在等離子體診斷模型中基準測試傳感器穩健性：基於TokaMark的系統性評估

摘要

托卡马克聚变装置的等离子体诊断模型几乎普遍基于干净、完整的传感器数据进行评估。然而在实际运行中，聚变诊断系统经常出现故障：采集系统启动延迟、单个传感器失效，且信号中断恰好在等离子体破裂临近时集中发生。我们首次提出针对等离子体诊断机器学习的系统性鲁棒性基准测试，使用包含11,573次MAST放电数据的TokaMark数据集，评估了XGBoost、LSTM、Transformer及TokaMark CNN基准模型在六种基于物理的故障场景和三种插补策略下的表现。我们引入鲁棒性评分（Robustness Score, RS）实现跨架构的标准化比较。核心发现是：临近破裂时的传感器故障（在最终窗口时间步注入的损坏）导致序列模型性能崩溃（LSTM的NRMSE增加212%），而统计特征模型保持相对稳定（XGBoost增加37%）。前向填充插补几乎消除了随机丢失对序列模型的所有退化影响（LSTM从+57%降至约0%），但在窗口末端数据损坏时几乎无效。基于真实破裂时间戳的放电级别报警评估显示，在临近传感器故障下LSTM报警检测性能崩溃（TPR=0.00），而均值填充插补可将其恢复至TPR=1.00，这与NRMSE中观察到的模式完全相反。在所有架构中，等离子体电流被证实为最关键的单通道诊断量（移除后误差增加73%至140%）。代码、数据和预训练检查点已公开于https://github.com/Neerav-Gupta/tokamark-robustness。

English

Plasma diagnostic models for tokamak fusion devices are almost universally evaluated on clean, complete sensor data. In practice, fusion diagnostics fail regularly: acquisition systems start late, individual sensors die, and signal dropouts cluster precisely when a plasma disruption is approaching. We present the first systematic robustness benchmark for plasma diagnostic ML using the TokaMark dataset of 11,573 MAST shots, evaluating XGBoost, LSTM, Transformer, and the TokaMark CNN baseline across six physically-grounded failure scenarios and three imputation strategies. We introduce the Robustness Score (RS) for standardized cross-architecture comparison. Our central finding is that disruption-proximate sensor failure (corruption injected in the final window timesteps) collapses sequence model performance (LSTM +212% NRMSE) while a statistical feature model remains comparatively stable (XGBoost +37%). Forward-fill imputation eliminates nearly all degradation from random dropout for sequence models (LSTM +57% to ~0%), but offers little help when the end of the window is corrupted. Shot-level alarm evaluation using ground-truth disruption timestamps reveals that LSTM alarm detection collapses to TPR=0.00 under proximate sensor failure, while mean-fill imputation recovers it to TPR=1.00, a reversal of the pattern observed in NRMSE. Plasma current emerges as the single most critical diagnostic across all architectures (+73% to +140% upon removal). Code, data, and trained checkpoints are available at https://github.com/Neerav-Gupta/tokamark-robustness.