A KPI-based experimental design strategy for bioprocess development

Abstract

Bioprocess development can benefit significantly from the use of mathematical models for prediction and optimization, yet the uncertainty in these models can hinder reliable early-stage decision-making for industrial-scale processes. This study introduces a telescopic model-based design of experiments approach that directly targets the reduction of uncertainty in key performance indicators (KPIs) at the optimum process conditions rather than focusing solely on model parameter precision. Using a sugarcane-to-ethanol biorefinery use case, the proposed approach is benchmarked against a traditional parameter-focused approach. Results demonstrate that the proposed strategy reduces KPI uncertainty more efficiently, identifies economically favorable process conditions faster, and prioritizes the estimation of parameters most influential on the KPI.