High-fidelity CFD
Days per scenario, and a specialist to run it.
Fermentation scale-up
Physics-informed AI, purpose-built for fermentation
Know how your process behaves at pilot and production scale before you pay for the pilot run. A scale-up answer for process engineers, without commissioning a modelling project.
BENCH2 L · characterised
PILOT50–1,000 L · unknown
PRODUCTION2,000 L +
The problem
Every failed scale-up iteration costs €50–250k and 3–9 months
A process that works at bench scale is well characterised. At 50–1,000 L, mixing and mass transfer change, and the batch is lost in ways the bench never showed.
€50–250kDirect cost per failed iteration
3–9 moDelay to commercialisation per iteration
€100/gCost of shortcuts at the 2,000 L scale
Where the batch is lost
- Poor mixing
- Substrate gradients
- Feeding-strategy misalignment
- Oxygen limitation
Three options exist today. Each fails on cost, access or realism.
Generic process simulators
Assume ideal mixing, so they overestimate productivity and mislead the business case.
Outsourced pilot runs
The very expenditure the decision is meant to de-risk, committed before the answer exists.
The solution
From reactor geometry to operating envelope in seconds to hours
FluidTwin couples hydrodynamics with your organism’s biokinetics, so you can test agitation, aeration and feeding strategies at scale before committing to them.
CFD characterisation
A geometry and impeller fingerprint, not a new study for every scenario.
Physics-informed surrogate
Hydrodynamics coupled natively to biokinetics, including transient feeding ramps.
Operating envelope
Agitation, aeration and feeding, with dead zones and gradients flagged.
1/700
of CFD runtime for an equivalent mixing evaluation, demonstrated.
Fed-batch
Dynamic, full-timespan simulation, where competing tools are mostly steady-state only.
Any vessel
Vendor-agnostic: works on any characterised fermenter.
Uncertainty is shown, never hidden. Every prediction comes with its uncertainty bounds, so you know how far to trust it before you act on it.
Interactive demo
See what changes when you scale up
Take the same process from a 2 L bench reactor to a 20 m³ production vessel, keeping the power per volume constant, and watch how mixing, substrate gradients and oxygen supply shift.
Concentration relative to vessel mean
0×1×2×≥3×
Elapsed 0.0 s · mixing…
Playback at 3× real time.
Mixing time (95 %)
Substrate gradient
Oxygen transfer / uptake
Impeller
Illustrative model, not FluidTwin output. Values come from textbook correlations for a geometrically similar stirred tank (Grenville and Cooke et al. for mixing time, van ’t Riet for kLa, Nienow for flooding) and a simple axial-dispersion picture of the feed gradient. Correlations like these ignore your vessel’s actual geometry, which is exactly the gap FluidTwin closes with CFD-derived predictions.
Who it’s for
Teams entering their first 10–1,000 L campaigns
Precision-fermentation startups, midcaps and CDMOs in the EU and UK that have a process working at bench scale and need to get it through pilot without burning a round on failed batches.
Practitioners tell us the barrier isn’t workflow integration. It’s the lack of an affordable, vendor-agnostic way to predict scale-up outcomes before committing to a pilot run.
150+EU precision-fermentation startups
300+SMEs in enzymes, biomaterials and natural ingredients
“We often observed they got stuck a second time, i.e. when going to industrial manufacturing scale.”
“If these tools can be improved further, then the industrial bioprocess practitioner would have more rational ways to scale-up and improve industrial process than are currently available.”
Team
Built at the Department of Biotechnology, TU Delft
FluidTwin grows out of research on bioreactor hydrodynamics and scale-up modelling at TU Delft.
Héctor Maldonado de LeónDepartment of Biotechnology, TU Delft
Dr. Cees HaringaDepartment of Biotechnology, TU Delft
Pilot programme
Scaling a fermentation process in the next 12 months?
We’re talking to precision-fermentation teams and CDMOs about how they make scale-up decisions today. Thirty minutes, no pitch. Early partners help shape what we build first.
Delft, the Netherlands