strathclyde logo

Finnish Case Study

In the south of Finland, not far from Helsinki, M. Tomi Hakari has installed in 2009 a Sanyo CO2 heat pump in his house. He monitors in-situ its performances and all the measurements are released every minute on his website: was a wonderful mine of information for our project and we thank Tomi for his contribution to our project.

refrig flow in cycle

Finland House with Sanyo Heat Pump

refrig flow in cycle

Map of house

Heating system in the House

The house is a big old log house from ~1920, poorly insulated, heated with radiators scaled for oil heating. The radiator requires quite high temperature in order to keep the House warm.

The Sanyo System, installed in 2009, was the old version of the tank: “Sanyo CO2 ECO SHP-TH22DDN” with a Sanyo CO2 heat pump 4,5KW.

CO2 Cycle

The 4.5 kW heat pump unit is not sufficient to heat the whole house. They use the 0-9kW direct electric heater inside the tank to provide the rest of the energy required. During wintertime all the heating capacity (back-up+heat-pump) seems to be almost constantly in use - on the other hand with mild temperature (0 degrees) 1.5 kW seems to be in use occasionally. At -20°C for the outside temperature, the heat pump struggles to supply 2500kW of Heat and stops completely below 25°C.

All the domestic hot water is provided by the Sanyo Eco Cute system. In terms of energy consumption, the Hot water demand is estimated to represent 10% of the system supply in winter if the comparison is made with the summer month. 2 big fireplaces are also used to burn around 3m3 of wood every year which represents around 4000kWh of Heat.

Monitoring Installation

The following parameters are monitored on the installation:

refrig flow in cycle

Mass flow meter

refrig flow in cycle

Power meter

refrig flow in cycle

Temperature Measurement

refrig flow in cycle


A PC with data capture is set up to store, process the data and release the information on the website every minute:

At the beginning the temperature sensors were measured from the top of the plastic pipes without any conducting paste. There was an important layer of insulation on the pipes but still a significant difference with the Sanyo’s readings. In terms of temperature difference, setting the 2 temperature sensors on the table next to each other gives +/- 0.1 °C of difference to their readings. The mass flow meter was not calibrated. The first data collected was quite difficult to analyze with a significant difference with the data from the manufacturer and the Swedish report.

After discussion with Tomi Hakari, this difference seemed to be explained by the heat losses in the pipes between the heat pump (outside unit) and the tank unit where the measurement are recorded. The pipes are indeed less than 5 meter long with a good layer of insulation.

Finally, the water flow meter has been calibrated with an ultra sound flow meter the 07/04/2011 and the measurements are now 14% bigger than previous. In addition the temperature sensors have been installed inside the tubes to give more accurate readings with faster reaction to the changes. The change on the measured value has been immediate. The reliability of the water flow meter is nevertheless a bit controversial: the values seem to be a bit high now according to a second test. No other changes has however been made on the installation.

Weather Data (relative humidity)

To study the defrost cycle, it is important to know the humidity level. A huge amount of the weather data for the area is easily accessible in Finland thanks to the project Helsinki Testbed whose main goal consists of weather research but also data distribution for research and public community.

The weather data of many stations around Helsinki is available. The closest station located at 8 km of M. Hakari‘s house was chosen (map below). The measurements are recorded at 4 meters height and stored every 5 minutes. The data can be collected online by XML queries.

refrig flow in cycle
The Data Collected

As mentioned previously, the performances are displayed online at tables for the current data and the tables for the last hour, the last 24 hours, the last month and the value of the COP for previous year are given and regularly updated (picture below).

refrig flow in cycle

The 07/04/2011, the calibration of the flow meter and the installations of the temperature sensors inside the tubes enabled to get much more accurate results. The screenshots below have been made before and after the change.

refrig flow in cycle
refrig flow in cycle
COP Estimated COP (Regression)
1.84 2.36

Before the calibration

COP Estimated COP (Regression)
2.44 2.5

After the calibration

The COP of the system is clearly lower than the values given by the Swedish report(KTH) and the manufacturer. Considering the outside temperature and the return water to the compressor, the COP should be higher. According to the multiple regressions that we did, the actual COP is 22% lower than expected. With the new calibration the results given are much more coherent.

Visualizing and Downloading the Data

M. Hakari sent us the data for the of months October, November, December, January and February. The data are available below, as well as the graph of the COP, the outside temperature, the return water temperature and the power to the compressor:

The data for November and January with the relative humidity data from the weather station are also available below. As the data for the humidity are available every 5 minutes while the data are recorded every minute, the humidity is considered as constant other the 5 minutes of the measurement.

The original data are available below:

Data file : file

The data of the relative humidity available on the website of the Helsinky testbed project:


Contact Us

Project Team


Home Page