标题：Big Data Analyzing for Control Valve Using the Spark Cluster Computing Engine
作者：Wang, Longhui; Wang, Yong; Xie, Yudong
通讯作者：Wang, Y;Xie, YD
作者机构：[Wang, Longhui; Wang, Yong; Xie, Yudong] Shandong Univ, Coll Mech Engn, Jinan, Shandong, Peoples R China.
会议名称：Chinese Automation Congress (CAC)
会议日期：NOV 30-DEC 02, 2018
来源：2018 CHINESE AUTOMATION CONGRESS (CAC)
关键词：Big Data; Spark MapReduce; parallel SDP; t-SNE
摘要：Control valve plays an important role in modern industry, where the production depends on distributing and controlling fluid accurately. With the development of automation and Internet technology, control valves can collect real-time signals and store large amount of historical data, which is difficult to analyze and utilize. In this paper, we use the Spark cluster computing engine (Spark) to realize a fault diagnosis approach based on big data for control valve. Firstly, we use the data visualizing algorithm t-Distributed Stochastic Neighbor Embedding (t-SNE) to observe the distribution of control valve data. Then we implement a parallel Searching Density Peaks (SDP) cluster algorithm combined with Spark MapReduce for fault detection. Experimental results demonstrate that the parallel algorithm based on Spark can process big data fast and efficiently.