《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (1): 314-321.DOI: 10.11772/j.issn.1001-9081.2025010122
收稿日期:2025-02-07
修回日期:2025-03-27
接受日期:2025-03-28
发布日期:2026-01-10
出版日期:2026-01-10
通讯作者:
李廷顺
作者简介:王婷婷(2000—),女,河北邯郸人,硕士研究生,主要研究方向:电缆温度数据挖掘、深度学习基金资助:
Tingting WANG1, Tingshun LI1(
), Wen TAN2, Bo LYU1, Yixuan CHEN1
Received:2025-02-07
Revised:2025-03-27
Accepted:2025-03-28
Online:2026-01-10
Published:2026-01-10
Contact:
Tingshun LI
About author:WANG Tingting, born in 2000, M. S. candidate. Her research interests include cable temperature data mining, deep learning.Supported by:摘要:
高压电缆长期过热可能导致绝缘热击穿,进而影响电网的稳定性。然而,当前研究主要集中在传统预测模型上,忽略了温度数据的复杂性和动态特征。为了解决此问题,提出一种基于多尺度Patch与卷积交互的电缆温度预测模型(MSP-CI)。首先,采用通道重组采样方法降低输入维度,并构建多尺度Patch分支结构,以实现复杂时间序列的解耦;其次,结合序列分解与卷积交互策略,分别提取粗粒度Patch的宏观信息与细粒度Patch的微观信息;最后,构建注意力融合模块,以动态平衡宏观与微观信息的权重,并得到最终的预测结果。在真实高压电缆温度数据集上的实验结果表明, MSP-CI相较于TimeMixer、PatchTST (Patch Time Series Transformer)和MSGNet(Multi-Scale inter-series Graph Network)等基线模型,在均方误差(MSE)上下降了7.02%~34.87%,在平均绝对误差(MAE)上下降了5.15%~32.04%。可见, MSP-CI能有效提升电缆温度预测的准确率,为电力调度运行提供依据。
中图分类号:
王婷婷, 李廷顺, 谭文, 吕博, 陈翼轩. 基于多尺度Patch与卷积交互的电缆温度预测模型[J]. 计算机应用, 2026, 46(1): 314-321.
Tingting WANG, Tingshun LI, Wen TAN, Bo LYU, Yixuan CHEN. Cable temperature prediction model based on multi-scale patch and convolution interaction[J]. Journal of Computer Applications, 2026, 46(1): 314-321.
| 参数 | HVCT1 | HVCT2 |
|---|---|---|
| batch_size | 32 | 32 |
| lr | 0.001 | 0.001 |
| r_steps | 103 | 286 |
| epoch | 10 | 10 |
| e_layers | 2 | 2 |
| M | 4 | 2 |
表1 实验超参数设置
Tab. 1 Experimental hyperparameter setting
| 参数 | HVCT1 | HVCT2 |
|---|---|---|
| batch_size | 32 | 32 |
| lr | 0.001 | 0.001 |
| r_steps | 103 | 286 |
| epoch | 10 | 10 |
| e_layers | 2 | 2 |
| M | 4 | 2 |
| 数据集 | 预测长度 | 本文模型 | PatchTST | TimeMixer | MSGNet | xPatch | iTransformer | DLinear | Autoformer | MICN | STHD | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | ||
| HVCT1 | 48 | 0.220 | 0.339 | 0.259 | 0.354 | 0.249 | 0.352 | 0.243 | 0.353 | 0.283 | 0.364 | 0.282 | 0.371 | 0.337 | 0.395 | 0.284 | 0.364 | 0.299 | 0.394 | ||
| 96 | 0.274 | 0.370 | 0.301 | 0.389 | 0.281 | 0.380 | 0.299 | 0.385 | 0.308 | 0.381 | 0.289 | 0.376 | 0.362 | 0.412 | 0.302 | 0.385 | 0.321 | 0.400 | |||
| 192 | 0.293 | 0.381 | 0.349 | 0.408 | 0.400 | 0.339 | 0.405 | 0.345 | 0.419 | 0.330 | 0.335 | 0.400 | 0.384 | 0.432 | 0.334 | 0.405 | 0.335 | 0.435 | |||
| 平均 | 0.262 | 0.363 | 0.297 | 0.382 | 0.290 | 0.379 | 0.296 | 0.386 | 0.307 | 0.380 | 0.302 | 0.382 | 0.361 | 0.413 | 0.307 | 0.385 | 0.318 | 0.410 | |||
| HVCT2 | 48 | 0.485 | 0.632 | 0.518 | 0.704 | 0.515 | 0.712 | 0.512 | 0.699 | 0.546 | 0.774 | 0.509 | 0.698 | 0.637 | 0.994 | 0.620 | 0.688 | 0.807 | 0.914 | ||
| 96 | 0.512 | 0.673 | 0.544 | 0.737 | 0.721 | 0.545 | 0.739 | 0.547 | 0.740 | 0.575 | 0.896 | 0.553 | 0.753 | 0.677 | 1.385 | 0.644 | 0.861 | 0.972 | |||
| 192 | 0.536 | 0.696 | 0.575 | 0.819 | 0.57 | 0.828 | 0.573 | 0.750 | 0.584 | 0.998 | 0.583 | 0.803 | 0.806 | 1.787 | 0.672 | 0.748 | 0.990 | 1.229 | |||
| 平均 | 0.511 | 0.667 | 0.543 | 0.748 | 0.543 | 0.760 | 0.544 | 0.730 | 0.568 | 0.889 | 0.548 | 0.751 | 0.707 | 1.389 | 0.645 | 0.886 | 1.038 | ||||
表2 实验结果对比
Tab. 2 Comparison of experimental results
| 数据集 | 预测长度 | 本文模型 | PatchTST | TimeMixer | MSGNet | xPatch | iTransformer | DLinear | Autoformer | MICN | STHD | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | MSE | MAE | ||
| HVCT1 | 48 | 0.220 | 0.339 | 0.259 | 0.354 | 0.249 | 0.352 | 0.243 | 0.353 | 0.283 | 0.364 | 0.282 | 0.371 | 0.337 | 0.395 | 0.284 | 0.364 | 0.299 | 0.394 | ||
| 96 | 0.274 | 0.370 | 0.301 | 0.389 | 0.281 | 0.380 | 0.299 | 0.385 | 0.308 | 0.381 | 0.289 | 0.376 | 0.362 | 0.412 | 0.302 | 0.385 | 0.321 | 0.400 | |||
| 192 | 0.293 | 0.381 | 0.349 | 0.408 | 0.400 | 0.339 | 0.405 | 0.345 | 0.419 | 0.330 | 0.335 | 0.400 | 0.384 | 0.432 | 0.334 | 0.405 | 0.335 | 0.435 | |||
| 平均 | 0.262 | 0.363 | 0.297 | 0.382 | 0.290 | 0.379 | 0.296 | 0.386 | 0.307 | 0.380 | 0.302 | 0.382 | 0.361 | 0.413 | 0.307 | 0.385 | 0.318 | 0.410 | |||
| HVCT2 | 48 | 0.485 | 0.632 | 0.518 | 0.704 | 0.515 | 0.712 | 0.512 | 0.699 | 0.546 | 0.774 | 0.509 | 0.698 | 0.637 | 0.994 | 0.620 | 0.688 | 0.807 | 0.914 | ||
| 96 | 0.512 | 0.673 | 0.544 | 0.737 | 0.721 | 0.545 | 0.739 | 0.547 | 0.740 | 0.575 | 0.896 | 0.553 | 0.753 | 0.677 | 1.385 | 0.644 | 0.861 | 0.972 | |||
| 192 | 0.536 | 0.696 | 0.575 | 0.819 | 0.57 | 0.828 | 0.573 | 0.750 | 0.584 | 0.998 | 0.583 | 0.803 | 0.806 | 1.787 | 0.672 | 0.748 | 0.990 | 1.229 | |||
| 平均 | 0.511 | 0.667 | 0.543 | 0.748 | 0.543 | 0.760 | 0.544 | 0.730 | 0.568 | 0.889 | 0.548 | 0.751 | 0.707 | 1.389 | 0.645 | 0.886 | 1.038 | ||||
| 模型 | MSE | MAE |
|---|---|---|
| w/o MSP | 0.255 | 0.354 |
| w/o AF | 0.261 | 0.352 |
| Model-1 | 0.295 | 0.374 |
| Model-2 | 0.304 | 0.401 |
| Model-3 | 0.241 | 0.350 |
| 本文模型 | 0.237 | 0.342 |
表3 消融实验结果
Tab. 3 Ablation study results
| 模型 | MSE | MAE |
|---|---|---|
| w/o MSP | 0.255 | 0.354 |
| w/o AF | 0.261 | 0.352 |
| Model-1 | 0.295 | 0.374 |
| Model-2 | 0.304 | 0.401 |
| Model-3 | 0.241 | 0.350 |
| 本文模型 | 0.237 | 0.342 |
| M | MSE | MAE | |
|---|---|---|---|
| 1 | {16} | 0.267 | 0.363 |
| {32} | 0.288 | 0.373 | |
| 2 | {16,8} | 0.263 | 0.356 |
| {32,16} | 0.277 | 0.367 | |
| 3 | {16,8,4} | 0.257 | 0.357 |
| {32,16,8} | 0.255 | 0.354 | |
| 4 | {32,16,8,4} | 0.251 | 0.355 |
| {64,32,16,8} | 0.237 | 0.343 | |
| 5 | {64,32,16,8,4} | 0.264 | 0.360 |
| {128,64,32,16,8} | 0.245 | 0.355 |
表4 多尺度Patch的组合结果
Tab. 4 Multi-scale patch combination results
| M | MSE | MAE | |
|---|---|---|---|
| 1 | {16} | 0.267 | 0.363 |
| {32} | 0.288 | 0.373 | |
| 2 | {16,8} | 0.263 | 0.356 |
| {32,16} | 0.277 | 0.367 | |
| 3 | {16,8,4} | 0.257 | 0.357 |
| {32,16,8} | 0.255 | 0.354 | |
| 4 | {32,16,8,4} | 0.251 | 0.355 |
| {64,32,16,8} | 0.237 | 0.343 | |
| 5 | {64,32,16,8,4} | 0.264 | 0.360 |
| {128,64,32,16,8} | 0.245 | 0.355 |
| [1] | 吴春芳,陈云,黄韬,等.电力电缆状态感知与检测技术研究综述[J].高压电器, 2024, 60(10): 86-103. |
| WU C F, CHEN Y, HUANG T, et al. Overview of condition sensing and detection technology for power cables [J]. High Voltage Apparatus, 2024, 60(10): 86-103. | |
| [2] | 孙俊峰,李志斌.基于LSTM的滚动预测算法的电缆缆芯温度的研究[J].电子测量技术, 2021, 44(21): 84-88. |
| SUN J F, LI Z B. Research on cable core temperature based on rolling prediction algorithm of LSTM [J]. Electronic Measurement Technology, 2021, 44(21): 84-88. | |
| [3] | 田洪亮,刘洋,韩文花,等.基于支持向量机的配电设备温度监测数据预测[J].电网与清洁能源, 2018, 34(1): 65-71. |
| TIAN H L, LIU Y, HAN W H, et al. Research on the prediction of temperature monitoring data of distribution equipment based on SVM [J]. Power System and Clean Energy, 2018, 34(1): 65-71. | |
| [4] | ZHOU H, WANG J, LIU K, et al. Temperature prediction study of cable joint conductor based on the PSO algorithms of BP neural network [C]// Proceedings of the 2015 IEEE International Magnetics Conference. Piscataway: IEEE, 2015: 1-1. |
| [5] | LI Z, LI Z, LI Z, et al. Application of GA-LSTM model in cable joint temperature prediction [C]// Proceedings of the 7th International Forum on Electrical Engineering and Automation. Piscataway: IEEE, 2020: 71-75. |
| [6] | XU Z, ZHANG Y, XUE F, et al. Short-term temperature forecasting of cable joint based on temporal convolutional neural network [J]. IEEE Access, 2024, 12: 132543-132551. |
| [7] | SHABANI M A, ABDI A, MENG L, et al. Scaleformer: iterative multi-scale refining Transformers for time series forecasting [EB/OL]. [2024-11-05]. . |
| [8] | WANG H, PENG J, HUANG F, et al. MICN: multi-scale local and global context modeling for long-term series forecasting [EB/OL]. [2024-10-05]. . |
| [9] | CHEN P, ZHANG Y, CHENG Y, et al. Pathformer: multi-scale transformers with adaptive pathways for time series forecasting [EB/OL]. [2024-11-15]. . |
| [10] | WANG S, WU H, SHI X, et al. TimeMixer: decomposable multiscale mixing for time series forecasting [EB/OL]. [2024-11-15]. . |
| [11] | NIE Y, NGUYEN N H, SINTHONG P, et al. A time series is worth 64 words: long-term forecasting with Transformers [EB/OL]. [2024-12-15]. . |
| [12] | STITSYUK A, CHOI J. xPatch: Dual-stream time series forecasting with exponential seasonal-trend decomposition [C]// Proceedings of the 39th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2025: 20601-20609. |
| [13] | GONG Z, TANG Y, LIANG J. PatchMixer: a patch-mixing architecture for long-term time series forecasting [EB/OL]. [2024-08-15]. . |
| [14] | 闫剑锋,刘香莲,武北松,等.瑞利散射辅助的分布式光纤测温系统[J].传感技术学报, 2024, 37(2): 360-364. |
| YAN J F, LIU X L, WU B S, et al. Distributed optical fiber temperature measurement system assisted by Rayleigh scattering [J]. Chinese Journal of Sensors and Actuators, 2024, 37(2): 360-364. | |
| [15] | 王翔,陈志祥,毛国君.融合局部和全局相关性的多变量时间序列预测方法[J].计算机应用, 2025, 45(9): 2806-2816. |
| WANG X, CHEN Z X, MAO G J. Multivariate time series prediction method combining local and global correlation [J]. Journal of Computer Applications, 2025, 45(9): 2806-2816. | |
| [16] | LIU M, ZENG A, CHEN M, et al. SCINet: time series modeling and forecasting with sample convolution and interaction [C]// Proceedings of the 36th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2022: 5816-5828. |
| [17] | 王慧斌,胡展傲,胡节,等.基于分段注意力机制的时间序列预测模型[J/OL].计算机应用, 2024: 1-9. |
| WANG H B, HU Z A, HU J, et al. Time series forecasting model based on segmented attention mechanism [J]. Journal of Computer Applications, 2024: 1-9. | |
| [18] | CAI W, LIANG Y, LIU X, et al. MSGNet: learning multi-scale inter-series correlations for multivariate time series forecasting [C]// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2024: 11141-11149. |
| [19] | WU H, XU J, WANG J, et al. Autoformer: decomposition Transformers with auto-correlation for long-term series forecasting [C]// Proceedings of the 35th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2021: 22419-22430. |
| [20] | ZENG A, CHEN M, ZHANG L, et al. Are Transformers effective for time series forecasting? [C]// Proceedings of the 37th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2023: 11121-11128. |
| [21] | LIU Y, HU T, ZHANG H, et al. iTransformer: inverted Transformers are effective for time series forecasting [EB/OL]. [2024-12-15]. . |
| [22] | ZHOU X, WANG W, BUNTINE W, et al. Scalable Transformer for high dimensional multivariate time series forecasting [C]// Proceedings of the 33rd ACM International Conference on Information and Knowledge Management. New York: ACM, 2024: 3515-3526. |
| [23] | ZHOU H, ZHANG S, PENG J, et al. Informer: beyond efficient Transformer for long sequence time-series forecasting [C]// Proceedings of the 35th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2021: 11106-11115. |
| [1] | 麦超云, 张洪燚, 秦传波, 曾军英, 王栋. 基于多尺度与空间频率特征的嗜铬细胞瘤图像分割网络[J]. 《计算机应用》唯一官方网站, 2026, 46(1): 280-288. |
| [2] | 郭伟, 王曼婷, 曲海成. 基于多尺度感知的多维空间融合水下图像增强算法[J]. 《计算机应用》唯一官方网站, 2026, 46(1): 224-232. |
| [3] | 马英杰, 覃晶滢, 赵耿, 肖靖. 面向物联网图像的深度压缩感知网络及其混沌加密保护方法[J]. 《计算机应用》唯一官方网站, 2026, 46(1): 144-151. |
| [4] | 梁一鸣, 范菁, 柴汶泽. 基于双向交叉注意力的多尺度特征融合情感分类[J]. 《计算机应用》唯一官方网站, 2025, 45(9): 2773-2782. |
| [5] | 颜承志, 陈颖, 钟凯, 高寒. 基于多尺度网络与轴向注意力的3D目标检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(8): 2537-2545. |
| [6] | 陈亮, 王璇, 雷坤. 复杂场景下跨层多尺度特征融合的安全帽佩戴检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2333-2341. |
| [7] | 陈丹阳, 张长伦. 多尺度去相关的图卷积网络模型[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2180-2187. |
| [8] | 王向, 崔倩倩, 张晓明, 王建超, 王震洲, 宋佳霖. 改进ConvNeXt的无线胶囊内镜图像分类模型[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 2016-2024. |
| [9] | 陈盈涛, 方康康, 张金敖, 梁浩然, 郭焕斌, 邱兆文. 基于多尺度空间特征的冠状动脉CT血管造影图像分割网络[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 2007-2015. |
| [10] | 周天彤, 郑妍琪, 魏韬, 戴亚康, 邹凌. 融合变分图自编码器与局部-全局图网络的认知负荷脑电识别模型[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1849-1857. |
| [11] | 郭诗月, 党建武, 王阳萍, 雍玖. 结合注意力机制和多尺度特征融合的三维手部姿态估计[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1293-1299. |
| [12] | 令狐鑫瑶, 陈燕, 张鹏程, 刘祎, 桂志国, 赵伟, 董展豪. 基于多尺度引导滤波的宫颈细胞核图像分割[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1333-1339. |
| [13] | 姜坤元, 李小霞, 王利, 曹耀丹, 张晓强, 丁楠, 周颖玥. 引入解耦残差自注意力的边界交叉监督语义分割网络[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1120-1129. |
| [14] | 袁宝华, 陈佳璐, 王欢. 融合多尺度语义和双分支并行的医学图像分割网络[J]. 《计算机应用》唯一官方网站, 2025, 45(3): 988-995. |
| [15] | 张众维, 王俊, 刘树东, 王志恒. 多尺度特征融合与加权框融合的遥感图像目标检测[J]. 《计算机应用》唯一官方网站, 2025, 45(2): 633-639. |
| 阅读次数 | ||||||
|
全文 |
|
|||||
|
摘要 |
|
|||||