Conference Paper

Proceedings of BauSim Conference 2016: 6th Conference of IBPSA-Germany and Austria

Combination of Monte-Carlo approach and artificial neural network for the determination of swiss typical building thermal load profiles

A. Seerig, A. Zakovorotnyi

Abstract: The comparison of the existed building thermal load profiles (heating and cooling) allows to identify buildings with the probable renovation potential. However, they depends on a variety of the factors: weather conditions, building properties, type of heating and cooling systems, user behavior etc. Therefore, there is no clear boundary that clarifies the extent since when similar buidlings become rather unsimilar. Hence, it is crucial to have a reliable procedure which adequetly separate buildings’ load profiles to the to main groups of typical and abnormal load profiles. The last group can imply renovation possibility and should be investigated further. Present research proposes an approach which fulfils the requirements mentioned above and is based on artificial neural clustering procedure. However, to obtain reasonable results neural network was trained on Monte-Carlo thermal building simulations which consider two-dimensional parameter variation that includes the variation of user behaviour from one hand and weather data from another.
Pages: 377 - 381
Paper: