Search site

Thermal modelling and building design

Objectives: The Chartered Institute of Building Services Engineers developed this mock case study (using dummy data)using the 11-member RCM as a source of simulated future climate that could be used to derive hourly weather years for thermal simulation modelling in building design.


How they used UKCP09 dummy data

1. The team used the 11-member RCM output because for the first year of their existence, the probabilistic projections and Weather Generator did not provide the required wind input data. Wind is an integral part of the information necessary in thermal simulations used for building design. It was also decided to use the existing CIBSE morphing methodology to develop the required hourly future design weather years (Test Reference Year TRY and Design Summer Year DSY).

2. The 11-member RCM information is available as daily absolute values. Change factors (relative values) were needed by CIBSE to apply its existing morphing methodology.

3. The baseline climates from the 11-member RCM needed to be subtracted from the future absolute values in order to get the relative values. To do this, the daily RCM data was averaged over the 30 year time period of 1961-1990, to get a year that represents the baseline, and over the 30 year future time period (e.g. 2010-2039) to get a year that represents the future climate. To get a relative change factor for each day, the average daily baseline data was subtracted from the average daily future data. This was repeated for each of the 11 model variants.

4. The standard CIBSE hourly design weather years (TRY and DSY) were morphed with the daily change factors. This involved stretching and shifting the present-day observed time series to produce a new time series that has the daily average statistics of the climate change scenario.  More information from CIBSE TM48 on this process.

5. This process was repeated for each of the 11 RCM variants and for all the emission scenarios, time periods and locations (as specified alongside).

6. Selected resultant future weather years or the whole range of the resultant future weather years were run through a thermal simulation model to examine the range of building performance for a variety of building types and assess levels of overheating risk.

Next steps

Following a study like this it would be possible to investigate ways in which the Weather Generator could provide the inputs needed to produce weather years for building thermal simulations. Exploring different options for producing the required weather years that bring into account the fuller consideration of uncertainties and the relative strength of evidence provided by the UKCP09 probabilistic projections would also be beneficial. This could include using a combination of the 11-member RCM and probabilistic projections, along with sensitivity analyses.

Lessons learned

· Getting robust information on wind data is a problem. In developing the UKCP09 projections it was decided not to include probabilistic projections for future wind due to the high level of associated uncertainty.

· Wind information, (although maintaining a high degree of uncertainty) is available from the 11-member RCM and thus will enable the building sector to have access to the required wind. Due to this high level of uncertainty in the wind projection, along with the fact that this uncertainty is not sufficiently explored using the 11-menber RCM data available, users should test the sensitivity of their results to wind.

Find out more

  • Contact details: Anastasia Mylona, CIBSE