Monday 26 October 2015

Self-Consumption of Solar PV Generated Electricity

The Amount Customers can use Themselves now Matters More than Ever

When the UK Feed in Tariff (FIT) launched, people investing in solar were paid the equivalent of 48p/kWh in today’s money for solar energy they generated (irrespective of whether they used it themselves or exported to the grid).  With electricity savings worth around 15p/kWh and export paid at 4.8p/kWh in today’s money the only number that really mattered was how much energy you would generate with your solar panels. Fortunately, solar professionals have accurate tools to forecast the annual yield for a solar, even relatively simplistic approaches such as the MCS calculation get pretty close.

As solar costs have fallen since the start of the FIT, the generation tariff has fallen too.  Now the generation tariff is worth about a quarter of the starting value - 12.47p.  If the government presses ahead with its reckless cutbacks on the FIT, then the generation tariff would be only 1.63p for domestic customers.

As these changes have occurred, the economic basis for installation of solar panels has become more and more driven by the value of the energy savings the system produces.  With the generation tariff at only 1.63p, the energy savings dominate (see the graph below).

Most solar companies have been using a value of 50% to estimate the amount of energy generated by the solar that would be used in the property (and therefore offset energy bills), so called self-consumption.  The justification for doing so is that this is a ‘government figure’ because the amount the FIT pays for export has to be deemed rather than metered, and government set the value of export at 50%.  So logic says that if the export is 50%, then the self-consumption must be 50% too, right?


Just because the government says it is willing to pay the export tariff on 50% of the energy generated, this is not the same as saying that all houses will use 50%, irrespective of the size of the array and energy consumption patterns of the house during the day.

In the past, errors in the estimate of self-consumption have not really mattered to the presentation of the economics.  Right now they are starting to matter.  As the generation tariff is reduced further, they will really matter.

The industry is going to have to develop ways to more accurately assess and predict self-consumption.

So let’s start by having a look at a few examples of real homes with solar.  Thanks to RBeeSolar and 4Eco for allowing me access to anonymised data on their systems.

Each graph shows a full 24 hour period running from midnight to midnight, with mid-day in the centre.  The day is divided into 10 minute sections and the energy flows are shown in watt-hours per 10 minutes.

Blue is electricity pulled from the grid for consumption in the house.  Orange is energy consumed in the house and provided by the solar panels.  Yellow is energy that cannot be used in the house and that is therefore exported to the grid for others to use.  The electricity consumption of the house is the sum of the blue and the orange.

House One

House 1 has relatively low total annualised energy use 2,250 kWh per year.  This figure represents about three quarters of the electricity use of a typical UK home (often taken to be 3,100kWh/year).    The use pattern shows a small morning peak and a larger evening peak.  There is little use above the baseload during the daytime on week-days, but additional electricity use at weekends, indicating a household where occupants are out during the working week.

Data was only available for July, August and September for this house, but the self-consumption rate during this period was only 22%, with 78% of generated energy exported.  The solar system is not especially large at 2.5kWp.

House 1 Self Consumption: 22%

House 2

By contrast, house 2 shows a consistent pattern of electricity use during the working week and weekend, indicating a household that is occupied during daylight hours all week.    The base load is a little higher than for house 1, and there are regular peaks of electricity consumption throughout the day.  

The pattern of electricity use is similar in winter, perhaps with a higher evening consumption.  The graphs clearly illustrate how on a gloomy winter day (Thursday), most of the solar is used in the house, but that there are still sunny days in winter and plenty of export going on. 

Total electricity consumption over the year was 2664kWh, so still a little lower than typical (86%).  The solar system on this house is 4kWp.

House 2 Self Consumption: 24%

House 3

House three has an annual electricity demand of 5,030kWh, comfortably higher than the typical UK house.  It also has a use pattern that indicates people are at home during the working week. When coupled with a  4kWp solar system, this results in a higher self-consumption level, but still only 37%.

House 3 Self Consumption: 37%


The withdrawal of Feed in Tariffs (whether sudden or gradual) is the clear direction of travel.  The result of this trend is that self-consumption of solar electricity becomes the dominant economic justification for installing solar PV.  Any error in the predicted level of self-consumption will have a larger impact on the overall financial returns than has previously been the case.

Based on the small sample considered above, the industry-standard use of a value of 50% for self-consumption of solar generated electricity in domestic installations looks generous.  With increasing availability of monitoring equipment householders will be able to check the accuracy of figures that were used in the sales process. 

Products that divert excess solar electricity to water heating may have a role to play in increasing self-consumption, but the economic savings will depend on the replaced energy that would have been used to heat the water.  The government’s announced intention to retrospectively pay only metered export once smart meters are installed means that if the house has gas-heating, the value of the gas use avoided is similar to the income from exporting the electricity. 

Diversion of excess solar electricity to charge electric vehicles or to battery systems that can store energy for evening use will become more possible as the price point of these technologies continues to fall.  In the meantime it could be that the economic optimum moves away from the current goal of maximising subsidy yield (aiming for 4kWp or as much as will fit) and we begin to offer slightly smaller solar PV systems that produce less excess on sunny days and a higher proportion of self-consumption. 

This maturing market could involve the solar installer fitting monitoring equipment in the house for a short period before making a recommendation about a right-sized solar installation.  From my experience of looking through data on these houses and others, the good news is that people appear to be real creatures of habit.  One or two weeks' worth of monitoring should be enough to give a good indication of the timing of people's energy use.

The industry needs to do more work to understand the relationship between the proportion of self-consumption and the size of the solar installation relative to the size of the annual electricity demand.  It may be that predictive tools, or at least rules of thumb can be developed to allow solar installers to size the solar system to achieve a level of self-consumption knowing the annual energy use of the household.

Thursday 22 October 2015

Scotland Shows the Way on UK Building Regulations

New Rules will Boost Deployment of Solar on New Homes

Ideal for New Build - Roof Integrated PV.  Image:  Viridian Solar

This month, new building regulations came into force in Scotland.  Let’s have a look at what they might mean for solar.

The way the building regulations for energy performance of houses work is that the designers have to keep the calculated carbon dioxide emissions associated with heating the house, providing hot water to occupants and running lights and pumps (but not electrical appliances) below a certain set level.  
The calculated carbon dioxide emissions are assessed using a government approved method (SAP2012), which is available in software form.

The level the designer has to keep below is arrived at by calculating the emissions from a house of the same shape, but with energy performance features defined in the regulations, a so-called ‘notional dwelling’.  The thermal insulation performance (U-value ) for the walls, floor, roof and openings is defined for this notional dwelling as well as other features such as values for air-tightness, the type of heating system and other energy saving measures such as use of low energy light fittings.

The reason for taking this approach is that it is not prescriptive.  It allows the building industry to experiment with combinations of measures that achieve the overall goal (carbon emissions) in the best way for them (which almost always means the cheapest way).

The table below shows a few key features of the notional dwelling for the new Scottish regulations and compares them with the same requirements for the previous regulations and also the current regulations in England.

Regulators also worry that cold, leaky homes might be built with low levels of thermal insulation and lots of bolt-on electricity generation, so they also set minimum levels of performance beyond which it is not possible to go.  These are called backstop values.

Comparing the notional values for 2013 with those for 2015, you can see that the thermal insulation has been tightened up somewhat, but not excessively.  You can also see that Scotland 2015 is not significantly better than England 2013 in this regard.

Where the two countries diverge massively is that the notional house in Scotland includes a PV system on the roof for homes heated with gas, LPG or oil, whereas the English regulations include no renewables at all in the notional dwelling.

The Scottish regulations call for the notional house to have a PV system sized as follows:

kWp = smaller of total floor area x 0.01  --- or---   30% of the roof area based on 0.12kWp/m2

If we take an average semi-detached house as an example, with total floor area of 85 m2 over two floors and a roof pitch of 35 degrees, this equates to a total roof area of 49m2

So the solar installation on the notional house would be the smaller of 0.85kWp or 1.8kWp
But solar installations on notional houses don’t help the solar industry.  What does this mean for real houses that are going to be built in Scotland from now on?

A developer in Scotland seeking to build this 85m2 semi could aim for any of the following:

1. Match the insulation levels in the notional values and install a 0.85kWp solar system
2. Exceed the insulation levels in the notional values and have no solar
3. Relax the insulation levels below the notional values and put a larger solar system on

As I discussed in a previous blog, insulation suffers from a diminishing return which means you need to pay for ever more insulation to make the next improvement.

By contrast, solar PV benefits from a falling marginal cost as you increase the size of the installation.  If you’re going to use solar, it’s more cost effective to use a larger system.  

The feedback I am hearing from housebuilders in Scotland is that they are embracing solar as a big part of their strategy for delivering homes to the new regulations.   Schemes we have seen so far indicate that the solar systems will be closer in size to those installed by householders when retrofitting, perhaps in the range of 2-3 kWp.  This is great news for the solar industry and also for the energy bills of people buying these homes.

If only England and Wales would implement such ambitious targets for new homes too.

Thursday 10 September 2015

The Dogs are Already Running

The Absurdities of the Feed in Tariff Review

The solarblogger has met  a number of officials from the Department of Energy and Climate Change (DECC) over the years and holds them in very high regard.  Make no mistake, these are smart people we’re dealing with.

Which makes the recent Feed in Tariff (FIT) review all the more perplexing.

Government is proposing to place a cap on the cost of any future deployment of solar under the FIT.  If events over the course of the consultation period indicate that this cap will be breached, government proposes to close the generation tariff to new entrants.

But at the same time it has created the perfect conditions for a ‘gold rush’ by announcing that the tariff payments will be cut by up to 87% infour months time. (January 2016)

A kind of self-fulfilling prophecy has been formed.

It doesn’t matter if industry provides evidence to support less draconian cuts to the proposed levels of the tariffs.  The cap has been set at such a low level that even a modest spike in solar installations during the consultation will ensure it is all spent.

They say the definition of madness is to repeat the same actions again and again and expect different results.  Well, back in 2011, DECC did almost exactly the same thing.  It announced a 50% stepped reduction in the FIT.  Installation rates exploded.  Within six weeks the industry was installing solar at a rate nearly 16 times higher than in the run up to the announcement.

The same thing has already started.   The dogs are already running.  And this time there’s four months for people to get their installations registered on the Feed in Tariff and claim the current payment levels.  Naturally, this is what the press has focused on, with headlines stressing that people need to get in now if they want to make money from solar panels.

If we rule out stupidity, and assume that the big brains at DECC are able learn from past experience, then there’s only one conclusion to draw.  DECC deliberately set things up for a gold rush, thereby creating an excuse to close the scheme entirely (or at least the generation tariff part, the export tariff appears to be slated to continue).

The most dismal part of this whole sorry episode is that under the guise of ‘controlling energy bills for hard working families’ the government has manufactured  conditions to ensure that the costs of the Feed in Tariff will be higher than ever, the country will get less solar installed, but at a far, far higher cost to those hard working families.

There was an alternative.  The Solar Trade Association, published  its Solar Independence Plan in the run up to the FIT review.  Clearly no-one at DECC read it.  It proposed a glide path to zero subsidy over the next four years by reducing the level of FIT payments to new entrants little and often.  This would have avoided the inevitable spike in installations that will now occur.  Because more of the installations would have occurred in the future (at low levels of Feed in Tariff) it would have ensured that the country got more for its money.

Government should act quickly to prevent the boom and bust and protect consumers from the higher energy bills this ill-considered proposal will inevitably produce.  The solar industry should demand the immediate withdrawal the consultation.  DECC should try again.

This article also appeared on the Solar Power Portal:

Monday 17 August 2015

Fabric First, but not Second and Third

3kWp integrated solar system on new build home.  Image: Viridian Solar

There’s an interesting shift going on in attitudes to solar in low carbon building at present.  Solar panels, previously seen by many in construction as a necessary evil, something to get you over the line if you couldn’t quite reach energy performance targets, are starting to feature more significantly in developments.  Instead of just one or two panels per house (0.25 - 0.5 kWp), we are seeing more and more developers installing solar arrays of a size someone might actually pay to have on their own home (typically in the range of 2 - 4 kWp)

Previously, voices in the sector have encouraged designers to ‘build in performance’ by using as much insulation as possible.  The argument, pitched under the catchy slogan ‘Fabric First’, goes that ‘bolt-ons’ such as solar panels can easily be removed from the building and only insulation can ensure the performance of the building for its lifetime.

This line of argument ignores the possibility that the performance of insulation can deteriorate over its lifetime (possibly a subject for a whole other blog, but see links 1, 2, 3 for starters).  It assumes also that when a solar array reaches the end of its life, or even before, the householder won’t replace it with a new one (with the benefit of 30 years of technical improvement to the technology in the meantime). 

It also ignores the inconvenient finding that some of the benefits of higher levels of insulation are lost to ‘temperature takeback”  (where people just run the house hotter so they can walk round in their underpants in winter, thus offsetting some of the expected energy efficiency gains). By contrast, the people with generating technologies such as solar have been shown to be more engaged in monitoring and managing their own energy use.

Building ever-more airtight homes has also raised concerns about indoor air quality and over-heating in summer.  For sure, there are technical fixes to both these issues, but doing it properly adds yet more to the cost of achieving very high levels of thermal performance.

And there’s that word, the thing the construction industry focuses on with relentless intensity, the thing that matters above all others. 


When it comes to reducing the carbon emissions from a building, the cost structure of solar is very different from the cost structure of insulation, and it’s all down to the different way their cost-benefits change with the amount you use on a building.

Cost Curves for Insulation and Solar

Double the thickness of your insulation material and you halve the heat loss conducted through it.  If your starting heat loss was 100 units the new heat loss is 50, a saving of 50.

Double it again and the heat loss goes from 50 to 25, saving 25.  The saving from the next doubling is 12.5, the next is only 6.25.  By now our insulation is 16 times thicker than our starting point.

In real life, of course, the insulation is only a part of the wall, floor or roof build up, with other elements making up the total U-value.  Kingspan Insulation provide an online U-value calculator to have a look at this effect in real walls and roofs.

You can see that doubling the insulation thickness reduces heat loss by only 30%.  Doubling a second time knocks a further 26% off. 

Insulation is a relatively cheap material, but the more you have the more you need to add to make a difference.  This is a classic example of a diminishing return.

Thicker insulation doesn’t just mean a higher spend on insulation, for example different wall ties are needed for thicker insulation in traditional cavity wall construction.  Window reveals become larger with cost implications.  Rooms become smaller for the same building footprint.

The cost structure for solar is very, very different.

First of all, the energy output from a solar array is linear with size.  If you double the number of panels and scale up the inverter, you’ll double the annual output of electricity - right up until the point you run out of available area on your southerly roof pitch and start having to use less good orientations.

Granted, doubling the energy output doesn't necessarily double the benefit for the homeowner - until battery storage technologies become more common, the larger the solar system, the more electricity will be exported to the grid.  However for the house-builder, the important measure is most often the treatment in the building regulations.  Currently the energy calculations (SAP) give exported electricity equal benefit to that used in the building, both for the carbon production and for the energy costs calculations.

Secondly, solar follows a diminishing cost per unit.  The larger you make a solar array the lower is the cost of the next increase in power.  The spend on the solar panels themselves will scale linearly with the size of the array, but other costs do not. 

Equipment costs per unit of output fall with system size.  The cost per watt-peak of power of a solar inverter (the electrical equipment that converts direct current solar generation to alternating current  for use in the home) falls as the units power rating increases.  Other equipment costs such as for generation meter and isolation switches are the same for any size of domestic system.

As an example, looking at the Viridian Solar price list, the cost of an electrical kit suitable to connect a 4kWp (16 panel) system is about 3 times the price of a kit suitable for a 0.5kWp (2 panel) system, despite offering 8 times the annual energy yield and carbon savings.

As for the labour costs, once you’re up on the roof the extra time to fix down a few more panels is small and the time to wire up a larger array is only marginally more than for smaller one.

Implications for Design

Until recently, a designer aiming for a particular energy performance would typically increase the level of insulation up to what they consider to be a practical limit and then turn to solar to get the last little bit of performance for the building.

However, it is becoming clear that building designers are starting to understand that the different cost structures of insulation and solar mean that considering embodiments with a larger energy contribution from solar may yield a more cost-effective overall design. 

The diagram above illustrates the concept.

The cost of solar power has fallen spectacularly in recent years and continues to trend lower. Cost-effective roof integrated solar systems have improved the aesthetic qualities of solar.  Is it time to reconsider your approach to low carbon design?

Thursday 28 May 2015

Solar Attrition Rates

An Analysis of the MCS Installer List

I recently had the opportunity to have a look in more detail at the list of installers registered with MCS, and what I found came as quite a surprise to me.

The number of solar PV installers registered with the Microgeneration Certification Scheme (MCS) has been on a declining path since the boom of 2010-11.  This is not news to anyone in the industry.

Right now, the number of solar PV installers registered with MCS (removing duplicates) is around 2,640, a fall of 24% since 2013.  But when you look at the actual companies that make up this headline figure you find that less than 50% of the solar PV installation companies on the list in 2013 remain two years later. 1,822 companies have left the market, but 980 new companies have joined the list in the last two years.

Churn Rates in Solar Installation Businesses

Turning to the list of solar thermal companies, we see that the decline in numbers has not been as severe as for the PV installer companies, a 13% drop from 1,298 companies in 2013 to 1,130 now.  However, the churn rate is just as eye-watering.  Nearly half of the solar thermal installers registered with MCS in 2013 are no longer on the list, but the 635 that have left have been replaced by 467 new companies.

What's going on?  Are these attrition rates normal for similar industries (home improvements, heating, electrical works)?  Or is there something 'special' about our solar industry?