基于Spark的水泥厂煤耗特性分析与实时诊断

Coal Consumption Performance Analysis and Real-time Diagnosis of Cement Plant Based on Spark

针对国内水泥厂煤耗较高、数据利用率低的问题,对大数据、数据挖掘技术应用于水泥厂煤耗管控进行了研究。搭建大数据计算平台,使用Spark计算引擎采集并清洗水泥厂生产设备的过往稳态数据,建立了煤耗特性的数学模型;研究实时诊断算法并运行于大数据计算平台上;对实时采集生产设备的运行参数进行计算,得出各运行参数当前的数值,并进行数据可视化,指导水泥厂改良将会造成能耗偏差的异常参数,以降低煤耗、改善生产。

In view of the problems of high coal consumption and low data utilization rate in domestic cement plants,the application of large data and data mining technology to coal consumption management and control in cement plants was studied.Build a big data computing platform,use Spark calculation engine to collect and clean the past steady-state data of cement plant production equipment,establish the mathematical model of coal consumption characteristics.Research on real-time diagnosis algorithm and run on large data computing platform.By calculating the operation parameters of real-time acquisition production equipment,the current values of each operation parameters are obtained,and data visualization is carried out to guide the improvement of abnormal parameters that may cause energy consumption deviation in cement plant,so as to reduce coal consumption and improve production.