G Model
EPSR-2965;
No. of Pages 9
ARTICLE IN PRESS
Electric Power Systems Research xxx (2009) xxx–xxx
Contents lists available at ScienceDirect
Electric Power Systems Research
journal homepage: www.elsevier.com/locate/epsr
Linking energy policy, electricity generation and transmission using strong
sustainability and co-optimization
Justin D.K. Bishop a,∗ , Gehan A.J. Amaratunga a , Cuauhtemoc Rodriguez b
a
b
University of Cambridge, Department of Engineering, 9 JJ Thomson Ave, Cambridge CB3 0FA, United Kingdom
Cambridge Consultants, United Kingdom
a r t i c l e
i n f o
Article history:
Received 6 March 2009
Received in revised form 6 September 2009
Accepted 24 October 2009
Available online xxx
Keywords:
Sustainable electricity system
Optimal power flow
Reduced transmission losses
Fuel mix diversity
a b s t r a c t
The design of a sustainable electricity generation and transmission system is based on the established
science of anthropogenic climate change and the realization that depending on imported fossil-fuels is
becoming a measure of energy insecurity of supply. A model is proposed which integrates generation
fuel mix composition, assignment of plants and optimized power flow, using Portugal as a case study.
The result of this co-optimized approach is an overall set of generator types/fuels which increases the
diversity of Portuguese electricity supply, lowers its dependency on imported fuels by 21.30% and moves
the country towards meeting its regional and international obligations of 31% energy from renewables
by 2020 and a 27% reduction in greenhouse gas emissions by 2012, respectively. The quantity and composition of power generation at each bus is specified, with particular focus on quantifying the amount
of distributed generation. Based on other works, the resultant, overall distributed capacity penetration
of 11.88% of total installed generation is expected to yield positive network benefits. Thus, the model
demonstrates that national energy policy and technical deployment can be linked through sustainability
and, moreover, that the respective goals may be mutually achieved via holistic, integrated design.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
A sustainable electricity system is one in which all stages of
the energy path are addressed, including the composition of the
generation mix, the allocation of centralized and distributed generation, and their subsequent deployment. This work aims to create
a link between a national, sustainable electricity policy and the performance of its transmission grid, recognizing that the latter may
better fit within the former if they are co-designed and optimized.
The goal of sustainable energy systems is to deliver affordable
energy services while raising the living standard for the global
population, chiefly through increased energy efficiency and deployment of renewables [1]. In particular, the latter can contribute to
mitigating the emissions of greenhouse gases (GHG), namely carbon dioxide (CO2 ), and enhancing energy security of supply and
independence.
From the beginning of the Industrial Revolution, carbon emissions have increased non-linearly per year to 38 Gt CO2 in 2004.
Three quarters of these emissions were due to human activities, of
which fossil-fuel combustion accounted for 56.60% [2]. The consequence of these emissions has been an increase in the concentration
∗ Corresponding author. Tel. +44 1223 655 406.
E-mail address: justin.bishop@cantab.net (J.D.K. Bishop).
URL: http://www.eng.cam.ac.uk (J.D.K. Bishop).
of atmospheric CO2 to the current value of 383.72 parts per million (ppm) [3], which is the highest recorded in the 650,000 years
preceding industrialization [4]. Among other effects, the mean surface temperature of the earth has been rising, with average 2007
temperatures being 0.91◦ higher than in 1907, making the former
the eighth warmest year recorded [5]. Various international agreements have been ratified to address this anthropogenic-induced
climate change. The most wide-reaching of them is the Kyoto Protocol to the United Nations Framework Convention on Climate
Change which requires Annex I countries to achieve an overall
target of at least a 5% decrease in emissions below 1990 levels
in the period 2008–2012 [6]. A number of regions and countries
have implemented national emissions policies, with examples of
the European Commission (EC) successor to Directive 2001/77/EC
which will require 20% energy from renewables by 2020 [7] and the
United States Government 2008 announcement of a halt to GHG
emissions by 2025, with sustained reduction from then.
Combining rising oil and gas prices with a recognized dependence of the developed world on foreign resources, the issue of
security of supply and energy independence is being raised more
commonly. For many countries, achieving the latter may require a
switch from imported fossil-fuels to domestic supplies of coal or the
stimulation of natural, indigenous wind, solar, geothermal, hydro
and biomass resource use. The research conducted in the field of
energy security of supply includes: deriving an underestimated cost
of 3 × 10−7 D /MWh for each barrel of oil not supplied to the market
0378-7796/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2009.10.014
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
G Model
EPSR-2965;
2
No. of Pages 9
ARTICLE IN PRESS
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
in the event of a supply disruption [8]; a cost of 4 × 10−9 D /MWh
due to a rise in energy prices [9]; the use of the Shannon–Weiner
Index (SWI)1 to assess the degree of dependence on a specific fuel
or technology type and the total system exposure to price perturbation, with a target value of at least two [10]; and metrics of net
energy import dependence and percentage of indigenous fuels used
as indicators of the energy independence [11].
Concurrent with emissions limits are the continued changes
in the domestic electricity markets in the form of deregulation.
As a result, independent power producing is becoming more
widespread and there is increased deployment of distributed generation (DG). Current drivers of DG are environmental, commercial,
and regulatory, however there are widely agreed negative network
effects of voltage rise, harmonic distortion and issues with islanded
operations [12].
A wealth of research is devoted to locating DG in order to maximize its benefits and minimize its negative grid impacts. Assigning
DG to buses with the highest locational marginal pricing has been
suggested as a method of increasing social welfare [13]. Conversely,
an analytical, weighted, multi-objective optimization proposed the
assignment of DG based on network performance impacts [14].
Additional specialized approaches have investigated deployment
of DG on feeders with particular load profiles [15]. The consensus
is that the type and location of DG, and its subsequent degree of
penetration, are critical to its benefits not being overshadowed by
decreased network performance.
Types of DG include conventional, dispatchable generators and
variable-output plants such as solar photovoltaics (PV) and wind.
While both types can provide grid support, only the renewables can
realize the environmental advantages of DG deployment. However,
to mitigate the network effects due to their fluctuating outputs, it
has been suggested to pair them with the dispatchable generation
[16].
DG is well-sited in high load areas [14], where branch congestion
[12] and power losses [17] may be reduced, particularly towards the
end of feeders or near to branch points [15,18]. However, multiple
voltage-support DGs in close proximity may work in opposition
with each other [19] requiring consideration of the spatial intensity of DG deployment. Since the presence of DG transforms the
traditional passive distribution network into, effectively, an active
transmission system [20], optimal power flow (OPF) analyses can
be used to assess the subsequent network performance once the
generator assignment has occurred [21].
In summary, the current level of carbon emissions is at odds with
binding national and international targets. Moreover, the continued
dependence on the fuels, from which the carbon emissions proceed, prolongs the risks of perturbing the energy supply. Switching
to indigenous fuels can mitigate the latter, while restructuring the
fuel mix composition to include more renewables can address both
challenges. Therefore, electricity system planners must determine
how much of each type of technology or fuel to include in their
power generation mix in order to achieve these goals. However,
given the benefits and potential drawbacks of DG, planners must
also consider where to locate their generation capacity and how
much should be centralized or distributed. This work proposes a
link between the emissions policy, fuel mix and generator assignment questions posed, using 2006 data for mainland Portugal as a
case study to illustrate the results.
1
The SWI is normally used in competition analysis, accounting for relative size
and distribution of market players. Analogously, a highly concentrated marketplace
(SWI < 1) represents a system which is dependent on one or two sources and could
be susceptible to sustained interruption. Alternatively, in a fully competitive case
(SWI > 2), there are multiple sources and consumers can be confident in continued
supply in the event of an disruption.
2. Model
The aims of the work were accomplished using a recursive optimization model, comprising the generation fuel mix, constrained
assignment of generators to buses, and OPF of subsequent network.
2.1. Portuguese transmission network
The network was modelled using the 2005 Union for the Coordination of Transmission of Electricity (UCTE) map of the 400 kV
and 220 kV transmission lines [22]. The model consisted of 57 buses,
with bus 1 being the slack bus, connected across 107 branches.
Additionally:
• generator power and fuel-type were derived from the transmission grid map of the network operator, Rede Eléctrica Nacional
(REN) [23];
• bus loads were determined using a variation of the method in
[24], where a weighted distribution of the population of each of
the five mainland administrative regions – Alentejo, Algarve, Centro, Lisboa e Vale do Tejo, and Norte– was used to assess regional
peak demand.2 For Norte and Lisboa e Vale do Tejo, the city buses
comprising Porto and Lisboa were assigned demand by population, with the remaining load in the respective regions divided
equally across the other buses;
• line characteristics were, in r/x/b format: 0.0417/0.1396/
0.0319 !/km for 400 kV and 0.0457/0.1417/0.0324 !/km for
220 kV, all at a base power of 100 MVA;
• a power factor of 0.95 was used, representing the demand seen
at each load bus by the transmission system; and
• a negative load of −1200 MW employed, representing maximum
demand from the single bus representing Spain at 1100 h on all
third Wednesdays per month in 2005 [25].
The total modelled transmission line length of 4311.43 km was
within 1% of the UCTE published value of 4355 km [25].
2.2. Power generation fuel mix optimization
A power generation fuel mix model was proposed which optimizes penetration of generator types using strong sustainability
principles as its core [26]. The objective is to achieve a high fuel
mix diversity, as a proxy for security of supply, subject to delivering the electricity at no more than the current retail price of
504.60 D /kW3 ; meeting current energy demand; reducing carbon
emissions by the quotient of the carbon footprint and forest sequestration capacity, yielding an emissions factor of 898.41 kg CO2 /kW;
and stabilizing material input per unit service (MIPS) at the current rate of 39.99 kg/kW (Table 2). The generation technologies
used in the model were wind, large hydro, geothermal, biomass,
PV, combined cycle natural gas (CCGT), coal, nuclear, and oil
(Table 1).4
2
Although peak demand is used to illustrate the co-optimized concept in this
work, it is acknowledged that a study focusing only on the peak demand can only
be simplistic by not representing the full suite of demand possibilities. However, a
peak demand analysis gives a reasonable worst case planning scenario under normal operating conditions (that is, in the absence of perturbations). Consequently,
the offpeak demand will not require any more installed generation at a particular
location, although the system may operate sub-optimally during those periods.
3
This retail price for the electricity constraint corresponds to an average residential price of electricity of 0.14 D /kWh for households consuming 3500 kWh annually,
of which 1300 kWh are at night. See Eurostat table NRG PC 204 H for details.
4
The selected nine generator fuels/types represent 93.71% of installed capacity,
with the remaining 6.29% comprising pumped hydro storage. See Eurostat table
NRG 113A for details.
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
G Model
EPSR-2965;
No. of Pages 9
ARTICLE IN PRESS
3
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Table 1
FAY for Portugal based on 2006 installed capacity and yield data from Eurostat.
Technology
Installed capacity (MW)
Annual energy yield (GWh)
FAY (kWh/kW) [32]
Wind
Hydro
Geothermal
Biomass
Solar PV (c-Si)
Gas CCGT
Coal
Nuclear
Oil
1.064
5.034
–
273
2
2.097
1.776
–
2.303
2.925
11.460
–
1.379
3
12.665
14.992
–
4.841
2.749.06
2.276.52
–
5.051.28
1.500
6.039.58
8.441.44
–
2.102.04
13.391
12.549
93.71%
51.700
48.265
93.36%
3.604.29
–
–
Total
Total represented
By selected nine proportion
The model was augmented to include:
• a space input per unit service (SIPS) constraint, recognizing that
space is a resource and therefore its use is not value neutral [27].
SIPS values were determined from the Global Emission Model for
Integrated Systems (GEMIS) database,5 using the current or projected data for plants commission in 2000 and 2010, respectively
(Table A.1). SIPS was to be stabilized at current levels of space
use.
T
!
st · xt ≤ maximum SIPS value
T
!
dispatchable generationt ≥ TDG
t=0
(1)
where xt is the fraction of technology, t in the mix; st is the SIPS
(km2 /kW) corresponding to t; and the maximum SIPS value for
the final mix is 14.45 km2 /kW.
• a requirement that the amount of dispatchable generation in the
current mix be at least preserved, if not enlarged, in the new
composition. All plants were dispatchable except for wind and
PV.
t=0
(2)
where TDG, the total proportion of dispatchable generation, is
0.86. It is comprised of hydro, biomass, oil, gas and coal.
The outputs of the augmented fuel mix routine are the new
generator composition vector and overall cost of electricity.
2.3. Assignment of generators to buses
Assigning generators to buses consisted of defining a resource
and space matrix and using a constrained RAS optimization [28],
where:
• A binary r × c resource matrix, R, was declared where r =
number of buses in the transmission system and c = generator
types. Ones indicated the resource was present and zeroes otherwise. For the nine generator types:
◦ The resource was considered always available for fossil-fuels
and biomass since the fuels could be transported to and from
any location.
◦ The solar resource was seen to yield a minimum of
1,697.25 kWh/m2 for fixed PV at 34◦ inclination relative to the
5
For information on GEMIS 4.4, see http://www.oeko.de/service/gemis/en/.
horizon [29]. Therefore, power generation via PV was deemed
feasible at all locations.
◦ The wind resource is seen to provide a minimum of 1800 annual
full load hours for a 2 MW turbine [30] across the country,
equivalent to a 20.50% capacity factor. Therefore, power generation from wind was deemed feasible on all buses.
◦ Hydro plants were only considered feasible where they were
already installed.
◦ Mainland Portugal does not posses power generation-quality
geothermal resources and has no nuclear power generation.
• A binary space matrix, S, of same size as R, was declared where
ones indicated enough space for generation, with zeroes denoting
the converse, where fossil-fuel, biomass, and hydro plants were
constrained to remain on buses to which they were already connected; wind farms were constrained to buses outside of Lisboa
and Porto; and PV could be assigned to any bus since it could be
either distributed or centralized.
• A r × c matrix Z denoted which buses in the transmission system
possessed the resources and the space to house a particular type
of generation, derived from the vector multiplication of R and S.
• The constrained RAS optimization is a matrix balancing technique
where a scale factor is applied to the elements of a m × n matrix,
alternating between rows and columns, so that their respective sums equal pre-determined values [28]. The constraints in
the algorithm were that all columns must sum to the generator
composition vector and all rows must sum to the total installed
capacity, as if it was placed on each bus. That is, the constraints
were
◦ for each row (bus), the sum of column elements (generation)
equals the total installed capacity, Gt .
∀r
c
!
1
Z(m, n) = Gt
(3)
◦ and for each column (generator type), the sum of row elements
(buses) equals the required fuel mix fraction *Gt .
∀c
r
!
1
Z(m, n) = fuel mix fraction∗ Gt
(4)
The sum of all elements per row quantified the new generation on that particular bus and constituted the generation vector to
replace the original one in the transmission model.
2.4. Optimized power flow
The standard OPF routine was run using the MATPOWER suite
[31]. 100 MVA was used as the base and voltage rise on buses limited to 1.00 ± 0.05 pu. Costs of generation were implemented using
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
G Model
EPSR-2965;
No. of Pages 9
4
ARTICLE IN PRESS
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Fig. 1. Flow graph illustrating recursive routine of fuel mix optimization, constrained generator assignment and OPF.
equal, linear relations for all generator types in the absence of specific cost generation curves and the desire not to bias the OPF by
the generation costs since those had already been accounted for in
the fuel mix routine.
2.5. Overall recursion routine
The power loss from the OPF was converted to a “grant,” which
could either be subtracted from the cost of generation or added to
the retail cost. The latter was chosen as this relaxed the cost constraint in the fuel mix optimization, permitting more renewables
into the mix, raising diversity. However, in practice, the retail cost of
electricity is unchanged from the customer perspective. Therefore,
grantk = power lossk × 1000 × generation COE
a solution when the required proportion of dispatchable penetration was lowered to 0.69. However, whereas the integrated model
of fuel mix plus RAS and OPF converged at this lower percentage,
the diversity of the mix fell from the current 1.51–1.22. Therefore, to avoid optimizing the fuel mix and transmission system
to a lower diversity, the dispatchable generation penetration was
reduced to 0.68 which preserved the diversity value of 1.51. Under
these conditions, diversity is preserved despite the presence of
only effectively six independent generator types in the final mix.
Additionally, whereas the electricity from fossil-fuels is reduced on
account of carbon emissions, that from hydro sources falls due to
the large space requirements of their water reservoirs (Table 2 and
Fig. 2). The trade-off between MIPS and SIPS (Table A.1) illustrates
(5)
where mix COE is the upper cost constraint for the model; power
loss is the difference between the power losses from the OPF at two
consecutive iterations; and the generation COE is the cost which the
model converges to, less than the mix COE.
The objective of the overall fuel mix plus RAS and OPF routine
was to minimize the difference between grant values in consecutive
iterations. That is, the recursion routine executed until either the
power loss at the k + 1 th iteration was greater than that at the k th,
or the difference between the two losses was less than 10−9 MW.
A flow graph illustrating the key data passed between modules
is illustrated in Fig. 1.
3. Results
An OPF on the Portuguese transmission system yielded network
losses of 89.65 MW for a peak demand of 8,673 MW. Compared to
published losses of 1.70% in 2003 and 1.50% in 2004 [33], the model
losses of 1.03% are in keeping with the observed trend.
The full recursion exited after two iterations. Although the fuel
mix model on its own failed to converge when the current dispatchable generation proportion of 0.86 was specified, it yielded
Table 2
Output from fuel mix optimization.
Case
Current
Proposed
Wind
Hydro
Geothermal
Biomass
Solar PV
Gas CCGT
Coal
Nuclear
Oil
0.0795
0.3759
0.0000
0.0200
0.0001
0.1566
0.1326
0.0000
0.1720
0.2126
0.3456
0.0000
0.0179
0.1124
0.2636
0.0474
0.0000
0.0006
SWI (Diversity)
COE (D /kW)
Emissions factor (kg CO2 /kW)
MIPS (kg/kW)
SIPS (m2 /kW)
1.5094
231.4939
1844.6051
39.9935
14.4536
1.5145
244.61
898.7985
39.9796
14.4554
0.8571
0.4755
0.6751
0.6885
0.0000
0.1188
Proportion of
Dispatchable plants
Electricity from indigenous
sources
Proportion of installed capacity as
distributed generation
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
G Model
EPSR-2965;
No. of Pages 9
ARTICLE IN PRESS
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Fig. 2. Current and proposed power generation fuel mix composition.
Table 3
Output of fuel-mix and transmission recursion optimization.
Iterations
Ploss (MW)
Loss relief (MW)
% Losses
0
1
2
89.6451
52.5397
52.5397
–
−37.1054
0
1.0336
0.6058
0.6058
that, although hydro has similar material inputs to conventional
generation, the total plant space requirements – on account of its
large reservoirs – are two orders of magnitude greater than conventional competitors. Therefore, when trying to preserve the current
SIPS, the proportion of hydro in the final mix is limited.6 Moreover,
the diversity of the final mix is sensitive to the number of independent generator types, requiring at least eight to pass the target of
two. Since mainland Portugal is without geothermal and nuclear
power resources, its best case diversity is 1.95.
The RAS optimization assigned wind farms to all non-city buses
in sizes of 4.70–541.90 MW throughout mainland Portugal. Since
PV was the only generator type which could be installed on city
buses, 94.96% of all PV required by the fuel mix model was located
in the urban environment as DG.
Executing an OPF, based on the new real and reactive generation
from the RAS output, yielded a power loss reduction of 37.11 MW
over the transmission network, equivalent to a 0.43% drop (Table 3).
The fuel mix composition remained unchanged when a grant of
0.77 D /kW was applied to the retail cost of electricity.
The results illustrate how a more diverse fuel mix, combined
with space and resource limitations, can yield transmission loss
relief.
The proposed increases in wind and PV are from 1,515 MW to
2,846.93 MW and from 2 MW to 1505.15 MW, respectively. The
Government of Portugal wind target is 5,100 MW by 2010 while the
European Photovoltaic Industry Association (EPIA) 2012 target for
Portugal is 50 MW. Moreover, the installed hydro capacity already
exceeds what the model requires. Therefore, while the wind target is both technically reasonable and implementable in the near
term, the proposed 1.51 GW of PV remains the most difficult goal
to achieve over time. However, Portugal should exceed the EPIA
target when the Moura 62 MW park, currently under construction,
is commissioned in 2010.
6
While it is not suggested that existing hydropower be prematurely decommissioned, the results serve to highlight that there are environmental costs associated
with all forms of power generation, including those that meet low carbon emissions
criteria. Therefore, it is worth it to Portugal to consider how it proceeds through time
with expanding its power production from hydro sources, given their large impacts.
5
Fig. 3. Carbon emissions (MT CO2 ) from current and proposed fuel mixes relative
to the Kyoto Protocol and the EU Burden Sharing Agreement for Portugal.
The Government of Portugal has as international target to reduce
its carbon emissions by 27% below 1990 levels as part of the Kyoto
Protocol and the European Union Burden Sharing Agreement. With
the proposed fuel mix, and assuming no carbon reductions in other
emitting industries, the effects of the new mix are insufficient
to meet the Kyoto Target (Fig. 3). Therefore, other emitters have
a clear, collective reduction target in order to achieve the 2012
goals.
Portugal currently generates 32.64% of its electricity from
renewables which is short of its 2001/77/EC requirement of 39%
by 2010 [34]. An expected 2008 EC Directive will require Portugal
to meet 31% of its energy from renewables by 2020. Deploying the
new fuel mix composition leads to 44.89% electricity from renewables, meeting the 2001/77/EC. Moreover, the new mix provides
10.09% of energy from renewables. This represents an increase
from 7.30% currently and brings Portugal closer to meeting its
obligation under the 2020 requirement of the new EC Directive
(Fig. 4).
By reducing the amount of fossil-fueled generation, the gross
inland consumption (GIC) is reduced by 1.97% (Fig. 5). This is largely
attributed to a reduction in coal imports since the optimized fuel
mix reduces the amount of coal-fired generation by 23.71%, while
increasing that generation from gas by 20.44%.
As a result of the distribution of generators via the RAS algorithm, almost all of the proposed PV fraction will be located in
the Lisboa and Porto urban centres. Relative to the cities’ respective demands, the DG PV is expected to meet 45.99% and 20.02%
Fig. 4. Proportion of total energy (ktoe) from renewables for current and proposed
fuel mixes relative to the 2008 EC RES Directive [7] requiring Portugal to attain 31%
energy from renewables by 2020.
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
ARTICLE IN PRESS
G Model
EPSR-2965;
No. of Pages 9
6
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Table 4
Allocation of DG PV between Lisboa and Porto.
City
Lisboa
Porto
Demand (MW)
2250.99
1110.35
DG cap. (MW)
% load met by DG
DG per household (kW)
1035.30
222.30
45.9932
20.0207
0.6573
1.5097
ments. In particular, new generation ought to maintain or improve
system adequacy and reduce congestion, towards maintaining a
stable system. The central planning approach advanced in this work
prescribes where and how much each type of generation ought to
be deployed in order to maximize transmission system efficiency.
As such, it is neither at odds with a market driven nor a vertically integrated utility service as such guidance may be present
nonetheless.
4. Conclusion
Fig. 5. Gross inland consumption (ktoe) for solid fuels, oil and gas under current
and proposed fuel mixes.
of the load (Table 4). The impacts of distributed PV have been
studied independently and over two cases in different countries.
It was concluded that, up to 30% penetration, any voltage rises
are within network tolerances [35,36]. A comparative study which
investigated large-scale PV deployment in Lisboa observed that net
positive voltage and network loss impacts accrue for all distribution
network types up to 1 kW PV per household [37]. The model results
indicate that for Porto, per household capacity installed capacity is
within the 1 kW of the above studies, at 0.66 kW/household. Conversely, Lisboa has 1.51 kW installed per household, representing a
27.48% DG penetration (Table 4). However, non-residential properties are not included in this study. Therefore, spreading some of
the excess generation or high per household capacity in the two
cities onto commercial properties may allow the results of previous
works to be realized.
Recalling total transmission and distribution losses of 9.50%
[33], the routine and results presented have focused on the less
lossy portion of the electricity transportation network. The strength
of the routine could be demonstrated further by modelling the distribution network as the mini-transmission network, highlighting
the positive impacts of DG.
Since Portugal is unable to support geothermal power generation, and will likely not deploy nuclear power plants, their
power needs can only be met by a less diverse fuel mix, though
greatly improved over their current generation regime. However,
only one DG technology is considered in this model. Further work
would require costing the use of other DG technologies, such
as household/neighbourhood-sized micro-turbines and small to
medium wind turbines, for example. The benefits of additional
technologies under consideration is that they would qualify as independent generator types, contribute to overall system diversity
and, particularly for the dispatchable plants, smooth the variations
of the non-dispatchable renewables to which they might be paired
[16].
Finally, this work has presented a central planning approach
to locating new generating plant. In deregulated markets, generation companies are responsible for investment in and deployment
of such facilities. Notwithstanding which party is responsible for
expanding generation, the location of new plant is subject to the
constraints of the transmission system and regulatory require-
Currently, many countries of the world are facing a dilemma
of increasing carbon emissions to the atmosphere, in the face of
legally binding targets to reduce pollution by GHGs. This work
advances the hypothesis that, in order to achieve an electricity system which performs well and meets the national targets,
the generation and transportation means must be evaluated
and co-optimized. This is accomplished by integrating three
specialist, focused, but normally disparate, aspects of power system planning. The advantage of such an approach is that the
objective of each sub-routine must be satisfied before the next
algorithm may start. This ensures that subsequent optimization is constrained by those preceding it and all objectives are
met.
In the case of Portugal, the overall model illustrates how a
diverse fuel mix can meet energy independence and security of
supply requirements, while working towards carbon emissions
targets, as dictated by the United Nations and European Union.
Specifically, the more diverse fuel mix meets the current Directive 2001/77/EC and moves Portugal appreciably towards satisfying
the new 2008 EC Directive on the promotion of the use of energy
from renewable sources. Kyoto Protocol targets remain elusive
as the emissions from electricity are generally small relative to
other industries. The percentage of indigenous generation rises by
21.30% and the gross inland consumption of fossil-fuels falls by
1.97%.
Assigning generation to particular buses in the transmission
network, based on resource and space allowances, incorporates
the practical questions of electricity generation, while quantifying
the amount which can be sustained in a distributed arrangement.
The results show that 11.88% of total installed capacity should be
located in the urban centres of Lisboa and Porta. The per-city penetration, relative to the demand of each location, should respect
established voltage limitations.
The current model however does not incorporate the distribution network explicitly, which is where most of the losses in
transporting power occur, while the consideration of PV as the
only DG technology results in the entire distributed capacity being
variable-output and time-of-day dependent. Thus, scope remains
to re-model the network with more DG technologies on an explicit
distribution network in order to illustrate the full potential of the
co-optimized approach.
Appendix A.
Tables A.1–A.6.
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
ARTICLE IN PRESS
G Model
EPSR-2965;
No. of Pages 9
7
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Table A.1
Material and space inputs per unit service (kW).
Technology
SIPS (m2 /kW)
MIPS (kg/kW)
Wind
Hydro
Geothermal
Biomass
Solar PV
Gas CCGT
Coal
Nuclear
Oil
2
37.8887
1
0.10
8.0968
0.0394
0.2857
0.1743
0.0311
34.8750
60.2308
37.50
465.3130
6.6667
6.5152
20
47.2403
9.20
Table A.3
Branch data for Portuguese transmission network model (2 of 2) using 100 MVA
base.
From bus
To bus
r (p.u.)
x (p.u.)
b (p.u.)
42
43
44
44
45
45
46
46
49
49
50
50
51
52
53
53
53
53
53
53
54
54
55
55
55
55
55
56
56
57
11
11
4
12
26
26
19
35
31
31
18
40
37
53
7
10
10
33
33
37
23
53
14
14
37
37
53
49
49
42
0.0029
0.0056
0.0007
0.0009
0.0003
0.0003
0.0014
0.0020
0.0039
0.0039
0.0026
0.0064
0.0011
0.0034
0.0032
0.0025
0.0025
0.0053
0.0053
0.0031
0.0046
0.0031
0.0011
0.0009
0.0021
0.0021
0.0047
0.0015
0.0015
0.0029
0.0094
0.0181
0.0022
0.0029
0.0011
0.0011
0.0050
0.0069
0.0127
0.0127
0.0083
0.0207
0.0036
0.0109
0.0105
0.0080
0.0080
0.0170
0.0170
0.0101
0.0149
0.0101
0.0036
0.0029
0.0069
0.0069
0.0152
0.0047
0.0047
0.0094
0.0022
0.0041
0.0005
0.0007
0.0013
0.0013
0.0060
0.0082
0.0029
0.0029
0.0019
0.0047
0.0008
0.0025
0.0024
0.0018
0.0018
0.0039
0.0039
0.0023
0.0034
0.0023
0.0008
0.0007
0.0016
0.0016
0.0035
0.0011
0.0011
0.0022
Table A.2
Branch data for Portuguese transmission network model (1 of 2) using 100 MVA base
From bus
To bus
r (p.u.)
it x (p.u.)
b (p.u.)
From bus
To bus
r (p.u.)
x (p.u.)
b (p.u.)
1
1
2
2
3
3
3
4
4
4
5
5
5
5
6
6
7
8
8
9
9
11
12
12
13
13
13
14
15
15
15
15
17
17
18
18
18
19
20
31
31
19
48
28
38
48
12
18
50
29
39
40
40
32
48
27
46
46
15
51
4
18
18
19
47
56
36
10
10
27
31
16
48
11
11
11
56
40
0.0031
0.0031
0.0011
0.0013
0.0008
0.0013
0.0007
0.0015
0.0015
0.0012
0.0015
0.0015
0.0030
0.0030
0.0009
0.0035
0.0046
0.0003
0.0003
0.0019
0.0006
0.0029
0.0009
0.0010
0.0034
0.0012
0.0036
0.0011
0.0036
0.0036
0.0023
0.0054
0.0080
0.0003
0.0016
0.0016
0.0016
0.0025
0.0016
0.0101
0.0101
0.0038
0.0047
0.0029
0.0044
0.0025
0.0047
0.0047
0.0040
0.0051
0.0051
0.0098
0.0098
0.0029
0.0112
0.0149
0.0012
0.0012
0.0062
0.0018
0.0094
0.0029
0.0033
0.0109
0.0040
0.0116
0.0036
0.0116
0.0116
0.0076
0.0174
0.0261
0.0012
0.0051
0.0051
0.0051
0.080
0.0055
0.0023
0.0023
0.0045
0.0056
0.0034
0.0053
0.0030
0.0011
0.0011
0.0009
0.0061
0.0061
0.0022
0.0022
0.0007
0.0026
0.0034
0.0015
0.0015
0.0014
0.0004
0.0022
0.0007
0.0007
0.0025
0.0009
0.0027
0.0008
0.0027
0.0027
0.0017
0.0040
0.0060
0.0015
0.0012
0.0012
0.0012
0.0018
0.0066
21
22
22
23
23
24
24
25
26
26
26
26
27
27
28
29
29
30
30
31
31
31
33
33
33
33
33
34
37
37
37
38
38
39
39
40
40
40
41
23
32
32
32
31
1
31
49
18
39
46
46
5
23
3
17
40
57
57
5
30
34
13
13
32
48
48
5
9
20
27
37
37
4
18
4
11
11
18
0.0026
0.0010
0.0010
0.0026
0.0023
0.0009
0.0025
0.0017
0.0014
0.0017
0.0024
0.0024
0.0021
0.0008
0.0004
0.0013
0.0016
0.0031
0.0031
0.0047
0.0022
0.0025
0.0044
0.0044
0.0054
0.0032
0.0027
0.0023
0.0019
0.0029
0.0016
0.0009
0.0009
0.0009
0.0005
0.0016
0.0030
0.0030
0.0010
0.0083
0.0033
0.0033
0.0083
0.0076
0.0029
0.0080
0.0054
0.0050
0.0060
0.0085
0.0085
0.0073
0.0025
0.0014
0.0045
0.0057
0.0101
0.0101
0.0152
0.0072
0.0080
0.0141
0.0141
0.0174
0.0105
0.0087
0.0076
0.0062
0.0100
0.0054
0.0031
0.0031
0.0032
0.0017
0.0054
0.0098
0.0098
0.0033
0.0019
0.0007
0.0007
0.0019
0.0017
0.0007
0.0018
0.0012
0.0060
0.0072
0.0102
0.0102
0.0087
0.0006
0.0017
0.0054
0.0068
0.0023
0.0023
0.0035
0.0017
0.0018
0.0032
0.0032
0.0040
0.0024
0.0020
0.0017
0.0014
0.0119
0.0065
0.0037
0.0037
0.0038
0.0020
0.0065
0.0022
0.0022
0.0007
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
G Model
EPSR-2965;
No. of Pages 9
8
ARTICLE IN PRESS
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Table A.4
Bus data for Portuguese transmission network model (1 of 2).
Bus no.
Bus type
pd (MW)
qd (MVAR)
Base (kV)
Vmax (p.u.)
Vmin (p.u.)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
3
2
2
1
1
2
1
2
1
2
1
1
1
1
1
1
2
1
1
1
2
1
1
1
2
1
1
2
1
2
1
2
2
1
1
1
1
1
141.29
0
0
204.64
141.29
0
141.29
0
277.59
0
261.91
204.64
141.29
431.35
141.29
448.16
0
204.64
204.64
141.29
0
277.59
431.35
141.29
0
204.64
141.29
0
204.64
0
141.29
0
0
141.29
366.83
277.59
277.59
431.35
68.43
0
0
99.11
68.43
0
68.43
0
134.44
0
126.85
99.11
68.43
208.91
68.43
217.05
0
99.11
99.11
68.43
0
134.44
208.91
68.43
0
99.11
68.43
0
99.11
0
68.43
0
0
68.43
177.66
134.44
134.44
208.91
220
400
400
220
400
220
400
400
220
220
220
220
220
220
220
400
400
220
400
400
220
220
220
220
220
400
400
220
400
220
220
220
220
220
400
220
220
400
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
Table A.5
Bus data for Portuguese transmission network model (2 of 2).
Bus no.
Bus type
pd MW
qd MVAR
Base (kV)
Vmax (p.u.)
Vmin (p.u.)
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
2
1
1
1
1
1
1
1
2
1
2
2
2
2
1
2
2
1
1
0
261.91
204.64
261.91
204.64
204.64
204.64
204.64
0
−1200
0
0
0
0
431.35
0
0
431.35
141.29
0
126.85
99.11
126.85
99.11
99.11
99.11
99.11
0
−581.19
0
0
0
0
208.91
0
0
208.91
68.43
220
400
220
220
220
220
400
400
220
400
220
220
220
220
220
220
220
220
220
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
1.05
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
.95
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014
ARTICLE IN PRESS
G Model
EPSR-2965;
No. of Pages 9
9
J.D.K. Bishop et al. / Electric Power Systems Research xxx (2009) xxx–xxx
Table A.6
Generator data for Portuguese transmission network model, with all buses at 100 MVA and status 1.
Bus no.
pg
qg
qmax
1
2
3
6
8
9
10
11
13
17
22
24
25
28
29
32
33
35
39
43
45
51
53
54
57
336
236
652
240
1278
117
341
710
40
254
369
24
1515
629
584
195
186
197
1170
56
946
990
238
240
311
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
235.2
165.2
456.4
168
894.6
81.9
238.7
497
28
177.8
258.3
16.8
1060.5
440.3
408.8
136.5
130.2
137.9
819
39.2
662.2
693
166.6
168
217.7
References
[1] P. Hennicke, S. Thomas, W. Irrek, Towards Sustainable Energy Systems: Integrating Renewable Energy and Energy Efficiency is the Key, Wupptertal
Institute for Climate, Environment and Energy, May 2004.
[2] IPCC, Climate change 2007: Synthesis report, Tech. rep., Intergovernmental
Panel on Climate Change, 2007, http://www.ipcc.ch/ipccreports/ar4-syr.htm.
[3] P. Tans, Trends in Atmospheric Carbon Dioxide-Mauna Loa. www.esrl.
noaa.gov/gmd/ccgg/trends/.
[4] U. Siegenthaler, T.F. Stocker, E. Monnin, D. Lüthi, J. Schwander, B. Stauffer, D.
Raynaud, J.-M. Barnola, H. Fischer, V. Masson-Delmotte, J. Jouzel, Stable carbon cycle–climate relationship during the late pleistocene, Science 310 (5752)
(2005) 1313–1317.
[5] P. Brohan, J.J. Kennedy, I. Harris, S.F.B. Tett, P.D. Jones, Uncertainty estimates in
regional and global observed temperature changes: a new dataset from 1850,
Journal of Geophysical Research 111 (D12106).
[6] UN, Kyoto Protocol to the United Nations Framework Convention on Climate
Change, United Nations, 1997.
[7] Proposal for a Directive of the European Parliament and of the Council on the
promotion of the use of energy from renewable sources, Commission of the
European Communities, 2008.
[8] A. Markandya, A. Hunt, ExternE-Pol Externalities of Energy: Extension of
Accounting Framework and Policy Applications, final edition, November 2004.
[9] S. Arnold, A. Hunt, WP5-Security and reliability of supply, University of Bath,
April 2007.
[10] M. Grubb, L. Butler, P. Twomey, Diversity and security in UK electricity generation, Energy Policy 34 (18) (2006) 4050–4062.
[11] IAEA, Energy Indicators for Sustainable Development: Guidelines and Methodologies, International Atomic Energy Agency, April 2005.
[12] J.A. Peças Lopes, N. Hatziargyrou, J. Mutale, P. Djapic, N. Jenkins, Integrating
distributed generation into electric power systems: A review of drivers, challenges and opportunities, Electricity Power Systems Research 77 (9) (2007)
1189–1203.
[13] D. Gautam, N. Mithulananthan, Optimal DG placement in deregulated electricity market, Electricity Power Systems Research 77 (12) (2007) 1627–
1636.
[14] L.F. Ochoa, A. Padilha-Feltrin, G.P. Harrison, Evaluating distributed generation
impacts with a multiobjective index, IEEE Transactions on Power Delivery 21
(3) (2006) 1452–1458.
[15] C. Wang, M.H. Nehrir, Analytical approaches for optimal placement of distributed generation sources in power systems, IEEE Transactions on Power
Systems 19 (4) (2004) 2068–2076.
[16] M. Ilic, J.W. Black, M. Prica, Distributed electric power systems of the future:
Institutional and technological drivers for near-optimal performance, Electricity Power Systems Research 77 (9) (2007) 1160–1177.
[17] D. Lima, A. Padilha-Feltrin, Allocation of the costs of transmission losses, Electricity Power Systems Research 72 (1) (2004) 13–20.
qmin
−235.2
−165.2
−456.4
−168
−894.6
−81.9
−238.7
−497
−28
−177.8
−258.3
−16.8
−1060.5
−440.3
−408.8
−136.5
−130.2
−137.9
−819
−39.2
−662.2
−693
−166.6
−168
−217.7
pmax
336
236
652
240
1278
117
341
710
40
254
369
24
1515
629
584
195
186
197
1170
56
946
990
238
240
311
pg min
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
[18] G.P. Harrison, A. Piccolo, P. Siano, A.R. Wallace, Distributed generation capacity
evaluation using combined genetic algorithm and OPF, International Journal of
Emerging Power Systems (2) 8.
[19] M.A. Rashem, G. Ledwich, Multiple distributed generators for distribution
feeder voltage support, IEEE Transactions on Energy Conversion 20 (3) (2005)
676–684.
[20] D. Popović, J. Greatbanks, M. Begović, A. Pregelj, Placement of distributed
generators and reclosers for distribution network security and reliability, International Journal of Electrical Power and Energy Systems 27 (5–6) (2005)
398–408.
[21] W. Rosehart, E. Nowicki, Optimal placement of DG, No. 14, PSCC, 2002.
[22] UCTE, UCTE Map. http://www.ucte.org/publications/uctemap/.
[23] REN, Rede Nacional de Transporte. http://www.ren.pt:8080/images/maparen.
gif.
[24] Q. Zhou, J.W. Bialek, Approximate model of european interconnected system as
a benchmark system to study effects of cross-Border trades, IEEE Transactions
on Power Systems 20 (2) (2005) 782.
[25] UCTE, Statistical Yearbook 2005, UCTE, 2006.
[26] J.D.K. Bishop, G.A.J. Amaratunga, C. Rodriguez, Using strong sustainability to
optimize electricity generation fuel mixes, Energy Policy 36 (3) (2008) 971–980.
[27] F. Schmidt-Bleek, The Fossil Makers, Factor 10 Institute, 1993, Ch. 4, 100–108.
[28] M.H. Schneider, S.A. Zenios, A comparative study of algorithms for matrix balancing, Operations Research 38 (3) (1990) 439–455.
[29] PVGIS, EU Photovoltaic Geographic Information System. http://re.jrc.
ec.europa.eu/pvgis.
[30] P. Costa, P. Miranda, A. Estanqueiro, Development and validation of the Portuguese Wind Atlas, in: European Wind Energy Conference, 2006.
[31] R.D. Zimmerman, C.E. Murillo-Sánchez, D. Gan, A MATLAB Power System Simulation Package. http://www.pserc.cornell.edu/matpower/.
[32] L. Moore, H. Post, H. Hayden, S. Canada, D. Narang, Photovoltaic power plant
experience at Arizona Public Service: a 5-year assessment, Progress in Photovoltaics: Research and Applications 13 (4) (2005) 353–363.
[33] L. Izquierdo, H. Cabal, R. Sáez, N. Caldés, EUSUSTEL WP1: country-wise analysis
for EU-25, Portugal, Tech. rep., EUSUSTEL (2005).
[34] EC, Directive 2001/77/EC of the European Parliament and of the Council of
27 September 2001 on the promotion of electricity produced from renewable energy sources in the internal electricity market, Official Journal of the
European Communities (2001).
[35] C. da Silva Jardim, R. Rüther, I.T. Salamoni, T. de Souza Viana, S.H. Rebechi,
P.J. Knob, The strategic siting and the roofing area requirements of buildingintegrated photovoltaic solar energy generators in urban areas in Brazil, Energy
and Buildings 40 (3) (2008) 365–370.
[36] M. Thomson, D.G. Infield, Network power-flow analysis for a high penetration
of distributed generation, IEEE Transactions on Power Systems 22 (3) (2007)
1157.
[37] J. Paatero, P. Lund, Effects of large-scale photovoltaic power integration on
electricity distribution networks, Renewable Energy 32 (2) (2007) 216–234.
Please cite this article in press as: J.D.K. Bishop, et al., Linking energy policy, electricity generation and transmission using strong sustainability
and co-optimization, Electr. Power Syst. Res. (2009), doi:10.1016/j.epsr.2009.10.014