Marco Savino Pasquadibisceglie (StM’03) received the degree in Electrical
       trasmissione nazionale – Published on Gazzetta Ufficiale della           Engineering from Politecnico di Milano (Italy) in 2002. Since 2003 he has
       Repubblica Italiana Supplemento ordinario n° 151 30/06/1999              been working for his Ph. D. in Electrical Engineering at Politecnico di Milano.
[14]   MAP: Decreto 23 dicembre 2002 Ampliamento dell’ambito della rete di      His areas of research include FACTS application and congestion management
       trasmissione nazionale dell’energia elettrica. Available on              in a deregulated system.

To top