Computational Chemical Engineering

Computational Chemical Engineering develops open, reusable software for modeling chemical reactors and the transport phenomena that govern them — from mass transfer and reaction in multiphase reactors to fluid flow, particle packing, and diffusion at the pore scale.

Our tools grow out of research and teaching in the Chemical Engineering & Chemistry department at Eindhoven University of Technology (TU/e), led by Frank Peters. His work centres on the modeling and simulation of (multiphase) reactors and transport phenomena — the same themes that shape everything you find here. Everything is released under permissive open-source licenses so that students, researchers, and engineers can learn from it, build on it, and contribute back.

Two active flagship projects anchor the organization today: pymrm for reactor-scale modeling, and Peclet for high-performance transport-phenomena simulation. Each ships with a companion project — a book and an examples gallery — so you can go from first principles to running code.

In development: OpenFlowsheet, an open process simulator for the flowsheet scale, built so that humans and coding agents work on the same explicit models and evidence.

What it is. pymrm is a Python package for building and solving multiphase reactor models. Born as a Matlab toolset for the Multiphase Reactor Modeling course at TU/e, it has been reborn as an open, Python-native library. The aim. Make reactor modeling approachable without sacrificing rigor — give students and practitioners composable building blocks for diffusion, convection, reaction, and mass transfer, assembled into sparse discretized systems that are easy to read, extend, and trust. Highlights 🧩 Building blocks for diffusion · convection · reaction · mass transfer 🧮 Sparse finite-volume construction with boundary-condition helpers 🎓 Tutorial- and example-driven, built for teaching and research 📦 One-line install: pip install pymrm Learn it — the pymrm-book. A companion Jupyter Book that walks through the theory and the package hands-on. 📖 Read it online: https://computational-chemical-engineering.github.io/pymrm-book See it work: the pymrm-gallery. Published reactor and transport models reproduced with pymrm, each a runnable notebook validated against the source and checked in CI. 📖 Browse it: https://computational-chemical-engineering.github.io/pymrm-gallery

The PyMRM logo, from the pymrm-book.

_{Rotating drum of cubic particles, computed with Peclet — from peclet-examples.}

What it is. Peclet is a suite of codes for simulating transport phenomena — Eulerian CFD (incompressible Navier–Stokes), Lagrangian DEM (particle packing), and mixed Voronoi methods — sharing one MPI block domain decomposition, signed-distance-field geometry, a common immersed-boundary methodology, GPU support, and Python bindings.

The aim. Provide a portable, high-performance foundation for pore- and particle-scale transport, where the same source runs multi-threaded on a laptop or across GPUs (CUDA / HIP) on a supercomputer. The name nods to the Péclet number — the ratio of advection to diffusion at the dimensionless heart of transport phenomena.

Highlights

• 🌊 CFD · 🧱 DEM · 🔷 Voronoi on a shared Kokkos + block-decomposition core
• ⚡ GPU (CUDA/HIP) and multi-rank MPI; multi-core wheels via pip install peclet
• 🧊 Cut-cell immersed boundaries over SDF-described solids
• 🐍 Unified peclet.* Python namespace

See it work — peclet-examples. A gallery of runnable, reproducible worked examples (channel MMS, cut-cell Poiseuille, packed-bed permeability, and more). 📖 Browse it: https://computational-chemical-engineering.github.io/peclet-examples

What it is. OpenFlowsheet is an open process simulator in which models, equations, numerical strategies, validity limits and evidence are explicit objects. Humans and coding agents work against the same scientific contracts, and each solve writes a bundle that records what was computed and how.

The aim. Connect the scales in this organization. A flowsheet can call a detailed model as one of its units, such as a pymrm reactor, or a surrogate trained on detailed runs, and say how far each answer can be trusted.

Where it stands (version 0.2.0).

• 🧪 Peng–Robinson property route and a reference ammonia-synthesis loop
• 🔗 A pymrm reactor model running as a unit inside a flowsheet solve
• 📈 Parametric sensitivities and studies; optimization with Ipopt or a trust-region method
• 🧾 Surrogate models admitted only with registered coverage evidence
• 🖥️ HTTP API with a diagnostic web shell; pip install openflowsheet

This is an early project. Its numerics are checked against independent reference values and agree with IDAES and DWSIM on shared cases. Version 0.2.0 is not validated against experiment. Distillation columns and liquid-mixture thermodynamics, with the first comparisons against measured phase-equilibrium data, are in development for v0.3. No part has had human numerical or process-modeling review. The repository records what has and has not been shown.

The Peclet suite is modular — peclet-core, peclet-flow, peclet-dem, peclet-voro, and peclet-morton each live in their own repository. Browse them all from the organization page.

• ⭐ Star the projects you find useful — it helps others discover them.
• 🐛 Open an issue to report bugs or request features.
• 🔧 Contribute — pull requests are welcome; see each repository's CONTRIBUTING guide.
• 🎓 Learn — start with the pymrm-book, the pymrm-gallery or the peclet-examples gallery.