Energy storage plays an essential role in stabilizing fluctuations in renewable energy sources such as wind and solar, enabling surplus electricity retention, and delivering dynamic frequency regulation. However, relying solely on a single form of storage often proves insufficient due to constraints in performance, capacity, and cost-effectiveness. To tackle frequency regulation challenges in remote desert-based renewable energy hubs—where traditional power infrastructure is unavailable—this study introduces a planning framework for an electro-hydrogen energy storage system with grid-forming capabilities, designed to supply both inertia and frequency response. At the system design stage, a direct current (DC) transmission network is modeled, integrating battery and hydrogen storage technologies. Using this configuration, the capacity settings for both grid-forming batteries and hydrogen units are optimized. This study then explores how hydrogen systems—comprising electrolyzers, storage tanks, and fuel cells—and grid-forming batteries contribute to inertial support. Virtual inertia models are established for each technology, enabling precise estimation of the total synthetic inertia provided. At the operational level, this study addresses stability concerns stemming from renewable generation variability by introducing three security indices. A joint optimization is performed for virtual inertia constants, which define the virtual inertia provided by energy storage systems to assist in frequency regulation, and primary frequency response parameters within the proposed storage scheme are optimized in this model. This enhances the frequency modulation potential of both systems and confirms the robustness of the proposed approach. Lastly, a real-world case study involving a 13 GW renewable energy base in Northwest China, connected via a ±10 GW HVDC export corridor, demonstrates the practical effectiveness of the optimization strategy and system configuration.
With the large-scale integration of renewable energy and power electronic devices, power quality disturbances exhibit strong nonlinearity and complex dynamic behavior. Traditional methods are limited by insufficient feature extraction and cumbersome classification, often failing to meet practical accuracy and robustness requirements. To address this issue, this paper proposes a multi-level ensemble method for power quality disturbance identification. A time–frequency dual-branch feature extraction module was designed, combining residual networks and bidirectional temporal convolutional networks to capture both local discriminative features and long-range temporal dependencies in the time and frequency domains. A cross-attention mechanism was further employed to fuse the time–frequency features, enabling adaptive focus on the most critical information for disturbance classification. The fused features were fed into fully connected layers and a Softmax classifier for multi-class identification. Experimental results demonstrated superior accuracy, robustness, and generalization capability compared with existing methods, validating the effectiveness of the proposed model.