基于张量分解的室内温度数据补全技术研究  

Research on indoor temperature data completion technology based on tensor decomposition

作　　者：孙龙 李海广[1] 秦杨 张超[3] SUN Long;LI Haiguang;QIN Yang;ZHANG Chao(School of Energy and Environment,Inner Mongolia University of Science and Technology,Baotou 014000,China;Baotou Shiguai Guangyuan Investment Co.,Ltd.,Baotou 014000,China;School of Mechanical Engineering,Inner Mongolia University of Science and Technology,Baotou 014000,China)

机构地区：[1]内蒙古科技大学能源与环境学院,包头014000 [2]包头市石拐广源投资有限公司,包头014000 [3]内蒙古科技大学机械工程学院,包头014000

出　　处：《通信与信息技术》2025年第2期70-74,共5页Communication & Information Technology

基　　金：国家自然科学基金资助项目(项目编号:50965052)。

摘　　要：室内温度是建设、运行、智能控制、智慧供热、节能降耗、日常管理、技术进步、满意服务的根本遵循。但室温数据获取困难、设备损坏等问题造成室温缺失,这给实时监测和精确调控带来了挑战,并导致对室内热环境的准确评估和分析产生困难。为了解决室温数据缺失问题,运用数据补全的概念,提出了一种基于塔克正则化(Tucker-Regularization)的张量补全方法,对缺失的室温数据集进行数据补全,通过在核心张量中加入L_2范数来提高鲁棒性,并采用交替最小二乘法求解模型。研究中使用真实的室温数据,研究如何在不同缺失场景下提高补全准确性。研究结果显示:所提出的算法的性能要优于对比算法,并且在高缺失率下具有良好的适用性和可扩展性。Indoor temperature is the fundamental principle for construction,operation,intelligent control,smart heating,energy con⁃servation and consumption reduction,daily management,technological progress,and satisfactory service.However,difficulties in obtain⁃ing room temperature data and equipment damage have resulted in missing room temperature,which poses challenges for real-time moni⁃toring and precise control,and leads to difficulties in accurately assessing and analyzing indoor thermal environments.In order to solve the problem of missing room temperature data,a tensor completion method based on Tucker Regularization is proposed using the concept of data completion.The missing room temperature dataset is completed by adding L2 norm to the core tensor to improve robustness,and the alternating least squares method is used to solve the model.In the study,real room temperature data was used to investigate how to im⁃prove completion accuracy in different missing scenarios.The research results show that the proposed algorithm outperforms the com⁃pared algorithms in terms of performance,and has good applicability and scalability under high missing rates.

关 键 词：室内温度 缺失数据补全 张量补全 L_2范数 

分 类 号：TN914[电子电信—通信与信息系统]