All of these projects are development work, developing small integrated process plants which are to be sold as products, which means that they have to meet Van Koolen's test of being as simple and robust as a washing machine.
The problems are however exactly the same as on larger scale plants, (though the lower skill level of the operators makes the design if anything more challenging):
Collaboration with clients and designers from other disciplines uses a lot more of my time than higher mathematics or leading edge science.
Geometry is often a lot more important than calculus.
Cost, safety and robustness are a lot more important than genuine novelty.
Novelty is always constrained by practical considerations.