Validation of fluid bed granulation utilizing artificial neural network.

Area of Science:

• Pharmaceutical Engineering
• Chemical Engineering
• Materials Science

Background:

• Fluid bed granulation is a critical process in pharmaceutical manufacturing.
• Conventional fluid bed granulators face challenges like nozzle clogging and spray loss.
• Improving granulation efficiency and granule properties is essential for drug formulation.

Purpose of the Study:

• To introduce and validate innovative components for fluid bed granulation.
• To assess the performance of a modified fluid bed granulator.
• To confirm the reliability of generalized regression neural network (GRNN) for process validation.

Main Methods:

• Replacement of conventional components with an annular gap spray system, booster bottom, and outlet filter.
• Granulation of sucrose under varied operating conditions (temperature, spray rate, air pressure, velocity, binder properties, batch size).
• Validation using generalized regression neural network (GRNN) to predict granule properties.

Main Results:

• The modified fluid bed granulator demonstrated improved performance.
• Granule properties (size, distribution, flow, angle of repose, bulk/tapped volumes) were measured.
• GRNN accurately predicted granule properties, showing good correlation with experimental data.

Conclusions:

• The developed innovative components effectively enhance fluid bed technology.
• Generalized regression neural network (GRNN) is a reliable method for validating modified fluid bed apparatus.
• The study provides a validated approach for optimizing granulation processes.