Data-aided collaborative planning of energy storage for remote renewable base

• Data-aided optimization-based planning framework considering operational robustness. • Bidirectional shareable flexibility reduces inter-regional storage investment costs. • Uncertainty interval prediction quantifies the operational flexibility demands. • A suitable conservative level achieves cost reduction while guaranteeing robustness. • A piecewise technique approximates the nonlinear unit variable operating conditions. Renewable bases are typically located far from load centers. High-voltage direct current tie-lines have therefore emerged as a promising solution for transmitting remote renewable power to load centers. Effective planning of energy storage and tie-line transmission power in inter-regional systems is crucial for promoting large-scale renewable accommodation. This work proposes a data-aided planning framework for inter-regional power systems that incorporates bidirectional flexibility from both energy storage and thermal units. A temporally correlated interval prediction method is developed to construct a data-informed operational envelope. By explicitly capturing temporal correlations in prediction errors, the resulting uncertainty intervals are tightened, thereby reducing planning conservatism. Flexibility-sharing and deliverability constraints are integrated into the framework to ensure operational robustness while lowering overall investment costs. The proposed data-aided planning framework enables flexibility deployed in the planning stage to accommodate operational uncertainties. To handle nonlinear terms arising from deep peak-regulation constraints of thermal units, a piecewise least-squares linearization technique is introduced, transforming the problem into a mixed-integer linear programming formulation. Simulation results indicate that bidirectional flexibility reduces investment and operational costs. Comparable robustness is attained under less conservative flexibility constraints, which further contributes to cost savings. The proposed linearization approach solves the problem effectively while maintaining acceptable optimality.