Geometric similarity is assumed: production tank diameter T and impeller diameter D grow with the cube root of the volume ratio, T/D held at 3. Four classic scale-up rules pick which quantity stays constant as N (impeller speed) is re-derived for the new diameter — and because they don't agree, they demand four genuinely different production speeds:
Power/volume: P/V ∝ N³D² → const. P/V: N₂ = N₁(D₁/D₂)^(2/3)
Tip speed: v_tip = πND → const. tip speed: N₂ = N₁(D₁/D₂)
Gas transfer: k_La ∝ (P/V)^0.4·v_s^0.5, v_s ∝ D (fixed vvm) → const. k_La: N₂ = N₁(D₁/D₂)^1.083
Mixing time: θ_m ∝ T^(2/3)/(P/V)^(1/3) → const. θ_m: N₂ = N₁ (impeller speed unchanged)
The bar chart on the right computes N₂ under all four rules at once from the same bench point and target volume — the spread between the bars is the scale-up dilemma. Choosing "constant tip speed" (to protect shear-sensitive cells) under-mixes and starves oxygen transfer at large scale; choosing "constant kLa" (to protect oxygen supply) can push tip speed and local shear far above the bench value. There is no rule that keeps every quality attribute constant at once — this is exactly why a cGMP tech-transfer package needs a comparability protocol and a defined "quality range" for each critical quality attribute, not a single formula.
- Rule buttons — highlight which quantity the readouts on the left describe; the bar chart always shows all four rules simultaneously.
- Production volume — sets the target scale (100 L–20,000 L); the production tank's diameter follows T₂ = T₁·(V₂/V₁)^(1/3).
- Bench speed N₁ — the starting agitation rate at 10 L bench scale that everything else is scaled from.
- Impellers — the two small schematic vessels spin at N₁ (bench) and at the highlighted rule's N₂ (production), so you can see the speed mismatch, not just read it.