Feed factor profile prediction model for two-component mixed powder in the twin-screw feeder

Area of Science:

• Pharmaceutical Manufacturing
• Chemical Engineering
• Materials Science

Background:

• Accurate control of powder flow rate is essential in continuous pharmaceutical manufacturing.
• The feed factor, representing powder amount per screw rotation, is a key parameter.
• Previous models primarily focused on single-component powders, limiting applications for mixtures.

Purpose of the Study:

• To develop predictive models for feed factor profiles (FFPs) of two-component pharmaceutical powder mixtures.
• To evaluate the impact of various input variables, including material properties, on prediction accuracy.
• To identify optimal modeling strategies for pharmaceutical powder feeding processes.

Main Methods:

• Experimental generation of four datasets with varying active pharmaceutical ingredient (API) mass fractions and powder weights.
• Development and comparison of multiple feed factor prediction models.
• Analysis of input variable significance, including API mass fraction, process parameters, and material properties.

Main Results:

• Models were developed to predict FFPs for two-component powder mixtures.
• The inclusion of material properties as input variables did not significantly enhance prediction accuracy.
• High-performance prediction models were achievable without laborious material property measurements.

Conclusions:

• Predictive models for FFPs of pharmaceutical powder mixtures can be effectively developed.
• Material properties are not essential for achieving high accuracy in FFP prediction models.
• Simplified models excluding material properties are preferable due to reduced measurement effort.