Master's thesis Marcel Nebeling


Investigation of uncertainties in thermal simulation of city districts by use of sensitivity analysis

Master's thesis Nebeling Copyright: EBC Schematic model of the energy system with energy supply by heat pump

The integration of large amounts of renewable energy increases the fluctuations in the residual load, which have to be compensated by conventional power plants or other facilities. The consideration of an entire urban quarter is an issue as a decentralized energy system that is able to act both as energy consumers and producers. Especially there are systems needed that generate heat and are also connected to the electricity grid. During a power surplus heat pumps can be used. These require electrical power and generate heat. In the opposite case, combined heat and power systems can be operated at a power deficit. These generate heat in addition to the required electrical energy. The heat is either used directly for the heating demand or stored in thermal storage.

In studies for city quarters various parameters are relevant which are fraught with uncertainty. These include for example the used materials or the insulation standards of each building. In addition, parameters dependent on the user's behavior should be considered, which may be different for various types of buildings. Another significant point is the analysis of the parameters based on the energy demand. The heat produced by heat pumps and CHP plants is affected with uncertain input data. The reason is that the supply of solar and wind energy remain difficult to predict accurately.

Due to the dynamics of the systems it is useful to build a city district inside a dynamic modeling tool for simulation and to get insight into the thermal and electrical behavior. The district includes several one family dwellings with certain different properties such as the year of construction.
In this work the influences of various parameters on energy efficiency and the storage capacity of a city quarter will be considered and classified according to relevance. In this regard a method has been developed which recognizes the influence of the various parameters. Through dynamic simulations in Dymola/Modelica results were generated for a sensitivity analysis. First of all a reference simulation of the city quarter was performed. Furthermore the parameters were modified to determine the changes in the simulation results. Finally sensitivity coefficients were calculated by using the differential method, which can be compared with each other. In addition, a factorial analysis was performed to detect interaction between the parameters. The parameters with the greatest impact were in general the indoor temperature, the wall thickness, the thermal conductivity of the wall material and the air exchange rate.