A fast monitoring method for multiple operating batch processes with incomplete modeling data types

作者：Wang Ya Jun; Jia Ming Xing^*; Mao Zhi Zhong

来源：Journal of Industrial and Engineering Chemistry, 2015, 21: 328-337.

DOI：10.1016/j.jiec.2014.02.042

• Multiple operating batch process
• Incomplete modeling data types
• Up-Down MDPCA
• 120 t Ladle furnace (LF) steelmaking process

For complex industrial processes with multiple operating conditions, a novel fast monitoring method, called up-down multi-model dynamic principal component analysis, is proposed with incomplete modeling data types in this paper. The method firstly classifies the process into several stages according to the operations. In each stage the data with similar operations are clustered together. Multiple model structures are constructed along with the variation of the cluster number. When on-line monitoring, the up-down monitoring method is proposed to monitor a new batch type. The effectiveness of the proposed monitoring method is demonstrated through the 120 t ladle furnace (LF) steelmaking process.

• 出版日期2015-1-25
• 单位东北大学; 辽宁工业大学