A multi-objective bilevel planning for data center integrated energy systems with waste heat utilization

As global data center energy demands surge, traditional energy systems face critical challenges in balancing multi-energy coupling and coordinating interactions among data centers and regional energy systems. To address these challenges, this study designs a transformative low-carbon data center integrated energy system and introduces a data center waste heat matching index, which can evaluate the alignment between data center waste heat utilization and data center stakeholders. Considering the economic cost, data center waste heat matching index, and exergy efficiency, this study establishes a multi-objective bilevel planning model for data center integrated energy systems, including the upper-level capacity allocation and lower-level scheduling optimization models. The proposed system achieves a 49 % primary energy saving rate and a 44 % green energy coefficient, with CO2 emissions reduced by 121,000 tons by comparing three different system configurations. This study further explores the relationship between data centers and regional energy systems, revealing the impact of the proposed index on data center stakeholders. Variations in electricity prices by ±20 % lead to a 12 % fluctuation in total economic costs, while reductions in hydrogen prices by ±20 % result in a 19.4 % increase in green energy coefficient. A comprehensive sensitivity analysis shows that while external market dynamics, especially hydrogen and electricity prices, are critical, the proposed performance index is most sensitive to internal load and renewable energy fluctuations. This work can provide actionable insights and a systematic framework for guiding data center integrated energy system planning to promote the sustainable development of green data centers.