基于重构变分自编码的缺失数据插补方法

杨昊东; 任少君; 朱保宇; 李清扬; 司风琪

doi:10.19805/j.cnki.jcspe.2026.240765

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于重构变分自编码的缺失数据插补方法

数字化与智能化 | 更新时间：2026-10-10

- 基于重构变分自编码的缺失数据插补方法
- Missing Data Imputation Method of Reconstructed-based Variational Autoencoder
- 动力工程学报 2026年46卷第4期页码：149-158,197
- 作者机构：
- 作者简介：
  
  杨昊东(2000—),男,安徽滁州人,硕士研究生,研究方向为发电机组的参数状态监测及故障诊断。任少君(通信作者),男,副教授,博士,E-mail:rsj@seu.edu.cn。
- 基金信息：
  
  国家重点研发计划资助项目(2022YFB4100700);国家自然科学基金(青年项目)资助项目(52306230)
- DOI：10.19805/j.cnki.jcspe.2026.240765
  中图分类号： TK39
- 纸质出版：2026
- 稿件说明：
移动端阅览
杨昊东, 任少君, 朱保宇, 李清扬, 司风琪. 基于重构变分自编码的缺失数据插补方法[J]. 动力工程学报, 2026, 46(4): 149-158,197.
杨昊东, 任少君, 朱保宇, 李清扬, 司风琪. 基于重构变分自编码的缺失数据插补方法[J]. 动力工程学报, 2026, 46(4): 149-158,197. DOI： 10.19805/j.cnki.jcspe.2026.240765.

DOI： 10.19805/j.cnki.jcspe.2026.240765

摘要

准确的测量数据是构建高精度数据驱动模型的重要基础。然而

电站实际运行数据受到测量、通讯、存储等过程的影响

不可避免地存在部分数据缺失的问题

影响后续基于运行数据的模型的有效性。为此

提出了一种基于重构变分自编码(RB-VAE)的缺失数据插补方法

该方法基于变分自编码模型构建重构指标

并通过Steffensen方法实现单个或多个缺失值的加速迭代更新

在保持插补精度的同时保证了计算效率

实现了对单个或多个缺失数据的高效准确插补。采用数学算例和工程实例进行分析

在缺失参数个数为1、2、3、6时

RB-VAE插补结果的决定系数(R2)均值分别为0.998 0、0.998 6、0.999 2、0.996 2

均能与真实值保持较小的偏差

插补效果明显优于missForest和生成对抗插值网络(GAIN)方法。

Abstract

Accurately measuring data are essential for constructing high-precision data-driven models. However

the actual operational data of a power station are subject to a number of influences

such as measurement

communication

storage and other processes

inevitably resulting in partial data loss

which can subsequently impact the effectiveness of modeling based on the operational data. To address this issue

a missing data imputation method based on reconstructed variational autoencoder was proposed. This method constructed reconstruction indexes based on the variational self-coding model and utilized the Steffensen method to accelerate the iterative updating of single value or multiple missing values. This approach ensures computational efficiency while maintaining interpolation accuracy

allowing for the efficient and accurate interpolation of single missing datum or multiple missing data. Mathematical calculations and engineering examples were employed to analyze the results. When the number of missing parameters is 1

3 and 6

the mean coefficients of determination (R2) of the reconstruction-based variational autoencoder interpolation results are 0.998 0

0.998 6

0.999 2 and 0.996 2

respectively. These values demonstrate minimal deviation from the actual values

and the interpolation effect is evidently superior to that of the missForest and generative adversarial imputation net (GAIN) methods.

关键词

Keywords

references

李永华, 陈伟华, 谢英柏. 基于数据驱动的燃气轮机性能监测方法[J]. 动力工程学报, 2024, 44(12): 1845-1853.LI Yonghua, CHEN Weihua, XIE Yingbai. Data-driven gas turbine performance monitoring method[J]. Journal of Chinese Society of Power Engineering, 2024, 44(12): 1845-1853.

张景欣, 周东华, 陈茂银, 等. 数据驱动的多工况过程异常监测方法: 综述与展望[J]. 中国科学: 信息科学, 2023, 53(11): 2087-2106.ZHANG Jingxin, ZHOU Donghua, CHEN Maoyin, et al. Data-driven anomaly monitoring methods for multimode processes: overview and perspectives[J]. Scientia Sinica Informationis, 2023, 53(11): 2087-2106.

王志敏, 黄骞, 王可轩, 等. 智慧电厂大数据云平台的架构建设与模型开发研究[J]. 动力工程学报, 2025, 45(2): 282-291.WANG Zhimin, HUANG Qian, WANG Kexuan, et al. Architecture construction and model development of big-data cloud platform for intelligent coal-fired power plants[J]. Journal of Chinese Society of Power Engineering, 2025, 45(2): 282-291.

MOU Miao, ZHAO Xiaoqiang. Incipient fault detection and diagnosis of nonlinear industrial process with missing data[J]. Journal of the Taiwan Institute of Chemical Engineers, 2022, 132: 104115.

YAO Zoujing, ZHAO Chunhui. FedTMI: knowledge aided federated transfer learning for industrial missing data imputation[J]. Journal of Process Control, 2022, 117: 206-215.

MUKAKA M, WHITE S A, TERLOUW D J, et al. Is using multiple imputation better than complete case analysis for estimating a prevalence (risk) difference in randomized controlled trials when binary outcome observations are missing?[J]. Trials, 2016, 17: 341.

XUE Jianing. Review on data imputation methods in machine learning[J]. Journal of Physics: Conference Series, 2023, 2646(1): 012034.

石丽. 多重插补在成分数据缺失值补全中的应用[D]. 太原: 山西大学, 2012.

王寓枫. 提高深度神经网络数据插补的若干方法研究[D]. 烟台: 烟台大学, 2023.

SUN Yige, LI Jing, XU Yifan, et al. Deep learning versus conventional methods for missing data imputation: a review and comparative study[J]. Expert Systems with Applications, 2023, 227: 120201.

遇茜, 钱政, 聂志鹏. 风机异常及缺失数据的填补方法研究[J]. 电测与仪表, 2020, 57(23): 1-8.YU Qian, QIAN Zheng, NIE Zhipeng. Research on filling method of abnormal and missing data of wind turbines[J]. Electrical Measurement & Instrumentation, 2020, 57(23): 1-8.

邵良杉, 赵松泽. 基于多模型融合的不完整数据分数插补算法[J]. 计算机工程, 2023, 49(9): 79-88, 98.SHAO Liangshan, ZHAO Songze. Fractional imputation algorithm for incomplete data based on multi-model fusion[J]. Computer Engineering, 2023, 49(9): 79-88, 98.

陈帅, 赵明, 郭栋, 等. 基于SVD-KDR算法的工业监测数据插补技术[J]. 机械工程学报, 2021, 57(2): 30-38.CHEN Shuai, ZHAO Ming, GUO Dong, et al. Missing data imputation using SVD-KDR algorithm in industrial monitoring data[J]. Journal of Mechanical Engineering, 2021, 57(2): 30-38.

华南, 马春萍, 朱彦霞, 等. 一种基于MissFroest的并行插补算法[J]. 河南科技, 2022, 41(3): 18-21.HUA Nan, MA Chunping, ZHU Yanxia, et al. A parallel interpolation algorithm based on MissFroest[J]. Henan Science and Technology, 2022, 41(3): 18-21.

ZHANG Shuo, HU Xingbang, CHEN Jinyi, et al. An effective variational auto-encoder-based model for traffic flow imputation[J]. Neural Computing and Applications, 2024, 36(5): 2617-2631.

LIU Wei, LI Guizhen, ZHOU Ling, et al. High-dimensional large-scale mixed-type data imputation under missing at random[J]. Science China Mathematics, 2025, 68(4): 969-1000.

LIU Xinyao, DU Shengdong, LI Tianrui, et al. A missing value filling model based on feature fusion enhanced autoencoder[J]. Applied Intelligence, 2023, 53(21): 24931-24946.

赵景启. 基于生成式对抗网络的多维时间序列修复算法研究[D]. 天津: 天津工业大学, 2023.

郑智泉, 陈妍, 王孟孟, 等. 不同缺失率下的数据填补算法稳定性研究[J]. 统计与决策, 2023, 39(8): 12-17.ZHENG Zhiquan, CHEN Yan, WANG Mengmeng, et al. Research on stability of data filling algorithms with different miss rates[J]. Statistics and Decision, 2023, 39(8): 12-17.

谷小兵, 牛少彰, 王茂森, 等. 基于Transformer的时间序列插补技术研究[J]. 图像与信号处理, 2024, 13(2): 151-162.GU Xiaobing, NIU Shaozhang, WANG Maosen, et al. Research on Transformer-based time series imputation technique[J]. Journal of Image and Signal Processing, 2024, 13(2): 151-162.

宋亮, 万建洲. 缺失数据插补方法的比较研究[J]. 统计与决策, 2020, 36(18): 10-14.SONG Liang, WAN Jianzhou. Comparative research on interpolation method of missing data[J]. Statistics & Decision, 2020, 36(18): 10-14.

WU Yangyang, WANG Jun, MIAO Xiaoye, et al. Differentiable and scalable generative adversarial models for data imputation[J]. IEEE Transactions on Knowledge and Data Engineering, 2024, 36(2): 490-503.

REN Shaojun, WU Shiliang, WENG Qihang, et al. Disentangled representation aided physics-informed neural network for predicting syngas compositions of biomass gasification[J]. Energy & Fuels, 2024, 38(3): 2033-2045.

ZIMMERMAN D L, CRESSIE N. Mean squared prediction error in the spatial linear model with estimated covariance parameters[J]. Annals of the Institute of Statistical Mathematics, 1992, 44(1): 27-43.

任少君. 热工过程海量数据流模型分析及诊断方法研究[D]. 南京: 东南大学, 2018.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

融合长短期依赖的双通道短期风电功率预测

分形特征对漂浮式风力机仿生平台动态响应的影响

基于全流道冷却集热器的PVT系统热电性能研究

基于Leckner模型的氨掺烧高水蒸气浓度烟气对燃煤锅炉传热特性的影响分析

基于WOA的循环流化床锅炉SO2排放浓度VMD-Transformer预测模型