Moisture soft sensor for agitated pan dryers using a hybrid modeling approach

Area of Science:

• Chemical Engineering
• Process Analytical Technology (PAT)
• Pharmaceutical Manufacturing

Background:

• Accurate real-time monitoring of moisture content in Active Pharmaceutical Ingredient (API) wet cake is crucial for agitated pan drying processes.
• Traditional methods for moisture determination can be time-consuming and may not provide real-time data for immediate process adjustments.
• Implementing Quality by Design (QbD) principles requires robust process understanding and control.

Purpose of the Study:

• To develop and validate a hybrid soft sensor model for real-time moisture content estimation in API wet cake during agitated pan drying.
• To assess the accuracy and reliability of the soft sensor using commercial-scale batch data.
• To investigate the influence of key process parameters on the soft sensor's estimation variability.

Main Methods:

• Development of a hybrid soft sensor model integrating process data.
• Calibration of the soft sensor model using data from 5 pilot batches.
• Validation of the soft sensor's performance with data from 21 commercial-scale batches.
• Application of dynamic global sensitivity analysis to identify significant input process parameters.

Main Results:

• The developed hybrid soft sensor model provides accurate real-time estimations of moisture content in the API wet cake.
• Validation across 21 commercial batches confirms the soft sensor's reliability and robustness.
• Sensitivity analysis identified critical process parameters influencing the soft sensor's estimations throughout the drying cycle.

Conclusions:

• The hybrid soft sensor is a powerful tool for real-time moisture content estimation in agitated pan drying.
• This technology enables enhanced monitoring and control of the API drying process.
• The soft sensor facilitates the implementation of a Quality by Design (QbD) approach, improving process consistency and product quality.