Solar Energy 217 (2021) 173–186

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Network impact of increasing distributed PV hosting: A utility-scale

case study
Gabriel Tévar-Bartolomé b, Antonio Gómez-Expósito a, *, Angel Arcos-Vargas a,
Manuel Rodríguez-Montañés b
a
University of Seville, Spain
b
Endesa, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: This paper presents a methodology to assess the cost of upgrading the LV network so that it can host increasing
Distributed energy levels of PV distributed generation. The cost is valued in terms of investment, operation and maintenance and
PV losses. A real scale case study (80,000 customers from urban and rural areas, 460 GWh of demand and 3,600 km
Grid congestion
of MV&LV lines/cables) is thoroughly analysed, using the actual network topology along with the hourly demand
Overvoltage
of each supply point, provided by smart meters. With the help of a planning-oriented load flow solver, technical
Investment
Grid losses violations (congestions and overvoltages) are quantified and classified for different PV penetration scenarios.
Then, a heuristic procedure is implemented aimed at identifying the most cost-effective investment solution that
sequentially eliminates the previously ranked excessive congestions, assuming that mild congestions and other
operational problems, can be addressed without extra investments, either through network switching practices or
through the implementation of local flexibility markets. Specifically, it has been found that the direct costs
(investment and O&M) arising from the massive deployment of PV generation amount to 2.73 €/kWp and 10.18
€/kWp for scenarios with 30% and 50% PV penetration levels, respectively. No additional investments needed in
digitization, monitoring, automation and systems have been considered.

1. Introduction energy-cost savings (Singh and Banerjee, 2017). In turn, this trend is
reducing the costs of distributed PV significantly (Sussams and Leaton,
One of the most important energy and environmental objectives of 2017).
the European Union (Proposal for a Directive of the European Parlia­ In this context, the role of distribution networks is of paramount
ment and of the Council on the promotion of the use of energy from importance in facilitating the integration of these new distributed re­
renewable sources (recast), 2016; Deloitte Monitor, 2050) is to achieve sources, while at the same time ensuring that current quality and con­
greater penetration of distributed electric generation, especially with tinuity of supply indices are kept intact (Deloitte Monitor, 2018).
photovoltaic (PV) technology. Some countries like Germany have This new scenario completely changes and complicates the way
significantly promoted this kind of thecnologies by implementing feed- distribution networks are operated, since they turn from “distribution”
in-tariff policies (Heesen et al., 2019). Other important objetives are the to “contribution” networks. As analysed by Haque and Wolfs in (Haque
promotion of electric mobility, increased energy efficiency and a greater and Wolfs, 2016), Barker and De Mello in (Barker and De Mello, 2000)
participation of consumers in demand response. All these levers will and Karimi et al. in (Karimi et al., 2016), dispersed PV generation gives
boost the objective of decarbonizing the economy and thus reducing rise to voltage fluctuations, imbalances, harmonics, reverse power flows,
CO2 emission levels. Although large PV power plants are more efficient, obsolescence of the protection systems etc. Moreover, beyond certain PV
due to the existence of significant economies of scale, other factors are penetration levels, it may lead to increasing losses, congestion of
driving individuals to install small PV facilities on their rooftops (Arcos, network elements and overvoltage at different supply points.
2016). Indeed, a large-scale rooftop PV system, integrated along with a In a previous work (Tévar et al. (Tévar et al., 2019), the authors have
conventional (hydrothermal) generation system can achieve great evaluated on a case study the impact of different rooftop PV penetration

* Corresponding author.
E-mail address: age@us.es (A. Gómez-Expósito).

https://doi.org/10.1016/j.solener.2021.01.066
Received 28 February 2020; Received in revised form 20 October 2020; Accepted 28 January 2021
Available online 19 February 2021
0038-092X/© 2021 International Solar Energy Society. Published by Elsevier Ltd. All rights reserved.
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

levels on network losses and congestions (ampacity and overvoltages). models to verify that losses can be reduced if generation facilities are
This was performed under the “constant network” hypothesis which, installed in some locations for certain penetration levels. Mandis et al.
being useful to draw certain conclusions (e.g. how network losses evolve (Mandiş Al et al., 2014) analyse the potential for improving reliability
with increasing PV power), does not seem to be acceptable if very high and reducing voltage drops, proposing an algorithm that identifies the
penetration levels are considered. Indeed, when the amount of rooftop connection points and power that optimize the overall system. Yang
PV exceeds a certain threshold, the underlying distribution system has to et al. (Quan et al., 2018) develop a stochastic power flow study, using the
be upgraded to a larger or lesser extent. standard IEEE 39 bus testing system and trhee case studies (8, 16 and
The primary objective of this paper is to evaluate the impact that 24% PV penetration), to conclude, among other things, that high
increasing levels of distributed PV generation have on the distribution penetration of renewable energy sources has a negative effect on
system, when it is reinforced to eliminate excessive technical constraints ampacities and overvoltages, and also for system power losses if gen­
violations. The assessment includes the evolution of losses with respect eration is far from load centers.
to the base case scenario (no PV) as well as the investment and operation (c) A different set of studies analyse the impact of renewable gen­
and maintenance costs (O&M) associated with the involved network eration on the low voltage network, using real demand and network
upgrading, aimed at maintaining the current level of service quality. The information. In this respect, it is relevant to highlight the contribution of
scope of this paper considers a long-term network planning perspective, Chiandone et al. (Chiandone et al., 2014) that verifies the positive
assuming that the DSO can resort to suitable operational tools in order to impact of PV generation on losses, at least for low penetration levels,
enforce customary network constraints. Therefore, the issues related to achieving up to 20% reduction in losses when the location of the gen­
short-term network operation and individual asset assessment are eration is optimized. Other works that use actual LV networks across
beyond the scope of this paper. A real and sufficiently representative extensive distribution areas are those by Cohen and Callaway (Cohen
distribution network, incorporating actual consumption and PV gener­ and Callaway, 2015; Cohen and Callaway, 2016), Westacott and Can­
ation data, is used as case study, thus complementing the previous work delise (Westacott and Candelise, 2016), Eguía et al. (Eguía et al., 2016)
(Tévar et al., 2019). In this work, the existing Spanish regulation has and García-Villalobos et al. (Garcia-Villalobos et al., 2017; García-Vil­
been taken as a reference (Decree, 1699; Decree, 2019). lalobos et al., 2017). Some of those works use very detailed generation
Several works have previously addressed this issue. In addition to the profiles, allowing for instance to assess the effect of passing clouds
aforementioned studies (Haque and Wolfs, 2016; Barker and De Mello, (Cohen and Callaway, 2016) or voltage fluctuations. Using a more
2000; Karimi et al., 2016), analysing the most significant effects on detailed demand model (hourly curves for eight typical days of the
voltage fluctuations, imbalances, reverse power flows, etc., other related year), Eguía et al. (Eguía et al., 2016) and García-Villalobos et al. (Gar­
works reviewed in the sequel deal with the following topics: cia-Villalobos et al., 2017; García-Villalobos et al., 2017) estimate losses
(a) Simplified high-voltage (HV) or medium-voltage (MV) models: by performing a full-scale analysis for a real LV network, extrapolating
works that resort to HV-MV case studies, with simplified demand the results to the rest of the market. The conclusions reinforce the pre­
models, to assess the impact on losses and other factors (Quezada et al., vious findings, namely that losses decrease for low concentrations of PV
2006; Bastiao et al., 2008; Wang and Lan, 2011; Morvaj et al., 2016). generation, but then increase significantly for higher levels, especially if
(b) Assessment of losses for different PV penetration levels, by PV production is concentrated.
using different network case studies (Borges and Falcäo, 2003; Osorio (d) Concerning the influence of PV generation on network operation,
et al., 2012; Mandiş Al et al., 2014; Quan et al., 2018). several works focus on the possible dynamic or operational effects that
(c) Analysis of the impact of PV penetration on real LV networks intermittent PV distributed generation can cause on the network,
(Chiandone et al., 2014; Cohen and Callaway, 2015; Cohen and Call­ including losses, as already discussed for previous works. For example,
away, 2016; Westacott and Candelise, 2016; Eguía et al., 2016; Garcia- Nguyen et al. (Nguyen et al., 2016) use a very detailed generation model
Villalobos et al., 2017; García-Villalobos et al., 2017). with solar radiation data obtained every 30 s by means of a sky imaging
(d) Research on the impact of PV penetration on the operation of system, which enables them to simulate the passage of clouds over areas
the network (Nguyen et al., 2016; Ebad and Grady, 2016; Rodriguez- with real networks of a certain size. In a nutshell, they conclude that a
Calvo et al., 2017; Kreuwel et al., 2020; Gastalver-Rubio et al., 2019). uniform generation model for an entire area tends to overestimate the
(e) Analysis of voltage and congestion violations for different PV amount of losses. Ebad and Grady (Ebad and Grady, 2016) model gen­
scenarios (Bottacciolia et al., 2017; Mateo et al., 2018). eration ramps that are comparable to clouds passing across broad areas,
(f) Assessment of the investment needs related to the increase in with production voltage drops of up to 70% in<10 s. They conclude that
PV generation (Xiang et al., 2016; Pudjianto et al., 2013; Schmalensee, large centralised PV plants can produce greater network voltage insta­
2015; Cohen et al., 2016; Gomes and Saraiva, 2017; Freitas et al., 2018). bility. From a different perspective, Rodríguez-Calvo et al. (Rodriguez-
(a) Simplified HV and MV models, such as those of Quezada et al. Calvo et al., 2017), by performing load flows by phases and considering
(Quezada et al., 2006), Bastiao et al. (Bastiao et al., 2008) and Wang and different connection scenarios for PV generation, conclude that LV
Lan (Wang and Lan, 2011), study the impact on losses caused by the network losses for an unbalanced system can be from 2.5 to 3.8 times
penetration of different renewable generation technologies, using higher than losses in a balanced system. Kreuwel et al. (Kreuwel et al.,
mainly case studies with simplified demand models. Those works 2020) use different direct measurements of high frequency fluctuations
already anticipate the importance of the specific location of PV facilities, of the PV power to demonstrate that, in short time intervals (some
as far as the resulting losses and the amount of energy that can be in­ seconds), short-term peaks or ramps of PV power can occur much higher
tegrated are concerned. A broader view is provided by Bastiao et al. than those traditionally calculated in long-term planning scenarios
(Bastiao et al., 2008), who examine how distributed PV generation can (hours on days of maximum radiation with clear skies), exceeding
affect the stability of the grid, the voltage variations, the useful life of voltage limits in short intervals of time and threatening the stability of
transformers and the protection system, among other topics. Morvaj et the network. Gastalver-Rubio et al. (Gastalver-Rubio et al., 2019) show
al. (Morvaj et al., 2016) use a simplified demand model with three how using the capability of today’s PV inverters to individually inject
residential buildings and two single-family dwellings to verify how, active and reactive power in each phase, network operators could
through both demand management and storage (hot water storage), reduce to a large extent some of the operational problems mentioned
greater levels of PV can be integrated into the grid, without the need for above, such as phase unbalances.
additional investment. (e) Regarding the analysis of voltage and ampacity constraints
(b) Regarding the influence of PV in network operation, the works by violation, in addition to the authors’ previous work (Tévar et al., 2019),
Borges and Falcäo (Borges and Falcäo, 2003) and Osorio et al. (Osorio other relatively recent works explore the capacity of the existing
et al., 2012) can be highlighted. They use different network and demand network to accommodate different levels of distributed generation

174
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

penetration. Bottacciolia et al. (Bottacciolia et al., 2017) use a PV scenarios are considered: a base case in which all available rooftop
generator simulator that incorporates digitized information of land surfaces, according to cartographic data and considering shadows, are
surfaces, orientations and shadows from an urban area of Milan, which covered with PV panels, and a second case in which well-oriented facade
is integrated into a real-time network simulator to analyse voltage surfaces are also included. The conclusion, albeit not quantified, is that
changes that would arise on different days of the year with different the capacity of current secondary substations may not be sufficient in
levels of PV penetration. One of their main conclusions is that the cur­ case all the PV potential for the urban area is materialized.
rent grid may not be prepared to incorporate all the PV generation that This paper completes the previous ones, by providing a more realistic
could potentially be installed, as this could lead to operational con­ and comprehensive quantitative assessment of the impact that the
straints violations. Mateo et al. (Mateo et al., 2018) evaluate the over­ different degrees of distributed PV penetration would have on the LV
voltages and congestion violations that take place in MV and LV distribution network. Instead of using theoretical models, the assess­
networks for different PV penetration scenarios, but they do not evaluate ment is performed with an actual distribution network that supplies over
possible solutions for those violations. To simulate the network behav­ 80,000 consumers, each located according to its actual geographic co­
iour they use a Reference Network Model (RNM), capable of building ordinates. Smart meters and solar irradiation data are used to carry out
from scratch idealized or theoretical MV and LV networks that serve real an energy balance (consumption minus generation) for each LV con­
distribution areas, and use hourly consumption curves to calculate load sumer over the 8,760 h of the year. Then, the impact of the PV gener­
flows for the day of the year with the highest solar radiation. Similar ation is quantified in detail, in terms of technical losses variations,
works, such as the one by EPRI (EPRI. Integration of hosting capacity network upgrading and expansion investments and the associated O&M
analysis into distribution planning tools. January 2016) and Wang and costs. For this purpose, beyond the methodology presented in (Tévar
Tisdale (Wang and Tisdale, 2018), develop methodologies to identify et al., 2019), a heuristic procedure has been developed to sort out the
network locations where the connection of a certain capacity of PV alternative upgrading solutions which can solve the unacceptable vio­
generation could be integrated without leading to voltage violations or lations detected for each PV penetration scenario. In order to assess and
congestions. rank the severity of technical constraints violations, a RNM has been
(f) The last category of studies deals with investment evaluation adopted (García-Conejo et al., 2007), mimicking the one used by the
models, derived from the potential network reinforcements that make Spanish Regulator (CNMC). In fact, as the results have been obtained
PV penetration possible. Xiang and others (Xiang et al., 2016) raise the with a large enough representative network, using actual hourly de­
need to modify the current planning criteria, both static and dynamic, in mand and PV, the main conclusions reached in this study should be of
order to accommodate this new type of generation, including tools such interest also for policy makers.
as demand management. Pudjianto et al. (Pudjianto et al., 2013) The rest of this paper is organized as follows: section 2 describes the
parameterize the PV generation effect through 15 more sophisticated notation, methodology and main procedures underpinning this study, as
theoretical models, quantifying the network improvements needed to well as the hypotheses adopted to construct different scenarios for the
assume different penetration levels. In the case of Spain, they estimate a grid, PV generation and demand. Section 3 outlines the data corre­
reinforcement cost of 0.28 €/MWh generated for 2% penetration, which sponding to the case study, while section 4 presents the main results
raises to 1.04 €/MWh for 4% penetration. Schmalensee et al. (Schma­ obtained, including the impact on the network in terms of variation of
lensee, 2015) use a RNM to design various prototype theoretical net­ technical losses, investment needs and O&M costs. Finally, a summary of
works where different predetermined PV penetration scenarios are the main findings is provided in section 5, followed by the most
considered. Through a brownfield network model they obtain the costs important conclusions and possible regulatory recommendations.
and losses of future upgradings required to accommodate the new
dispersed generation. They found that, for 20% PV generation with 2. Methodology and hypotheses
respect to the annual demand, the total distribution cost increase could
vary between USD 10 and USD 30 per installed kWp of PV capacity, This work further elaborates on the methodology developed in
depending on the network type. This cost is exclusive of losses, but does (Tévar et al., 2019), intended to evaluate the losses and congestions for
include investment and operation costs for LV networks, including increasing degrees of rooftop PV penetration on an hourly basis. More
substations. specifically, for all the scenarios considered, a heuristic, RNM-based
Other works use real distribution networks or areas to quantify the procedure is proposed to determine cost-effective reinforcement mea­
impact on network investments owing to the presence of distributed PV sures, to be implemented on the base-case network so that major over­
generation. From a different perspective, Cohen et al. (Cohen et al., voltages and congestions are eliminated. Then, a cost assessment of the
2016) estimate the network investment savings that can be achieved by necessary investments and resulting losses is performed.
implementing different levels of PV penetration for a real size feeder in Starting from the baseline (0% PV) case, the methodology adopted in
California. The evaluation method for investment savings is based on a this work sequentially applies the following five steps for each scenario
previous quotation provided by the company to forecast future demand, of increasing PV penetration (the first two steps embed the methodology
and the evaluation of investment delay due to the adjustment of peak presented in (Tévar et al., 2019)
demand as PV penetration grows. The evaluation is performed just up to (1) Determination of hourly energy balances for each point of supply:
the PV generation that matches the peak demand for each feeder. The hourly energy balances are determined at each point of supply,
Therefore, as reverse power flows seldom happen, the expected impact including the electricity generated, the self-consumed energy, the en­
on network investments is negligible. ergy taken from the network and the excess energy injected into the
Gomes and Saraiva (Gomes and Saraiva, 2017) propose a method­ network. By properly aggregating the individual hourly balances, the
ology that, using a Transmission Expansion Planning (TEP) model on an energy delivered to the LV circuits by each MV/LV secondary substation,
IEEE study network, and assuming four PV penetration scenarios (0%, or injected to the MV subsystem in case of excessive LV generation, is
10%, 15% and 20% PV), search for optimal constructive solutions by obtained, and hence the energy flows through each MV feeder section.
means of an Evolutionary Particle Swarm Optimisation (EPSO) model. The customers connected to LV grids who decide to deploy a rooftop PV
By only evaluating the impact on the transmission network, the work installation for self-consumption are randomly selected from among
confirms that there is virtually no impact at these voltage levels, for the those located in the zone of interest.
PV penetration considered. To conclude this review, Freitas et al. (2) Load flow solution at the feeder and hourly level to individually
(Freitas et al., 2018) perform power balances at the MV/LV transformer identify technical constraints violations (overvoltages and/or
level for an urban area of Lisbon to quantify the number of transformers congestions):
that may need upgrading under different scenarios. Essentially two Using the ANETO RNM tool (García-Conejo et al., 2007), a load flow

175
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

is run for each of the 8,760 h of the year, assuming that all LV loads are analysis be focused on, and consequently which upgrading choices are
three-phase balanced supplies. Then, all supply points facing over­ most effective.
voltages above regulatory limits (>7%) and all congested feeder sections Therefore, after trying and discarding other alternatives, a normal­
(loading > 100%) are identified. Yang et al. (Quan et al., 2018) propose ized severity index indicator, computed at the feeder level and intended
a set of indices to quantify the system performance including active and to guide the selection and assessment of upgrading alternatives, is
reactive power losses, voltage deviation and line loading indices. introduced as follows:
However, as our work prioritizes network reinforcing investments
intended to solve severe violations (assuming mild or moderate viola­ - Let F denote the set of all LV feeders (or circuits) under study, Fx the
tions have been addressed through operational actions), we have subset of feeders for which excessive violations are detected ac­
adopted the severity metrics developed in (Tévar et al., 2019), cording to the annual indices (LSSI, EVVI), and hm(j) the hour of the
comprising the Length-cumulated Section Saturation Index (LSSI) and year at which the LSSI indicator for feeder j reaches its maximum
the Energy-cumulated Voltage Violation Index (EVVI, defined for each value.
LV feeder j, as follows: - Then the index EX j hm(j), representing the normalized (weighted)
sum of ampacity and overvoltage violations at hour hm(j), is defined
• Length-cumulated Section Saturation Index (LSSI): for each LV circuit j as follows, where the normalization is based on
∑H,NS ( ) the whole set F of LV feeders.
LSSI j = h,=1,i=1
Li,h *%ΔSi,h (1)
LSSIhm(j) EVVIhm(j)
(7)
j j
EXhm(j)
j =∑ hm(i)
+ ∑ hm(i)
i∈F LSSIi i∈F EVVIi
oi, h: Feeder section and hour indices
oNS: Total number of feeder sections The introduced severity index is computed also for the entire set of
o%ΔSi,h : Overload in % of section i recorded at hour h. feeders F and the critical subset Fx:
oH: Hours of the year when section i of the feeder j is overloaded ∑ hm(j)
oLi,h : Length of the section i (in meters) where an overload exists at EXF = EXj (8)
j∈F
hour h.
• Energy-cumulated Voltage Violation Index (EVVI): ∑ hm(j)
EXFx = EXj (9)
∑H,NP ( )
Ei,h *%ΔVi,h (2)
i∈Fx
EVVI j = h=1,i=1
As long as EXFx > 0, a series of candidate reinforcement actions are
considered, aimed at reducing EXF under the condition that EXhm(j)
j does
o i, h: Point of supply and hour indices not increase for any feeder in F. The candidate upgrading actions and the
o NP: Total number of supply points fed by the feeder j way the most cost-effective action is chosen are explained in the
o %ΔVi,h : Overvoltage in % at point i recorded at hour h. following step.
o H: Hours of the year when the voltage is above 7% (regulatory limit) (4) Assessment and selection of reinforcement actions required to
at supply point i correct excessive violations of technical constraints:
o Ei,h : Energy demanded by supply point i at time h in which an As findin g the truly optimal sequence of upgrading measures,
overvoltage takes place capable of solving all technical constraints at minimum cost, is a very
hard combinatorial problem, a much simpler heuristic procedure is
where j denotes the respective LV feeder, h is the number of hours for proposed that prioritizes and sequentially solves all excessive violations
which overvoltages/congestions are detected, and H is the set of 8,760 h arising in the studied area. For this purpose, the following additional
(the indices are computed for a whole year). Next, the LV feeders are notation is introduced:
preliminarily ranked according to their separate violation indices and
classified as being subject to either excessive or moderate violations. As - “Spotlighted feeder” (SF) is the feeder for which upgrading solutions
in the previous work (Tévar et al., 2019), the following thresholds are to solve excessive violations are being sought. The buses with un­
adopted: acceptable overvoltages, or those fed by congested sections of the SF
are termed the “involved nodes”. Note that all circuits initially
• Constraints violations are deemed as excessive if: included in the set Fx will eventually become “spotlighted feeders”,
o EVVI ≥ 1MWh*10% or, (3) unless their violations have vanished as a consequence of previously
o LSSI ≥ 100m*100h*120% (4) implemented reinforcement actions.
• Constraints violations are deemed as moderate if - Given a SF, a subset of “candidate feeders” (CF) is identified to be
o 0 < EVVI < 1MWh*10% potentially involved in the solutions required to solve the SF viola­
and (5) tions. This subset corresponds to all subsidiary LV circuits of the MV/
o 0 < LSSI < 100m*100h*120% (6) LV secondary substation from which the SF is fed, plus the LV circuits
of the three closest secondary substations.
(3) Characterization of the overall severity of technical constraints
violations at the feeder level: For each involved node the following upgrading/reinforcement al­
Assume there are two nearby LV feeders, j and k, for which excessive ternatives are just considered:
violations are detected. Each of them can face the most severe violation
at different hours of the year, or they can be subject to different degrees 1. Reinforcement of the upstream feeder section: this consists of
of violation severity depending on whether the overvoltage or conges­ upgrading the master section that feeds the involved node, all the
tion index is considered. Furthermore, an upgrading solution initially way from the secondary substation, by increasing the conductor
intended to solve the violations of circuit j, might also solve to a certain ampacity.
extent the violations of circuit k. For these and other reasons, it is not 2. Additional link (or interconnection) from neighbour feeders: a new
obvious or straightforward to determine, based solely on the yearly section that connects the involved node, and its downstream sec­
feeder indices defined by (1) and (2), which constraint violations tions, to any other node belonging to the CF set is considered (to keep
(overvoltage or ampacity), or which hours of the year, should the the radiality of operation, it is assumed that the involved node is

176
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 1. Methodology flow chart.

disconnected from its original master section). All possible options are deemed acceptable.
are analysed and compared. Once the list of candidate actions is available, and the associated
3. Splitting of the upstream feeder section: a new feeder, running from investment costs are computed, the one that offers a greater reduction of
the MV/LV secondary substation in parallel with the congested one, the ΔEXFx /cost ratio is chosen and implemented. Then, the 8,760 load
is added to take on just the load of the downstream buses (the flow solutions are run again for the upgraded network, the parameters
original section remains feeding only the involved node). EXFx and EXF are recomputed and a new list of network adaptation
candidates is generated. This process is repeated until EXFx is zero or no
The cost of each alternative measure is evaluated based on reference suitable upgrading actions are found to further reduce EXFx .
unitary investment costs for each network asset type. For this purpose, Finally, once the whole set of LV feeders is upgraded for each PV
the reference unitary costs stablished by the Spanish regulator in penetration scenario, it is checked whether any MV/LV secondary sub­
(Orden, 2011) are adopted. Moreover, for each candidate upgrading station remains serving a load that exceeds 120% of its rated power
action, the benefit in terms of EXF reduction (ΔEXF) is also obtained. over>100 h per year. In that case, the secondary transformer is replaced
This implies recalculating the 8,760 load flows for the affected feeders by another one with the minimum power strictly needed to eliminate the
(SF and eventually CF), by duly considering the network model modi­ congestion, among those available in the standardised catalogues for
fications involved in the reinforcement measures being analysed. Note that type of transformer.
also that, while handling the constraints violations of the SF, only those (5) Economic assessment of the solution:
actions that do not worsen the violation indicators of other LV circuits In order to assess the economic impact of the grid reinforcement

177
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 2. Geographical scope and case study characteristic data.

solutions found in step (4), at least three cost components should be for increasing PV penetration levels are determined. This may
considered: 1) the variation of losses with respect to the scenario include new feeder sections, interconnections among neighbour
without PV generation, 2) the investment costs associated with the feeders or secondary substations. Therefore, a list is created of all
network upgrading actions required to maintain the current level of physical units that have been added to the network in the area (km of
service quality, and 3) the O&M costs related with them. A brief dis­ new LV lines, km of sections with upgraded conductor, kVA of more
cussion on each of those three factors follows. transformation capacity). Applying the reference unitary investment
costs provided by the Spanish regulator (Orden, 2011) to the
• Economic assessment of losses: once all LV circuit violations are resulting physical units yields the associated investment in the whole
solved for each PV penetration scenario, the solution of 8,760 load area.
flows provides the hourly losses at the feeder level and, through • Evaluation of O&M costs: the Spanish regulation also provides
accumulation, the resulting annual losses. Also, multiplying hourly reference unitary costs for operation and maintenance of each type of
losses by the hourly price of the wholesale energy market, and network asset (Orden, 2011). Applying these unitary costs to the
adding for all hours of the year, the cost of annual losses is obtained. physical units obtained in points (3) and (4) yields the annual O&M
The difference between the resulting losses and those registered in costs associated to each PV scenario. The cost saving arising from the
the baseline scenario (0% PV), in both kWh and euros, yields the LV network sections that have been removed and replaced with new
impact (positive or negative) on the LV network losses for every PV ones is also taken into account. It is worth noting the recurrent nature
penetration level. It is worth noting that the results shown below are of those annual costs, until the end of the useful life of the
different from those obtained in (Tévar et al., 2019), as the network installations.
has been adapted in this work to remove violations, whereas in
(Tévar et al., 2019) the grid was always the same. In order to obtain the total cost, the three cost components have to be
• Economic assessment of investment costs needed to upgrade the referred to the same time basis. While the investment cost is incurred
network: as a result of the procedure described in points (3) and (4) only once, for each increasing PV penetration, O&M costs and losses are
above, the reinforcements required to prevent excessive violations recurrent and repeat annually. Given that, from a regulatory point of
view, the distribution activity considers a useful life of network assets of
40 years, and assuming the hypothesis that the demand remains con­
Table 1 stant for all PV penetration scenarios, the Total Cost (TC) is defined as:
Network distribution elements considered for new developments, and unit costs
of investment and O&M. TC%PV = I%PV + OM 40 %PV + L40 %PV (10)
Network Wire Wire Investment O&M Ampacity
where:
element section section (€/km) cost (A)
description (mm2) range (€/km
I%PV: Investment to adapt the distribution network for an increase in
(mm2) yearly) the %PV level from the previous one.
OM40%PV: Present value of increased annual operation and mainte­
LV single 50 S < 75 26,620 276 150
circuit 95 S ≥ 75 32,536 338 230 nance costs for an increase in the %PV level from the previous one,
underground calculated for 40 years with an internal rate of return (IRR).
line L40%PV: Present value of the increase in annual losses due to the in­
LV single 95 0<S< 48,384 502 260
crease in %PV level from the previous one, calculated for 40 years with
circuit 150
overhead 150 S ≥ 150 59,136 614 330
an internal rate of return (IRR).
line Fig. 1 provides a flow chart summarising the methodology outlined

178
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Table 2
Comparison of three different PV penetration definitions applied to increasing scenarios of PV generation.
PV Penetration Scenario

1 2 3 4 5 6 7 8 9 10

PV Installed Power (MW) 31 62 92 123 154 185 216 246 277 308
Scale %
SMaxP 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
PZeroD 15.2 30.3 45.5 60.7 75.9 91.0 106.2 121.4 136.6 151.7
SPeakP 21.8 43.7 65.5 87.4 109.2 131.1 152.9 174.8 196.6 218.4

Fig. 3. Examples of grid upgrading to get rid of excessive non-compliances.

above. the reference ranges and costs1 considered.

For the economic assessment of losses, the hourly rates recorded by
3. Case study the Spanish wholesale market2 in 2017 have been taken into account.

As a case study, the same real distribution network tested in (Tévar 4. Results and discussion
et al., 2019), serving a large area in the province of Barcelona, is used in
this work. Fig. 2 shows a (zoomed out) view of the one-line diagram In order to quantify the PV penetration level for each scenario, in
along with the most important network and load data. The reader is such a way that the results are comparable with those previously pub­
referred to (Tévar et al., 2019) for further details. lished, this work resorts to three out of the five different bases found in
In addition, this work needs extra data related to ranges of elements the literature (see (Tévar et al., 2019). For the sake of completeness,
to be considered in the development of reinforcing measures, as well as those reference values are described herein: 1) SMaxP is the sum of
their corresponding unitary investment and O&M costs. Table 1 shows

1
Unitary costs published in [43].
2
Costs published at www.esios.ree.es

179
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 4. LV feeders length with technical violations before (left bars) and after (right bars) grid upgrading.

maximum PV power installable at each supply point, which takes into Circuit A consists of a 61-meter feeder exhibiting excessive ampacity
account both the limit imposed by the available roof surface and the congestions for 780 h a year, while circuit B, also presents excessive
ampacity of the lateral feeder allocated to each home or building; 2) saturations for 311 h a year in its first 202 m. At 13:00 on May 25, both
PZeroD is the PV peak power that produces over the year the same en­ circuits present maximum saturation (178% for circuit A and 143% for
ergy supplied to the load by the network under study; 3) SpeakP is the circuit B). The upgrading procedure begins by considering the addition
sum of the peak demands served by the set of LV feeders. Table 2 shows of a new circuit (61 m, 95 mm2, 3,000 €), running in parallel with the
the equivalence among these indicators for increasing PV penetration existing one, from the secondary substation to the first bifurcation point
scenarios in the case under study. In the sequel, the PV penetration levels of circuit A. Then, 55 customers originally fed by circuit A downstream
will be referred to SMaxP, unless otherwise indicated. this point are isolated (switch s1) and reconnected to the new added
section, giving rise to a new subcircuit (A’). This action, selected among
other choices owing to its highest benefit/cost ratio, reduces the satu­
4.1. Illustration of network upgrading choices ration of circuit A to 135%.
The next selected action consists of shifting part of circuit B (32
Before presenting the aggregated results for the whole studied area, a loads) to the new circuit (A’). This involves the deployment of a new
tutorial example will be discussed showing in detail how a real underground section (53 m, 95 mm2, 2,565 €) and the installation of a
congestion problem is solved according to the methodology described in switching device (s2), which fully eliminate the congestion on the
section 2. resulting circuit B. Then, the focus is put back on the remaining
Fig. 3 shows a window of the studied area, where an increase in the congestion of circuit A, which is addressed by transferring 13 customers
PV penetration level, from 50% to 60%, leads to new technical con­ from circuit A to circuit A’, for which a switching device (s3) along with
straints being violated (in this particular case, only ampacity conges­ a new underground circuit section (19 m, 95 mm2, 919 €) are
tions, shown in red, arise).

180
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Table 3
Reinforcements and new distribution network extensions needed to eliminate excessive violations (in physical units).
PV penetration scenario (SMaxP)

10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Overhead LV network reinforcement

Section reinforcement (km) 0.11 0.81 2.16 5.43 3.39 3.39 2.52 6.09 6.70 7.53
Network development (km) – 0.15 2.07 1.10 3.66 2.47 7.41 9.81 8.21 12.42
Underground LV network reinforcement
Section reinforcement (km) – 0.01 – 0.09 0.05 0.07 0.59 0.20 0.41 0.29
Network development (km) 0.03 0.15 1.06 3.69 6.30 9.19 12.21 18.17 15.62 17.99
Secondary substations repowering
Transformer replacement (MVA) 0.10 – 0.70 2.28 5.69 19.88 19.03 31.22 44.12 42.71
Additional installed power (MVA) 0.05 – 0.35 1.00 2.47 7.63 7.41 11.46 15.83 14.30

Fig. 5. Necessary investment in the distribution network to eliminate excessive congestions.

considered. As a result, the saturation of circuit A drops to 122%, which SMaxP). For each PV penetration level, the left bar refers to the initial
is finally removed by increasing its section to 150 mm2 (2,886 €)3. network scenario, before upgrading actions are undertaken. As can be
Overall, the required upgrading actions imply an investment cost of observed, the fraction of LV network with technical constraints viola­
9,370 €. It was verified that the upgraded network does not present tions, and their severity, grow with PV generation, as new LV feeders
ampacity violations or overvoltage at any time during the year. appear with excessive violations. The right bar refers to the network
with the same level of PV penetration, after having been upgraded to
4.2. Technical constraints violations before and after network upgrading remove excessive violations. Note that the yellow part (feeders with
excessive violations) disappears, remaining only moderate violations
Starting from the baseline scenario without PV generation and the which, hopefully, can be handled via customary network operation
initial network, the methodology presented in section 2 and illustrated measures.
above is applied every 10% PV generation increase. Fig. 4 shows the
length of LV feeders with excessive (yellow) and moderate (green) vi­ 4.3. Investment on network upgrading and resulting losses
olations for all PV penetration scenarios considered (up to 100% of
Table 3 summarizes, for the studied area, the new assets added as a
consequence of the LV network upgrading, in physical units (re­
3
Taking advantage of the existing circuit leads to a 45% reduction of the inforcements and new distribution network extensions), for every 10%
investment cost compared to the greenfield construction. increase in PV penetration.

181
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 6. Annual losses for increasing PV penetration scenarios in two cases with and without grid upgrading.

The resulting reinforcement costs (Fig. 5) are obtained by applying scenario with 90% PV penetration the total reduction would be 10%
reference unitary investment costs to the values of the above table. This (from 6.62% to 5.93%). As a result, while approximately 30% of PV
represents the required investments so that the distribution network can penetration was the threshold point at which losses began to exceed
withstand increasing degrees of PV generation, without reducing the those of the baseline scenario under the constant network hypothesis
existing service quality levels. (Tévar et al., 2019), when the network upgrading is considered such a
It can be noticed that, up to 30% of PV penetration (approx. 45% threshold shifts to slightly above 40% PV penetration.
according to the PZeroD base), there are very few structural actions Fig. 7 separately shows the evolution of losses for LV circuits and
(upgrades or new developments) that have to be actually implemented MV/LV secondary substations. It can be noticed that all components of
on the LV distribution network. However, starting at around 40–50% PV losses grow at a slower rate when the grid is gradually adapted, except
penetration, the need for investments increases more than linearly, for the transformer core losses, which are independent of the trans­
owing primarily to the upgrading of LV lines and existing MV/LV sec­ former charge but increase with its capacity.
ondary substations, but then, beyond 50% penetration, the need to carry
out new network extensions becomes notable. This implies to find new
locations and new spaces to install substations, lines and cables. 4.4. Impact of network upgrading in terms of overall network cost
Overall, taking on 50% of PV penetration requires upgrading and/or
expanding 1.4% of the existing network in the area, with an investment As explained in section 2, all cost components (losses, investment
cost representing 1.5% of the replacement value (value as new) of the and O&M costs) associated to every level of PV penetration, and the
existing assets. Accommodating 90% PV generation, requires that 6.1% resulting reinforced grid, can be translated into uniform units (euros)
of the LV network be upgraded, with an investment cost of 9.3% of the and then added up on the same time basis (40 years of useful life for
value as new. installations). Fig. 8 shows the evolution of the three cost components
The investment required to adapt the network implies also an in­ versus the PV penetration degree, when a 5.58% internal rate of return
crease of the network capacity and, consequently, a reduction in tech­ (IRR) for distribution assets4 is considered in expression (6).
nical losses. This can be seen in Fig. 6, representing the total LV network As clearly seen, the total direct costs are negative for small PV
losses (secondary substations included) for increasing PV generation penetration levels, which means that rooftop PV has a positive impact on
scenarios, both for the baseline network (i.e., assuming the violations the distribution network. The maximum profit for the network occurs at
seen in previous sections), and for the gradually upgraded network. approximately 18% penetration level, whereas the impact is neutral
Losses are expressed in % over the total energy delivered. Logically, the again for 33% of PV penetration. Beyond this point, direct costs grow
losses increase at slower rates with PV generation when the network is nearly quadratically with PV penetration, investment costs being the
reinforced. For example, for a scenario with 50% PV penetration,
adapting the network to correct excessive violations reduces the losses
from 4.86% to 4.57% (6% reduction of total losses), whereas for a 4
This value is proposed by the Spanish Regulator (CNMC, 2018).

182
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 7. Annual losses split by main network components for increasing PV scenarios in two cases with and without grid upgrading.

largest component followed by the costs of losses. The O&M costs benefit characterized by its hourly consumption curves throughout a year,
from the fact that, while new installations are considered, the cost of including approximately 4,300 LV feeders (2,200 km), fed by 1,800 MV/
dismantled assets is subtracted. LV secondary substations, has provided a bunch of results regarding the
It is worth stressing that this work has not considered other relevant impact of different rooftop PV generation scenarios on the LV distribu­
issues and associated costs that could also grow as the distributed PV tion network, which can be summarised as follows:
generation flourishes. These include several operational problems, such
as rapid voltage changes due to passing clouds, phase imbalances, - Up to approximately 30% of PV penetration (45% in terms of
generation of harmonics and flickers, power flow reversals and the need PZeroD), no significant structural actions (upgrading or expansion)
to readjust protection co-ordinations, premature ageing of certain are needed to maintain the ampacity violations and overvoltage
network components, etc. In addition, unresolved moderate violations levels on the LV network within acceptable limits, which can be
will involve other measures not quantified in this paper, such as network hopefully handled through operational control actions or by
control actions or the implementation of local flexibility markets5 deploying local flexibility markets.
(Institute, 2019; Tabone et al., 2016). Altogether, in order to properly - Beyond 40–50% PV penetration, the cost of structural actions
address all those issues, additional investments in digitization, moni­ required to maintain the LV network saturation and overvoltage
toring, automation and systems are needed. levels within acceptable limits increases quadratically. For instance,
Fig. 9 shows the total direct costs (ignoring losses) per installed PV to be able to reach 50% of PV penetration requires an investment
for each PV penetration scenario. Those values range from approxi­ representing 1.5% of the value as new of the base case network. In
mately 0.50 €/kWp for 10–15% PV penetration degrees (15–20% in order to incorporate 90% of PV generation the required investment
terms of PZeroD), to 10.18 €/kWp for PV penetration levels as large as would be 9.3% of that value.
50%. - Investment in the upgrading and network expansion contributes also
to reducing the technical losses with respect to those that would arise
5. Summary of findings in the base case network, for the same PV penetration. As a result,
while approximately 30% of PV penetration was the point, under the
The methodology developed in this work, when applied to an actual constant network hypothesis (Tévar et al., 2019), at which losses
case study comprising approximately 80,000 customers, each began to exceed those of the baseline scenario, such threshold point
lies beyond 40% PV penetration when the network is gradually
upgraded.
5 - Regarding the impact in terms of total cost (investment, O&M and
Decentralized flexibility resources such are electric vehicle, storage devices,
losses), it is concluded that, for low penetration levels, PV generation
electrical heating and cooling, public lighting loads, etc.

183
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Fig. 8. Total direct costs split by its components for increasing PV penetration scenarios.

Fig. 9. Reinforcement and repowering plus O&M cost vs. PV installed.

produces a positive impact (profit) on the distribution network, PV penetration whereas, beyond that point, the additional distribu­
which reaches a maximum at approximately 18% PV penetration. tion costs grow quadratically with PV generation. This excludes
Such positive impact vanishes (null total cost) at approximately 33% other operational costs that could be incurred to address related

184
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

problems, such as rapid voltage fluctuations, imbalances, generation Declaration of Competing Interest
of harmonics, readjustment of protections to safely accommodate
reverse power flows, etc. The authors declare that they have no known competing financial
- Neglecting losses, the total direct costs per installed PV power interests or personal relationships that could have appeared to influence
amounts to 2.73 €/kWp for scenarios with 30% PV and 10.18 €/kWp the work reported in this paper.
for 50% PV penetration.
- No additional investments needed in digitization, monitoring, Acknowledgements
automation and ICT systems have been valued, nor the need to
implement local flexibility markets or alternative control actions This work has been performed within and supported by the
intended to solve the remaining moderate violations. Department of Regulations of Endesa, under the direction of Dr. Juan J.
Alba-Ríos. This has made it possible to gather and access the valuable
As pointed out in the introduction, it is worth stressing that those information on actual load curves and network models. Thanks also to
findings stem from a long-term planning perspective and for a large- Endesa, for allowing the use of the Network Reference Model (Aneto), a
scale geographical area. Local or short-term effects, such as phase im­ crucial tool to carry out all of the required simulations, on a massive
balances, cloud induced PV fluctuations or spatial load concentrations, scale, for the diversity of scenarios considered in this work. The leading
may have also detrimental effects, which could be solved through the author owes special thanks to his colleagues from the Department of
implementation of local flexibility markets, by network switching/ Regulation, and very specially to the Distribution Regulation team for
operational actions, or even by resorting to the features provided by their guidance, orientation and, of course, their patience. The other
current PV inverters. We implicitly assume that those solutions would authors acknowledge the financial support of Consejeria de Economia y
not involve significant additional investments, provided suitable regu­ Conocimiento within the EU- FEDER Andalusia 2014–20 program
latory measures are adopted. (project S2V, US-1265887, 2020 PAIDI plan).

6. Conclusions Appendix A. Supplementary data

In this work a methodology is presented to assess the impact on the Supplementary data to this article can be found online at https://doi.
LV distribution network arising from the expansion of distributed PV org/10.1016/j.solener.2021.01.066.
generation, mainly from rooftop self-consumption installations. The
assessment considers the direct investment costs associated to the References
expansion/upgrading of the network, which is required to maintain the
network congestions and overvoltage levels within acceptable limits, as Proposal for a Directive of the European Parliament and of the Council on the promotion
of the use of energy from renewable sources (recast), Winter Package. European
well as the corresponding O&M costs and the variation in losses. The Commission November 2016.
methodology has been applied to a large-scale, real distribution network Deloitte Monitor. A sustainable energy model or Spain in 2050 energy policy
comprising approximately 80,000 customers with an annual demand of recommendations for transition March 2016.
Heesen, H., Herbort, V., Rumpler, M., 2019. Performance of roof-top PV systems in
460 GWh, almost 2,200 km of LV lines and over 1,800 secondary Germany from 2012 to 2018. Sol Energy 194 (October), 128–135.
substations. Arcos A. After the Paris agreement. DYNA 2016;92(1):12–4. https://doi.org/10. 6036/
For low PV penetration levels, the average energy flowing through 8229.
Singh, R., Banerjee, R., 2017. Impact of large-scale rooftop solar PV integration An
the LV network tends to decrease, even though local congestions or algorithm for hydrothermal-solar scheduling (HTSS). Sol Energy 157 (September),
overvoltages may still arise to a lesser extent. Starting at around 30% PV 988–1004.
(45% using the base PZeroD), investments are increasingly necessary to Sussams, L., Leaton, J., 2017. Expect the unexpected the disruptive power of low-carbon
technology, carbon tracker. Grantham Institute at Imperial College, London.
solve unacceptable overvoltages or congestions, the growing rate being
Deloitte Monitor. Towards the decarbonization of the economy the contribution of
nearly quadratic beyond that point. electricity grids to the energy transition February 2018.
Regarding the losses, thanks to the reduction in transmitted energy Haque, M.M., Wolfs, P., 2016. A review of high PV penetrations in LV distribution
from the initial stages of PV penetration, they get reduced and reach networks present status, impacts and mitigation measures. Renew Sustain Energy
Rev 62 (September), 1195–1208.
their lowest level at approximately 20% PV penetration. The in­ Barker PP, De Mello RW. Determining the impact of distributed generation on power
vestments necessary to eliminate excessive congestions and voltage vi­ systems. I. Radial distribution systems. 2000 Power Engineering Society Summer
olations also favour the reduction of losses. As a result, the global losses Meeting July 2000.
Karimi, M., Mokhlis, H., Naidu, K., Uddin, S., Bakar, A.H.A., 2016. Photovoltaic
do not exceed again those of the baseline scenario until the PV pene­ penetration issues and impacts in distribution network – a review. Renew Sustain
tration level is beyond 40%. Energy Rev 53 (January), 594–605.
Overall, considering the total direct costs, including losses, low PV Tévar, G., Gómez-Expósito, A., Arcos-Vargas, A., Rodríguez-Montañés, M., 2019.
Influence of Rooftop PV Generation in Net Demand, Losses and Network Congestions
penetration levels lead to a distribution cost saving that peaks at 18%. A Case Study. Electrical Power and Energy Systems 106 (March), 68–86.
Beginning at 33%, the distributed generation leads to an increase in Royal Decree 1699/2011 of November 18th that regulates connection to a small power
distribution cost that grows more than proportionally with PV pene­ electrical power production installations network. MINETAD November 2011.
Royal Decree 244/2019 of April 5th that regulates administrative, technical and
tration. Specifically, without losses, the total costs per installed PV economic conditions of self-consumption of electrical energy. MITECO April 2019.
power amounts to 2.73 €/kWp for scenarios with 30% PV and 10.18 Quezada, V.H.M., Abbad, J.R., Gómez, T., 2006. Assessment of energy distribution losses
€/kWp for 50% PV penetration. for increasing penetration of distributed generation. IEEE Transactions on Power
Systems 21 (May), 533–540.
From a regulatory standpoint, until more efficient dynamic rates are
Bastiao F, Cruz P, Fiteiro R. Impact of distributed generation on distribution networks.
developed, based on remote information provided by smart meters, our 2008 5th International Conference on the European Electricity Market May 2008.
findings suggest that PV-based self-consumers should be subject to extra Wang B, Lan K. Analysis of the Distributed Generation System and the Influence on
charges only in those areas where the degree of PV penetration is >30%. Power Loss. 2011 Asia-Pacific Power and Energy Engineering Conference March
2011.
Other indirect costs not evaluated in this work, such as those arising Morvaj, B., Evins, R., Carmeliet, J., 2016. Optimization framework for distributed energy
from the need of greater monitoring and remote control, amongst others, systems with integrated electrical grid constraints. Applied Energy 171 (June),
should be somehow recognised under the remuneration scheme for the 296–313.
Borges CLT, Falcäo DM. Impact of distributed generation allocation and sizing on
distribution activity. reliability, losses and voltage profile. In IEEE Bologna powertech conference, June
2003.

185
G. Tévar-Bartolomé et al. Solar Energy 217 (2021) 173–186

Osorio L, Patiño S, Hernández J. Methodology to assess the influence of PV systems as a Gastalver-Rubio, A., Romero-Ramos, E., Maza-Ortega, J.M., 2019. Improving the
distributed generation technology. 2012 IEEE International Symposium on Performance of Low Voltage Networks by an Optimized Unbalance Operation of
Alternative Energies and Energy Quality (SIFAE) October 2012. Three-Phase Distributed Generators. IEEE Access 7 (December), 177504–177516.
Mandiş Al. C, Manoloiu A, ŞtefanaNeagoe AG, Leonida T, Neagoe AC. Impact of Bottacciolia, L., Estebsarib, A., Pattia, E., Ponsb, E., Acquaviva, A., 2017. A novel
distributed generation on steady state of electrical networks. 2014 International integrated real-time simulation platform for assessing photovoltaic penetration
Symposium on Fundamentals of Electrical Engineering (ISFEE) November 2014. impacts in smart grids. Energy Proc 111 (March), 780–789.
Quan H, Yang D, Khambadkone A, Srinivasan D. A Stochastic Power Flow Study to Mateo, C., Cossent, R., Gómez, T., Prettico, G., Frías, O., Fulli, G., Meletiou, A.,
Investigate the Effects of Renewable Energy Integration. 2018 IEEE Innovative Smart Postigo, F., 2018. Impact of solar PV self-consumption policies on distribution
Grid Technologies-Asia (ISGT Asia);May 2018. networks and regulatory implications. Solar Energy 176 (December), 62–72.
Chiandone M, Campaner R, Massi Pavan A, Sulligoi G, Mania P, Piccoli G. Impact of Xiang, Y., Liu, J., Li, F., Liu, Y., Liu, Y., Xu, R., Su, Y., Ding, L., 2016. Optimal Active
Distributed Generation on power losses on an actual distribution network. 2014 Distribution Network Planning A Review. Electric Power Components and Systems
International Conference on Renewable Energy Research and Application (ICRERA) 44 (May), 1075–1094.
October 2014. Pudjianto D, Djapic P, Dragovic J, Strbac G. Grid Integration Cost of PhotoVoltaic Power
Cohen, M. A., & Callaway, D. S. (2015). Physical Effects of Distributed PV Generation on Generation. Direct Costs Analysis related to Grid Impacts of Photovoltaics. Imperial
California’s Distribution System. arXiv preprint arXiv:1506.06643. College London September 2013.
Cohen, M.A., Callaway, D.S., 2016. Effects of distributed PV generation on California’s Schmalensee, R., 2015. The future of solar energy an interdisciplinary MIT study. Energy
distribution system, Part 1engineering simulations. Sol Energy 128 (April), 126–138. Initiative MIT.
Westacott, P., Candelise, C., 2016. Assessing the impacts of photovoltaic penetration Cohen, M.A., Kauzmann, P.A., Callaway, D.S., 2016. Effects of distributed PV generation
across an entire low-voltage distribution network containing 1.5 million customers. on California’s distribution system, Part 2 economic analysis. Sol Energy 128 (April),
IET Renewable Power Generation 10 (March), 460–466. 139–152.
Eguía P, Torres E, García J. GEDISPER project. Study of the impact of distributed Gomes PV, Saraiva JT. Transmission system planning considering solar distributed
generation on electrical network losses. Informe final. Energy and Society July 2016. generation penetration. 2017 14th International Conference on the European Energy
Garcia-Villalobos J, Eguia P, Torres E, Etxegarai A. Impact of photovoltaic self- Market (EEM) June 2017.
consumption on power losses and voltage levels of MV distribution networks. In IEEE Freitas, S., Santos, T., Brito, M.C., 2018. Impact of large scale PV deployment in the
PES innovative smart grid technologies conference Europe (ISGT-Europe) September sizing of urban distribution transformers. Renewable energy 119 (April), 767–776.
2017. García-Conejo JA, Tévar-Bartolomé G, Gómez-Expósito A, Rodríguez-Montañés M.
García-Villalobos J, Eguía P, Torres E, Etxegarai A. Assessing the impact of photovoltaic ANETO a system for the automatic generation of theoretical network models. In 9th
self-consumption support policies on energy losses. In International conference on International conference on electrical power quality and utilisation (EPQU’07)
renewable energies and power quality (ICREPQ’17) April 2017. October 2007.
Nguyen, A., Velay, M., Schoene, J., Zheglov, V., Kurtz, B., Murray, K., et al., 2016. High Orden IET/2660/2015, de 11 de diciembre, por la que se aprueban las instalaciones tipo
PV penetration impacts on five local distribution networks using high resolution y los valores unitarios de referencia de inversión, de operación y mantenimiento por
solar resource assessment with sky imager and quasi-steady state distribution system elemento de inmovilizado y los valores unitarios de retribución de otras tareas
simulations. Sol Energy 132 (July), 221–235. reguladas que se emplearán en el cálculo de la retribución de las empresas
Ebad, M., Grady, W.M., 2016. An approach for assessing high-penetration PV impact on distribuidoras de energía eléctrica, se establecen las definiciones de crecimiento
distribution feeders. Electr Power Syst Res 133 (April), 347–354. vegetativo y aumento relevante de potencia y las compensaciones por uso y reserva
Rodriguez-Calvo, A., Cossent, R., Frías, P., 2017. Integration of PV and EVs in de locales. MIET December 2011.
unbalanced residential LV networks and implications for the smart grid and EPRI. Integration of hosting capacity analysis into distribution planning tools. January
advanced metering infrastructure deployment. Electr Power Energy Syst 91 2016.
(October), 121–134. Wang, L., Tisdale, M., 2018. Advanced capacity analysis in O’ahu. Gridworks.
Kreuwel, F.P.M., Knapc, W.H., Visser, L.R., van Sark, W.G., Vilà-Guerau, J., van The Oxford Institute for Energy Studies. (2019). The electricity market design for
Heerwaarden, C., 2020. Analysis of high frequency photovoltaic solar energy decentralized flexibility sources. July 2019.
fluctuations. Solar Energy 206 (August), 381–389. Tabone, M., Goebel, C., Callaway, D., 2016. The effect of PV siting on power system
flexibility needs. Sol Energy 139 (November), 776–786.

186