When scaling up a mammalian cell culture process, engineers must simultaneously manage mixing time, oxygen transfer, and carbon dioxide removal. After scale-up, an engineering batch is run at manufacturing scale specifically to demonstrate that the process transfers cleanly to the facility and by then, it is expensive to discover a problem. At Graphtal, we believe Computational Fluid Dynamics (CFD) is the tool that turns bioreactor scale-up from an educated guess into a data-driven decision.
A single-number scale-up target says nothing about what's happening inside the vessel. A CFD contour plot shows it directly: cool, well-mixed zones near the top, and hot, high-shear or high-energy-dissipation zones concentrated right around the impeller blades and sparger exactly where cells spend time and take damage.
Why Scale-Up Is Hard: Finding the Design Space
The conventional basis for scale-up leans on holding certain parameters as close to constant as possible between scales. But these targets pull in different directions push aeration too high and you risk foam formation and impeller flooding; push power too high and you risk shear damage. Too little of either causes poor mixing, oxygen limitation, or CO₂ accumulation. Somewhere in the middle sits a narrow design space, and every scale-up strategy is really an exercise in finding it.
Flooding of impeller
CO₂ accumulation
The Trouble with Traditional Scale-Up Criteria
Each classic scale-up rule has a blind spot. Choosing one in isolation means accepting the risk that its blind spot creates at the new scale.
| Scale-Up Criteria | Process Parameter | Benefit | Limitation |
|---|---|---|---|
| Tip Speed | Stirring | Simple, linear scale-up | Considers only the impeller, not overall bioreactor design |
| Power per Unit Volume (P/V) | Stirring | Accounts for bioreactor design, impeller design, and fluid properties | Does not account for gassing |
| kLa | Stirring & Gassing | Holistic approach ties stirring and gassing together | Still doesn't reveal dead zones, local shear rate, or energy dissipation hotspots, CO₂ stripping, or air/oxygen ratio effects |
The Hidden Risks in a "Well-Mixed" Reactor
A bioreactor can look fine on paper acceptable global shear, acceptable average mixing time while still hiding dangerous local conditions. None of these six issues show up in a single scale-up number like tip speed or average P/V.
| Limitation | Description |
|---|---|
| Dead Zones | Cause hypoxia (low oxygen) and increased mixing time, reducing cell viability and process efficiency. |
| Shear Rate | High local shear rates near impeller/sparger, despite acceptable global shear rates, can severely damage cells, reducing viability and productivity. |
| Energy Dissipation | High local energy dissipation rates near the impeller and sparger can cause excessive mechanical stress, damaging cells and reducing viability and productivity. |
| CO₂ Accumulation/Stripping | Accumulation increases base consumption, causing osmolality spikes that stress cells, reducing productivity and product quality. |
| Air/O₂ Ratio | Critical for bioreactor scale-up ensures effective CO₂ stripping to prevent accumulation while maintaining optimal oxygen levels for cell growth. |
| Sparger Selection | Key for optimal kLa and CO₂ stripping, ensuring adequate oxygen supply and CO₂ removal for cell growth and bioreactor efficiency. |
Why averages aren't enough: Global shear and average mixing time can look completely acceptable while local hotspots near the impeller or sparger quietly damage cells throughout a run. These hotspots are invisible to any single-number scale-up criterion they only become visible once you can see the full spatial profile inside the vessel.
A bioreactor can pass every average check on paper and still damage cells in the 5% of the vessel volume no one is looking at.
Where CFD Comes In
This is exactly the gap CFD fills. Rather than relying on bulk averages, CFD lets you visualize and quantify what is actually happening at every point in the vessel and tune the parameters that control it.
| Monitored Parameter | CFD Role | Advantage | Tuning Parameter |
|---|---|---|---|
| Shear & Energy Dissipation | Identifies high shear zones and turbulence hotspots | Avoids lethal transient shear exposure | Stirring, aspect ratio |
| O₂ Transfer | kLa distribution, gas holdup, bubble residence, O₂-starved zone profile | Proper OTR in the bioreactor; avoids hypoxia | Stirring, gassing, air/oxygen ratio |
| CO₂ Removal | CO₂ accumulation profile | Avoids over-consumption of base and osmolality spikes | Stirring, gassing, air/oxygen ratio |
| Mixing Time | Full mixing time profile | pH/osmolality spikes don't hit cells | Stirring |
| Flow Patterns & Dead Zones | Circulation loop profile | Avoids dead zones | Stirring, reactor volume |
| Probe Siting | Eddies profile | Avoids probe placement in recirculating eddies | Depends on bioreactor design |
| Sparger Selection | Mass transfer profile, CO₂ stripping profile | Avoids hypoxia and CO₂ accumulation; delivers the right kLa | Micro, drilled-hole sparger |
A Concrete, Low-Risk Workflow Bench to 1000L
In practice, we recommend a structured six-step workflow that turns scale-up from a single leap of faith into a series of checkpoints, each backed by simulation data rather than assumption.
Characterize the Small Scale
Establish baseline P/V, mixing time, kLa, shear rate, flow pattern, energy dissipation rate (EDR), and CO₂ profile at bench scale.
Pick Scale-Up Targets
Define acceptable bands not single numbers for average P/V, kLa, mixing time, shear rate, and EDR.
Build the At-Scale CFD Baseline
Optimize stirring speed and gassing at the larger scale to match target kLa and mixing time.
Run a Hotspot Audit
Examine shear rate and EDR distributions, and calculate exposure frequency × duration for fluid elements passing through high-stress zones.
Optimize Sparger Strategy
Split gas flow across ring, micro-sparger, and headspace overlay to balance shear and CO₂ stripping.
Audit Dead Zones
Eliminate poorly mixed regions without inadvertently spiking energy dissipation rate elsewhere.
Powered by Partnership: Graphtal + SimVantage
To deliver this at speed, Graphtal has partnered with SimVantage, a cloud-based CFD bioreactor simulation platform, to accelerate bioprocess scale-up through CFD simulation.
DoE, MVDA, SPC, hypothesis testing
Kinetic + data + CFD digital twins
CMC, CDMO, cell therapy, mAb
Explainable AI for bioprocess
See mixing, shear & mass transfer inside the vessel
Results same day, no hardware needed
Reduces test runs from ~5–6 down to 1–2
Optimize conditions before the first batch
Combined outcome: Fewer experiments · faster scale-up · data-proven CFD decisions.
Cloud-based access to a state-of-the-art simulation tool, secured by the Kaleidosim Framework built to simulate your reactors and answer your design questions.
Not Every Parameter Scales the Same Way
Bioreactor scale-up parameters fall into three categories, and confusing them is a common source of failed scale-up strategies.
Batch medium, inoculum volume, feed volume, supplement, filter area. These scale volumetrically straightforward to carry across scales.
Seed density, pH, DO, temperature, feed condition (%). These should hit a similar set point at both scales, regardless of vessel size.
Non-linear parameters impeller agitation and oxygen/air aeration are the hard part. Tip speed, mixing time, kLa, and vvm airflow cannot all be held constant simultaneously. The comparable target is usually a "happy medium" between tip speed, mixing time, kLa, and air flow (vvm), not a perfect match on any single one.
Scaling up process parameters is genuinely tricky: every scale-up parameter is dependent on another, no single scale-up process is universally correct, and it falls to the process engineer to determine which parameter is most critical and find a workable middle ground among the rest. CFD doesn't remove that judgment call it replaces guesswork with visibility.
How Graphtal Supports Your Scale-Up Programme
Our scientists combine rare ability in cell culture and computational fluid dynamics, deeply understanding bioprocesses and their limitations with a dedicated expert assigned to your project for seamless collaboration.
Tell Us About Your Project
Our team of experienced professionals at Graphtal is ready to transform your project from idea to reality, ensuring alignment with your organisation goals through advanced data analytics and predictive modelling.
Let's simplify your work