03/11/17 1

An Investigation Into Solar Cells:

“What realistic contribution could roof-mounted photovoltaic cells make towards domestic electricity consumption
in Exeter and what would be the impact on greenhouse gas emissions from the city?”
by Thomas Turner

Abstract
Relying on sound assumptions and data from reputable sources, the number of houses and flats are considered and the
total energy that could be produced from their roof area is calculated. It is found that solar cells with a 16.2% efficiency
could produce between 56.54% and 87.06% of Exeter’s electricity demand depending on the geometry of the houses.
When the efficiency of the cells is increased to 20.1%, it is found that Exeter could generate 70.09% to 107.93% of its
energy demand. Using 16.2% efficient solar cells would reduce Exeter’s CO2 emissions by 41.2 - 63.4 kilotonnes per year.
The project is estimated to cost £251.3 million in total, but could be reduced as low as £200.7 million when considering a
discount due to the volume of the order. An investment strategy to fund the scheme is proposed that estimates the scheme
could be funded with an initial investment of just £80 million over 15 years, paid in instalments. The discounted scheme
would be complete by year 30.

I. I NTRODUCTION The geometry of flats is different to the geometry of houses.

The UK is heavily reliant on nuclear and wind power to In this report, a set of flats is modelled as a 4 storey building
supply its electricity, accounting for 21.2% and 11.5% respec- with the same area per floor as a house. Under this assumption,
tively in 2016 [1]. Solar cells on the other hand accounted for there is one roof for every four flats in a building.
only 3.4% of total electricity generated [1]. In 2011, Exeter had 49,242 households. 76% of household
This report assesses the impact on electricity demand, if spaces in Exeter are houses and 23% are flats [7]. 1% of
solar photovoltaic cells are mounted onto the roof of every households are caravans and will be ignored in this report
domestic household in Exeter. It discusses the feasibility of assuming they are too small to mount photovoltaic cells onto.
the project and the effect on greenhouse gas emissions from The final 1% of properties are flats in commercial buildings
the city. and will be ignored, as this investigation is only interested in
domestic consumption.
The number of households has increased in the last 6 years.
II. M ETHOD & R ESULTS Projects such as those at Cranbrook have built 1000 new
A. Domestic Roof Area In Exeter houses from 2011 to 2015 [8]. In this report it is assumed
The average size of a 3 bed house in the South West is 1500 houses and 500 flats have been built in Exeter since the
87m2 and the average size of a 1 bed house in the South 2011 census. Thus, it is estimated there are 38,924 houses and
West is 46m2 [2]. This study by RIBA estimates that the ratio 11,826 flats in Exeter in 2017.
of 3 bed homes to 1 bed homes in the South West is 5:11. Using the house and flat geometry described earlier, there
Using this ratio, the average house size in the South West are 41,880 household roofs in Exeter, with a power producing
can be estimated as 74m2 (Appendix A). The UK loft height area of 15.50m2 to 23.87m2 . The total area of domestic power
is in the region of 1.9-2.4m [3]–[5]. Using this range, the producing roof space is 649×103 m2 to 1000×103 m2 .
average roof area is calculated in Appendix B. Using the lower
bound for loft height, an estimate for average house roof area
B. Total Energy Production From Photovoltaic Cells
would therefore be 38.75m2 , using the higher bound this value
becomes 39.78m2 . The average solar irradiance received by Exeter each month
However, not all of the roof space is suitable for photo- is summed [9], and the yearly figure is calculated to be 1,060
voltaic cells. A clearance of 0.5m is required between the kWh/m2 . The total retrievable irradiance on roofs per year is
edge of a roof and the edge of a solar panel and 0.2m between 688GWh to 1,060GWh, depending on the geometry of the
chimneys and other obstacles [6], making the number of panels roof.
which can be installed reliant on the shape of the roof (see The average efficiency of 14 of the UK’s most popular solar
Figs. 1 and 2 on pg. 2). Furthermore, it is assumed only a panels [10] was found to be 16.2%, with a range of 14.1% to
proportion of sunlight will reach the roof. The calculation takes 20.1%. Using the average solar panel efficiency, Exeter could
this into account by estimating that the area of roof producing produce 111.6×106 kWh year−1 to 171.8×106 kWh year−1
power is 40% to 60% of its total area. Thus, the total power of electricity.
producing area per roof of a house in Exeter is estimated at The average energy consumption per house last year in the
15.50m2 to 23.87m2 . The minimum value is obtained by using UK was 3.89MWh [11]. There are a 50,750 households in
the smallest loft height, 1.9m and smallest percentage of power Exeter, giving a total usage 197.4×106 kWh year−1 .
producing roof area, 40%. The maximum value is obtained This means that using the average efficiency of a photo-
using the highest bounds for these parameters. voltaic cell Exeter could generate between 56.54% (roof area
03/11/17 2

38.75m2 and 40% usable) to 87.06% (roof area 39.68m2 and roof area producing electricity. A more detailed analysis of
60% usable) of its domestic electricity demand. error will occur in section III.
By increasing the efficiency of the average solar panel from
16.2% to 21.1% this range increases to 70.09% - 107.93%; C. Reduction In Emissions
thus, if the dimensions of the houses in Exeter are large enough The extra energy produced from photovoltaic cells would
it is possible to produce a 7.93% excess of domestic electricity. reduce the need to generate energy from fossil fuels, signif-
A full table of results can be seen in Fig. 3. icantly reducing Exeter’s CO2 emissions. In 2016, the UK’s
The range in electricity supplied for 21.1% efficiency cells electricity was produced from CO2 emitting fuel sources as
is 30.52%, which is high. This arises as the percentage of follows: 42.4% from gas, 9.1% from coal and 2.9% from oil
roof producing electricity is varied from from 40% to 60%. [1]. The remaining sources did not add to CO2 emissions. Each
If the usable percentage of roof area is set to 50% and only fuel produces a different amount of CO2 per unit energy: coal
house geometry is varied, the range decreases to just 1.88%. 0.870kg kWh−1 , oil 0.650kg kWh−1 and gas 0.487kg kWh−1
Therefore, the dominant source of error is the percentage of [12].

Fig. 1. A trapezium shape roof with area 40m2 which can fit ten, 1.65m x 1.00m solar panels.

Fig. 2. A rectangular shape roof with area 40m2 which can fit fourteen, 1.65m x 1.00m solar panels.
03/11/17 3

Fig. 3. Energy production per year as a proportion of total energy demand. The parameters: average roof area, % of producing power and cell efficiency are
varied.

would account for a 30% reduction in electricity production

per year.
2) Geometry of housing: In reality, the shape and number
of floors of the houses are not the same. Assuming the shape
to be cubic maximises the size of the roof area. Furthermore,
it is hard to estimate how many flats there are per roof in
Exeter. The size of blocks range massively, large blocks may
have over 10 floors whereas small ones may have only two.
More detailed surveying of house geometry must be completed
to improve the accuracy of total roof area.
3) Percentage of roof that generates power: This percent-
age will vary greatly depending on the geometry of the housing
and the direction it faces. The effect of roof shape on the
Fig. 4. Calculated reduction in CO2 emissions for varying cell efficiency and number of solar panels installed is shown in Fig. 1 and Fig. 2.
power producing roof area. In these diagrams it is shown that a rectangular roof of 40m2
can fit fourteen solar panels, whilst a trapesium shaped roof
of equal area can only fit ten. The error in this comparison is
The calculated drop in CO2 with varying cell efficiency 4 solar panels. Using the average panel dimensions [19] as in
and power producing area per house is displayed in Fig. 4. Figs. 1 and 2, this would account for a 6.60m2 reduction in
Using the average cell efficiency would account for a 41.2 - area producing power.
63.4 kilotonne drop in CO2 emissions per year (Appendix C). There is also a random error in this calculation due to
Increasing the photovoltaic efficiency to 20.1% would decrease shade from chimneys, other buildings, trees and raised areas
emissions by 51.0 - 78.6 kilotonnes per year. Using the highest on the roof reducing the area receiving sunlight. This random
estimated decrease, this could drop the UKs CO2 emissions error is difficult to quantify and should be measured in future
by up to 0.016%. The change in sulphur dioxide and nitrous surveying.
oxides would be negligible.
B. Feasibility
D. The Cost The prices used in the calculation are quotes for individual
2
The cost of 21m of solar panels and installation is esti- orders of solar cells. In practice, for a project of this size a
mated at £4,000 to £7,000 [14]–[18]; the most common quote supplier would put a discount on each solar panel due to the
is £6,000 per roof. Considering all 41,880 roofs, this project volume of orders and scale of the project. Depending on the
would cost £251.3 million. number of businesses competing for the contract, it is plausible
this could reduce the price per unit by up to 20%; reducing
III. A NALYSIS the cost of the project by up to £50.6 million.
This project would require a huge amount of funding from
A. Sources of Error the government. The scale of the project is realised when it
There are three main sources of error in this report. They is noted that the Exeter City Council received £5.82 million
come from the efficiency of the cells, the percentage of roof from the government in 2016 [20], £194.9 million short of the
that can be used to generate power and the geometry of the discounted cost of the project.
housing. Even with a discount, for the government to fund a city-
1) Efficiency of cells: The error in efficiency is the dif- specific project costing £200.7 million is a huge investment.
ference between the maximum and minimum cell efficiency, To put this in perspective, in 2016 the UK’s budget for Major
which is 6%. Using 14.1% efficient cells instead of 20.1% cells Projects was £24.6 bilion [21], making the Exeter solar project
03/11/17 4

Fig. 5. Investment strategy for the project. Values are in £million. PH1 = Phase 1, PH2 = Phase 2 etc. With only an £80 million investment over 15 years
£224 million can be spent by year 30.

1% of the annual budget - whilst only benefiting 0.2% of the Using the investment scheme laid out in Fig. 5 would reduce
UK’s population [7], [22]. This would come at a time when this to £0.57 million per MW, as only £80 million capital
the government has cut the solar feed-in tariff subsidy by 65% is required, however this includes the significant trade off of
[23]. spreading the cost up to 35 years.
The estimated cost of the Exeter project is validated, as it
C. Funding The Project has a similar price per MW of electricity with the UK’s current
To roll out such a scheme, it would be necessary to fund it largest solar farm.
over many years. It may be wise to fund panels with a 20.1%
efficiency. These would see a quicker recuperation in initial IV. C ONCLUSION
investment and allow revenue from the panels to be reinvested The model in this report estimates that the cost of funding
into more panels sooner. the scheme is £256 million, but could be reduced to £200.7
One plausible investment strategy is shown in Fig. 5. This million when high-volume discounts are considered. The like-
strategy invests four instalments of £20 million over 15 years lihood of this project receiving funding in one instalment is
into the project. After twenty years the base-investment would extremely low, but the investment plan laid out in Fig. 5 is a
only be £80 million, but if the revenue of each solar cell more plausible way to spread the cost over 30 to 35 years. This
installed is re-invested into buying more cells, the scheme would be a slow, long-term project, but would greatly benefit
would accrue the non-discounted price of £256 million re- the city - supplying 70.09% to 107.93% of the city’s electricity
quired by year 35 and the discounted price by year 30. This provided 21.1% efficient panels are installed. As cells become
model assumes a revenue of £400 per roof per year and more efficient in the future, this percentage could increase even
the £20 million investment can purchase 4,000 solar panels higher. A reduction in CO2 of 51.0 - 78.6 kilotonnes every year
(£5000 per roof). However, this is a very simple model and would reduce UK emissions by up to 0.016% and hopefully
doesn’t take into account the depreciating value of money the scale of the project would inspire other households across
with time. Including a discount rate would increase the length the UK to invest in panels too.
of time until the required £200.7 million - £251.3 million is
raised. Further inaccuracy is realised if the investment strategy R EFERENCES
accounted for the 1% decrease in efficiency of solar panels
[1] Department for Business, Energy & Industrial Strategy, UK Energy
each year [24], [25] reducing revenue per roof in later years Statistics 2016 & Q4 2016, p. 9 (March 2017)
significantly. [2] RIBA, The case for space: The size of England’s new homes, p. 23
It should be carefully considered how to choose which [3] [Link]
(accessed 25/10/17)
houses receive solar panels first, to avoid complaints. A fair [4] [Link]
manner may be a ‘post code lottery’. However, installing what-is-the-minimum-height-size-needed-in-my-loft-to-convert-it
panels on larger roofs first would accrue returns quicker; whilst (accessed 25/10/17)
[5] [Link]
it could be seen as unfair, installing panels on optimised houses loft-conversion-guide-in-depth-information/
first could speed the project along considerably. on-how-to-successfullytackle-a-loft-conversion (accessed 25/10/17)
[6] Poweri, Guide to measuring your roof for solar,
D. Comparing With Other Projects [Link]
[Link] p.
The largest solar project in the UK is the 300 acre ‘solar 5 (accessed 28/10/2017)
energy farm’ in Norfolk. The project cost £50 million and can [7] Census Profile, Exeter City, p. 10 (2011)
[8] [Link]
produce 32 MW of electricity [26], which is £1.56 million per hundreds-of-homes-complete-at-devons-sustainable-new-community
MW produced. (accessed 26/10/2017)
If the Exeter project was paid for upfront, the price per MW [9] [Link]
(accessed 25/10/17)
of electricity would be £1.78 million - including the discount [10] [Link]
this could be reduced to £1.42 million. how-much-should-i-expect-pay (accessed 25/10/17)
03/11/17 5

[11] L. Waters, Energy Consumption in The UK, p.21 (July 2017) A PPENDIX C
[12] [Link] (accessed 26/10/2017) C ALCULATING CO2 E MISSIONS D ROP
[13] Department for Business, Energy & Industrial Strategy, 2015 UK
Greenhouse Gas Emissions, Final Figures, p. 8 (Feb 2017)
[14] [Link]
#home-cost (accessed 26/10/2017)
[15] [Link] (accessed 26/10/2017)
[16] [Link] (accessed 26/10/2017)
[17] [Link]
(accessed 26/10/2017)
[18] [Link]
what-is-the-installation-cost-for-solar-panels (accessed 26/10/2017)
[19] [Link] (accessed Fig. 6. CO2 drop for 16.2% efficiency cells for minimum roof area.
26/10/17)
[20] Revenue and Capital Estimates 17-18 p. 16
[21] Infrastructure and Projects Authority, ‘Annual Report on Major Projects
2016-17’, p. 4 (2017)
[22] Office for National Statistics, 2011 Census: ‘Population Estimates for
the United Kingdom’ (March 2011)
[23] [Link] (accessed 26/10/17)
[24] [Link]
ArticleID/7475/[Link] (accessed
26/10/17)
[25] [Link] (accessed
26/10/17)
[26] [Link] (accessed Fig. 7. CO2 drop for 16.2% efficiency cells for maximum roof area.
29/10/17)

A PPENDIX A
C ALCULATING AVERAGE H OUSE S IZE

Ratio of 3 bed (87m2 ) to 1 bed houses(46m2 ) is 5:11,

therefore,
Fig. 8. CO2 drop for 20.1% efficiency cells for minimum roof area.
5(87) + 11(46)
Average house area = = 74m2 (1)
5 + 11

A PPENDIX B
C ALCULATING ROOF AREA

Assuming most houses have two floors, the average area of

a house in the South West is 37m2 . Assuming the living√space
in each house to be a cube, the roof width is simply 37m. Fig. 9. CO2 drop for 20.1% efficiency cells for maximum roof area.
The UK loft height is in the region of 1.9-2.4m [3]–[5].

√
q
2
Min. Roof Length = 37 + 1.92 = 6.37m (2)

√
q
2
Max. Roof Length = 37 + 2.42 = 6.54m (3)

√
Min. Roof Area = 6.37 × 37 = 38.75m2 (4)

√
Max. Roof Area = 6.54 × 37 = 39.78m2 (5)