基于快速傅立叶变换-卷积神经网络的医疗设备多维生命周期管理系统的构建研究  

Research on the construction of FFT-CNN-based MDLM system of medical equipment

作　　者：张苏 张民 史亚香[2] 黄丽亚[3] Zhang Su;Zhang Min;Shi Yaxiang;Huang Liya(Department of Information,Baoying People’s Hospital,Yangzhou 225800,China;Network Information Center,Zhongda Hospital,Southeast University,Nanjing 210009,China;College of Electronic and Optical Engineering,College of Flexible Electronics(Future Technology),Nanjing University of Posts and Telecommunications,Nanjing 210023,China)

机构地区：[1]宝应县人民医院信息科,扬州225800 [2]东南大学附属中大医院网络信息中心,南京210009 [3]南京邮电大学电子与光学工程学院、柔性电子(未来技术)学院,南京210023

出　　处：《中国医学装备》2025年第4期117-122,共6页China Medical Equipment

摘　　要：目的:构建基于快速傅立叶变换-卷积神经网络(FFT-CNN)的医疗设备多维生命周期管理(MDLM)系统,为医疗设备生命周期管理及多维度延伸提供数据支撑。方法:MDLM系统架构设置综合信息展示层、多维业务轨迹数据延伸层、设备生命周期数据层和设备数据采集层,采用FFT-CNN算法将医疗设备运行特征数据映射为医院业务数据,推衍出碳轨迹、运维轨迹、效益轨迹、安全轨迹、绩效轨迹和考核轨迹6个维度的医院运营管理指标,采用智能物联网(AIoT)数据探针、FFT-CNN模式识别和指标推衍技术与方法对医疗设备生命周期全面数字化转型管理。分析MDLM系统自2023年8月应用于宝应县人民医院医疗设备管理后的数据,对比系统应用前后的医院核磁共振设备维护效率。结果:医疗设备MDLM系统实现了医疗设备的需求、采购、安装、归档、运行、监控、维修和报废8个阶段的生命周期管理,同时由推衍出的6个维度指标延伸出医疗设备能耗管理、运维管理、效益管理、安全管理、绩效管理和考核管理数据对医院医疗设备进行日常管理。系统应用后医院核磁共振设备的平均故障修复时间由应用前的6.8 h缩短为2.3 h,提升了运行维护效率和设备可用率。结论:基于FFT-CNN的医疗设备MDLM系统建立了医疗设备生命周期管理专属数据库,能够为医院医疗设备精益管理提供多维度的客观化、完整化和关联化的数据支撑,提升设备运行维护效率。Objective:To construct a multi-dimensional life-cycle management(MDLM)of medical equipment based on fast Fourier transform-convolutional neural network(FFT-CNN),so as to provide data support for the life-cycle management of medical equipment and its multi-dimensional derivation.Methods:The MDLM system of medical equipment was constructed by setting display layer of comprehensive information,data extension layer of multi business,data layer of life cycle of equipment,data acquisition layer.The FFT-CNN algorithm was adopted to mapping characteristic data of operating medical equipment as business data of hospital,and then,they were derived to operational management indicators of hospital from 6 dimensions included carbon trajectory,operation and maintenance trajectory,benefit trajectory,safety trajectory,performance trajectory and assessment trajectory.The data prober of artificial internet of things(AIoT),and pattern recognition and indicator derivation of FFT-CNN were adopted to conduct comprehensively digital transformation management for life-cycle of medical equipment.The management data that MDLM system were applied in the medical equipment of Baoying People's Hospital from August 2023 to April 2024 were selected to compare the efficiency of maintaining magnetic resonance(MR)equipment of hospital before and after the system was applied.Results:The MDLM system of medical equipment realized the life-cycle management for medical equipment in eight stages included demand,procurement,installation,archiving,operation,monitoring,maintenance and scrapping,and realized the use of“the extended data of energy consumption management,operation and maintenance management,efficiency management,safety management,performance management and assessment management that derived from the indicators of 6 dimensions”in conducting routine management for medical equipment of hospital.The average repair time of the failure of MR equipment of hospital has been decreased from 6.8 hours before the system was applied to 2.3 hours after

关 键 词：快速傅立叶变换 卷积神经网络 生命周期管理 多维度 

分 类 号：R197.39[医药卫生—卫生事业管理]