strathclyde logo


Monitoring in Athelstaneford and Oban

The under floor heating system in Oban offers the system a relatively low return temperature to the heat pump of 36.3C on average. It is however significantly higher than the average returning temperature of the heating system (31.7C on average) mainly because of the sterilisation cycles.

The underfloor also presents a certain inertia which requires an early start of the system in the morning so that the house becomes warm around 8am. The heat pump has thus to run earlier in the morning, in general at lower outside temperatures. During this period in the morning, the underfloor requires an important energy supply. The system runs therefore during an important part of the morning at a quite steady state (with nevertheless a few defrosts cycles) until the ambient temperature in the house reaches the set point. The good quality of the house insulation and the inertia of the under floor heating system enable the house to maintain its warmth almost throughout the whole day.

The radiators used in Athelstaneford require a much higher temperature in order to heat the house efficiently. As a result the return temperature to the heat pump is much higher, at 40.2C on average during the week. Actually, with the temperatures used in the radiators, the ambient temperature set point of 22C in the House can not be reached. The radiators are then operating in permanence on the periods specified in the controls (from 7 to 11am and from 4 to 10pm). During their operations, the radiators do not draw enough energy from the tank to enable a steady operation of the heat pump. The house is indeed not so big with only a few radiators. As in the design of the tank, the heat supply from the heat pump is almost at the same level than the main connection for the radiators. Thus when the heat pump does not work, the temperature tends to decrease quickly and there is a relatively fast increase of the temperature in this region when the outside unit start running. As the system tries to maintain a sufficient temperature in the middle of the tank to provide the energy to the radiators, the heat pump is always starting and stopping. The control's parameters could be improved to avoid this problem in the tank design.

Finally if both installations are compared over each week, the system in Oban has a bit better performances mainly due to the lower return temperatures to the heat pump. The COP in Oban is 2.25 against 2.05 in Haddington for the heat pump alone. If the whole system with the tank and the electric back up is considered, there are around 7kWh of energy for the control system of the tank and the pumps, 0.67 kWh of back-up electric heating for Haddington and 20.91 kWh for Oban. The average COP of the whole Sanyo system becomes 1.99 for Haddington and 2.04 for Oban.

A high domestic hot water demand does not enable a decreased return temperature to the heat pump since the cold water from the city arrives in the middle of the tank. However it tends to draw the heat at the top of the tank which activates the electric heater if the heat pump can not meet the heat demand at the top of the tank within a certain delay. As the hot water coming from the heat pump is released in the middle of the tank, it takes in general some time to heat the top of the tank. When the electric heater is used, it is in general for a few minutes since the system detects directly a sufficient temperature at the top which deactivate the electric heater. The control and design of the system is still questionable on this point.

It is interesting to notice that the multiple starts and stops of the system at temperatures which are above 0C limit the number of required defrost cycles. Indeed the ice has time to melt when the system stops. In contrast to this, the steady operation of the system in the morning as in Oban will require more defrost cycles. However this particularity is only verified when the outside temperature is above 0C.

The large number of starts and stops of the system is questionable. It would probably tend to reduce the performance of the thermodynamic cycle. As the inlet of the water coming from the heat pump is positioned at the middle of the tank, the fluctuations of temperature in this part of the tank are quite important. The sensors and/or the controls of the temperature are not properly set up to deal with this fluctuation of temperatures and avoid consecutive start and stop of the systems. The tank is able to store 227 litres of water which should be used as buffer storage to avoid this kind of behaviour. One of the reasons given to explain the position of the water inlet from the heat pump is that the whole part of the tank needs to be quite hot to provide a significant amount of domestic hot water if it is required. As keeping an important stratification in the tank is also crucial for the performances of the system, the solution could to have a bigger tank.

The large number of sterilisation cycles realized during the week for both installations is really surprising and contribute to reduce the performances of the installations. A frequency of 4 sterilisation cycles every 4 starts of the compressor seems far too high. It seems that one defrost cycles every two weeks would be sufficient. But the real efficiency of the sterilisation is also questionable since the heat pump deliver some water at 64C so that the middle of the tank reach a temperature of 62C. However the top of tank never reaches more than 56C according to the monitoring. It would be however necessary that the whole coil with the domestic hot water inside reaches a temperature of 60C during 2 minutes to kill 90% of the germs according to the recommendations the World Health Organization (2007). However some other sources claim that a temperature of 60C should be maintained during 30min (ASHRAE standard, 2000 - HSE, 2001 - Armstrong International, 2003) to have an efficient sterilisation. Finally a study on the occurrence of legionnella in hot water systems of single family residences (Werner Mathys et Al., 2007) found that raising hot water temperature to 60C periodically and for very short time intervals seems to favour growth of legionnella and cannot be recommended. The efficiency of these sterilisation cycles are thus more than questionable.

The design, and necessarily the controls of the tank could be modified as suggested in the KTH report with the outlet from the heat pump at the top of the tank. A better use of the cooling capacity of the water coming from the city should also be considered. Instead of pumping the energy in the hot part of the tank, the cold water should be first heated with the return water to the heat pump, still warm. The return water temperature to the heat pump would be then lower, which would increase the COP of the heat pump.

Monitoring in Ballencrief

This new tank does not have a design which enables it to decrease more efficiently the return temperatures to the heat pump. This one is indeed higher than the return temperatures from the radiators and the high domestic hot water demand does not improve the situation.

There are also a lot of quick fluctuations of the temperatures in the tank. The volume of the outer tank is indeed quite small especially in the bottom part. In addition the water for the radiators is mainly taken at the bottom of the tank (1/3 of the height), where the cold water from the city comes. When the heat pump does not work, the heat is mainly drawn from the bottom of the tank. The sensors detect quickly the decrease of temperatures and start again the heat pump. But as the volume of the outer tank is quite small and that the top of the tank is quite hot, new hot water quickly reaches the bottom of the tank and stops the heat pump operations.

The problem with this tank design is that the heat is mainly taken from the bottom of the tank whose temperature has also to be quite high to provide the energy to the radiators. When the heat pump starts running, the bottom of the tank is quickly heated since all the top of the tank is already at a quite high temperature. The storage provided by the volume of the tank is not used for the space heating demand. As the heating load is quite low for this day, the system has to start and stop many times. When the heating load must be higher, this problem may be corrected.

This new tank provides bigger heat storage but no real improvement in term of design compared to the Sanyo tank. It is still not really adapted for an operation with a CO2 heat pump.

The Finnish installation

The heat pump installation in Finland is clearly undersized for the size of the house. The electrical capacity of the tank is thus used most of the time. The climate in Finland is also quite cold and the outside temperature reached -25C during January. The heat pump is not performing very well at such low temperatures. Thus an important part of the power is directly provided by the electric heater in the tank. If the whole installation is considered, the overall COP for the installation would be quite low.

Radiators are also used in the house which does not help to have low return water temperature to the heat pump. However, for this installation, the return temperature is not so high (37.5C on average during the month of January) but it can be noticed that no sterilisation cycles are implemented in the controls of the first version of the Sanyo Eco Cute system.


1. World Health Organization, egionella and the prevention of legionellosis, 2007,

2. Armstrong International, Inc., Controlling Legionella in Domestic Hot Water Systems, 2003

3. HSE (Health and Safety Executive), Legionnaires disease: A guide for employers, 2001,

4 ASHRAE Guideline 12-2000 - Minimizing the Risk of Legionellosis Associated with Building Water Systems, 2000,

5. Werner Mathys, Juliane Stanke, Margarita Harmuth, Elisabeth Junge-Mathys, Occurence of Legionella in hot water systems of single-family residences in suburbs of two German cities with special reference to solar and district heating, february 2007


Contact Us

Project Team


Home Page