A coordinated dispatch strategy for fluctuation mitigation in data centers considering flexible computing load
Next-generation green power direct-connection data centers face the dual challenges of the random nature of renewable energy output and short-term load fluctuations, which can cause significant power fluctuations at the grid connection point. This not only places enormous strain on grid operations but also poses security risks to the data centers themselves. Traditional scheduling paradigms often treat data centers as rigid loads, overlooking the inherent flexibility of computing tasks. In light of this, this paper proposes a coordinated optimization strategy for data center microgrids. First, a thermodynamic model incorporating temperature coupling is established to quantify the interactive energy consumption between IT systems and cooling systems, then a mixed-integer linear programming (MILP) model is constructed with the objective of minimizing operational costs, comprehensively considering grid power purchases, energy storage degradation, and penalty terms for workload shifting under time-of-use pricing. Simulation results indicate that, compared to traditional rigid operation modes, the proposed strategy can effectively shift peak loads, significantly reduce the curtailment rate of photovoltaic power, and demonstrate good economic performance.