Testing on continuous production of mefenamic acids-Design of experiment through simulation and process optimisation.

Area of Science:

• Pharmaceutical Manufacturing
• Chemical Engineering
• Process Systems Engineering
• Machine Learning

Background:

• Continuous manufacturing offers advantages over batch production in pharmaceuticals, including flexibility and quality.
• Simultaneous optimization of multiple sub-processes in continuous manufacturing remains under-researched.
• Mefenamic acid production via wet milling (WM) and mixed-suspension mixed-product removal (MSMPR) serves as a case study.

Purpose of the Study:

• To explore and optimize continuous pharmaceutical production processes, specifically mefenamic acid synthesis.
• To apply data-driven evolutionary optimization algorithms to many-objective optimization problems (MaOPs).
• To develop a robust framework for pharmaceutical process optimization using integrated high-fidelity and surrogate models.

Main Methods:

• Utilized General Process Modelling System (gPROMS) for high-fidelity model-generated data.
• Developed Radial Basis Function Neural Network (RBFNN) based surrogate models for faster simulations.
• Employed evolutionary optimization algorithms for model-based process optimization of WM and MSMPR sub-processes.

Main Results:

• Demonstrated the viability of integrating high-fidelity and surrogate models for process optimization.
• Achieved approximated solutions that are, on average, 58% better than those from Latin hypercube sampling.
• Identified optimal solutions suitable for parameter setting in future experimental campaigns.

Conclusions:

• The many-objective optimization (MaOP) approach using surrogate models is effective for continuous pharmaceutical production.
• This study provides a novel framework for optimizing complex pharmaceutical manufacturing processes.
• Findings highlight the potential of machine learning in enhancing pharmaceutical production efficiency and quality.