Novel method for data center placement using multiobjective optimization with thermohydraulic models of existing district heating systems

Integrating data centers into district heating systems is on the rise, due to the low cost of the waste heat, and energy efficiency targets for both data center and district heating companies. This study presents a novel method for data center placement using multiobjective optimization with thermohydraulic models. A 1.5 MW data center producing 75 °C waste heat was simulated across the entire distribution lines of existing district heating systems at a 100 m interval. Location suitability index (LSI) was formed to compare each location based on three performance indicators: total production costs, average and maximum supply temperature reductions for customers. The objective was to determine the data center waste heat utilization potential based on how the waste heat mixes into other high temperature production in each specific location. The results show that the data center placement has a major impact on the waste heat utilization potential and overall profitability. The best locations near other production plants had a utilization potential of 93–99 % compared to full power potential, while the worst locations near demand had less than 60 %. The yearly production cost reduction was up to 5 % at Valkeakoski and 12 % at Kangasala, with a waste heat price of 10€/MWh.