Dynamic thermal simulation models of buildings: a new method for empirical validation

A new empirical validation methodology for dynamic thermal models has been devised and illustrated using the three models ESP, SERIRES and HTB2, and a simple data set collected from the Polytechnic of Central London test cells. The method hinges on making an initial base case prediction and then taking account of the errors and uncertainties in a systematic way. The method proved capable of revealing internal errors in thermal models. The relationship of this method to other validation techniques is discussed.