LLaTSA improves power grid stability predictions using language model alignment
LLaTSA: Large Language Model-Aligned General-Purpose Transient Stability Analysis
Artificial Intelligence
Summary
Predicting how power grids stay stable after disturbances is important but hard because models often only work for specific setups. The authors created LLaTSA, which uses a large language model to understand and predict these stability scenarios more generally. It adds extra information to help the model know what kind of problem it’s solving and looks at how different grid parts interact. Their tests show it predicts future behavior well and can work on setups it wasn’t trained on.
What this means in practice
- •For power grid operators: Predict system stability under varied conditions without retraining for each network change.
- •For energy system modelers: Simulate dynamic responses in different generation mixes using aligned language model embeddings for better accuracy.
Authors
Chao Shen, Hongwei Zhen, Junyan Shao, Zhenghao Yang, Yifan Zhang, Mingyang Sun
Abstract
Dynamic trajectory prediction has become an important paradigm for data-driven transient stability analysis (TSA), yet most existing predictors remain system-specific and require substantial retraining when network configurations, generation mixes, or state-variable sets change. Uni-TSA introduced a general-purpose TSA framework that combines channel-independent modeling with a pretrained large language model (LLM) predictor. Nevertheless, its application to heterogeneous systems is limited by ambiguity in short observations, a mismatch between numerical trajectories and LLM embeddings, neglected coupling among state variables, and the high inference cost of dense backbones. This paper proposes LLaTSA, an LLM-aligned framework for general-purpose trajectory-based TSA. LLaTSA first incorporates operating conditions, disturbance attributes, and state-variable identity through a structured textual prefix. It then aligns normalized temporal patches with a TSA-related vocabulary before processing them with a pretrained sparse decoder-only mixture-of-experts (MoE) backbone. A state-variable coupling module captures coordinated post-fault evolution, while teacher forcing and rollout-based training support iterative long-horizon prediction. Case studies on multiple test systems demonstrate accurate trajectory prediction, reliable stability discrimination, and effective adaptation across unseen scenarios.