Remaining Useful Life prediction on C-MAPSS FD002
12.98RMSEHSMGNN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HSMGNN2025.12 | 12.98 | 10.05 | 168.64 | — | |
| FCSTGNN2025.12 | 13.04 | 10.16 | 170.04 | — | |
| MAGNN2025.12 | 13.09 | 10.76 | 171.35 | — | |
| MsFormer2026.03 | 13.64 | — | — | 766.7 | |
| Bi_cLSTM (Ours, 4 blocks)Year=2025, Pre-processing steps=OS norm; z-score by mode; RUL cap=1252026.02 | 13.96 | — | — | — | |
| MHAF2026.04 | 13.96 | — | — | 904.88 | |
| LSTM + multi-layer self-attnYear=2021, Pre-processing steps=Windows; norm; self-attn2026.02 | 14.02 | — | — | — | |
| RMTF-Transformer2026.04 | 14.02 | — | — | 970.84 | |
| DFormerYear=2024, Pre-processing steps=Transformer variant2026.02 | 14.11 | — | — | — | |
| TS-MLLM2026.03 | 14.22 | — | — | 929.81 | |
| DVGTformer2026.03 | 14.28 | — | — | 797.26 | |
| Neural ODEYear=2023, Pre-processing steps=ODE layer; norm2026.02 | 14.3 | — | — | — | |
| AMSF-AMoE2026.04 | 14.42 | — | — | 1,077.44 | |
| KDnet2026.03 | 14.47 | — | — | 929.2 | |
| MGLSNYear=2023, Pre-processing steps=Multi-granularity; norm2026.02 | 14.57 | — | — | — | |
| PE-Net2026.03 | 14.69 | — | — | 881.73 | |
| TF-SCN2026.03 | 14.71 | — | — | 1,358 | |
| AMR-Net2026.03 | 14.72 | — | — | 1,085.03 | |
| SAGDFN2025.12 | 14.74 | 10.33 | 217.29 | — | |
| MLEAN2026.03 | 14.74 | — | — | 914 | |
| GAT-DAT2026.03 | 14.81 | — | — | 1,163.8 | |
| MSAN2026.03 | 14.91 | — | — | 1,259 | |
| Variational EncodingYear=2022, Pre-processing steps=Variational reg.; norm2026.02 | 14.92 | — | — | — | |
| TATFA-TransformerYear=2024, Pre-processing steps=Transformer + attention2026.02 | 15.07 | — | — | — | |
| CBHRL (Regression)2026.04 | 15.55 | — | — | 1,215 | |
| EAPNYear=2023, Pre-processing steps=Embedded attention; norm2026.02 | 15.68 | — | — | — | |
| EAPN2026.03 | 15.68 | — | — | 1,126.49 | |
| MR-LSTM2026.03 | 15.71 | — | — | 1,434.27 | |
| NSD-TGTN2026.03 | 15.87 | — | — | 1,477 | |
| DiffRULYear=2024, Pre-processing steps=Diffusion-based RUL2026.02 | 15.9 | — | — | — | |
| Bi-LSTM based Attention methodYear=2022, Pre-processing steps=RUL target function2026.02 | 15.94 | — | — | — | |
| GA+PredictorYear=2023, Pre-processing steps=GA tuning; norm2026.02 | 15.99 | — | — | — | |
| C-Transformer2026.03 | 16.11 | — | — | 2,214.59 | |
| Hybrid DL prognosticsYear=2021, Pre-processing steps=Fusion; norm2026.02 | 16.19 | — | — | — | |
| MSTSDNYear=2024, Pre-processing steps=Multi-scale two-stream; norm2026.02 | 16.28 | — | — | — | |
| DeeBERT2026.03 | 16.32 | — | — | 1,429.15 | |
| CNN-SSEYear=2024, Pre-processing steps=CNN + squeeze-excite2026.02 | 16.54 | — | — | — | |
| One Fits AllBackbone=GPT-22026.03 | 16.62 | — | — | 1,309.36 | |
| MDRNN2026.04 | 16.64 | — | — | 2,231 | |
| BLS+TCNYear=2022, Pre-processing steps=Feature sel.; norm; piece-wise RUL2026.02 | 16.87 | — | — | — | |
| BiLSTM2026.03 | 16.91 | — | — | 1,316.97 | |
| DMHA-ATCNYear=2024, Pre-processing steps=Hybrid CNN2026.02 | 16.95 | — | — | — | |
| LSTM (with auto piece-wise)Year=2022, Pre-processing steps=Corr. analysis; median filter; norm; auto piece-wise RUL2026.02 | 17.04 | — | — | — | |
| BiGRU2026.03 | 17.34 | — | — | 1,345.92 | |
| DTW-GPR2026.03 | 17.38 | — | — | 1,714.3 | |
| IMDSSN2026.03 | 17.4 | — | — | 1,775.15 | |
| CDSGYear=2023, Pre-processing steps=Graph structure; norm2026.02 | 18.13 | — | — | — | |
| Siamese2026.04 | 18.18 | — | — | 1,618 | |
| Attention-DCNNYear=2021, Pre-processing steps=Conv features; attention; norm2026.02 | 18.34 | — | — | — | |
| BiGRU + Temporal AttnYear=2022, Pre-processing steps=Norm; temporal attn2026.02 | 18.94 | — | — | — | |
| AEQRNN2026.04 | 19.1 | — | — | 3,220 | |
| GA–RNN/LSTMYear=2021, Pre-processing steps=GA tuning; norm2026.02 | 19.33 | — | — | — | |
| Multi-Scale CNNYear=2020, Pre-processing steps=Multi-scale conv; norm2026.02 | 19.35 | — | — | — | |
| AGCNNYear=2020, Pre-processing steps=Feature sel.; norm; RUL target2026.02 | 19.43 | — | — | — | |
| MT-CNN2026.04 | 19.77 | — | — | 2,023 | |
| Multi-head CNN+LSTMYear=2020, Pre-processing steps=Feature sel.; RUL target2026.02 | 19.93 | — | — | — | |
| CNN+LSTMYear=2019, Pre-processing steps=Var. threshold; norm; HI2026.02 | 20.44 | — | — | — | |
| BiGRU-AS2026.04 | 20.81 | — | — | 2,454 | |
| MegaCRN2025.12 | 21.37 | 16.23 | 454.42 | — | |
| One Fits AllBackbone=Qwen3-0.6B2026.03 | 21.38 | — | — | 4,941.72 | |
| Hybrid modelYear=2021, Pre-processing steps=Feature sel.; norm; piece-wise RUL2026.02 | 22.26 | — | — | — | |
| Semi-Supervised DL+GA2026.04 | 22.73 | — | — | 3,366 | |
| GCU-Transformer2026.04 | 22.81 | — | — | — | |
| FedLSTMYear=2022, Pre-processing steps=Federated learning; norm2026.02 | 22.83 | — | — | — | |
| Bi-level LSTMYear=2022, Pre-processing steps=Hierarchical LSTM2026.02 | 23.14 | — | — | — | |
| BiLSTMYear=2018, Pre-processing steps=Feature sel.; norm; RUL target2026.02 | 23.18 | — | — | — | |
| BiLSTM2026.03 | 23.18 | — | — | 4,130 | |
| MATT2025.12 | 23.9 | 19.38 | 571.1 | — | |
| QAQLstatus=proposed2026.06 | 24.35 | — | — | — | |
| LSTMYear=2017, Pre-processing steps=Data normalization; RUL target2026.02 | 24.49 | — | — | — | |
| RUL-RNN2026.04 | 24.67 | — | — | — | |
| TFSCL2026.03 | 25.05 | — | — | 2,852.54 | |
| Crossformer2025.12 | 28.8 | 25.56 | 839.19 | — | |
| CNNYear=2016, Pre-processing steps=Data normalization; RUL target2026.02 | 30.29 | — | — | — | |
| Quantum VQL2026.06 | 32.89 | — | — | — | |
| Quantum DQN2026.06 | 33.78 | — | — | — | |
| Quantum AOA2026.06 | 36.65 | — | — | — | |
| Transformer2026.06 | 39.38 | — | — | — | |
| Quantum Eigensolver2026.06 | 39.94 | — | — | — | |
| PPO2026.06 | 40.07 | — | — | — | |
| Deep Q-Network2026.06 | 40.22 | — | — | — | |
| Quantum SGD2026.06 | 40.7 | — | — | — | |
| SARSA2026.06 | 41.86 | — | — | — | |
| SGD2026.06 | 43.47 | — | — | — | |
| Quantum Decision Tree2026.06 | 43.48 | — | — | — | |
| Quantum LSTM2026.06 | 43.69 | — | — | — | |
| GRU2026.06 | 48.31 | — | — | — | |
| Bi-LSTM2026.06 | 52.42 | — | — | — |