Modular plants built from Process Equipment Assemblies (PEAs) according to VDI 2776 and VDI/VDE/NAMUR 2658 can improve the time-to-market, flexibility and efficiency in the process industry to adapt to faster moving markets and decreasing product batch sizes. Therefore, this plant concept is interesting for the production of products that often fall under GMP-conditions. Especially for this product class, but also in general, the assurance of product quality through Process Validation is an important goal. However, in the literature on modular process development and modular plant engineering, Process Validation has hardly been considered up to this point.
Therefore, this research work investigates the question of how Process Validation in modular plants can be supported and accelerated with the help of digital twins. The focus is on the phases of “Process Design” and “Process Performance Qualification” as a sub-step of process qualification in Process Validation. For this purpose, aspects of process development, Process Validation, the Quality by Design paradigm, modular plants, and digital twins are jointly discussed. Starting from the initial situation for Process Validation in modular plants defined within the framework of a reference model, the introduction of digital twins, model-based design of experiments, and Quality by Design in modular plants are defined as support and examined in more detail in three prescriptive studies. The main result is the Smart PEA (sPEA) concept, which extends the concept of the intelligent PEA by introducing a digital twin in the form of a capability model consisting of semantically linked descriptive, operational and behavioral models. The introduction of this capability model is worthwhile, since considerable portions in particular of the behavioral models are reusable in multi-product plants and thus efforts in modeling and experiments for the description of the design space can be reduced in the context of Quality by Design.
As an overarching research method, this work uses the Design Research Methodology (DRM) according to Blessing and Chakrabarti (2009). Within the studies, design with requirements analysis, concept development and evaluation, as well as case studies based on simulation experiments with simple reaction examples utilizing dosing and reactor PEAs.