Efficient Optimization of Process Strategies with Model-Assisted Design of Experiments

Area of Science:

• Biotechnology
• Process Engineering
• Statistical Modeling

Background:

• Conventional design of experiments (DoE) demands extensive expertise and often results in time-consuming, expensive trials.
• Limited mechanistic understanding can hinder process optimization in biological and chemical manufacturing.

Purpose of the Study:

• To introduce and detail the workflow of model-assisted design of experiments (mDoE) for efficient process development.
• To demonstrate how mDoE enhances mechanistic understanding and process optimization.
• To reduce experimental burden in process development.

Main Methods:

• Adaptation of a mathematical process model using initial experimental data or prior knowledge.
• Integration of model-assisted simulations as experimental responses within statistical DoEs.
• Evaluation of DoEs using simulated data for constrained-based optimization of the experimental space.

Main Results:

• mDoE facilitates a deeper mechanistic understanding of investigated processes.
• Simulations derived from mathematical models can effectively substitute for experimental data in DoE evaluations.
• Iterative application of the mDoE loop significantly minimizes the number of required experiments.

Conclusions:

• Model-assisted design of experiments (mDoE) offers a powerful strategy to streamline process development.
• mDoE enhances both process understanding and optimization efficiency, leading to reduced experimental costs and time.
• This integrated approach provides a robust framework for optimizing complex biological and chemical processes.