Flexible Demand–Driven LCOE Optimization of Off-Grid Data Centers with Hybrid Renewable, Battery, and Hydrogen Systems

Abstract

This contribution presents an optimization framework for computing the optimal Levelized Cost of Energy (LCOE) of an off-grid data center that incorporates demand flexibility as a decision variable. The total electricity consumption of the data center is endogenously optimized, reflecting the ability of digital workloads to adapt to energy availability. This feature leads to a nonlinear LCOE formulation, since both total system costs and total delivered energy depend on decision variables. To address this challenge, the model applies the Charnes–Cooper transformation, enabling a reformulation of the fractional objective into a linear optimization problem. The framework jointly optimizes investment and operation decisions, capturing both CAPEX and OPEX within a single model. It determines optimal capacities of solar PV and wind power using normalized generation profiles, short-term electrical storage based on Li-ion batteries, and long-term energy storage through hydrogen technologies. An optional gas-fired generator is included to evaluate the cost impact of different renewable penetration levels. The resulting LCOE estimates provide a quantitative basis to support data center siting decisions.