Optimierung der Auslegung einer resilienten erneuerbaren Energieversorgung für Rechenzentren

The growing demand for computing power and data processing, driven by increasing digitalization and use of artificial intelligence, raises the energy consumption of data centres (DC). As a result, efficient and sustainable energy supply concepts are becoming the focus of research. This thesis examines the integration of renewable energy sources such as photovoltaics (PV) and wind energy, as well as storage systems in data centres. Moreover, their energy and economic impacts are evaluated. The basis of the analysis are the real load profiles of a data centre, the generation profiles of PV and wind energy systems and the charging and discharging profiles of battery storage systems. By using these, several supply scenarios are examined and compared to reference scenarios without on-site generation. Initially, the sizing of the energy systems are analysed to achieve self-sufficient operation of the DC. In the next step, for an economic evaluation and design of the energy systems, the net present value (NPV) method is used. In addition to the initial investment costs for PV and wind power plants as well as battery storage systems, the ongoing operating and grid electricity costs are included. To generate an extra revenue stream, a theoretical waste heat utilization is modeled. The NPV calculation analyses the system dimensions at which a positive NPV (i.e., profitability) is achieved over the system¿s lifetime. The results show that achieving self-sufficiency in the data centre requires a significant oversizing of the energy systems. However, even the smallest scales increase the profitability compared to the reference scenario. A particularly favourable development of the NPV results from a balanced combination of energy sources as well as from a PV-only supply. The Wind-only scenario improves the NPV, however cannot reach full profitability over the project lifetime. Furthermore, the revenue ratio is improved when reusing the waste heat. Nevertheless, all scenarios require high system scales to achieve a positive NPV. In conclusion, this thesis enables an evaluation of energy and economic aspects of several power supply concepts for data centres, and their contribution to reducing CO2 emissions.