基于Hadoop的轴承监测海量数据的存储与管理

摘要

针对高速机车转向架滚动轴承监测数据的格式多样化、数据量激增带来的管理和存储问题,在分析监测数据特征和数据量大小的基础上,构建了树形索引结构对滚动轴承状态监测海量数据进行管理.利用研究所当前的软硬件环境,搭建了分布式集群系统存储海量监测数据,并对集群系统在存储和管理海量数据时的可用性和稳定性进行了探讨.通过对实际数据的测试和分析,结果表明基于树形索引结构的分布式集群系统能在一定程度上满足滚动轴承状态监测非结构化海量数据存储和管理的要求.%The monitoring data of rolling bears in high speed locomotive bogie presents a variety of formats.The monitoring data volume grows rapidly.It causes management and storage problems.Based on the analysis of the monitoring data characteristics and data size,a tree index structure is put forward to manage the massive data of rolling bearing detection in high speed locomotive bogie.Using the current software and hardware environment of the institute,a distributed cluster system is built to store massive monitoring data.Then,the availability and stability of cluster system in storing and managing massive data are discussed.Through the test and analysis of the actual monitoring data,the results show that distributed cluster tree index structure can satisfy the management and storage requirements for massive data of rolling rnonitoring.