House: Thermal Store


A key to making the best use of intermittently available energy is to be able to store it until it can be used. Obviously electrical batteries play an important role but they are expensive and don't last forever so it is best to minimise their use. In particular, using stored electrical energy for heating is so uneconomic as to be insane; it is far cheaper to store the energy as heat if at all possible.

As well as being intermittent, solar thermal energy also comes in a variety of different temperatures. E.g., when the sun is very bright the output of an evacuated tube array can be near boiling (the system must be designed to deal with boiling if the water stagnates, e.g., due to a pump failure). On the other hand, on cool overcast days or early and late on clear days the temperature of the output of the panels may be much lower.

Though the power available in somewhat overcast conditions is obviously much less than that in bright sunshine the total amount of energy which might be extracted is likely to be significant simply because at times of year when energy resources are low anyway there are, almost by definition, more hours of overcast conditions. The low grade heat from overcast conditions is suitable for initial pre-heating of water which is then to be heated to higher temperatures and also for space heating.

Typical solar thermal systems have only a single tank sized to deal with the consumption of a few days at most. There are two problems with this: the small tank means that there is much more need for additional heat sources (e.g., an immersion heater) after a run of cloudy days and, once the tank has reached its highest temperature for the day, any further insolation, which is only capable of heating to a lower temperature, is wasted. The typical responses to these problems are to make the tank reasonably large (for a practical single tank) and to design it to encourage stratification so that the top of the tank is hotter than the bottom.

The system I want to build would take this idea a step further. By having a very large volume of water (perhaps 10 m³, i.e., 10 tonnes) a lot more heat would be stored. To make this practical it will be done with a number of separate, interconnected, tanks. Instead of relying on stratification within a single tank the plumbing will be arranged so that different tanks are at different temperatures.

This thermal store will be within the insulated envelope of the house along the north side, somewhat insulated from the rest of the house. The warmest tanks would be behind the bathroom (probably with an airing cupboard) with cooler tanks north of the office and behind the mechanical room.


The tanks will be plumbed in a line with the two warmest at one end being conventional hot water (copper or steel) cylinders. The remainder would be much cheaper and simpler, probably IBCs so having a lower maximum temperature. The order shown in the diagram is, therefore, a logical one as physically the warmest tanks will be near the middle of the line.

Hover the pointer over plumbing components for a label. This diagram is a simplification, as described below.


There are two separate water systems here. Most of the plumbing shown contains the primary heating water. This is pretty icky stuff as it contains inhibitors and spends a lot of its time at nice warm but not hot temperatures encouraging the growth of bacteria. You wouldn't want to take a shower in it.

Domestic Hot Water

On the right of the diagram is the output to the domestic hot water system (taps, shower, whatever). This is fed from the cold tank and kept separate from the primary water by the plate heat exchanger.

Primary Hot Water

Within the primary hot water system there are essentially two separate flows. The first is clockwise on the diagram from the coolest (left-most) thermal store tank, through the solar pump then the solar panels and back to the thermal store inlet manifold across the bottom of the diagram. The second flow is clockwise from the warmest (right-most) cylinder, through the heat exchanger then through the heating pump before joining the solar loop in the inlet manifold.

Tank Choice

The key to the operation of the system is the choice of which valve from the inlet manifold is opened to allow flow to which tank in the thermal store. The basic principle is that water should be fed into the warmest tank which is cooler than the available water in the manifold.

Feeding it to a warmer tank would, obviously, be silly as that would dilute the temperature available in that tank. Feeding it to a cooler tank would, similarly, dilute the available solar water unnecessarily.

The only exception to this scheme would be when the warmer tanks are already up to their maximum safe temperatures. In that case, feeding hot solar water into the cooler tanks for dilution there would make sense. Some care would be needed to ensure that the hot water entering the tank doesn't cause immediate damage, of course.

Note that, as shown in the diagram, these two flows are both clockwise. However, the flows through the tanks are in opposite directions. Solar hot water entering the tanks flows right to left in the tank line whereas the flow to the heat exchanger goes left to right.

Flow Rate Choice

Somehow a selection must be made of the flow rate through the solar panels. If the flow rate is low the water will reach a higher temperature and will therefore be fed into a warmer tank. If the flow rate is faster then the temperature will be lower and a cooler tank will be heated.

Note that the actual amount of energy transferred is not likely to vary greatly with the flow rate; a higher flow rate and lower output temperatures will, of course, result in lower temperatures in the collector and therefore lower losses resulting in more energy being harvested but with evacuated tubes this should not be a large effect.

This is not a simple choice as it depends on the expected usage and expected weather over the next few days. There's some opportunity for careful tuning here.


As already noted, the diagram shown above omits important details. Missing parts include:

Also, depending on the site, it could be geometrically inaccurate in that the solar panels might well be physically located below the level of the thermal store. This will take some care to deal with the volume of steam in the case of stagnation (e.g., when the pump fails) but has advantages:


As well as sorting out the simplifications noted above there are various modifications which can be made to the system with the view to improving performance. It's not obvious which are useful and which are either a waste of time or actually harmful. Therefore, the plan is to get a basic system up and running and then, in the light of practical experience and measurements in operation, make changes of an experimental nature.

Panel Return Source

As shown above the return pipe to the the solar panels, via the pump, is always taken from the coolest tank. This is likely to be the best strategy in spring and autumn but may not be the best idea in summer and winter.

In summer when there is likely to be high-grade heat available in the next few days and when extra heat leaking from the thermal store to the rest of the house is not welcome it would be better to have, in effect, a smaller thermal store. This can be achieved by taken the return water from a tank to the "right" of the coolest tank, cutting off the coolest tanks as a temporary dead-end.

Conversely, in winter periods with little available energy it would be better not to overly dilute that energy which is available in too large a store. A similar reduction in effective volume might well be found to help.

Panel Arrangement

As shown in the diagram the panels are simply arranged in series. In cooperation with varying the solar pump flow rate there may sometimes be advantages to being able to switch to a series/parallel arrangement to allow larger flows of lower temperature water to achieved.

Dual Plate Heat Exchangers

As shown there is a single plate heat exchanger fed from the warmest cylinder. In a way this is a waste as the high grade heat is used for the first stages of warming the cold water (within, of course, the contra-flow arrangement of the exchanger). It might be better to have another heat exchanger inserted in the system between that shown and the cold tank which is heated from a cooler tank, say the second cylinder.

The primary effect of this would be to allow less of the high grade hot water to be used to create any given amount of domestic hot water.

Heat Pump

A small heat pump could be used to take heat from one of the cooler tanks and deliver it to the warmest cylinder. This would be operated as a dump load for the wind turbines (and, in principle, the PVs though this would be less useful) to make good use of available energy in cases where the batteries are already charged.

In rough numbers, this might be sized to take about 1 kW of electrical power and 2 kW of heat flow from the cool tank and deliver just under 3 kW (about equivalent to a typical immersion heater) to the hot cylinder.

By cooling the water in the leftmost tanks it would make any subsequent flow from them through the solar panels be more efficient in delivering energy into the house, albeit at a lower temperature.

Cold Tank Heating

If low-grade heat is available, say below 20°C, but surplus to tank heating requirements it could put to good use taking the chill off the water in the cold tank. While it's important that the cold tank doesn't get too warm raising it from, perhaps, 5°C to 15°C is a worthwhile aim as it means that further heating to produce hot water is reduced, any mixing with other hot water will use more of the cold and the cold tank is not absorbing heat from the house.

Possibly this could be arranged by having valves available that allow the return water from the thermal store to the solar panels to be routed through a heat exchange coil immersed in the cold tank - making the cold tank effectively the end-most element in the thermal store.

Heat Dump

Any system which has much chance of working reasonably well in the winter will have too much heat available in the summer. One simple method of resolving this is to cover a few of the panels. However, this seems a pity if there are better possible uses. Perhaps, particularly late in the summer, it would be worthwhile directing this heat into the greenhouse for longer term storage either in the ground or in a water tank to carry over into the autumn.