Recently,the big data technology has been applied in many field widely,such as finance,trade and medical healthy. But the applications in process industries are only in the beginning stages. In this paper,the characteristics,analyzing methods and applications of the data in process industries are introduced. The data obtained by the process industry have the characteristics of high dimension,strong nonlinearity, uneven sample distribution and low signal-to-noise ratio except from the characteristics of volume, variety,velocity and variability. The big data technology has emerged and developed to be available in analyzing data from the process industries. The analyzing methods based on the industrial data include dimension reduction analysis,cluster and classification analysis,correlation analysis and prediction analysis according to their functions. In this paper,the applications of the big data technology in process industries are summarized from three aspects including process optimization,process monitoring and fault diagnosis and prediction of product properties and yield. It is found that the big data will play a more important role if the production data in the process industries can be combined with the market data of raw material and product.%近些年，大数据技术在金融、贸易和医疗健康等行业也得到了较好的应用，但大数据技术在过程工业中的应用还处于起步阶段。本文分别从过程工业大数据的特点、分析方法以及应用现状3个方面进行介绍，简述了过程工业数据除了具有一般大数据海量性、多样性、高速性和易变性的4V特点外，还具有高维度、强非线性、样本分布不均和低信噪比的特点。基于过程工业数据的分析方法，按照功能划分可以分为降维分析、聚类和分类分析、相关性分析和预测分析四大类。在此基础上，综述了近些年大数据技术在过程工业上的应用，分别从过程工业优化、过程监测与故障诊断以及产品性能和产率预测3个方面介绍了其在过程工业中的应用情况，并指出未来应该将企业内部的生产数据和原料与产品的市场数据等相结合进行分析和挖掘，这样能够更大程度地发挥大数据的价值。
展开▼
