Feasibility assessment for high penetration of distributed photovoltaics based on net demand planning

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 1 / 19 www.brianazzopardi.eu

Feasibility assessment for high penetration of distributed photovoltaics based on net demand

planning

Brian Azzopardi a, b, *, Alejandro Gabriel-Buenaventura b

* Corresponding Author

a Institute of Electrical and Electronics Engineering, Malta College of Arts, Science and Technology

(MCAST), Malta

b Department of Mechanical Engineering and Mathematical Sciences, Faculty of Technology,

Design and Environment, Oxford Brookes University, Oxford, United Kingdom

Highlights

Literature review addressing high penetration of renewable energy sources.

A Net Demand Planning (NDP) feasibility assessment developed.

A case study is based on a building development in Oxfordshire.

Options for high penetration of photovoltaics (PVs) integration are discussed.

Keywords: Photovoltaic; Storage; Energy; Demand; Load; Management.

ABSTRACT

Well integrated distributed photovoltaics (PVs) will help tomorrow's energy system become more

sustainable. This paper highlights the net demand planning feasibility assessment to understand

these dynamics. These dynamics underpins the technical feasibility when setting up high penetration

of PVs within a new development area. Real measurements of disaggregated domestic electricity

demand from three low-carbon homes and simulated PV output are used. A case study is based on

a new building development in North West Bicester Eco Town, Oxfordshire, consisting of 393

homes, community and commercial units. Using simulation models and real domestic load

measurements, the results show potential for demand management and energy storage. With a

sensitivity analysis, the paper discusses also related energy policy aspects such as options for PV

orientation aspects, the use of Electric Vehicles (EVs), the potential of Direct Current electrical

installations and schemes that encourage load shifting. The paper concludes, within the context of

the case study, assumptions and scenarios that, contrary to what is commonly believed, high

penetration of PVs in a development area is feasible and may even reduce grid infrastructure costs.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 2 / 19 www.brianazzopardi.eu

1. Introduction

Distributed generation (DG) photovoltaics (PVs) bring about an attractive investment to domestic

users, especially when supported by generous financial support schemes. Additionally, the pressure

on developers to integrate substantial PV systems for new developments has increased to meet

benchmarks for zero-carbon buildings and other planning requirements. Integrating high penetration

of DG PVs is a challenge, as domestic electricity demand and insolation are mostly negatively

correlated on a daily basis. Hence if not properly assessed through Net Demand Planning (NDP), it

may result in increase of energy cost for the final customer due to extra grid investment as well as

increased risk of power outages across the development. This issue provides an incentive to cap the

level of installed PV capacity in contradiction to government policies objectives. Design may be

necessary for energy storage and demand management systems and consumer engagement for

behavioural change on electricity consumption.

High penetration of DG is when the installed capacity of DG connected to the distribution networks

already exceeds the area's total peak demand. In 2012, global PV installations have reached close to

the 100 GW mark and during 2013 it is expected that 110 TWh of energy is produced. While Europe

still represents the majority of installations globally, the Asian and American continents are picking

up rapidly [1]. In some regions, such as Bavaria in Germany, high penetration of DG, mostly PVs,

is already experienced. Most likely these have been installed on the ‘fit-and-forget’ approach [2].

The European Union energy strategy proposal reflect the “20–20–20” objectives: a 20% cut in

greenhouse gas (GHG) emissions; a 20% increase in use of renewable energy; and a 20% cut in

energy consumption through improved energy efficiency – all by 2020. The increase in use of

renewable energy can be achieved by large increase in the quantity and proportion of electricity

generated from variable renewable sources, such as wind, wave, tidal and solar power. Aggregating

these intermittent generation sources over large areas mitigates their variability [3]. However, the

potential for high grid exports in the grid from high-regional-penetration of PV generation during

periods of low energy demand and high solar irradiance still exists. During these periods, if electric

energy demand is not flexible, the grid infrastructure requires major investment to cope with the

transfer of energy from one area to another [2].

On a distribution feeder level, the major limiting factors are voltage rise caused by mismatching

production and demand, which is mainly dependent on the load demand, the network topology and

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 3 / 19 www.brianazzopardi.eu

cable impedances [4]. A number of studies analysed embedded generation [5], distribution networks

with photovoltaic generators using stochastic techniques [6] and its impacts [7]. However, voltage

rise caused by mismatching production and demand, may supplement the already existing voltage

drops on the feeders and thereby improve the customer voltages and reduce network losses. This

has been identified in optimal allocation of PV generators on low-voltage (LV) feeders in India [8].

The debate on options to integrate high penetration of PVs is ongoing with studies covering a wide-

range of options for high latitude locations [9], existing and future support schemes [10], smart

integration with emerging technologies [11], cost boundaries for possible third generation of

photovoltaics [12], using real options theory for PV investment [13], optimising the integration of

emerging PV technologies [14], analysing alternatives for decision making both as utility and

consumer perspectives [15] and decision support systems for PV investment [16].

The challenge is to encourage the local use of energy generation by load matching. These may

include considering the PV system design in previous studies or the alternative use of DC power

and schemes that encourage behaviour change [17] and preparation to the influx of possible

emerging responsive loads such as electric vehicles (EVs) [18]. In detail some local use strategies

may include simple domestic water demand management [19], stochastic programming on random

local demand [20], price-based electricity demand management using time-of-use tariffs [21], user

expected demand response [22], and day-ahead electricity pricing model [23] and energy storage

[24].

1.1. Aim of the study

The aim of this study is to present a methodology based on NDP for the technical feasibility analysis

to accommodate high-penetration of PV without the detrimental and limiting factors on the

generators by increasing the load matching capability. The NDP methodology is general and is

useful to any feasibility studies on regional developments of any latitude on the energy system level.

In this paper, the NDP methodology is applied to an imminent new building development at North

West Bicester Eco Town, Oxfordshire, consisting of 393 homes, community and commercial units

using measured disaggregated domestic load demands from three low-carbon home and national

average commercial demand data in the United Kingdom. The paper is based on the energy system

level of the building development and therefore does not take into account the network topology,

which would be for further studies. The priority of the NDP methodology is local consumption,

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 4 / 19 www.brianazzopardi.eu

while considering the options of (i) demand management (DM), which is the active load shifting of

domestic loads by households, and (ii) energy storage, where the overproduction of electricity is

shifted to periods of net demand that may power electronic devices via direct current (DC) circuits

or energy storage units such as EVs. The technical potential is analysed over the economic situation

of grid infrastructure costs that may be required with high-penetration of PV.

1.2. Outline of the paper

The paper is structured as follows: first, in Section 2, the methodology is described. Afterwards, in

Section 3, the case study and the respective results are presented. Later, in Section 4, the key findings

are discussed through a sensitivity analysis. Finally, in Section 5, the main conclusions are drawn.

2. Methodology

Regional developments for high penetration of PV shall be assessed considering the technical

energy level system of the whole complex. Hence, net demand planning (NDP) based on the

difference of the buildings energy demand and PV generation is proposed.

The methodology may be used to assess the potential of regional developments for high penetration

of PV based on the circumstances of the load demand profiles and grid connection. Additionally, it

may provide insights about the drivers for local consumption. This is achieved by applying the

methodology to multiple scenarios, as in a sensitivity analysis, and recording and analysing the

results.

The methodology framework shown in Fig. 1 is divided into three different levels: (i) household

(domestic) level, (ii) commercial level and (iii) community level. Time-series data and models have

been averaged or calculated for half-hourly time series. The framework is applicable to any site with

available meteorological data and appropriate electric demand data.

2.1. Data input

The data required to perform the methodology comprise:

i. Insolation data - measurements or satellite estimations of ground level and extraterrestrial

insolation, and ground reflectance estimation.

ii. Demand data - disaggregated measurements of domestic low-carbon housing data or

national estimates of demand profiles of commercial units.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 5 / 19 www.brianazzopardi.eu

Fig. 1. Flowchart of the methodology. Net demand planning (NDP).

iii. PV system data - available area or peak power rating for the system.

iv. Energy storage data - battery energy storage is considered with the following parameters,

discharge/charge rate, discharge/charge efficiency and self-discharge.

v. Financial data - grid infrastructure costs.

In this paper, the insolation and demand profiles are taken as half-hourly data series for a year and

it is assumed that insolation and demand are the same every year.

2.2. Data processing

The input data are used to compute the expected grid infrastructure costs, total insolation on the PV

system, and expected generation, load shifting (if any), storage (if any), imports and exports at

household level as well as at the community level. The case study presented in this paper is mainly

based on UK data.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 6 / 19 www.brianazzopardi.eu

The domestic demand data are based on 5 min measured circuit-disaggregated UK low-carbon

households by Carnego Systems Ltd., which were filtered, sorted and averaged in half-hourly time

steps. Meanwhile the commercial demand data are based on hourly national profiles for winter and

summer, modelled on the profile from the Sustainability First [25] and data is assumed in a

simplified form with an average load of 12 VA/m2 which have been extrapolated for the planned

5311 m2 roof area of commercial premises and school.

2.3. Photovoltaic systems modelling

A model for computation of the half-hourly energy output, surplus and deficit of a PV system within

a building environment was developed based on [13]. Subsequently the average power was

computed.

Insolation (I) can be separated into beam (IB) and diffuse (ID) insolation on the earth surface using

the clearness index (kT) method [27]. This process is performed on half-hourly computed with (1),

(2) and (3).

The insolation received on the surface of a PV collector (Ic) is determined by the beam (IBC), diffuse

(IDC), and reflected (IRC) components of the radiation on the collector which is a function of the

shape, face, and tilt angles of the collector, and the use of tracking devices [13]. For this paper, only

fixed flat and tilted PV modules are addressed, as these are the most common development types in

urban areas and the development under consideration in the case study. Therefore, the total

insolation on a collector can be estimated with (4), (5), (6), (7), (8) and (9).

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 7 / 19 www.brianazzopardi.eu

where β is the tilt angle of the collector, ρ is the ground reflectance, α is the solar altitude angle, ϕ

is the latitude, δ is the solar declination, ω is the hour angle, θ is the incident angle between the sun

and the face of the collector, and φS and φC are the solar and collector azimuth angles. All angles are

expressed in degrees.

The half-hour energy in kilowatt hour generated by a PV system (EPV) can be computed with (10).

where Prating is the PV system rating (kWp) and PR is the performance ratio of the PV system.

Therefore the expected time-series energy of the PV system (EPV) within every building demand

load (EL), being domestic household with PV system or commercial PV system, the amount of

surplus energy which can be exported to the grid or stored for a later use (Esurplus), the energy deficit

which would require imports from the grid or load shifting (Edeficit), and the Net Demand Profile

(ENDP) in kilowatt hour per half-hour can be determined with (11), (12) and (13).

Subsequently the average power in W of the half-hourly energy time-series calculations can be

determined with (14).

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 8 / 19 www.brianazzopardi.eu

2.4. Solar fraction

The PV performance can be defined by the term Solar Fraction (SF), the fraction of load met directly

by a PV system [28], given in (15) for an arbitrary time interval [to, tmax].

The index I refers to a general energy imports and the index L refers to a general load demand

respectively. is also known as the ‘loss of load probability’, ‘deficit of energy’ or ‘loss

of power probability’. The SF is also known as ‘autonomy’ or ‘load coverage rate’. These terms are

usually used for stand-alone systems. However, such term also quantifies the reliability in grid-

connected PV systems in respect to a thorough techno-economic analysis. A negative factor implies

imported energy is stored for later use, while a positive factor implies the fraction that the system

contributes directly to the local load, the load matching indicator. For purpose of this paper and

since storing grid energy from ‘unknown’ source is not yet economically feasible, imports for

storage is restricted.

2.5. Options for load matching

The impacts of demand management and energy storage under NDP helps to manipulate the PV

production timing or performance for load matching, without significantly impact on the PV

maximum power delivery. The PV maximum power delivery is an important variable for solar

electricity price and solar electricity environmental benefits, unless there are other non-quantitative

criteria such as aesthetics such as building integrated developments.

2.5.1. Demand management

The most applied demand management (DM) strategies [9] are:

i. peak clipping, contributing towards load peak reductions by for example load shedding

ii. valley filling, contributing to load build-up during off-peak periods

iii. load shifting, combining the previous two strategies into one by moving loads from on-peak

period to off-peak periods.

Load shifting strategy is used in this paper to evaluate demand management impacts both at

household level and community level that applies DM within households.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 9 / 19 www.brianazzopardi.eu

The approach is based on [9], and the procedure is repeated for each chosen time frame of the

production and demand data, in this paper for each day. The domestic load (DL) is divided into

shiftable EDL,sh and fixed EDL,fix parts, where the shiftable part is considered distributable over the

whole analysed time frame.

The NDP scheme applied is EDL(t) − EPV(t), where net PV energy surplus is a negative demand. The

DM strategy fills the valleys of the net demand curve with the shiftable load. An energy level, Eopt,

is computed to fill energy in the valley which is equal to the shiftable energy as in (16).

The new net load pattern, (16) is solved numerically for Eopt, that is iterating Eopt until the left-hand

side of the equation equals the right-hand side. With the discovered Eopt value, the new distribution

of the shiftable energy is (17).

The new net demand curve after the DM can be obtained by adding the new distribution of the

shiftable load to the fixed load (18).

2.5.2. Options for energy storage

The option storing intermittent energy sources such as solar energy is on debate whether it is

economically feasible or not. Here it is wise to acknowledge that a number of energy storage

alternatives may be considered even on the electricity electro-chemical storage level only. Without

going into details beyond the scope of this paper, these alternatives may include small local storage

for electronic as well as now LED lighting devices, and on a community scale the charging of

electric vehicles. The scope of this paper is to evaluate the impact of energy storage for NDP which

goes beyond the economic feasibility of the technology or device in use. Therefore, a realistic

storage model, considering charge/discharge rate and state-of-charge limits as well as self-discharge

is developed on the domestic household that can absorb excess solar energy and used at time of

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 10 / 19 www.brianazzopardi.eu

solar deficit and on the community level that can absorb excess solar energy for transportation

purposes such as the charging of electric vehicles.

The household level energy storage (S) algorithm has priority over exports and imports from the

grid. During surplus of energy, the energy is stored at a charging rate level until maximum capacity

(Scap) and excess energy is exported as expressed in (19).

where h is the building/household number, CR is the charging rate, 0.999 is self-discharge and 0.90

is the energy storage efficiency. Similarly during deficit of energy, which model is only used at

household level, as community storage is not used on the grid, the energy is discharged at a

discharging rate (DR) level until minimum capacity, 20% of State of Charge (SOC), and further

energy deficit is imported as expressed in (20).

3. Case study

An illustrative case study is presented to exemplify the application of the methodology. A schematic

is shown in Fig. 2. Afterwards an extensive case study is used as a sensitivity analysis.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 11 / 19 www.brianazzopardi.eu

Fig. 2. Schematic of the illustrative case study.

3.1. Input data

The insolation data are obtained from satellite measurements [29]. The performance ratio of the PV

system is 85%. Fig. 3 illustrated the measured disaggregated domestic demand data of three UK

low-carbon households by Carnego Systems Ltd., the community demand profile which includes

the commercial load and the PV generation summer profiles. It is clear that there is significant

surplus of energy during the summer period.

3.2. Results

Net load duration curves for the three households and on community level are shown in Fig. 4.

Negative values represent power exported to the grid. Net load duration curves illustrate the

variation of the power transfer at the household and community level grid connection. The greatest

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 12 / 19 www.brianazzopardi.eu

Fig. 3. Average power summer profiles.

imports are plotted in the left and the greatest exports are in the right, therefore giving a clear

relationship between capacity requirements and capacity utilisation.

The plots show the baseline analysis curves (i) no PV systems, (ii) with PV systems only, (iii) with

PV systems and 30% relative Load Shifting on household level, (iv) with PV systems and Energy

Storage and (v) with PV systems, Load Shifting and Energy Storage.

Fig. 4. Net load duration curve.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 13 / 19 www.brianazzopardi.eu

On the community level the household profile proportion was assumed equal and community energy

storage for transportation was considered. The baseline for load shifting is a shiftable load of 30%

of the total household load and for household level energy storage is a 3 kWh energy storage. The

baseline for community level energy storage is 3 MWh which is equivalent to around 150 battery

electric vehicle (BEV) assuming a storage of 20 kWh per vehicle. Within a whole community, BEVs

may be available for the grid to utilities during periods of high insolation, low demand and high

demand, low insolation. This may be achieved through different incentivised schemes but also due

to patterns of work, shopping and home activities within the region.

The results are similar in the three houses, with the business as usual (no PV) case presenting the

highest imports and of course no exports. Adding PV systems reduces moderately the imports but

adds massive overproduction. The implementation of load shifting and energy storage levels out

those values, achieving significant reductions in energy connections. The best results are obtained

with a combination of load shifting and energy storage.

4. Sensitivity analysis

Since future developments in technology and scale-up are likely to affect the components within a

utility grid, the sensitivity of the maximum power transfer, was analysed. The community grid

power transfer of 890 kW was taken as the baseline for this sensitivity analysis, assuming possible

30% load shifting on the household level, 3 kWh energy storage per household with

charge/discharge rate at 15%, 3000 kWh community slow charging (6% charge rate) per storage for

use other than within the grid such as in BEV for transportation. The sensitivity analysis was

performed by adjusting the variables from −80% to 80% from the baseline as shown in Fig. 5.

The factors which have the most significant impacts on the Maximum Grid Power Transfer are the

community storage and its rate of charge, as shown in Fig. 5. This is mainly due to the fact that the

community storage, exhibited by the EVs charged during the day, would have absorbed most of the

excess energy as well as any exaggerated spikes of excess energy. Meanwhile the community charge

rate baseline was close to optimal as reducing or increasing it would not have used the community

storage levels matched with the load and PV outputs. The load shifting and community energy

storage have linear effect on the Maximum Grid Power Transfer, as these factors are directly

correlated with energy transfers.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 14 / 19 www.brianazzopardi.eu

Fig. 5. Significant variables effecting grid power transfer.

In practice, load shifting may be difficult to achieve higher than 30% and high capacity of

community energy storage may only be useful in few days which spike over the imports as shown

in Fig. 6. Though, it is suggested that about 40% of total consumption by domestic appliances may

be dynamic demand [30]. This initiates the economic debate beyond the scope of this paper whether

grid storage is necessary or not. As experience develops in this field, a more sophisticated sensitivity

analysis could be done, for example, employing statistical methods.

5. Conclusion

The paper has illustrated a feasibility assessment for high penetration of distributed PVs based on

NDP. A comprehensive technical feasibility of integrating PV systems on all houses and commercial

buildings within an imminent new build development of 393 homes in North West Bicester Eco

Town, Oxfordshire, was performed. The analysis showed that a mix of technical initiatives, load

shifting, energy storage and energy storage for other use, are required for high penetration of PV.

The baseline results are summarised in Fig. 7.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 15 / 19 www.brianazzopardi.eu

Fig. 6. Snapshot of the whole year grid power transfer time-series at community level.

Fig. 7. Summary of baseline case scenario results.

Since the current electricity tariffs do not incentivise demand management or energy storage, Fig. 7

illustrates the current scenario on the electricity tariffs based on the current data from SSE Standard

Energy at 13.42 p/kWh UK in 2012 [31] and 4.5 p/kWh for exported PV electricity [32]. The night

electricity or the economy 7 were not considered as the electricity load profiles would require to

reflect this while these is also the possibility for new electricity tariffs being put in place within the

NW Bicester Development to reflect the surplus generation from PV micro-generators.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 16 / 19 www.brianazzopardi.eu

For large developments with high penetration of PVs, additional copper cable, sub-stations and local

grid infrastructure reinforcement may be required to accommodate the expected electricity output.

This is a costly investment and therefore acts as disincentive to the extensive use of PVs. A typical

value for annuitised grid infrastructure capital costs is taken at £7.07/kW/year [33]. Annuatised costs

are usually given in £/kW/year and converted to £/kVA/year using appropriate power factors. The

£7.07/kW/year is an illustration at a capital discount rate of 6.5% and a notional investment lifetime

of 40 years, for a £100/kW lifetime cost. Fig. 7 shows the estimated annuatised infrastructure cost

which is lowered at every option for load matching.

Meanwhile, in this case study, the domestic loads of the low-carbon houses and commercial loads

had some correlation with PV output, as shown in Fig. 3. This led to the Grid Maximum Power

Transfer for imports reduction while the increase in export was not above the imports levels. Hence,

a reduction in infrastructure may occur only with PVs. However this may not always be the case,

especially if the homes are standard old types with no energy efficiency and habitants did not have

correlation of demand during the day time.

The load switching may be achieved both through behaviour and technical means. Behaviour means

that the system provides intelligent prompts to the householders based on time of year, weather

(forecast and actual), tariff rates etc to recommend / encourage effective load switching. This may

be via for example the touch-screen, email, facebook and twitter. Technical means are where

switching is completed automatically. Currently this is often to heat hot water, but as appliances

become more intelligent this may include other dynamic loads of switchable appliances.

Costs for typical Load Management Units, such as Shimmy from Carnego Systems Ltd., are

currently approximately £650 to £850 per home depending on functionality. The cost is expected

this to fall in the future, although it may be seen from Fig. 7 that investment per household may be

easily achieved in about one year. This does assume that internet connection is available in the house

for full functionality. The costs include all of the functions (transport, community info etc.) as well

as the energy functions. The cost of the energy functions only would be less but it is hard to break

out separately. Installation is also required. If this is done as part of the house build then estimated

at approximately £90 per house. If a single retro-fit to an existing house then £165.

Using simulation models, the results show potential for setting optimal demand in residences, which

is then compared to business as usual models on electrical energy demand. With a sensitivity

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 17 / 19 www.brianazzopardi.eu

analysis, the paper discusses also related energy policy aspects such as options for PV orientation

aspects, the potential of Direct Current electrical installations and schemes that encourage load

shifting. The paper concludes, within the context of the case study and all relevant options

considered in this paper that, contrary to what is commonly believed high penetration of PVs in a

development area is feasible and may even reduce grid infrastructure costs.

Acknowledgements

The work has been partly carried out under the project “Feasibility Study for Optimising Demand

in Residential Developments with High Density Photovoltaic Arrays” funded by the Technology

Strategy Board Smart Power Distribution and Demand Programme. The authors would like to thank

William Box at Carnego Systems Ltd. for providing complementary measured low-carbon

household disaggregated data and for the coordination in the project. The authors would also like to

thank the other project partners Nicole Lazarus at Bioregional and Phil Shadbolt at Zeta Group.

References

[1] IEA. PVPS report: a snapshot of global PV 1992-2012. International Energy Agency; 2013.

IEA-PVPS-T1-22:2013.

[2] EURELECTRIC. Active distribution system management: a key tool for the smooth integration

of distributed generation e findings and recommendation. EURELECTRIC; 2013.

[3] National_Grid. Operating the electricity transmission networks in 2020 initial consultation.

National Grid; 2009.

[4] PV-UPSCALE. Publications review on the impacts of PV distributed generation and electricity

networks; 2007.

[5] Jenkins N. Embedded generation. The Institution of Engineering and Technology, Michael

Faraday House, Six Hills Way, Stevenage SG1 2AY, UK: IET; 2000.

[6] Conti S, Raiti S. Probabilistic load flow using Monte Carlo techniques for distribution networks

with photovoltaic generators. Sol Energy Dec. 2007;81(12):1473e81.

[7] Povlsen AF. Impacts of power penetration from photovoltaic power systems in distribution

networks. International Energy Agency; 2002. IEA-PVPS-T5-10: 2002.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 18 / 19 www.brianazzopardi.eu

[8] Davda AT, Azzopardi B, Parekh BR, Desai MD. Dispersed generation enable loss reduction and

voltage profile improvement in distribution network case study, Gujarat, India. Power Syst IEEE

Trans May 2014;29(3):1242e9. http:// dx.doi.org/10.1109/TPWRS.2013.2292117.

[9] Widen J, Wackelgård E, Lund PD. Options for improving the load matching capability of

distributed photovoltaics: methodology and application to high latitude data. Sol Energy Nov.

2009;83(11):1953e66.

[10] Azzopardi B, Mutale J. Analysis of renewable energy policy impacts on optimal integration of

future grid-connected PV systems. In: 2009 34th IEEE photovoltaic specialists conference (PVSC),

Philadelphia, United States; 2009. pp. 000865e70.

[11] Azzopardi B, Mutale J. Smart integration of future grid-connected PV systems. In: 2009 34th

IEEE photovoltaic specialists conference (PVSC), Philadelphia, United States; 2009. pp. 002364e9.

[12] Azzopardi B, Mutale J, Kirschen D. Cost boundaries for future PV solar cell modules. In: IEEE

international conference on sustainable energy technologies, 2008. ICSET 2008, Singapore; 2008.

pp. 589e94.

[13] Martinez-Cesena EA, Azzopardi B, Mutale J. Assessment of domestic photovoltaic systems

based on real options theory. Prog Photovolt Res Appl Mar. 2013;21(2):250e62.

[14] Azzopardi B, Mutale J. Optimal integration of grid connected PV systems using emerging

technologies. In: 24th European photovoltaic solar energy conference, Hamburg, Germany; 2009.

pp. 3161e6.

[15] Azzopardi B. Green energy and technology: choosing among alternatives. In: Hossain J,

Mahmud A, editors. Renewable energy integration. Singapore: Springer; 2014. pp. 1e16.

[16] Azzopardi B, Martinez-Cese~na EA, Mutale J. Decision support system for ranking

photovoltaic technologies. IET Renew Power Gener 2013;7(6): 669e79.

[17] Breukers SC, Heiskanen E, Brohmann B, Mourik RM, Feenstra CFJ. Connecting research to

practice to improve energy demand-side management (DSM). Energy Apr. 2011;36(4):2176e85.

[18] Finn P, Fitzpatrick C, Connolly D. Demand side management of electric car charging: benefits

for consumer and grid. Energy Jun. 2012;42(1):358e63.

Accepted Author Manuscript

How to Cite: Azzopardi, Brian, and Alejandro Gabriel-Buenaventura. “Feasibility Assessment for High Penetration of

Distributed Photovoltaics Based on Net Demand Planning.” Energy 76 (November 1, 2014): 233–40.

doi:10.1016/j.energy.2014.06.052.

Received 12 November 2013, Revised 13 June 2014, Accepted 13 June 2014, Available online 9 July 2014

[email protected] 19 / 19 www.brianazzopardi.eu

[19] Atikol U. A simple peak shifting DSM (demand-side management) strategy for residential

water heaters. Energy Dec. 2013;62:435e40.

[20] Ji L, Niu DX, Huang GH. An inexact two-stage stochastic robust programming for residential

micro-grid management-based on random demand. Energy Apr. 2014;67:186e99.

[21] Torriti J. Price-based demand side management: assessing the impacts of time-of-use tariffs on

residential electricity demand and peak shifting in Northern Italy. Energy Aug. 2012;44(1):576e83.

[22] Li XH, Hong SH. User-expected price-based demand response algorithm for a home-to-grid

system. Energy Jan. 2014;64:437e49.

[23] Doostizadeh M, Ghasemi H. A day-ahead electricity pricing model based on smart metering

and demand-side management. Energy Oct. 2012;46(1): 221e30.

[24] Prüggler N, Prüggler W, Wirl F. Storage and demand side management as power generator's

strategic instruments to influence demand and prices. Energy Nov. 2011;36(11):6308e17.

[25] Hesmondhalgh S. GB electricity demand e 2010 and 2025. Initial brattle electricity demand-

side model e scope for demand reduction and flexible response. Sustainability First 2012.

[27] Masters GM. Renewable and efficient electric power systems. John Wiley & Sons; 2004.

[28] Azzopardi B. Integration of hybrid organic-based solar cells for micro-generation. PhD.

Manchester, United Kingdom: The University of Manchester; 2010.

[29] Menard L, Wald L. SoDa Service: knowledge in solar radiation. In: Armines/ MINES

ParisTech. Sophia-Antipolis, France: Centre Energ_etique et Proc_ed_es; 2010.

[30] Pina A, Silva C, Ferr~ao P. The impact of demand side management strategies in the

penetration of renewable electricity. Energy May 2012;41(1):128e37.

[31] SSE electricity prices. [Online]. Available: http://www.southern-electric.co.uk/ OurPrices/;

[accessed 04.01.13].

[32] FIT tariffs. [Online]. Available: http://www.fitariffs.co.uk/eligible/levels/future/ ; [accessed

04.01.13].

[33] Williams P, Andrews S. Distribution network connection: charging principles and options. Ilex

Energy Consulting, K/EL/00283/REP e DTI/Pub URN 02/1147.