Domestic Combined Heat and Power

Ceres Power, a £150m AIM-listed company, recently demonstrated its new Combined Heat and Power product. This power plant is targeted at ordinary domestic homes. Combining an efficient central heating boiler with a fuel cell that converts gas to electricity, the new product has excited the City. Ceres is extremely optimistic about sales of the device, based on the cash and carbon dioxide savings it says can be achieved.

The Ceres fuel cell (on the left) is incorporated into an ordinary domestic condensing boiler (on the right)

Ceres promises reductions in utility bills of £300 a year and 2.5 tonnes savings in carbon dioxide for the typical UK house. Our short report shows why we think that these savings are unlikely even in the most appropriate UK installation. In fact, the emissions reductions are likely to be minimal and the reductions in the electricity bill will not easily justify the approximately £1,000 extra cost of the CHP cell.

Micro CHP is a difficult proposition. Other companies have found that it is hard to make substantial savings in domestic installations. CHP is not well suited to rapidly fluctuating and unpredictable demand for electricity and hot water.


What is micro CHP? The logic behind CHP is that almost all electricity generation methods involve gross inefficiency. A large coal-fired generating station might only convert 35% of the energy in the fuel into electricity. The rest is lost as heat. An average of 7% of all electricity is then lost in the transmission networks, emitted as heat or electromagnetic radiation. A combined heat and power plant captures the lost heat and uses it to warm buildings or heat water. Big CHP plants exist in the UK, and in much greater numbers in northern Europe. Micro CHP devices are attempts to replicate these plants at the scale of a single household. The core financial proposition is that the householder generates valuable electricity using inexpensive gas.

Why is it difficult to make CHP work, either in terms of reduced emissions or lower bills? Home electricity demand fluctuates every second. All small generation technologies find it difficult to adjust to this. Unless surplus power can be sold to the electricity networks, CHP devices lose money from using gas to make electricity with no value. In the UK, unlike Germany and other countries, the money paid for feeding electricity into the grid is negligible. There is very little prospect of so-called 'feed-in' tariffs improving in the UK. Secondly, the household has to have a use for the heat that is produced as a by-product of generation. In most micro-CHP, the heat is dumped into the hot water tank, which may or may not need it. Thirdly, most CHP devices, including Ceres, are slightly less efficient (converting fuel into useful outputs) than a new condensing boiler. There just isn't much of a gain from installing a tiny CHP plant.

Past experience Field trials of other micro CHP devices have proved inconclusive. In some cases, fuel bills went up as a result of installing the technology. In the large majority of cases, the savings in carbon emissions were low. Ceres claims its new technology, which is based on a relatively new type of fuel cell, is much better. I accept that its fuel cells may be excellent at delivering efficient generation of electricity, but I think that in domestic homes the savings are likely to be as disappointing as previous technologies.

The ideal Ceres CHP installation The new Ceres CHP power plant delivers between 300w and 1 kW of electric power, responding to the level of electricity demand in the household. 300 W is equivalent to a two large TVs operating simultaneously. 1 kW is a typical power use for an electric heater.

To have maximum value to the householder, the electricity demand of the home should be a constant 1 kW. This would enable the CHP plant to deliver a high and consistent efficiency and maximise the savings.

When it was generating 1 kW of electricity, the CHP plant would also be delivering approximately 1 kW of heat to the hot water tank. About 0.4 kW would be wasted, partly in the form of unused heat and partly in electricity used to drive the CHP plant itself.

  • The input cost of 2.5 kWh gas is about 6.25p.
  • The hourly output of the CHP appliance (1 kWh electricity and 1 kWh heat to the water tank) would cost about 13.8p.
  • So the hourly saving would be about 7.45p. Grossed up, this is £653 a year.
  • The saving in CO2 is about 1.3 tonnes a year.

These are the absolute maximum savings attainable with this boiler. To get these savings the CHP plant needs to be working at 100%, or exactly 1 kW all the time. This means the house has to have an absolutely constant electricity demand. This is, of course, an unreasonable assumption; typical domestic demand fluctuates every second and falls to very low levels at night and when the house is empty. If the CHP plant is working flat out it will deliver 8,760 kWh a year, or over two and a half times the UK average for a domestic property. A large house might well have aggregate demand this high, but not consistently. Secondly, the house also has to need about 8.760 kWh of heating for the hot water tank. This is a very high level for a UK property and unlikely to be used except in houses with a large number of occupants.

What about the benefits of installing the CHP cell in the typical UK property? The typical UK house is thought to take about 3,300 kWh of electricity a year, far lower than the level discussed in the preceding paragraphs. (This conventional assumption is used in all advertising and in product comparisons. It is out-of-date and 3,700 would be a better figure.) If this demand was exactly constant, it would mean that the CHP plant would need to deliver 377 watts of electricity all the time, and provide a similar amount of hot water heat. I calculate that the annual savings from using the Ceres CHP plant, above and beyond those installing a good condensing boiler are still quite large, though not as great as if the CHP cell worked constantly at 1 kW.

  • £249 in reduced electricity charges, net of a smaller increase in gas bills
  • 0.55 tonnes of CO2

These figures are much lower than those provided by Ceres, which suggests figures of £300 and 2.5 tonnes of CO2 for a typical installation. I want to stress that the savings that I estimate also assume absolutely constant electricity demand and a perfect match between the amount of heat provided and household needs for hot water. Of course, real households have rapidly and erratically varying electricity needs and more consistent, but still highly unpredictable, hot water needs.

The other reasons why the Ceres forecasts of carbon and cash savings are unlikely to be met I have suggested so far that the maximum saving from using the CHP stack would be achieved in a very big house with stable electricity demand and very high water needs. In the average UK house, needing 3,300 kWh of electricity a year, stable electricity demand and full use of the hot water heat, the CHP plant would save some carbon dioxide and a reasonable fraction of the home energy bill, though not as much as Ceres suggests.

Now I want to go to show that the rapidly varying electricity demand in an ordinary household, combined with the likely pattern of hot water need, will mean that a Ceres CHP plant will not save the householder significant sums of money or avoid much carbon emission.

There are six principal reasons for my scepticism:

  1. The Ceres CHP plant does not react instantaneously to changes in electricity demand. According to a company spokesperson, the stack may take over 5 minutes to adjust its output to the level of demand in the house. This is an extremely important failing. The chart below shows how demand in a typical house might vary over the course of an hour.
  2. The swings in demand are sharp. The 'baseload' of the typical house might be 150-200 W or so, depending on the number of appliances on standby and other factors. When major appliances are switched on, the amount of electricity taken by a house will instantaneously rise to 3 kW or more. A kettle, for example, uses 3 kW, and is on for about 3 minutes. A dishwasher has quite low demand for part of its cycle, but uses large amounts of electricity when it is heating water. Refrigerators are using electricity some of the time, but not at other points. To give the most obvious example, if the Ceres CHP plant adjusts its output with a lag of five minutes, it will completely fail to meet any of the extra demand ever created by a kettle or a toaster. These two devices alone might be as much as 7% of electricity demand in some houses.

    The output pattern of a CHP stack based on a fuel cell is likely to look very roughly like the dotted green line in the drawing below. The stack is often supplying power when it is not needed, but at other times is failing to meet the household's demand.

    This reduces the household's savings in electricity consumption, perhaps significantly. (Please note: the relatively small Ceres plant can only ever produce 1 kW, so not only would the output lag demand, but it will not meet the peaks.)

  3. The Ceres power-generating stack will generally not generate electricity at a level below 300 W. In many houses, but not all, the baseload electricity demand is well below this minimum level. The Ceres machine can either be programmed to turn itself off when demand is lower than 300 W, or it can simply 'spill' the excess production back into the electricity network (for which the householder will generally not get paid). When I questioned Ceres about this, I was told that in most households, for most of the time, background demand is 300 watts or more. Ceres has done substantial work on the profile of household demand, so I defer to their findings. Nevertheless, it is worth pointing out that a household taking 300 watts all the time uses over 2,600 kWh a year on baseload alone, out of a total electricity use of 3,300 or 3,700 kWh.
  4. In my view it is inherently unlikely that baseload demand is such a very high fraction of total demand. (People often mention the high 'standby' demand of consumer electronics when discussing baseload, but total use from this source is very unlikely to exceed 50 W in the average house, or perhaps 60 W in houses with a Sky box. The underlying point is this – a household that is careful and economical with its electricity is unlikely to have a base demand as much as 300 watts. For households, buying a Ceres CHP plant means that a lot of electricity is going to be lost to the outside network or the Ceres machine will simply not be working most of the time.

  5. And, perhaps more importantly, the Ceres power plant cannot generate more than 1 kW. Most domestic households will frequently exceed this level during the average day. Any domestic appliance that heats water (kettle, dishwasher, tumble dryer, washing machine) uses two or three kilowatts for a substantial part of their operating cycle. At those times, the Ceres CHP will be able to fulfil only a small fraction of total electricity demand. Similarly, any machine with a large motor (a vacuum cleaner) or heating elements (toasters) will generally use much more than 1 kW. The Ceres unit will be unable to cope with these peaks. I don't have accurate figures but I suspect that electricity consumed in a typical house at times when total demand is over 1 kW may be as much as 30% of total usage. This demand is completely unmet by the Ceres device.
  6. The CHP plant responds to electricity demand; the useful heat produced goes to the hot water tank. On days when not much electricity is produced, not enough heat will get to the hot water tank. Households use hot water primarily for washing. This does need does not fluctuate much each day. But electricity use varies greatly. It varies by time of year (winter higher than summer), it varies by the day of the week and by the hour of the day. At times, the heat produced when the CHP plant is generating electricity will not be enough to cover the hot water needs of the house. When I asked about this point, Ceres responded by saying that the central boiler will make up any deficiency. And indeed it can in most circumstances. But imagine the following scenario: water is taken from the tank at 10.30pm for showers. From that point on, electricity generation is very limited, largely because the house is on baseload use, probably less than 300 watts. The CHP plant may not be operating at all until the household gets up again in the morning. Therefore the central heating boiler will heat the water instead of the CHP device, so that the tank is ready for any morning draw of water. This is fine, but then the CHP plant is needed to generate electricity during the day. If the water is already hot, there will be nowhere for the heat to go and it will have to be dumped, or the CHP plant turned off. Once again, this will reduce the cash and carbon savings of the device.
  7. An analogous position arises when the house needs a lot of electricity and little hot water. Imagine a big house with only two people in it who only take brief showers. Electricity demand is high, but hot water need is low. In these circumstances, the CHP plant will have to dump the heat or turn itself off. Savings will be small.
  8. When the household is not present, the CHP machine will have to be turned off, or waste all its heat. A house has continuing demand for electricity when unoccupied (perhaps when the household is taking a holiday) but not for hot water. The continuing demand may be lower than the 300-watt minimum, in which case the CHP plant may be off anyway, but it will still be wasting heat.

Our conclusions The Ceres plant is an extraordinarily impressive piece of technology. It has solved major problems in applying fuel cell technology to small-scale installations and at low cost. Unfortunately, I think that the household savings are unlikely to be anywhere near as large as the company hopes. This will severely restrict potential demand.