Optimization of drying-end-points measurements for the automation of a fluidized-bed dryer using FT-NIR sprectroscopy

Abstract

Automatic control of a fluidized-bed drying process depends on the availability of an in-line sensor to provide accurate measurements of product moisture content. Nearinfrared (NIR) spectroscopy technology provides a potentially non-invasive and nondestructive analytical method that could serve as a sensor option for a wide range of applications. The purpose of this research was to investigate the use of NIR spectroscopy for accurate in-line moisture measurements during fluidized-bed drying process and to integrate the NIR set-up as part of drying automation. Five powder mixtures consisting of lactose anhydrous, lactose monohydrate, povidone, blue color additive, and distilled water were dried in a bench-scale fluidized-bed dryer (FBD). Samples were withdrawn from the FBD for the calibration phase. A NIR moisture calibration and validation using partial least squares (PLS) was developed by analyzing statically these samples in conjunction with Karl Fisher Titration. Three probe axial positions were designed and installed in the FBD to take in-line NIR measurements. Due to fluidization effects (segregation and sample density distribution along the bed), a mixed-level factorial experimental design was performed to determine the significance of factors such as mass load, air flow and fiber optic probe axial position in the NIR prediction. The response variable to be analyzed was the residual between in-line measurements and static samples taken immediately after. Data analysis indicated that all factors were significant with residuals ranging from 0.04–2.32. A mathematical correlation was determined to predict future residuals as a function of the operating conditions.