(679f) Digitalise Process Assets for Improving Productivity: Enhance Process Insights, Metric and Parameters Via Simulations

Enhancing resource efficiency and decarbonisation are global priorities of manufacturing sector. There is a significant excitement about digitalisation, Industry 4.0 and their impact on manufacturing in process industries. It has a potential to become the most transformative era for chemical manufacturing (Industry 4.0: Top Challenges for Chemical Manufacturing, Elsevier White Paper, 2017). Digitalisation of process assets with the help of high-fidelity computational models is essential for realising this potential. However, development of computational models is only the first step towards it. It is essential to develop appropriate tools which will allow systematic digitalisation of process assets, derive new insights and synergistically evolve process understanding, key process metric and optimal process parameters for enhancing productivity. In this work, we present our approach towards developing such a tool for enabling enhanced productivity via digitisation of process assets.

Our approach uses science-based fundamental understanding as the focal point of process asset digitalization. Process understanding is most effectively represented using quantitative relationships between process parameters or PPs (such as impeller speed, liquid volume, solids loading etc.) and “Process Metrics” (PM). A PM is a (typically dimensionless) physical variable that describes some aspect of the physical process internal to a unit operation equipment [1]. For instance, in an agitated vessel bioreactor, the shear rates (and mass transfer coefficients) near the impeller is a PM - related to oxygen transfer. Similarly, the “dimensionless spray flux” in a high-shear wet granulator is a PM that relates to the effectiveness of the liquid binder spray with respect to the granulation process [2]. The aforementioned PP-PM relationships are the most appropriate means to digitally represent an asset. When a process asset has well defined PP-PM relationships, it is said to be “fully characterized” from a process performance perspective. For such digital characterization of performance assets, it is necessary to develop high-fidelity computational models. However, this is not sufficient. It is important to develop appropriate framework which goes beyond mere computational modeling and allows development of quantitative relationships between PP and PM and facilitates gaining new insights into process. In this talk, we present such a framework, database-driven web-based software framework - called “SimSight” which enables process insights via digital characterization of assets.

The web-based software framework, “SimSight” is generic in that it can be used in conjunction with any computational model. Here, we demonstrate this digitalization framework using computational fluid dynamics (CFD) models for agitated vessels - a widely employed class of process equipment. Other computational models such as discrete element method (DEM), Lattice Boltzmann method (LBM), chemical reaction engineering models etc. can be easily ‘connected’ to this framework.

SimSight consists of an asset database at its core - each asset being an object in the database. Attributes for each asset
include geometry information and simulation data/results. The simulation data/results are obtained via a “full computational characterization” of the asset. Such a characterization is enabled by running a set of simulations with process parameter settings dictated by a statistical DoE matrix. Using a set of calculation utilities provided along with this database of fully characterized process assets/equipment, SimSight empowers process engineers to create/share fundamental process knowledge in a more streamlined way.

With its easy-to-use features, SimSight enables science-driven digitalization of process equipment and reduces time-to-scale for process development and transfer activities. We highlight key learnings of our development process. The application of SimSight is illustrated by a few case studies to bring out the inherent power and potential of such digitalization of process assets in improving productivity.

References

1. ICH Harmonized Tripartite Guideline, Pharmaceutical Development Q8 (R2)
2. Litster, J.D., K. P. Hapgood, J. N. Michaels, A. Sims, M. Roberts, S.K. Kamineni, T. Hsu, “Liquid Distribution in Wet Granulation: Dimensionless Spray Flux”, Powder Technology 114 (1-3), pp 32-29