In situations where it is necessary to keep a building’s outer doors open, you will sometimes find “air curtains”, fans which blow a sheet of air down across the width of the doorway. These are an effective way of preventing dust and insects getting in through the door: they are entrained in the outer layer of airflow, and where the jet hits the floor it splits, with the outer layer discharging the contaminants back outside.

Some suppliers of air curtains claim that they conserve energy as well. The basis of this claim lies in what would naturally happen in an open doorway in still conditions, namely convective circulation in which warm air at high level flows out to be balanced by cold air flowing inwards at low level (right). This effect will be especially marked with high doorways. The claim for air curtains is that they disrupt the flow of escaping warm air, forcing it down to floor level where the jet splits, with the warm inner layer returning inside.

However, even in still conditions there is a problem here, because the fan is drawing air from high level inside and at floor level only half of it returns inside. 50% of the internal air drawn into the fan is diverted outside when the jet splits at floor level (left).

A further problem with pedestrian doorways particularly is that the air curtain usually needs heating to prevent the perception of cold that the air’s velocity would create. If the building actually doesn’t need that heat, it is all a waste of money. Even if it does need the heat, half of what is put in goes straight outside.

In windy conditions the argument for air curtains as heat barriers really breaks down. A moving sheet of air is simply not as effective as a door. If there is any differential pressure whatever, that sheet of air will be displaced, and the problem is exacerbated if there are open doors or windows on the far side of space – or extract fans. In one instance I visited a restaurant that operated an open-door policy. Their air curtain had a 20kW heater that ran continuously, but the downjet did not reach the floor: about 60cm above the floor it turned inwards along with a layer of cold air at floor level, thanks to the kitchen extract depressurising the space.

The exhaust from a natural gas appliance contains about 0.15 litres of water per kWh of gas input, and about a tenth of the thermal output is lost because that water is emitted as vapour. Condensing boilers are a good idea in theory because they can condense the vapour and recover latent heat from the products of combustion, boosting output by around a tenth.

In practice, too few condensing boilers achieve their potential because they cannot cool the flue gas below its dew point (around 59C ). Result: plumes of vapour outside. This one resembles what you’d see boiling a 2-3 kW kettle in the open air, and that’s a measure of how much energy is being wasted.

The truth is that so-called condensing boilers need to be installed in heating systems with low return water temperatures. Underfloor heating, or systems with oversized radiators for example. Only then will they get sufficiently-low temperatures in their heat exchangers to get the exhaust vapour to condense.

Just how big a saving is it possible to achieve with a product which improves heat transfer in a ‘wet’ heating system (one which uses circulating water to feed radiators, heater batteries or convectors)?  It is an important question to answer because suspect additives claiming to reduce losses through water treatment are becoming prevalent, making claims in the range of 10-20%, while air-removal devices have been claiming up to 30%. It is possible to show that the plausible upper limit is of the order of 7%  and that this would be achievable through good routine maintenance anyway.

To work this out we first break the system into its two major components: the heating boiler (which in reality may be two or more plumbed in parallel) and the building, which represents the heat load. The first thing we can say is that if the heating in the building is maintaining the required temperatures, the thermal load which it presents to the boiler will not be affected by internal heat transfer coefficients. If heat transfer in the heat emitters is impeded, then either the circulating water temperature will rise or control valves will be open for a greater percentage of time in order to deliver the required heat output, or both; either way, the net heat delivered (and demanded from the boiler) is the same.  So water treatments will not affect the heat demanded from the boiler; their only effect will be to improve the efficiency with which the boiler converts fuel into useful heat.  Let us consider how this can be done. Consider the routes by which energy is lost in the boiler:

1. Standing losses from the boiler casing and associated pipework and fittings;
2. Sensible heat loss in the exhaust gases. This is the energy that was needed to elevate the temperature of the dry products of combustion (i.e. excluding latent heat);
3. Latent heat losses, e. the energy implicitly used in converting water to vapour in the exhaust (it is this heat which is recovered in a condensing boiler);
4. Unburned fuel (carbon monoxide or soot).

Which of these could be affected by water treatment and which would not?  Standing heat loss is sensitive only to the extent that the external surface temperature of the boiler might differ with and without water-side scaling. As such losses would only be about 2% of the boiler’s rated output in the first place, we can safely take the effect of variations to be negligible. Latent heat losses would not be affected because they are solely a function of the quantity of water vapour in the exhaust, and that is fixed by the chemistry of combustion and in particular the amount of hydrogen in the fuel. Unburned fuel losses will not be affected either. They are determined by the effectiveness of burner maintenance in terms of air/fuel ratio and how well the fuel is mixed with the combustion air.

That just leaves sensible heat losses.  Two things can cause higher-than necessary sensible heat loss. One is to have excessive volumes of air fed through the combustion process, and the other is having a higher-than-necessary exhaust gas temperature.  Excess air is self-evidently totally unrelated to poor water-side heat transfer, but high exhaust temperatures will definitely occur if the heat transfer surfaces are dirty or scaled up.  With impaired heat transfer the boiler cannot absorb as much of the heat of combustion as it should, or to look at it a different way, higher combustion-product temperatures are needed to overcome the thermal resistance.

Elevated stack temperature, then, is the only significant symptom of water-side scaling.  So how high could that temperature go, and what are the implications?  Most people would agree that an exhaust temperature of 250°C or more would be highly exceptional and values of 130°C to 200°C more typical.  Now let us suppose for the sake of argument that the exhaust gases in a reasonably well-maintained boiler contain 4% residual oxygen in the exhaust and have a temperature of 130°C, with (to make it realistic) 200 parts per million of carbon monoxide. The stack losses under these conditions will be:

4.2% sensible heat in dry flue gases

11.2% enthalpy of water vapour

0.1% unburned gases.

This leaves a net 84.5% as “useful” heat but we should deduct a further 2% for standing losses, giving 82.5% overall thermal efficiency as our benchmark.

Now let’s suppose that the same boiler had badly fouled heat transfer surfaces, raising the exhaust temperature to 300°C —  way in excess of what one might normally expect to encounter.  Under these conditions the stack losses become:

10.4% sensible heat in dry flue gas

12.7% enthalpy of water vapour

0.1% unburned gases

So we now have only 76.9% “useful” heat which, after again deducting 2% standing losses, means an overall efficiency of 74.9%, compared with the 82.5% benchmark.  The difference in efficiency between the dirty and clean conditions is

(82.5 – 74.9) / 82.5 = 6.8%

and this figure of about 7% is the most, therefore, that one could plausibly claim as the effect of descaling a heating system whose boilers are otherwise clean and reasonably well-tuned. In fact if the observed stack temperature before treatment is lower, the headroom for savings is lower too.  At 200°C the overall efficiency would work out at 81.4% and the potential savings would be capped at about 3%.

Three points need to be stressed here. Firstly, just measuring the flue gas temperature will tell you accurately the maximum that a boiler-water additive alone could conceivably save. Secondly, fireside thermal resistance is orders of magnitude higher so even a relatively huge reduction of waterside resistance will have little impact.

Thirdly, all these potential savings should be achievable just with good conventional cleaning and descaling.

The vapour-compression cycle at the heart of most air-conditioning systems consists of a closed loop of volatile fluid. In the diagram below the  fluid in vapour form at (1) is compressed, which raises its temperature (2), after which it passes through a heat exchanger (the “condenser”) where it is cooled by water or ambient air. At (3) it reaches its dewpoint temperature and condenses, changing back to liquid (4). The liquid passes through an expansion valve. The abrupt drop in pressure causes a drop of temperature as some of the fluid turns to vapour: the resulting cold liquid/vapour mixture passes through a heat exchanger (the “evaporator”) picking up heat from the space and turning back to vapour (1).

The condenser has two jobs to do. It needs to dump latent heat (3->4) but first it must dump sensible heat just to reduce the vapour’s temperature to its dewpoint. This is referred to as removing superheat.

It has been claimed that it is possible to improve the efficiency of this process by injecting heat between the compressor and condenser (for example by using a solar panel). Could this work?

The claim is based on the idea that injecting heat reduces the power drawn by the compressor. It is an interesting claim because it contains a grain of truth, but there is a catch: the drop in power would be inextricably linked to a drop in the cooling capacity of the apparatus. This is because we have now superheated the vapour even more than before, so the condenser now needs to dump more sensible heat. This reduces its capacity to dump latent heat. The evaporator can only absorb as much latent heat as the condenser can reject: if the latter is reduced, so is the former. Any observed reduction in compressor power is the consequence of the cooling capacity being constrained.

The final nail in the coffin of this idea is that reduced power is not the same as reduced energy consumption: the compressor will need to run for longer to pump out the same amount of heat. Thus there is no kWh saving, whatever the testimonials may say.

View a vendor’s response