Hybrid process modelling
THE STATUS QUO
Critical process insights often remain in the head of a few individuals.
Relationships between operating conditions and final product properties are complex and often difficult to interpret.
As a result, process design remains largely empirical, relying on intuition, experience, and trial and error.
This is the challenge we address.
Our approach
We combine first-principles models of particle formation with data-driven methods. Physics lowers data requirements while improving interpretability and extrapolation; the data-driven part keeps model complexity manageable and captures unknown effects.
Hybrid modeling
Hybrid models work on the small, imperfect datasets real production generates — not the mountains of clean data that pure machine learning demands.
Full distributions, not averages
We model full size, shape and composition distribution, because that — not a mean value — is what sets product properties.
Done for you
From particle technology to mathematical optimization, we provide more than 30 years of expertise so your team can focus on developing products — not modeling them.
Core OFFERS
Three ways we help companies in particle design.
How an engagement works
01
Initial assessment
We evaluate if we can help and define the goals you want to achieve.
02
Data
We start from the runs, parameters and measurements you want to share.
03
Hybrid model
We establish a hybrid model for your specific synthesis, validated against your own data.
04
Insights
You gain clear process understanding and the parameters that put you on spec.
Track record
Our methods come out of seven years of research in the Collaborative Research Centre CRC 1411 at FAU Erlangen-Nürnberg, on some of the most demanding particle systems in existence: quantum dots, where colour is tuned by size down to the nanometer, and shape-controlled plasmonic metals.
Founders
Get in touch
One conversation is usually enough to see whether we can help. Let's explore what's possible.
team@prtcl.eu