基于PHM的电容退化状态可视化系统的设计与实现

发布时间：2018-04-10 21:16

本文选题：故障预测与健康管理 + 剩余使用寿命　；参考：《中国科学院大学(中国科学院沈阳计算技术研究所)》2017年硕士论文

【摘要】：在工业电子系统中,PHM可以提前预知将要发生故障的时间和位置,预测整个系统的剩余使用寿命(URL),为未来将要发生故障的时间进行预测,提高维修准时性。美国辛辛那提大学的李杰教授对诊断健康管理系统(PHM)多年研究后提出了一套完整的数据驱动的PHM研究路线:数据分析建模,数据可视化,数据等平台的搭建与维护。本文根据这一研究路线的进行工业电容退化状态的研究,并为其设计可视化系统:1.数据分析建模。基于工业电容充电周期下获得的EIS频谱数据信息,寻找随电容退化而变化的隐藏变量。整理隐藏变量进行电容的健康状态的量化建模分析。最后根据这些量化的状态数据,建立以时间状态序列为基础的电容的退化预测模型。2.数据可视化系统。把电容退化预测模型进行可视化展示,并将展示系统整合成一个面向非技术类管理人员的专业设备的状态管理系统。完成适应于更多仪器设备的PHM可视化Web系统。该部分注重界面的直观性与交互性。3.数据等平台的搭建与维护。针对PHM系统的运行特性设计集群配置。从冗余的思想出发,设计基于Web服务器与数据库服务器的集群网络,并整合离线大数据处理spark平台,用于大数据的计算。根据上述的方案开发的工业电容PHM系统,适用于非技术类的工业设备管理人员,为管理人员提供及时更换过度损耗设备的辅助决策支持与知识服务,避免因流水线意外中断而造成的停工损失,提升了企业的利益与工作的效益。
[Abstract]:In the industrial electronic systems, PHM can predict in advance the location and time to fail, the remaining service life prediction system (URL), to predict the future will be the time to failure, improve the maintenance of time. University of Cincinnati professor Li Jie on the diagnosis of health management system (PHM) after years of research put forward a complete set of data driven PHM research process: data analysis, data visualization, build and maintain the data platform. Based on the research of industrial capacitance degradation state of this line of research, design and visualization system for data analysis: 1.. Industrial capacitor charging EIS spectrum data obtained under cycle based on the finding with latent variable capacitor degradation changes. Quantitative analysis of modeling latent variable capacitance finishing health. Finally according to the amount of the number of States According to the established, capacitor to time state sequence based degradation prediction model of.2. data visualization system. The capacitance degradation visual display model, and display system integrated into the state of a management system for non technical management personnel professional equipment. Complete the PHM Web visualization system adapted to the more equipment. Some focus on building and maintaining the interface intuitive and interactive.3. data platform. The operating characteristics of the cluster configuration design for PHM system. Starting from the thought of redundancy design of cluster network, Web server and database server based on off-line data processing and integration of large spark platform, used to calculate the data. According to the industrial capacity PHM system the development plan, industrial equipment management personnel for non technology, provide auxiliary equipment in time to replace excessive loss management The decision support and the knowledge service can avoid the shutdown loss caused by the accidental interruption of the pipeline, and improve the benefit of the enterprise and the work.

【学位授予单位】：中国科学院大学(中国科学院沈阳计算技术研究所)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52

【参考文献】