Predicting granule size distribution of a fluidized bed spray granulation process by regime based PLS modeling of acoustic emission data

Granule size distribution (GSD) is generally regarded as one of the most critical quality-affecting attributes of a fluidized bed spray granulation process. Several process analytical techniques have been applied for in- and on-line monitoring of the GSD during the process. In this study, an acoustic emission technique was used for off-line prediction of the GSD. Acoustic spectra were recorded during 24 batch fluidized bed spray granulations. An off-line analysis of the acoustic spectra was performed using partial least squares regression (PLSR). First, the acoustic data from the spraying phase was divided into separate regimes, each corresponding to distinctive physico-chemical conditions occurring during different stages of the process. Then, a synthetic calibration set was used to build, predictive models for GSD. The GSD could be predicted with good accuracy and precision. In addition, the option for using the acoustic emission technique as an early-warning system to detect process deviations was evaluated.