The imaging of phase distributions in vessels using tomographic techniques offers benefits in relation to improved and more efficient control schemes arising from comprehensive data acquisition. However, such image measurements are multivariate in nature and require statistical treatment. Therefore, the modelling of a multiphase process by tomography is different from the conventional process modelling where the measurements are univariate. In this paper, principal component analysis is used to extract useful information from tomographic images and to estimate the dynamic behaviour of mixing processes.
Simons, S. J. R., Ni, X., West, R. M., & Williams, R. A. (1999). Reduction of tomographic data for use in the control of multiphase processes. Chemical Engineering Communications, 175(1), 99-115. https://doi.org/10.1080/00986449908912141