Real-time design space description in pharmaceutical manufacturing

In this paper, we present a methodology to assist the offline description and online maintenance of the design space of a pharmaceutical product. The methodology is intended as a proactive risk-management tool to be included in a design space submission. First, classical feasibility analysis is exploited to obtain an offline model-based description of the design space (e.g., at the process development stage). Then, a dynamic state estimator is used to continuously adapt the model during plant operation, using measurements available from plant sensors. Surrogate-based feasibility analysis is used to continuously update the model-based description of the design space based on the up-to-date model returned by the state estimator. The tracking effectiveness of the methodology is shown for a co-milling process of a pharmaceutical production.