Renewable integration strategies for data centers : opportunities and challenges for efficiency and scalability

Abstract

The thesis investigates the way in which renewable energy, battery storage, UPS operation, and flexible IT workloads can be combined in a grid-connected data center to reduce the grid electricity consumption and to use renewable energy efficiently. An hourly operational optimization model was developed using Python and Pyomo and then simulated for the whole year of 2025 based on weather data which was obtained from the Finnish Meteorological Institute. The model considers photovoltaic and wind power generation, battery storage, UPS operation, both flexible and non-flexible IT workloads, cooling demand, and grid electricity. Six operating scenarios were analyzed by varying renewable-generation capacity, battery capacity and the permitted level of workload flexibility. Their effects on grid electricity consumption, renewable-energy utilization, storage operation and workload scheduling were then evaluated.

The results show that the largest reduction in grid electricity imports occurs when renewable-generation capacity is increased. A larger battery is most useful when sufficient surplus renewable electricity is available for charging. In high-renewable scenarios, shifting flexible workloads reduces grid electricity use and enables more of the available renewable energy to be utilized. The study finds that for renewables to be effectively and massively used in the data centers, they need to be balanced with flexible demand, renewable generation and storage capacity. Instead of increasing each of these elements independently, they must all be raised simultaneously.