Master's Thesis Markus Schraven

 

Optimal design of energy supply systems for local neighbourhoods in consideration of passive building elements

Example realization Copyright: EBC Example realization of a local neighborhood

Compared to renowned single building optimizations an inspection of coupled buildings, as within the meaning of local neighborhoods where buildings are linked via a local electricity grid, reveals additional energetic as well as monetary advantages but also entails model-related disadvantages. Thus, synergy effects occur whereby, for instance, power generators in some buildings supply electrical-supplied heat generators in others. On the other hand, computing times and storage requirements increase with the coupled consideration of all variables and this compact model formulation is barely applicable for case studies with even minor numbers of buildings. However, this type of model where most of the constraints are independent and only few linking relations exist can be reformulated by Dantzig-Wolfe decomposition, hence becoming an easier problem, which can be solved by an iterative column generation approach.

A verification shows that the column generation model in general is scalable, however the solution quality depends on the defined termination criteria. The necessary number of iterations for an accurate result is depending on the problem itself. Most of the iterations barely lead to any progress and the convergence shows an incoherent behavior. To accelerate the convergence the setting of a fine starting solution or the application of stabilizing methods could contribute.

  caseStudy Copyright: EBC Synergy effects between electricity generation and electricity-powered heat generation for two summer days and a winter day and linked buildings: The high power production on summer days comes along with a high heat generation by solar-thermal collectors

A verification shows that the column generation model in general is scalable, however the solution quality depends on the defined termination criteria. The necessary number of iterations for an accurate result is depending on the problem itself. Most of the iterations barely lead to any progress and the convergence shows an incoherent behavior. To accelerate the convergence the setting of a fine starting solution or the application of stabilizing methods could contribute.

A case study for a local neighborhood comprising 20 buildings shows an increased ratio of self-consumed electricity of 40.54 % compared to 31.73 % in the case with the electricity exchange prohibited. This results in cost and emission savings of 0.21 % and 6.22 % respectively. However, higher synergies are not achieved, as the heat demand on summer days is covered by solar-thermal collectors making the high PV power production not quite usable. Besides, without consideration of a heating network, each building requires its own heat supply, which inhibits the installation of big central devices.