W2UP3 – Co-heating Tests in the Field and Laboratory
23rd October 2015 Samuel Stamp

Co-heating Tests in the Field and Laboratory

Samuel Stamp, UCL Energy Institute


A co-heating test is a method of measuring the whole building heat loss or heat loss coefficient (W/K). Tests to date have shown a significant difference between design and as-measured thermal performance.

A co-heating test essentially requires a building to heated to a constant internal temperature of 25C. There is then an energy balance between the electricity supplied to heat the building (with the addition of any solar gains) and the heat lost through the building fabric or air infiltration. Plotting 24 hour averages of the power supplied (W) and internal/external temperature difference (K) allows the thermal loss coefficient (W/K) to be calculated from the gradient.


The test typically takes 2-3 weeks and needs to be carried out over the heating season. This, combined with its dependence on the weather and uncertainty in the result mean the test has been used in a limited capacity to date.

To understand both practical and physical issues with the test method two field tests were performed, along with some controlled laboratory testing.

Field Tests

Two co-heating tests were performed in high performance new builds. Whilst one was conducted in December, experiencing cold and dull weather conditions associated with a reliable result the other, in March, experienced warm and sunny weather. This led to overheating and a large amount of scatter in the daily data (see below). However, a correction for this resulting thermal mass contribution could be made using heat flux measurements.


Laboratory Testing

Controlled co-heating tests allowed the complex weather effects from real world testing to be removed and areas of the tests accuracy explored. Results at steady state heat loss shown below allowed a HLC to be measure within 5%.


Variations in external temperature showed higher deviations in daily measurements (10% compared to 5%). An underestimation in the W/K was also seen if averages were taken midnight- midnight as opposed to 6am-6am; supporting the later method.

These laboratory tests plan to be scaled up by changing the heat loss coefficient and thermal mass of the wall.


  • Low heat loss/ highly glazed properties need to be tested in a restricted period. Inputting weather files with building heat loss and glazing properties can give an idea of the probability of overheating and error.
  • Overheating and subsequent internal thermal mass effects can in part be corrected from heat flux measurements.
  • The required period of testing depends on weather conditions. In good test conditions answers to within 5% can be achieved within just 6 days.
  • In controlled conditions, removing solar gains and at steady-state heat loss the coheating method is accurate and reliable.
  • External thermal mass effects due to varying external temperatures cause errors, but these small compared to the effect of the variation in solar gains.

Project Team

Samuel Stamp
Tadj Oreszczyn