工业实时数据在线分析平台分析-通信与信息系统专业论文.docxVIP

华 华 中 科 技 大 学 硕 士 学 位 论 文 II II Abstract As the increasing efforts and applications on industrial information system, the widely spread devices of networks declare more demand on data acquisition size, acquisition precision and acquisition speed. With the rapid increasing speed of massive data, as a special database system which can collect, store and manage real-time data, the industrial real-time database system has widely used in information application system such as supervisory information system and distribution network automation system. The massive data means massive information, and it is a very important function for industrial real-time database to discover useful knowledge by data analysis. However, current data analysis method of industrial real-time database is not online, and could only find some rough information. In the condition of streaming database, user need an online dynamic data analysis method to response the arriving data in real-time. Based on our industrial real-time database and online analysis model and method of data stream management systems, this thesis designed an online analysis platform of industrial real-time database. The platform provides a declarative language to answer analysis query from user and can be used to do many analysis jobs as multi-scale analysis, continuous querying and data abstraction etc. the platform extend relation algebra for time series data, and make query to like in relation database. In our analysis platform, we proposed a new kind of bloom filters called Probability Bloom Filters (PBF) to represent set with attributes and design a PBF based synopsis structure called Multi-PBF (MPBF) for industrial real-time data. Furthermore, we provide theoretical analysis of the effectiveness of MPBF and demonstrate it by simulation. Keywords: industrial real-time database, data stream management system, online analysis, synopsis structure, Bloom filters III III 目 录 摘 要 I ABSTRACT II 1 绪论 1.1 课题研究背景 (