Master's Thesis Omar Daouk

 

Optimal retrofitting of residential buildings accounting for LCA and time variant factors

With over a third of Germany's total end energy consumption, the building sector plays an important role in achieving the set CO2 reduction targets. One way of reducing greenhouse gas emissions is to refurbish the existing building stock. In that context, optimization models are used, in order to efficiently determine the optimal retrofitting approach among the large pool of renovation options for the building's envelope and energy system.

In this work, an innovative building optimization model is used, to comprehensively compare the emission-saving potential of the different retrofitting possibilities for a given residential building. The model builds on a base optimization model which includes multiple discrete upgrades of the building's energy system and envelope, and accounts for the thermal storage capacity of the various envelope components. In the scope of this work, the base model is further enhanced to account for the embodied emissions of the building components and the dynamic grid emission factor. The model's performance is also improved by modeling a continuous energy system and clustering of the input data. A detailed analysis showed that the the weekly clustering approach outperforms daily clustering due to it's ability to account for energy flows over multiple days.

The results of the optimization show that energy storage systems, and intelligent energy system operation are necessary to achieve the full CO2 saving potential of building retrofitting. Gas-based heating is shown to have economic benefits, but at a given point, an overall emission threshold is reached beyond which the shift towards an electricity based heat supply becomes unavoidable. The combined use of heat pumps, electric heaters , photo-voltaic panels, batteries and thermal storage systems dominates the optimization configurations that result in the lowest emissions. Envelope retrofitting is restricted to moderate wall and window upgrades in most cases. The building envelope is only fully retrofitted in the scenario that provides the lowest possible emissions, but results in a considerable increase in costs.

Using simulated grid emissions factors, a comparison between the years 2017, 2030 and 2050 is carried out, the results of which reveal that a direct comparison of the sort is not possible due to limitations set by the used clustering method. The results also show that an accurate comparison is only possible if a detailed correlation analysis precedes the clustering and optimization steps.