Interval prediction of current transformer measurement error based on DTCN and residual Bootstrap method
Received date: 2026-04-27
Revised date: 2026-07-27
Online published: 2026-09-03
To address the problems that the ratio-error series of current-transformers,affected by environmental disturbance, operating condition variations and random noise,exhibit nonstationary and multi-scale characteristics,and it is difficult to balance point prediction and interval estimation,this paper proposes an interval forecasting method integrating complete ensemble empirical mode decomposition(CEEMD),dual-scale temporal convolutional network(DTCN)and bias-corrected residual bootstrap(BCRB).Firstly,The CEEMD is adopted to decompose and reconstruct the original ratio-error series to suppress high-frequency noise and mode mixing.Secondly, the DTCN is used to extract short-term fluctuation and long-term drift features to realize point prediction of ratio errors.Finally,prediction intervals under different confidence levels are constructed based on the residual bootstrap method. Case studies show that the proposed method achieves a high interval coverage probability at the confidence level of 0.99. This work is supported by National Key Research and Development Program of China(No.2023YFB2405903).
Herong ZHU , Shaodong LI , Senlin ZHAO , Wei LU , Jindong DENG , Zhenhua LI . Interval prediction of current transformer measurement error based on DTCN and residual Bootstrap method[J]. Electric Power, 2026 , 59(8) : 49 -60 . DOI: 10.11930/j.issn.1004-9649.202604070
表 1 未经 CEEMD 处理的不同模型点预测性能对比Table 1 Point forecasting performance comparison of different models without CEEMD preprocessing |
| 模型 | 评估指标 | ||
|---|---|---|---|
| RMSE/ 10-3 | MAE/ 10-3 | R2 | |
| CNN | 2.278 | 2.032 | 0.938 |
| TCN | 1.602 | 1.330 | 0.969 |
| BiTCN | 1.515 | 1.229 | 0.973 |
| DTCN | 1.418 | 1.153 | 0.976 |
表 2 基于CEEMD分解的不同模型点预测性能对比Table 2 Point forecasting performance comparison of different models based on CEEMD decomposition |
| 模型 | 评估指标 | ||
|---|---|---|---|
| RMSE/ 10-3 | MAE/ 10-3 | R2 | |
| CEEMD-CNN | 1.573 | 1.313 | 0.970 |
| CEEMD-TCN | 1.213 | 0.909 | 0.982 |
| CEEMD-BiTCN | 1.166 | 0.864 | 0.984 |
| CEEMD-DTCN | 0.882 | 0.772 | 0.991 |
图6 CEEMD分解前后不同模型拟合结果对比Fig. 6 Comparison of prediction results of different models before and after CEEMD decomposition |
表 3 不同 Bootstrap 区间方法的区间预测性能对比Table 3 Comparison of interval forecasting performance among different Bootstrap-based methods |
| 置信水平 | 方法 | 评估指标 | ||
|---|---|---|---|---|
| PICP | PINAW | 区间得分 | ||
| 0.80 | PB | 0.7630 | 0.0615 | 0.1014 |
| BC-PB | 0.7986 | 0.0676 | 0.1008 | |
| 0.90 | PB | 0.8863 | 0.0855 | 0.1278 |
| BC-PB | 0.8981 | 0.0890 | 0.1277 | |
| 0.95 | PB | 0.9431 | 0.1079 | 0.1623 |
| BC-PB | 0.9455 | 0.1187 | 0.1618 | |
| 0.99 | PB | 0.9858 | 0.1991 | 0.2533 |
| BC-PB | 0.9905 | 0.2170 | 0.2499 | |
| [1] |
王海元, 李恺, 张军号, 等. 电能计量装置二次回路误差分析与应对[J]. 中国测试, 2021, 47(9):133-137.
|
| [2] |
赵晓明, 余志慧. 电流互感器5%误差曲线测试数据的矫正[J]. 浙江电力, 2014, 33(5):45-47.
|
| [3] |
|
| [4] |
许灵洁, 李航康, 周琦, 等. 电子式互感器的现场误差试验及问题分析[J]. 浙江电力, 2012, 31(11):5-7.
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
田金虎, 汪金刚, 徐郁, 等. 计及互感器融合偏差的变电站站用电剩余电流监测系统研究[J]. 电工电能新技术, 2024(8):69-77.
|
| [9] |
|
| [10] |
邢作霞, 马岩溪, 郭珊珊, 等. 基于 WOA-GRU 的风电机组发电机故障预警方法[J]. 电机与控制学报, 2025, 29(6):54-62.
|
| [11] |
宁少慧, 戎有志, 董振才. 基于 PCA 和 DRSN-TCN 的滚动轴承的剩余寿命预测[J]. 科学技术与工程, 2025, 25(35):15096-15104.
|
| [12] |
吴江雄, 刘千宽, 阳国燕, 等. 应用深度学习网络的变电站二次测量回路误差评估[J]. 中国电力, 2025, 58(6):206-212.
|
| [13] |
张勇, 柯梦晗, 刘伟.220 kV 母差保护电流互感器断线异常原因分析与处理[J]. 浙江电力, 2017, 36(5):16-19, 33.
|
| [14] |
李成雲, 杨东升, 周博文, 等. 基于数字孪生技术的新型电力系统数字化[J]. 综合智慧能源, 2024, 46(2):1-11.
|
| [15] |
毕贵红, 张靖超, 赵四洪, 等. 基于小波散射变换与 IRCA-ICA-Res 结合的电压源控制型静止同步补偿系统的故障诊断[J]. 电力系统保护与控制, 2025, 53(8):144-158.
|
| [16] |
黄红桥, 谈丛, 汪凤娇, 等. 拆回智能电能表烧表故障分析及对策研究[J]. 电测与仪表, 2026, 63(3):185-192.
|
| [17] |
武延林, 卢健强, 朱禹泓, 等. 基于物理信息神经网络的电网连锁故障风险评估[J]. 电测与仪表, 2025, 62(12):160-167.
|
| [18] |
|
| [19] |
|
| [20] |
|
| [21] |
邢宇, 徐云武, 翟晓卉, 等. 计及一次电压波动影响的 CVT 内绝缘状态在线评估方法[J]. 电测与仪表, 2025, 62(6):102-110.
|
| [22] |
尹元亚, 潘文虎, 赵文广, 等. 基于 CEEMDAN 和 BiLSTM-AM 的超短期风速预测方法[J]. 电测与仪表, 2024, 61(9):77-84.
|
| [23] |
张安安, 谢琳惺, 杨威. 基于 CNN-GRU 组合神经网络的锂电池寿命预测模型研究[J]. 电测与仪表, 2025, 62(7):77-84.
|
| [24] |
|
| [25] |
|
| [26] |
韩宇昊, 刘鸿鹏. 基于 FMD-LSTM-Bootstrap 的电动汽车充电负荷区间预测[J]. 东北电力大学学报, 2025, 45(6):51-59.
|
| [27] |
方砚兵, 封坤, 李博, 等. 基于 Bootstrap-AIC 的隧道围岩参数不确定性表征与最小样本数研究[J]. 中国公路学报, 2024, 37(11): 26-37.
|
| [28] |
刘立, 朱健成, 韩光洁, 等. 基于1D-CNN联合特征提取的轴承健康监测与故障诊断[J]. 软件学报, 2021, 32(8):2379-2390.
|
| [29] |
周航, 程泽, 弓清瑞, 等. 基于 TCN 编码的锂离子电池 SOH 估计方法[J]. 湖南大学学报(自然科学版), 2023, 50(4):185-192.
|
| [30] |
赵长啸, 方玉麟, 汪克念. 基于 BiTCN 的无人机指挥控制链路DoS 攻击检测方法[J]. 航空学报, 2026, 47(1):249-265.
|
/
| 〈 |
|
〉 |