(39a) Data-Driven, Model-Based Approaches to the Scale-up, Transfer and Alignment of Processes

A very powerful approach to the scale-up of processes based on models built from data collected on processes at the different scales is presented and illustrated with several industrial examples. The technology is also applicable to the transfer of production between manufacturing sites, and to the alignment of existing plants producing similar slates of products. If the scale-up, transfer or alignment involves processes that are almost identical in terms of design and sensors, the technology is quite straightforward. However, the technology is most powerful when the scale-up, transfer or alignment is between processes of different design with different sensors and actuators and yet requires all the processes to manufacture a similar slate of products or modifications of them.

Although the models are built using data from the processes at the different scales using empirical multivariate latent variable methods, theoretical calculations involving engineering calculations on mixing characteristics, energy balances, etc. are readily included.

Examples given will include: (i) the scale-up of a roller compaction process from a laboratory compactor to a full-scale compactor in a pharmaceutical company; (ii) the scale-up of a batch pulp digester process; (iii) the transfer of production between batch reactors in a specialty chemical process; and (iv) the alignment of batch reactors in a biopharmaceutical plant to achieve similar yields and quality.