Predicting the Particle Size Distribution in Twin Screw Granulation through Acoustic Emissions

Abstract

A non-destructive process analytical technology for monitoring the complex particle size distributions inherent to twin-screw granulation (TSG) was presented, based on ultrasonic acoustic emissions (AE). AE spectra were collected by discrete signal acquisition during the continuous impacts of granules on an inclined plate positioned below the exit of an extruder. The paper outlines the setup considerations associated with the impact plate, based on an examination of its location, thickness (0.7, 1.0, 1.5 mm) and angle of inclination (10-60°) and the resulting particle behavior at the plate, as determined by high-speed image analysis and AE monitoring. Subsequently, AE spectra were collected during the wet granulation of lactose monohydrate at different liquid-to-solid ratios from 8-14% and correlated with the particle size distributions (PSD) to train a neural network model. Predicted PSD for particle sizes from 400 to 7000 m based on the AE spectra of validation trials showed the largest root mean squared error (RMSE) of 4.25 wt% at 2230 μm. After transforming the AE data with a newly created digital filter based on particle impact mechanics to address auditory masking, the error for predicting fractions of each particle size was significantly reduced to below 1 wt%. The technology shows great promise as a monitoring method for TSG, being capable of predicting its complex size distributions in real time.