JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Process Validation and Scale-Up

A process that works once might just be lucky. Validation proves it works reliably; scale-up proves it still works when the batch grows a thousandfold. Both are about replacing hope with evidence.

What validation actually proves

Process validation is documented evidence that a process, run within its defined parameters, reliably produces a product meeting all its CQAs — not once, but consistently. The modern view treats validation as a lifecycle in three stages, not a single one-time event.

  1. Stage 1 — Process design. Use development and QbD knowledge to define the process and its CPPs. This is where the design space and control strategy are written down.
  2. Stage 2 — Process qualification. Run full-scale batches under tight scrutiny to demonstrate the process consistently delivers. Extra sampling — for example, testing content uniformity across the whole blend — confirms it works everywhere, not just on average.
  3. Stage 3 — Continued verification. Keep monitoring routine production forever, watching trends in in-process controls and CQAs so any slow drift is caught early.

Why scale-up is genuinely hard

Scale-up takes a process that works in a 5 kg lab batch and makes it work in a 500 kg production batch. The trap is assuming you can simply multiply everything by a hundred. You cannot, because physics does not scale linearly. A bigger mixer's blades travel faster at the tip; a bigger dryer has less surface area per kilogram; heat takes longer to reach the centre of a larger mass.

Why "just multiply by 100" fails: blending tip speed
---------------------------------------------------
Tip speed = pi * diameter * rotations-per-second

Lab blender:    diameter 0.30 m, 30 rpm
  tip speed = 3.14 * 0.30 * (30/60) = 0.47 m/s

Production blender (same rpm), diameter 1.0 m:
  tip speed = 3.14 * 1.0 * (30/60)  = 1.57 m/s   (3.3x faster!)

Same rpm does NOT mean same mixing intensity.
To keep shear similar, scale by tip speed, not rpm:
  target rpm = 0.47 / (3.14 * 1.0) * 60 ~= 9 rpm

Lesson: hold the relevant PHYSICS constant (tip speed,
Froude number, drying air per kg), not the dial setting.
A worked reminder that holding the dial (rpm) constant changes the physics; you must scale the parameter that the CQA actually depends on.

This is where QbD pays off again. If you understood your CPPs and the design space at small scale, you scale up by holding the right physical quantity constant, not the dial. And with PAT you can confirm in real time that the big batch is behaving like the small one — turning scale-up from a leap of faith into a measured, evidence-backed transition.