Decision tree for early-stage design of hybridGEOTABS office buildings

Abstract

Hybrid GEOthermal heat pump system coupled to Thermally Activated Building Systems (hybridGEOTABS) utilises the high thermal capacity of TABS to smooth out the building thermal loads and downsize the production units. Moreover, hybridGEOTABS has achieved remarkable carbon emissions saving. However, the optimal design of hybridGEOTABS is not achieved with current design methodologies. This article provides a decision tree for early-stage design of hybridGEOTABS office typology. To derive the decision tree, a design methodology which has been previously developed and verified was applied on nearly 40,000 office building case studies with variety of parameters such as climate, insulation level, and internal gains. The methodology exploits multi-zone dynamic simulation of building energy performance and optimal control of TABS for peak-shaving to offer an optimal sizing of the HVAC components. To analyse the results of the numerous simulations and to drive the decision tree, supervised machine learning, specifically a classification technique, was deployed. The application of the decision tree is exemplified in this article using three case studies. The decision tree also enables architects to practice the influence of different parameters on the sizing and performance of the HVAC system. Thus, designers may use it to optimise the building physical design to increase the possible share of geothermal system as a sustainable core for providing thermal comfort in buildings.