Bachelor's thesis Tobias Beckhölter


Clustering of Buildings on City District Level for Energetic Optimization

To optimize the distribution of energy plants on city district level an optimization model, based on computer language Python, was developed at the Institute for Energy Efficient Buildings and Indoor Climate (EBC) at RWTH Aachen University. This model uses different data, e. g. building type, age, energy consumption, position, street and other networks, to solve a mixed-integer linear programming (MILP) problem. Its aim is the minimization of total costs or greenhouse gas emissions. The applicability of this optimization model to city districts with more than eight buildings turns out to be very time consuming or even not feasible. Because of the rising complexity due to an increasing number of buildings the run time of the optimization is approximately exponential.
To analyze the potential for an optimized distribution of energy plants in city districts with a higher number of buildings, the district can be divided into smaller groups. In order to prove the thesis, multiplemethods of clustering will be depicted and explained by using an authentic city district. The developed methods consider the existent street network and the energy consumption of each building so that they are adapted to the features of a city district analysis. By implementing several tests it can be shown that a computing time reduction of over 80% is reached by the use of the developed clustering methods. Finally, further tests concerning the optimization of the calculated clustering results should determine whether the quality of the optimization is sufficient.