Early Detection of Agglomeration in Conical Spouted Beds Using Recurrence Plots

Stability of spouted bed during spray cold coating on the surface of carrier particles

The rapid freezing of droplets on the surface of cold carrier particles can reduce the spray freeze-drying time. The coating freezing process can be treated in batch by using a spouted bed and the bed stability is the key. The freezing process of droplets would agglomerate the carrier particles. By analyzing the standard deviation of the bed pressure drop and the fluctuation of the bed pressure, the stable operating conditions of the spouted bed under cold coating were studied, and a dynamic adjustment method of the inlet air velocity to maintain stable spouting was proposed. The results show that the stable operating air velocity range can be described useing the standard deviation of pressure fluctuations. Before bed collapse occurs, it will experience an unstable spouting process, which causes the bed pressure to fluctuate a standard deviation of less than 15 Pa after more than 40 Pa.

Development of a PAT tool for monitoring the Wurster coater performance

Real-time process analytical technology (PAT) is proposed as an effective approach for monitoring the performance of a Wurster coater. The coater was used for coating of 0.78 mm pharmaceutical pellets. The coating solution consisted of Hydroxypropyl methylcellulose/Poly ethylene glycol. During the coating process, small amounts of pellets (∼2 g) were sampled at 10 min. intervals and the pressure fluctuations were recorded at the same time. The signals were analyzed using the wavelet transform (WT) and decomposed into different sub-signals. Principal component analysis (PCA) was employed to the energies of these sub-signals, whose outliers were eliminated using the Mahalanobis distance method. The reference coating thickness data were obtained via spectrophotometry during the coating process. The partial least squares (PLS) regression was incorporated with the PCA for the development of a model for prediction of the coating thickness. The PLS regression revealed that the pressure fluctuations can be used to evaluate the thickness with a good precision. This study demonstrated the applicability of pressure fluctuations for the prediction of the coating thickness. This method can be regarded as a new robust, fast and non-intrusive PAT approach for monitoring the coating process which can be easily used by engineers and practitioners.