Monitoring of metallurgical plants by using biplots

Although principal component analysis has been applied widely for monitoring plant performance in a broad range of industrial processes, the use of accompanying biplots has not received similar attention. Moreover, principal component analysis is a linear technique that tends to break down when processes exhibit significant nonlinear behavior. In this paper biplot methodology is introduced. This methodology allows for projecting high-dimensional data to a low-dimensional subspace that can be visualized by a human operator. It provides management with sophisticated tools, highly graphical in nature, to extract information regarding variation in process variables, correlations among these variables, as well as class separation, taking into account the multidimensional character of the data. Biplot methodology can also be applied to data sets exhibiting nonlinear behavior. As is shown by way of two case studies, operating regions can be quantitatively explored by superimposing alpha bags on biplots, which facilitates the automatic detection and visualization of process disturbances. © 2004 American Institute of Chemical Engineers.