Rooftop solar power in the UK – real world data

Twelve building-integrated PV (BIPV) projects in Britain were monitored under a UK government contract. Here are the cost and performance data.

St. Mary’s Church Hall, Osterley, London.	Terry Haigh

In my previous post, I had a go at estimating the maximum amount of solar power that could fit on England’s roofspace. It turns out there’s enough roofspace for all our summertime electricity use, but only about a fifth of what’s needed for wintertime needs.

This time round, I’ll take a look at the real world. What is the performance of rooftop solar power in the UK today?

In the industry jargon, rooftop solar is often called “building-integrated photovoltaic” (BIPV) technology. “BIPV” includes wall-mounted and window-mounted solar panels as well. BIPV technology has had a recent large-scale field trial in the UK ^{[1] [2] [3]}. BIPV installations on twelve large buildings had their performance monitored for about 24 months. Monitoring started mid-2003. The projects were spread across the UK from Cornwall to the Hebrides.

The field trial was conducted for the DTI (now BERR). It provides a useful check on how modern solar photovoltaic technology operates under real world conditions in the UK. The study also provides data on the costs of the projects. Table 1 lists the power data for the twelve projects.

Table 1.  Performance of Large Scale Field Trial PV Systems

Project Name	Size (a)	Annual Output	Mean Power	Mean Power/ Peak Power
	kWp	MWh/year	kW
Gaia Energy Centre	63.0	56.1	6.4	10.2%
Delabole, Cornwall
St. Mary’s Church Hall	29.6	18.7	2.1	7.2%
Osterley, London
The Columba Centre	19.7	8.1	0.9	4.7%
Islay, Scotland
West Oxfordshire District Council	23.1	10.8	1.2	5.4%
Witney, near Oxford
University of Gloucestershire	64.4	37.9	4.3	6.7%
Gloucester
Cotswold Water Park	51.0	42.8	4.9	9.6%
South Cerney, Cotswolds
The Insolvency Service	25.4	11.8	1.3	5.3%
Bloomsbury, London
OpTIC Centre	85.3	60.1	6.9	8.1%
St Asaph, North Wales
Birmingham City Council	102.1	44.0	5.0	4.9%
Birmingham
University of East Anglia	33.9	22.7	2.6	7.6%
Norwich, Suffolk
Belfast Education and Library Board	46.5	28.8	3.3	7.1%
Belfast, Northern Ireland
North Devon District Council	56.6	39.3	4.5	7.9%
Barnstaple, North Devon
Data are derived from Table 4.1 of Ref.[2]. Notes: (a)  The size of the project is the rated peak capacity.

The first point is that the mean power output is between 10 and 20 times lower than the rated “peak” power. That’s because of the way that the rated peak capacity of a solar panel is defined. I’ve covered this in the post “Insolation, and a solar panel’s true power output”. A solar panel’s peak output power is defined as its power when it is illuminated by 1000 Wm⁻² of solar radiation. The actual insolation is typically several times lower than this. For example, the annually averaged insolation for London ^[4] is 109 Wm⁻², and for Edinburgh it is 94 Wm⁻². That accounts for much of the difference between mean power and peak power. System downtime and shading by adjacent structures account for most of the rest.

Costs

Cotswold Water Park Visitor Centre. Halcrow

The report lists the costs of all the projects. That large set of cost data is a useful thing to have, because costs often stay confidential, and information on true project costs can be hard to find. On the downside, these costs are now getting a bit out of date, so these figures have to be seen as a “snapshot” of the time when these projects were undertaken, circa 2002–4.

The key figures are the price per kilowatt-hour of electricity and the financial payback times of the projects. (The energy payback ratio is covered in this post). For the cost per kilowatt-hour, I’ve assumed a 25 year operating life for the PV installations. The cost is derived from the capital outlay alone. It doesn’t include ongoing maintenance and operating costs, and ignores interest on the capital.

I’ve assumed an electricity price of 11.4p/kWh to calculate the payback times. That was the average electricity price to UK domestic consumers in 2008 ^[5]. Table 2 lists the cost data for the twelve projects. I’ve highlighted the highest and lowest payback times.

Table 2.  Costs of Large Scale Field Trial PV Systems

Project Name	Size (a)	Annual Output	Project Cost	Cost per peak W	Cost per  kWh (b)	Payback  time (c)
	kWp	MWh/year	£	£/Wp	p/kWh	years
Gaia Energy Centre	63.0	56.1	288,100	4.6	20.6	45
Delabole, Cornwall
St. Mary’s Church Hall	29.6	18.7	122,236	4.1	26.2	57
Osterley, London
The Columba Centre	19.7	8.1	274,190	13.9	135.0	296
Islay, Scotland
West Oxfordshire District Council	23.1	10.8	140,253	6.1	51.8	114
Witney, near Oxford
University of Gloucestershire	64.4	37.9	428,860	6.7	45.3	99
Gloucester
Cotswold Water Park	51.0	42.8	265,000	5.2	24.8	54
South Cerney, Cotswolds
The Insolvency Service	25.4	11.8	318,760	12.6	108.2	237
Bloomsbury, London
OpTIC Centre	85.3	60.1	549,972	6.4	36.6	80
St Asaph, North Wales
Birmingham City Council	102.1	44.0	417,852	4.1	38.0	83
Birmingham
University of East Anglia	33.9	22.7	345,613	10.2	61.0	134
Norwich, Suffolk
Belfast Education and Library Bd.	46.5	28.8	284,451	6.1	39.5	87
Belfast, Northern Ireland
North Devon District Council	56.6	39.3	289,100	5.1	29.4	65
Barnstaple, North Devon
Data are derived from Table 2.2 and Table 4.1 of Ref.[2]. Notes: (a)  The size of the project is the rated peak capacity. (b)  Cost is for capital outlay alone, assumes a 25 year operating life and 0% interest on the capital outlay. (b)  Payback time assumes an electricity price of 11.4p/kWh. Ref.[5].

The most cost-efficient installation was at the Gaia Energy Centre in Cornwall. The least cost-efficient was at the Columba Centre on Islay in the Hebrides.

The Gaia Energy Centre, in Delabole, Cornwall, during construction of the solar roof. This solar roof provides the lowest-cost electricity of the projects monitored. Halcrow

The Gaia Energy Centre project used a low cost photovoltaic product, and the system was installed in a part of the country – Cornwall – with high insolation. It has a payback time of 45 years.

The Islay project used an exceptionally expensive solar panel product called the Atlantis Sunslate™. This product has a cosmetic finish designed to resemble a roof slate, and is used where architectural and aesthetic considerations are paramount. The combination of an exceptionally expensive product, a location with exceptionally low insolation, and an east-west roof created a triple-whammy for this project, resulting in a financial payback time of 296 years.

At the time these projects were commissioned, the costs of photovoltaic solar were still well above grid parity, even on the best UK sites, and payback times were over a factor of two greater than the expected lifetimes of the systems. On semiconductor cost curves however, this is a bridgeable gap. There is light at the end of the tunnel for solar PV in Britain.

More Information:  Domestic Photovoltaic Field Trials

The study above looked at large scale BIPV systems, but an earlier DTI field trial looked at domestic photovoltaic installations ^[6]. There were 33 sites in that study, spread across the UK. Here’s a summary of the key numbers.

The Pinehurst Estate, Liverpool.

System costs averaged £7.28 per watt-peak, and the average cost of the electricity worked out at 47.5p/kWh, assuming a 25 year system lifetime. The projects were commissioned between 2002 and 2005, so the prices are somewhat dated.

The “yield” of a photovoltaic system is the amount of energy it produces in a year divided by the rated peak capacity. The units of yield are usually kilowatt-hours per kilowatt-peak. This quantity is just another way of expressing the ratio of mean power-to-peak power.

Yields on the domestic PV systems in the field trial varied from below 400 kWh/kW_p to over 900 kWh/kW_p, but the largest number of sites was in the yield range 701–800 kWh/kW_p. This range of yields is equivalent to a mean power-to-peak power ratio of 8.0%–9.1%, about what you’d expect for UK insolation levels.

When Does Solar PV Become Cost-Effective in the UK?

At what price point will solar PV become cost-effective in the UK?

The cost of solar panels is generally quoted as the cost per peak watt. We have to know the panels’ operating environment – the average insolation where they are installed – to work out what the payback time and electricity cost per kWh are going to be. Those are the numbers we’d need to have to assess cost-effectiveness.

From the field trial data above, a well installed solar panel in southern England, free from shading and with high reliability, can expect a mean/peak power ratio of over 7%, equivalent to an annual yield of over 610 kWh/kW_p. If we adopt an electricity price to the consumer of 11.4p/kWh (that’s the UK average for 2008), and if we require a 5 year payback time, the breakeven cost of solar PV, stated as the total installation cost per peak watt, is  £0.35/W_p .

How do current costs compare with that breakeven value? There are solar PV manufacturers now claiming manufacturing costs below $1/W_p. From the UK domestic field trial data ^[6], the average cost of installing a system (module plus electrical) was £1.24/W_p, but the minimum cost was just £0.13/W_p, so installation costs need not stand in the way of cost-effective solar in the UK.

According to the UK Energy Saving Trust ^[7], current UK prices for domestic rooftop PV are about £5–£7.50/W_p, 14–20 times higher than the breakeven value. Given that some manufacturers are quoting manufacturing costs of $1/W_p, and installation costs can be as low as £0.13/W_p, I’m having trouble understanding the spread between the costs of manufacturing and installating the systems and the price charged to the UK consumer.

References

1. Large-Scale Building Integrated Photovoltaics Field Trial:  First Technical Report – Installation Phase, URN Number 04/1955, BERR, 2004  (WebCite cache)
2. Large-Scale Building Integrated Photovoltaic Field Trial:  Second Technical Report – Monitoring Phase, URN Number 07/1316, BERR, 2007  (WebCite cache)
3. Large-Scale Building Integrated Photovoltaic Field Trial:  Third Technical Report – Case Studies, URN Number 07/1376, BERR, 2007  (WebCite cache)
4. Release 3 NASA Surface Meteorology and Solar Energy Data Set for Renewable Energy Industry Use, C.E. Whitlock et al., Rise & Shine 2000, the 26th Annual Conference of the Solar Energy Society of Canada Inc. and Solar, Oct. 21–24, 2000, Halifax, Nova Scotia, Canada. (requires login)
5. Average annual domestic electricity bills for selected towns and cities in the UK and average unit costs (QEP 2.2.3), BERR, December 2008  (WebCite cache)
6. Domestic Photovoltaic Field Trials:  Final Technical Report, URN Number 06/2218, BERR, 2006  (WebCite cache)
7. Solar electricity, Energy Saving Trust, 2009  (WebCite cache)