中国机械工程 ›› 2026, Vol. 37 ›› Issue (1): 209-222.DOI: 10.3969/j.issn.1004-132X.2026.01.022
• 智能制造 • 上一篇
崔硕1,2(
), 刘秀丽1,2(
), 李相杰3, 吴国新1,2
收稿日期:2024-10-22
出版日期:2026-01-25
发布日期:2026-02-05
通讯作者:
刘秀丽
作者简介:崔硕,男,1999年生,硕士研究生。研究方向为机电装备寿命预测与健康管理。发表论文2篇。E-mail:1317754704@qq.com基金资助:
CUI Shuo1,2(
), LIU Xiuli1,2(
), LI Xiangjie3, WU Guoxin1,2
Received:2024-10-22
Online:2026-01-25
Published:2026-02-05
Contact:
LIU Xiuli
摘要:
针对高端旋转机械剩余使用寿命预测中的准确性和不确定性量化问题,提出了基于变分深度高斯过程(VDGP)的预测方法。通过构建深度高斯过程更新模型实现不确定性的递推量化,并采用诱导点和变分推断提高大数据的处理能力。C-MAPSS和风机行星齿轮数据集的实验表明,VDGP比高斯过程方法具有更高的预测准确度和更窄的置信区间,在C-MAPSS的 FD002数据集上,均方根误差、评分函数分别比现有最佳的对比方法减小0.21%和45.3%。
中图分类号:
崔硕, 刘秀丽, 李相杰, 吴国新. 高端旋转机械剩余使用寿命预测及其不确定性量化评估方法[J]. 中国机械工程, 2026, 37(1): 209-222.
CUI Shuo, LIU Xiuli, LI Xiangjie, WU Guoxin. Prediction of RUL and Uncertainty Quantification Evaluation Methods for High-end Rotating Machinery[J]. China Mechanical Engineering, 2026, 37(1): 209-222.
| 核函数 | 函数形式 |
|---|---|
| RBF | |
| Matern5/2 | |
| Matern3/2 | |
| Linear | |
| Polynomial | |
| RQ | |
| Periodic |
表1 核函数的类型
Tab.1 Types of kernel functions
| 核函数 | 函数形式 |
|---|---|
| RBF | |
| Matern5/2 | |
| Matern3/2 | |
| Linear | |
| Polynomial | |
| RQ | |
| Periodic |
| 层与核函数 | 最后周期 | ERUL | |||||
|---|---|---|---|---|---|---|---|
| 第1层 | 第2层 | ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE |
| RBF | RBF | 17.06 | 553.55 | 11.87 | 3.40 | 2.96 | 2.76 |
| Matern5/2 | Matern5/2 | 16.80 | 578.33 | 11.74 | 3.84 | 3.49 | 3.22 |
| Matern3/2 | Matern3/2 | 17.04 | 589.11 | 11.96 | 4.44 | 3.91 | 3.61 |
| Linear | Linear | 21.27 | 1084.26 | 16.57 | 16.69 | 54.16 | 13.38 |
| Polynomial | Polynomial | 18.04 | 666.45 | 12.95 | 7.72 | 7.53 | 6.47 |
| RQ | RQ | 17.00 | 609.10 | 11.90 | 3.54 | 3.20 | 2.96 |
| Periodic | Periodic | 17.51 | 603.14 | 12.62 | 5.34 | 4.67 | 4.45 |
表2 核函数的实验结果
Tab.2 Experimental results of kernel functions
| 层与核函数 | 最后周期 | ERUL | |||||
|---|---|---|---|---|---|---|---|
| 第1层 | 第2层 | ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE |
| RBF | RBF | 17.06 | 553.55 | 11.87 | 3.40 | 2.96 | 2.76 |
| Matern5/2 | Matern5/2 | 16.80 | 578.33 | 11.74 | 3.84 | 3.49 | 3.22 |
| Matern3/2 | Matern3/2 | 17.04 | 589.11 | 11.96 | 4.44 | 3.91 | 3.61 |
| Linear | Linear | 21.27 | 1084.26 | 16.57 | 16.69 | 54.16 | 13.38 |
| Polynomial | Polynomial | 18.04 | 666.45 | 12.95 | 7.72 | 7.53 | 6.47 |
| RQ | RQ | 17.00 | 609.10 | 11.90 | 3.54 | 3.20 | 2.96 |
| Periodic | Periodic | 17.51 | 603.14 | 12.62 | 5.34 | 4.67 | 4.45 |
| 层与核函数 | 最后周期 | ERUL | |||||
|---|---|---|---|---|---|---|---|
| 1层 | 2层 | ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE |
| RBF | Matern5/2 | 17.61 | 599.46 | 12.28 | 3.58 | 3.18 | 3.03 |
| Matern5/2 | RBF | 17.66 | 624.40 | 12.31 | 3.19 | 2.80 | 2.51 |
| RBF | RQ | 17.67 | 619.58 | 12.39 | 3.90 | 3.48 | 3.35 |
| RQ | RBF | 17.93 | 658.38 | 12.55 | 4.07 | 3.40 | 3.17 |
| Matern5/2 | RQ | 17.89 | 634.52 | 12.56 | 3.67 | 3.36 | 3.27 |
| RQ | Matern5/2 | 17.96 | 641.72 | 12.68 | 4.64 | 3.92 | 3.71 |
表3 不同核函数的组合实验结果
Tab.3 Experimental results of different kernel function combinations
| 层与核函数 | 最后周期 | ERUL | |||||
|---|---|---|---|---|---|---|---|
| 1层 | 2层 | ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE |
| RBF | Matern5/2 | 17.61 | 599.46 | 12.28 | 3.58 | 3.18 | 3.03 |
| Matern5/2 | RBF | 17.66 | 624.40 | 12.31 | 3.19 | 2.80 | 2.51 |
| RBF | RQ | 17.67 | 619.58 | 12.39 | 3.90 | 3.48 | 3.35 |
| RQ | RBF | 17.93 | 658.38 | 12.55 | 4.07 | 3.40 | 3.17 |
| Matern5/2 | RQ | 17.89 | 634.52 | 12.56 | 3.67 | 3.36 | 3.27 |
| RQ | Matern5/2 | 17.96 | 641.72 | 12.68 | 4.64 | 3.92 | 3.71 |
| 核函数 | 最后周期 | ERUL | ||||
|---|---|---|---|---|---|---|
| ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE | |
| RBF | 18.47 | 650.63 | 13.31 | 3.86 | 3.82 | 3.13 |
| Matern5/2 | 19.40 | 744.54 | 14.07 | 3.62 | 3.20 | 2.70 |
| RQ | 18.51 | 631.68 | 13.20 | 4.34 | 4.47 | 3.57 |
表4 3层的核函数实验结果
Tab.4 Experimental results of 3-layer kernel functions
| 核函数 | 最后周期 | ERUL | ||||
|---|---|---|---|---|---|---|
| ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE | |
| RBF | 18.47 | 650.63 | 13.31 | 3.86 | 3.82 | 3.13 |
| Matern5/2 | 19.40 | 744.54 | 14.07 | 3.62 | 3.20 | 2.70 |
| RQ | 18.51 | 631.68 | 13.20 | 4.34 | 4.47 | 3.57 |
| 模型 | 第一层 | 第二层 |
|---|---|---|
| VDGP-R-R | RBF | RBF |
| VDGP-M-M | Matern5/2 | Matern5/2 |
| VDGP-R-M | RBF | Matern5/2 |
| VDGP-M-R | Matern5/2 | RBF |
表5 VDGP模型配置
Tab.5 VDGP model configurations
| 模型 | 第一层 | 第二层 |
|---|---|---|
| VDGP-R-R | RBF | RBF |
| VDGP-M-M | Matern5/2 | Matern5/2 |
| VDGP-R-M | RBF | Matern5/2 |
| VDGP-M-R | Matern5/2 | RBF |
| 模型 | 指标 | FD001 | FD002 | FD003 | FD004 |
|---|---|---|---|---|---|
| MLP[ | ERMSE | 37.56 | 80.03 | 37.39 | 77.37 |
| ESCORE | 1.80×104 | 7.80×106 | 1.74×104 | 5.62×106 | |
| CNN[ | ERMSE | 18.45 | 30.29 | 19.82 | 29.16 |
| ESCORE | 1286.70 | 13570 | 1596 | 7886 | |
| LSTM[ | ERMSE | 16.14 | 24.49 | 16.18 | 28.17 |
| ESCORE | 338 | 4450 | 852 | 5550 | |
LSTM with attention[ | ERMSE | 14.53 | 27.08 | ||
| ESCORE | 322.53 | 5649.14 | |||
| HMC[ | ERMSE | 13.84 | 20.74 | 14.41 | 22.73 |
| ESCORE | 427 | 19400 | 2977 | 10376 | |
| Transformer[ | ERMSE | 11.27 | 22.81 | 11.42 | 24.86 |
| ESCORE | 213.80 | 2589 | 271.30 | 3029 | |
| VarSeqLSTM[ | ERMSE | 15.23 | 22.87 | 14.53 | 26.11 |
| ESCORE | 250 | 4532 | 1523 | 5627 | |
| BiGRU-AS[ | ERMSE | 13.68 | 20.81 | 15.53 | 27.31 |
| ESCORE | 284 | 2454 | 428 | 4708 | |
| AE-MSEN[ | ERMSE | 14.57 | 20.18 | 15.58 | 23.72 |
| ESCORE | 342.38 | 1781.81 | 1231.5 | 3698.01 | |
Bi-level LSTM Scheme[ | ERMSE | 11.80 | 23.14 | 12.37 | 23.38 |
| ESCORE | 194 | 3771 | 224 | 3492 | |
| VAE+RNN[ | ERMSE | 15.81 | 24.12 | 14.88 | 26.54 |
| ESCORE | 326 | 4183 | 722 | 5634 | |
| EAGDE-SVM[ | ERMSE | 14.10 | 20.60 | 18.86 | 26.40 |
| ESCORE | 253.11 | 3122.20 | 514.50 | 4795.02 | |
| C2D2M2[ | ERMSE | 12.32 | 20.81 | 15.32 | 22.43 |
| ESCORE | 221 | 3062 | 369 | 3118 | |
RCNN+ ABi-LSTM[ | ERMSE | 12.98 | 19.16 | 13.24 | 22.29 |
| ESCORE | 258 | 2980 | 246 | 3795 | |
Attention- LSTM[ | ERMSE | 15.45 | 20.91 | 14.67 | 24.01 |
| ESCORE | 455.92 | 3602.94 | 473.97 | 6841.82 | |
| MACNN[ | ERMSE | 18.45 | 19.77 | 18.05 | 22.80 |
| ESCORE | 704.50 | 1633.65 | 1045.22 | 3430.29 | |
| CBLSTM[ | ERMSE | 17.40 | 19.80 | 19.11 | 22.47 |
| ESCORE | 588.96 | 2590.20 | 2059.90 | 3523.17 | |
| LSTM-Z[ | ERMSE | 17.77 | 19.30 | 18.74 | 22.63 |
| ESCORE | 637.02 | 1678.59 | 2345.60 | 2950.77 | |
| VDGP-R-M | ERMSE | 17.61 | 21.82 | 18.25 | 25.43 |
| ESCORE | 599.46 | 3221.40 | 1407.07 | 5188.19 | |
| VDGP-M-R | ERMSE | 17.66 | 21.62 | 18.42 | 25.38 |
| ESCORE | 624.40 | 3053.05 | 1628.08 | 5766.34 | |
| VDGP-R-R | ERMSE | 17.06 | 19.50 | 18.66 | 21.92 |
| ESCORE | 553.55 | 1720.84 | 2826.08 | 3082.37 | |
| VDGP-M-M | ERMSE | 16.80 | 19.12 | 17.68 | 22.52 |
| ESCORE | 578.33 | 1631.69 | 640.69 | 2855.57 |
表6 VDGP与其他方法的实验结果
Tab.6 Experimental results of VDGP and other methods
| 模型 | 指标 | FD001 | FD002 | FD003 | FD004 |
|---|---|---|---|---|---|
| MLP[ | ERMSE | 37.56 | 80.03 | 37.39 | 77.37 |
| ESCORE | 1.80×104 | 7.80×106 | 1.74×104 | 5.62×106 | |
| CNN[ | ERMSE | 18.45 | 30.29 | 19.82 | 29.16 |
| ESCORE | 1286.70 | 13570 | 1596 | 7886 | |
| LSTM[ | ERMSE | 16.14 | 24.49 | 16.18 | 28.17 |
| ESCORE | 338 | 4450 | 852 | 5550 | |
LSTM with attention[ | ERMSE | 14.53 | 27.08 | ||
| ESCORE | 322.53 | 5649.14 | |||
| HMC[ | ERMSE | 13.84 | 20.74 | 14.41 | 22.73 |
| ESCORE | 427 | 19400 | 2977 | 10376 | |
| Transformer[ | ERMSE | 11.27 | 22.81 | 11.42 | 24.86 |
| ESCORE | 213.80 | 2589 | 271.30 | 3029 | |
| VarSeqLSTM[ | ERMSE | 15.23 | 22.87 | 14.53 | 26.11 |
| ESCORE | 250 | 4532 | 1523 | 5627 | |
| BiGRU-AS[ | ERMSE | 13.68 | 20.81 | 15.53 | 27.31 |
| ESCORE | 284 | 2454 | 428 | 4708 | |
| AE-MSEN[ | ERMSE | 14.57 | 20.18 | 15.58 | 23.72 |
| ESCORE | 342.38 | 1781.81 | 1231.5 | 3698.01 | |
Bi-level LSTM Scheme[ | ERMSE | 11.80 | 23.14 | 12.37 | 23.38 |
| ESCORE | 194 | 3771 | 224 | 3492 | |
| VAE+RNN[ | ERMSE | 15.81 | 24.12 | 14.88 | 26.54 |
| ESCORE | 326 | 4183 | 722 | 5634 | |
| EAGDE-SVM[ | ERMSE | 14.10 | 20.60 | 18.86 | 26.40 |
| ESCORE | 253.11 | 3122.20 | 514.50 | 4795.02 | |
| C2D2M2[ | ERMSE | 12.32 | 20.81 | 15.32 | 22.43 |
| ESCORE | 221 | 3062 | 369 | 3118 | |
RCNN+ ABi-LSTM[ | ERMSE | 12.98 | 19.16 | 13.24 | 22.29 |
| ESCORE | 258 | 2980 | 246 | 3795 | |
Attention- LSTM[ | ERMSE | 15.45 | 20.91 | 14.67 | 24.01 |
| ESCORE | 455.92 | 3602.94 | 473.97 | 6841.82 | |
| MACNN[ | ERMSE | 18.45 | 19.77 | 18.05 | 22.80 |
| ESCORE | 704.50 | 1633.65 | 1045.22 | 3430.29 | |
| CBLSTM[ | ERMSE | 17.40 | 19.80 | 19.11 | 22.47 |
| ESCORE | 588.96 | 2590.20 | 2059.90 | 3523.17 | |
| LSTM-Z[ | ERMSE | 17.77 | 19.30 | 18.74 | 22.63 |
| ESCORE | 637.02 | 1678.59 | 2345.60 | 2950.77 | |
| VDGP-R-M | ERMSE | 17.61 | 21.82 | 18.25 | 25.43 |
| ESCORE | 599.46 | 3221.40 | 1407.07 | 5188.19 | |
| VDGP-M-R | ERMSE | 17.66 | 21.62 | 18.42 | 25.38 |
| ESCORE | 624.40 | 3053.05 | 1628.08 | 5766.34 | |
| VDGP-R-R | ERMSE | 17.06 | 19.50 | 18.66 | 21.92 |
| ESCORE | 553.55 | 1720.84 | 2826.08 | 3082.37 | |
| VDGP-M-M | ERMSE | 16.80 | 19.12 | 17.68 | 22.52 |
| ESCORE | 578.33 | 1631.69 | 640.69 | 2855.57 |
| 数据集/模型 | 最后周期 | ERUL≤15 | |||
|---|---|---|---|---|---|
| ERMSE | ESCORE | ERMSE | ESCORE | ||
| FD001 | VDGP-R-M | 17.61 | 599.46 | 3.58 | 3.18 |
| VDGP-M-R | 17.66 | 624.40 | 3.19 | 2.80 | |
| VDGP-R-R | 17.06 | 553.55 | 3.40 | 2.96 | |
| VDGP-M-M | 16.80 | 578.33 | 3.84 | 3.49 | |
| FD002 | VDGP-R-M | 21.82 | 3221.40 | 20.89 | 990.35 |
| VDGP-M-R | 21.62 | 3053.05 | 20.96 | 976.32 | |
| VDGP-R-R | 19.50 | 1720.84 | 14.83 | 196.02 | |
| VDGP-M-M | 19.12 | 1631.69 | 13.72 | 156.61 | |
| FD003 | VDGP-R-M | 18.25 | 1407.07 | 7.04 | 8.88 |
| VDGP-M-R | 18.42 | 1628.08 | 6.88 | 8.70 | |
| VDGP-R-R | 18.66 | 2826.08 | 3.40 | 2.96 | |
| VDGP-M-M | 17.68 | 640.69 | 3.66 | 3.53 | |
| FD004 | VDGP-R-M | 25.43 | 5188.19 | 30.63 | 2467.67 |
| VDGP-M-R | 25.38 | 5766.34 | 31.47 | 3170.79 | |
| VDGP-R-R | 21.92 | 3082.37 | 19.67 | 642.89 | |
| VDGP-M-M | 22.52 | 2855.57 | 19.65 | 415.02 | |
表7 四种VDGP模型在C-MAPSS测试集上的性能
Tab.7 Performance of four VDGP models on the C-MAPSS test set
| 数据集/模型 | 最后周期 | ERUL≤15 | |||
|---|---|---|---|---|---|
| ERMSE | ESCORE | ERMSE | ESCORE | ||
| FD001 | VDGP-R-M | 17.61 | 599.46 | 3.58 | 3.18 |
| VDGP-M-R | 17.66 | 624.40 | 3.19 | 2.80 | |
| VDGP-R-R | 17.06 | 553.55 | 3.40 | 2.96 | |
| VDGP-M-M | 16.80 | 578.33 | 3.84 | 3.49 | |
| FD002 | VDGP-R-M | 21.82 | 3221.40 | 20.89 | 990.35 |
| VDGP-M-R | 21.62 | 3053.05 | 20.96 | 976.32 | |
| VDGP-R-R | 19.50 | 1720.84 | 14.83 | 196.02 | |
| VDGP-M-M | 19.12 | 1631.69 | 13.72 | 156.61 | |
| FD003 | VDGP-R-M | 18.25 | 1407.07 | 7.04 | 8.88 |
| VDGP-M-R | 18.42 | 1628.08 | 6.88 | 8.70 | |
| VDGP-R-R | 18.66 | 2826.08 | 3.40 | 2.96 | |
| VDGP-M-M | 17.68 | 640.69 | 3.66 | 3.53 | |
| FD004 | VDGP-R-M | 25.43 | 5188.19 | 30.63 | 2467.67 |
| VDGP-M-R | 25.38 | 5766.34 | 31.47 | 3170.79 | |
| VDGP-R-R | 21.92 | 3082.37 | 19.67 | 642.89 | |
| VDGP-M-M | 22.52 | 2855.57 | 19.65 | 415.02 | |
| C-MAPSS航空发动机数据集 | 模型 | 最后周期 | ERUL≤15 | ||||
|---|---|---|---|---|---|---|---|
| ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE | ||
| FD001 | GP(RBF) | 17.86 | 586.03 | 44.21 | 5.33 | 5.62 | 4.11 |
| GP(Matern 5/2) | 18.05 | 607.59 | 44.12 | 5.13 | 5.17 | 3.67 | |
| VDGP-R-R | 17.06 | 553.55 | 11.87 | 3.40 | 2.96 | 2.76 | |
| VDGP-M-M | 16.80 | 578.33 | 11.74 | 3.84 | 3.49 | 3.22 | |
| FD002(100) | GP(RBF) | 40.63 | 25688.53 | 46.52 | 28.63 | 1578.14 | 23.06 |
| GP(Matern 5/2) | 36.16 | 10221.29 | 44.08 | 30.23 | 1807.18 | 25.04 | |
| VDGP-R-R | 20.90 | 2629.83 | 17.37 | 21.33 | 552.60 | 17.44 | |
| VDGP-M-M | 20.25 | 2534.57 | 16.67 | 20.78 | 438.53 | 17.57 | |
| VDGP-R-R(full data) | 19.50 | 1720.84 | 15.75 | 14.83 | 196.02 | 11.73 | |
| VDGP-M-M(full data) | 19.12 | 1631.69 | 15.22 | 13.72 | 156.61 | 10.66 | |
| FD003 | GP(RBF) | 19.45 | 2949.21 | 43.95 | 9.30 | 16.49 | 5.66 |
| GP(Matern 5/2) | 19.83 | 3488.08 | 44.00 | 8.52 | 13.66 | 5.38 | |
| VDGP-R-R | 18.66 | 2826.08 | 13.38 | 5.20 | 5.42 | 3.47 | |
| VDGP-M-M | 17.68 | 640.69 | 12.36 | 3.66 | 3.53 | 3.08 | |
| FD004(100) | GP(RBF) | 36.82 | 10555.82 | 43.40 | 36.16 | 2161.26 | 32.41 |
| GP(Matern 5/2) | 37.12 | 10790.38 | 43.38 | 35.63 | 2047.18 | 31.65 | |
| VDGP-R-R | 23.85 | 4738.03 | 18.68 | 20.60 | 454.33 | 14.74 | |
| VDGP-M-M | 24.25 | 4615.70 | 19.25 | 27.03 | 1187.90 | 20.91 | |
| VDGP-R-R(full data) | 21.92 | 3082.37 | 17.18 | 19.67 | 642.89 | 13.68 | |
| VDGP-M-M(full data) | 22.52 | 2855.57 | 17.72 | 19.65 | 415.02 | 14.90 | |
表8 GP和VDGP的实验结果
Tab.8 Experimental results of GP and VDGP
| C-MAPSS航空发动机数据集 | 模型 | 最后周期 | ERUL≤15 | ||||
|---|---|---|---|---|---|---|---|
| ERMSE | ESCORE | EMAE | ERMSE | ESCORE | EMAE | ||
| FD001 | GP(RBF) | 17.86 | 586.03 | 44.21 | 5.33 | 5.62 | 4.11 |
| GP(Matern 5/2) | 18.05 | 607.59 | 44.12 | 5.13 | 5.17 | 3.67 | |
| VDGP-R-R | 17.06 | 553.55 | 11.87 | 3.40 | 2.96 | 2.76 | |
| VDGP-M-M | 16.80 | 578.33 | 11.74 | 3.84 | 3.49 | 3.22 | |
| FD002(100) | GP(RBF) | 40.63 | 25688.53 | 46.52 | 28.63 | 1578.14 | 23.06 |
| GP(Matern 5/2) | 36.16 | 10221.29 | 44.08 | 30.23 | 1807.18 | 25.04 | |
| VDGP-R-R | 20.90 | 2629.83 | 17.37 | 21.33 | 552.60 | 17.44 | |
| VDGP-M-M | 20.25 | 2534.57 | 16.67 | 20.78 | 438.53 | 17.57 | |
| VDGP-R-R(full data) | 19.50 | 1720.84 | 15.75 | 14.83 | 196.02 | 11.73 | |
| VDGP-M-M(full data) | 19.12 | 1631.69 | 15.22 | 13.72 | 156.61 | 10.66 | |
| FD003 | GP(RBF) | 19.45 | 2949.21 | 43.95 | 9.30 | 16.49 | 5.66 |
| GP(Matern 5/2) | 19.83 | 3488.08 | 44.00 | 8.52 | 13.66 | 5.38 | |
| VDGP-R-R | 18.66 | 2826.08 | 13.38 | 5.20 | 5.42 | 3.47 | |
| VDGP-M-M | 17.68 | 640.69 | 12.36 | 3.66 | 3.53 | 3.08 | |
| FD004(100) | GP(RBF) | 36.82 | 10555.82 | 43.40 | 36.16 | 2161.26 | 32.41 |
| GP(Matern 5/2) | 37.12 | 10790.38 | 43.38 | 35.63 | 2047.18 | 31.65 | |
| VDGP-R-R | 23.85 | 4738.03 | 18.68 | 20.60 | 454.33 | 14.74 | |
| VDGP-M-M | 24.25 | 4615.70 | 19.25 | 27.03 | 1187.90 | 20.91 | |
| VDGP-R-R(full data) | 21.92 | 3082.37 | 17.18 | 19.67 | 642.89 | 13.68 | |
| VDGP-M-M(full data) | 22.52 | 2855.57 | 17.72 | 19.65 | 415.02 | 14.90 | |
| 模型 | 最后周期 | ERUL≤15 | ||
|---|---|---|---|---|
| ERMSE | ESCORE | ERMSE | ESCORE | |
| LSTM[ | 5.64 | 181.25 | 3.01 | 4.48 |
| MLP[42] | 10.86 | 253.40 | 15.31 | 12.01 |
| CNN[ | 4.37 | 179.94 | 3.48 | 6.38 |
| LSTM with attention[ | 12.60 | 1117.89 | 22.55 | 141.39 |
| Bi-level LSTM Scheme[ | 10.99 | 695.51 | 1.94 | 2.51 |
| CBLSTM[ | 9.67 | 510.71 | 2.77 | 4.24 |
| LSTM-Z[ | 9.58 | 417.48 | 12.96 | 42.66 |
| VDGP-R-M | 30.92 | 101 378.96 | 13.09 | 53.64 |
| VDGP-M-R | 24.76 | 105 515.92 | 8.57 | 18.30 |
| VDGP-R-R | 4.70 | 166.71 | 1.85 | 2.23 |
| VDGP-M-M | 3.88 | 155.33 | 5.65 | 12.09 |
表9 VDGP与其他方法的实验结果
Tab.9 Experimental results of VDGP and other methods
| 模型 | 最后周期 | ERUL≤15 | ||
|---|---|---|---|---|
| ERMSE | ESCORE | ERMSE | ESCORE | |
| LSTM[ | 5.64 | 181.25 | 3.01 | 4.48 |
| MLP[42] | 10.86 | 253.40 | 15.31 | 12.01 |
| CNN[ | 4.37 | 179.94 | 3.48 | 6.38 |
| LSTM with attention[ | 12.60 | 1117.89 | 22.55 | 141.39 |
| Bi-level LSTM Scheme[ | 10.99 | 695.51 | 1.94 | 2.51 |
| CBLSTM[ | 9.67 | 510.71 | 2.77 | 4.24 |
| LSTM-Z[ | 9.58 | 417.48 | 12.96 | 42.66 |
| VDGP-R-M | 30.92 | 101 378.96 | 13.09 | 53.64 |
| VDGP-M-R | 24.76 | 105 515.92 | 8.57 | 18.30 |
| VDGP-R-R | 4.70 | 166.71 | 1.85 | 2.23 |
| VDGP-M-M | 3.88 | 155.33 | 5.65 | 12.09 |
| 指标 | GP(RBF) | GP(Matern 5/2) | VDGP-R-R | VDGP-M-M |
|---|---|---|---|---|
| ERMSE | 13.48 | 10.22 | 4.70 | 3.88 |
| ESCORE | 1376.15 | 659.54 | 166.71 | 155.33 |
| EMAE | 123.32 | 124.29 | 2.87 | 3.39 |
| ERMSE(ERUL≤15) | 10.43 | 8.58 | 1.85 | 5.65 |
| ESCORE(ERUL≤15) | 35.37 | 22.58 | 2.23 | 12.09 |
| EMAE(ERUL≤15) | 7.64 | 5.97 | 1.68 | 5.60 |
表10 GP和VDGP的实验对比结果
Tab.10 Experimental comparison results between GP and VDGP
| 指标 | GP(RBF) | GP(Matern 5/2) | VDGP-R-R | VDGP-M-M |
|---|---|---|---|---|
| ERMSE | 13.48 | 10.22 | 4.70 | 3.88 |
| ESCORE | 1376.15 | 659.54 | 166.71 | 155.33 |
| EMAE | 123.32 | 124.29 | 2.87 | 3.39 |
| ERMSE(ERUL≤15) | 10.43 | 8.58 | 1.85 | 5.65 |
| ESCORE(ERUL≤15) | 35.37 | 22.58 | 2.23 | 12.09 |
| EMAE(ERUL≤15) | 7.64 | 5.97 | 1.68 | 5.60 |
| [1] | GAO Pengjie, WANG Junliang, SHI Ziqi, et al. Long-term Temporal Attention Neural Network with Adaptive Stage Division for Remaining Useful Life Prediction of Rolling Bearings[J]. Reliability Engineering & System Safety, 2024, 251: 110218. |
| [2] | LIN Ruiguan, YU Yaowei, WANG Huawei, et al. Remaining Useful Life Prediction in Prognostics Using Multi-scale Sequence and Long Short-term Memory Network[J]. Journal of Computational Science, 2022, 57: 101508. |
| [3] | CHEN Chuang, LU Ningyun, JIANG Bin, et al. Prediction Interval Estimation of Aeroengine Remaining Useful Life Based on Bidirectional Long Short-term Memory Network[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 3527213. |
| [4] | 刘小峰, 亢莹莹, 柏林. 轴承自驱式独立退化轨迹构建与剩余寿命灰色预测[J]. 中国机械工程, 2024, 35(9): 1613-1621. |
| LIU Xiaofeng, KANG Yingying, BO Lin. Self-driven Independent Degradation Trajectory Construction and Remaining Life Gray Prediction for Bearings[J]. China Mechanical Engineering, 2024, 35(9): 1613-1621. | |
| [5] | CHEN Zuoyi, HUANG Hong-Zhong, DENG Zhongwei, et al. Shrinkage Mamba Relation Network with Out-of-distribution Data Augmentation for Rotating Machinery Fault Detection and Localization under Zero-faulty Data[J]. Mechanical Systems and Signal Processing, 2025, 224: 112145. |
| [6] | ZHU Ting, CHEN Zhen, ZHOU Di, et al. Adaptive Staged Remaining Useful Life Prediction of Roller in a Hot Strip Mill Based on Multi-scale LSTM with Multi-head Attention[J]. Reliability Engineering & System Safety, 2024, 248: 110161. |
| [7] | CAO Wei, MENG Zong, LI Jimeng, et al. A Remaining Useful Life Prediction Method for Rolling Bearing Based on TCN-transformer[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 74: 3501309. |
| [8] | PENG Peng, LI Yonghua, GUO Zhongyi. High-performance Remaining Useful Life Prediction for Aeroengine Based on Combining Health States and Trajectory Similarity[J]. Engineering Applications of Artificial Intelligence, 2025, 141: 109799. |
| [9] | CHEN Zesheng, TU Xiaotong, HU Yue, et al. Real-time Bearing Remaining Useful Life Estimation Based on the Frozen Convolutional and Activated Memory Neural Network[J]. IEEE Access, 2019, 7: 96583-96593. |
| [10] | 夏然, 苏春. 基于健康因子和混合Bi-LSTM-NAR模型的锂离子电池剩余寿命预测[J]. 中国机械工程, 2024, 35(5): 851-859. |
| XIA Ran, SU Chun. Remaining Useful Life Prediction for Lithium-ion Batteries Based on Health Indicators and Hybrid Bi-LSTM-NAR Model[J]. China Mechanical Engineering, 2024, 35(5): 851-859. | |
| [11] | WANG Zhe, YANG Lechang, FANG Xiaolei, et al. Image-based Remaining Useful Life Prediction through Adaptation from Simulation to Experimental Domain[J]. Reliability Engineering & System Safety, 2025, 255: 110668. |
| [12] | LUO Peien, YIN Zhonggang, ZHANG Yanqing, et al. A Novel Whale Optimization Algorithm Based on Music Theory Knowledge for RUL Prediction of Motor Bearing[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 3534211. |
| [13] | 曹现刚, 叶煜, 赵友军, 等. 基于KPCA-LSTM的旋转机械剩余使用寿命预测[J]. 振动与冲击, 2023, 42(24): 81-91. |
| CAO Xiangang, YE Yu, ZHAO Youjun, et al. Remaining Useful Life Prediction of Rotating Machinery Based on KPCA-LSTM[J]. Journal of Vibration and Shock, 2023, 42(24): 81-91. | |
| [14] | LI Biao, TANG Baoping, DENG Lei, et al. Self-attention ConvLSTM and Its Application in RUL Prediction of Rolling Bearings[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 3518811. |
| [15] | 刘华开, 丁康, 何国林, 等. 联合故障机理和卷积神经网络的齿轮剩余使用寿命预测方法研究[J]. 机械工程学报, 2024, 60(12): 116-125. |
| LIU Huakai, DING Kang, HE Guolin, et al. Research on the Prediction Method of Remaining Useful Life of Gears by Combining Fault Mechanism and Convolutional Neural Network[J]. Journal of Mechanical Engineering, 2024, 60(12): 116-125. | |
| [16] | LIU Xiaofei, LEI Yaguo, LI Naipeng, et al. RUL Prediction of Machinery Using Convolutional-vector Fusion Network through Multi-feature Dynamic Weighting[J]. Mechanical Systems and Signal Processing, 2023, 185: 109788. |
| [17] | WANG Biao, LEI Yaguo, LI Naipeng, et al. Multiscale Convolutional Attention Network for Predicting Remaining Useful Life of Machinery[J]. IEEE Transactions on Industrial Electronics, 2020, 68(8): 7496-7504. |
| [18] | DUVENAUD D. Automatic model construction with Gaussian processes[D]. Cambridge, UK: University of Cambridge, 2014. |
| [19] | LI Yanhui, YANG Pengfei, BAI Lu, et al. BRDF Modeling and Optimization of a Target Surface Based on the Gradient Descent Algorithm[J]. Applied Optics, 2023, 62(36): 9486-9492. |
| [20] | HEIMES F O. Recurrent Neural Networks for Remaining Useful Life Estimation[C]∥2008 International Conference on Prognostics and Health Management. Denver: IEEE, 2008: 1-6.. |
| [21] | SAXENA A, GOEBEL K, SIMON D, et al. Damage Propagation Modeling for Aircraft Engine Run-to-failure Simulation[C]∥2008 International Conference on Prognostics and Health Management. Denver: IEEE, 2008: 1-9. |
| [22] | LI Han, ZHAO Wei, ZHANG Yuxi, et al. Remaining Useful Life Prediction Using Multi-scale Deep Convolutional Neural Network[J]. Applied Soft Computing, 2020, 89: 106113. |
| [23] | LIU Hui, LIU Zhenyu, JIA Weiqiang, et al. A Novel Deep Learning-based Encoder-decoder Model for Remaining Useful Life Prediction[C]∥2019 International Joint Conference on Neural Networks (IJCNN). Budapest: IEEE, 2019: 1-8. |
| [24] | THANG B, HERNANDEZ-LOBATO D, HERNANDEZ-LOBATO J, et al. Deep Gaussian Processes for Regression using Approximate Expectation Propagation [C]∥Proceedings of the 33rd International Conference on Machine Learning (ICML). New York, 2016: 1472-1481. |
| [25] | DAMIANOU A. Deep Gaussian Processes and Variational Propagation of Uncertainty[D]. Sheffield: University of Sheffield, 2015. |
| [26] | XIE Yucen, ZOU Jianxiao, PENG Chao, et al. A Novel PEM Fuel Cell Remaining Useful Life Prediction Method Based on Singular Spectrum Analysis and Deep Gaussian Processes[J]. International Journal of Hydrogen Energy, 2020, 45(55): 30942-30956. |
| [27] | SATEESH BABU G, ZHAO Peilin, LI Xiaoli. Deep Convolutional Neural Network Based Regression Approach for Estimation of Remaining Useful Life[C]∥Database Systems for Advanced Applications. Cham: Springer, 2016: 214-228. |
| [28] | ZHENG Shuai, RISTOVSKI K, FARAHAT A, et al. Long Short-term Memory Network for Remaining Useful Life Estimation[C]∥2017 IEEE International Conference on Prognostics and Health Management (ICPHM). Dallas: IEEE, 2017: 88-95. |
| [29] | CHEN Zhenghua, WU Min, ZHAO Rui, et al. Machine Remaining Useful Life Prediction via an Attention-based Deep Learning Approach[J]. IEEE Transactions on Industrial Electronics, 2020, 68(3): 2521-2531. |
| [30] | BENKER M, FURTNER L, SEMM T, et al. Utilizing Uncertainty Information in Remaining Useful Life Estimation via Bayesian Neural Networks and Hamiltonian Monte Carlo[J]. Journal of Manufacturing Systems, 2021, 61: 799-807. |
| [31] | GAO Hui, LI Yibin, ZHAO Ying, et al. Dual Channel Feature Attention-based Approach for RUL Prediction Considering the Spatiotemporal Difference of Multisensor Data[J]. IEEE Sensors Journal, 2023, 23(8): 8514-8525. |
| [32] | ElDALI M, KUMAR K D. Fault Diagnosis and Prognosis of Aerospace Systems Using Growing Recurrent Neural Networks and LSTM[C]∥2021 IEEE Aerospace Conference (50100. SkyBig: IEEE, 2021: 1-20. |
| [33] | DUAN Yuhang, LI Honghui, HE Mengqi, et al. A BiGRU Autoencoder Remaining Useful Life Prediction Scheme with Attention Mechanism and Skip Connection[J]. IEEE Sensors Journal, 2021, 21(9): 10905-10914. |
| [34] | XIA Tangbin, SHU Junqing, XU Yuhui, et al. Multiscale Similarity Ensemble Framework for Remaining Useful Life Prediction[J]. Measurement, 2022, 188: 110565. |
| [35] | SONG Tao, LIU Chao, WU Rui, et al. A Hierarchical Scheme for Remaining Useful Life Prediction with Long Short-term Memory Networks[J]. Neurocomputing, 2022, 487: 22-33. |
| [36] | LIN Lin, WU Jinlei, FU Song, et al. Channel Attention & Temporal Attention Based Temporal Convolutional Network: a Dual Attention Framework for Remaining Useful Life Prediction of the Aircraft Engines[J]. Advanced Engineering Informatics, 2024, 60: 102372. |
| [37] | LI Yasong, ZHOU Zheng, SUN Chuang, et al. Life-cycle Modeling Driven by Coupling Competition Degradation for Remaining Useful Life Prediction[J]. Reliability Engineering & System Safety, 2023, 238: 109480. |
| [38] | 闫啸家, 梁伟阁, 张钢, 等. 基于RCNN-ABiLSTM的机械设备剩余寿命预测方法[J]. 系统工程与电子技术, 2023, 45(3): 931-940. |
| YAN Xiaojia, LIANG Weige, ZHANG Gang, et al. Prediction Method for Mechanical Equipment Based on RCNN-ABiLSTM[J]. Systems Engineering and Electronics, 2023, 45(3): 931-940. | |
| [39] | WANG Huan, LIU Zhiliang, PENG Dandan, et al. Understanding and Learning Discriminant Features Based on Multiattention 1DCNN for Wheelset Bearing Fault Diagnosis[J]. IEEE Transactions on Industrial Informatics, 2020, 16(9): 5735-5745. |
| [40] | YOU Dazhang, CHEN Linbo, LIU Fei, et al. Intelligent Fault Diagnosis of Bearing Based on Convolutional Neural Network and Bidirectional Long Short-term Memory[J]. Shock and Vibration, 2021, 2021(1): 7346352. |
| [41] | 庄雨璇, 李奇, 杨冰如, 等. 基于LSTM的轴承故障诊断端到端方法[J]. 噪声与振动控制, 2019, 39(6): 187-193. |
| ZHUANG Yuxuan, LI Qi, YANG Bingru, et al. An End-to-end Approach for Bearing Fault Diagnosis Based on LSTM[J]. Noise and Vibration Control, 2019, 39(6): 187-193. |
| [1] | 李鑫, 程军圣, 吴小伟, 王健, 杨宇, . SPCA和OCHD相结合的旋转机械早期微弱故障检测方法[J]. 中国机械工程, 2023, 34(15): 1805-1812. |
| [2] | 张龙, 胡燕青, 赵丽娟, 张号. 多通道信息融合与深度迁移学习的旋转机械故障诊断[J]. 中国机械工程, 2023, 34(08): 966-975. |
| [3] | 姜万录, 李满, 张培尧, 赵亚鹏, 张淑清. 基于全矢增强深度森林的旋转设备智能故障诊断方法[J]. 中国机械工程, 2022, 33(11): 1324-1335. |
| [4] | 蒋佳炜, 胡以怀, 方云虎, 张陈, 芮晓松, 汪猛. 基于多尺度时域平均分解和模糊熵的船用风机故障诊断方法[J]. 中国机械工程, 2022, 33(10): 1178-1188. |
| [5] | 石明宽, 赵荣珍. 基于局部平衡判别投影的旋转机械故障诊断方法[J]. 中国机械工程, 2021, 32(14): 1653-1658,1668. |
| [6] | 严建文, 钟小虎, 范煜, 郭三敏, . 基于整群抽样和支持向量回归模型的高功率半导体激光器剩余使用寿命预测[J]. 中国机械工程, 2021, 32(13): 1523-1529. |
| [7] | 李锋1;程阳洋1,2;陈勇1;汤宝平3. 用于旋转机械状态趋势预测的量子注意力循环编码解码神经网络[J]. 中国机械工程, 2020, 31(21): 2573-2582. |
| [8] | 曾大懿1;杨基宏2;邹益胜1;张继冬1;宋小欣1. 基于并行多通道卷积长短时记忆网络的轴承寿命预测方法[J]. 中国机械工程, 2020, 31(20): 2454-2462,2471. |
| [9] | 孔帅;崔洪宇;季顺迎. 船体结构冰载荷反演方法及试验验证[J]. 中国机械工程, 2020, 31(03): 281-288. |
| [10] | 杨宇;李鑫;潘海洋;程军圣. 基于非线性模式分解的旋转机械复合故障特征提取方法[J]. 中国机械工程, 2018, 29(24): 2947-2953. |
| [11] | 王凌, 陈长骏, 潘静, 许宏, 陈锡爱, 那文波. 地铁车辆电动塞拉门的剩余寿命预测[J]. 中国机械工程, 2016, 27(22): 2994-3000. |
| [12] | 陈换过, 陈培, 陈文华, 蔡丽, 沈建洋, 吴建伟, 吴明建, . 采样频率对聚合经验模态分解的影响研究[J]. 中国机械工程, 2016, 27(18): 2472-2476. |
| [13] | 陈宏, 雷文平, 陈磊, 韩捷. 一种转子动态不对中量计算方法[J]. 中国机械工程, 2016, 27(17): 2379-2383. |
| [14] | 时培明, 苏翠娇, 赵娜, 韩东颖. 基于解析模态分解的机械故障诊断方法[J]. 中国机械工程, 2016, 27(05): 674-679. |
| [15] | 曹劲然, 冯毅, 陆宝春, 张登峰, 吴建, 石胜征, 关德壮. 基于组合核函数OSVR算法的起重机减速齿轮箱磨损趋势预测[J]. 中国机械工程, 2015, 26(5): 641-646. |
| 阅读次数 | ||||||
|
全文 |
|
|||||
|
摘要 |
|
|||||