An empirical or theoretical relationship that predicts how key performance indicators (flow, heat and mass transfer rates, mixing time, conversion, yield, stresses, etc.) change when a process or device is enlarged or multiplied from laboratory/pilot scale to commercial scale, typically expressed by preserving geometrical similarity or relevant dimensionless groups.

Principle

Performance does not scale linearly with size because transport phenomena, mixing, and boundary conditions change with geometry and operating parameters; therefore scale‑up must preserve the dimensionless criteria (e.g., Reynolds, Froude, Peclet, Damköhler, etc.) or explicitly correct for departures from similarity to maintain the same dominant physics across scales.

Demonstration

Illustrative scenario — Stirred tank mixing: A laboratory reactor and a production reactor are geometrically similar. If power per unit volume is kept constant, mixing time, shear and heat removal scale differently than if tip speed is held constant (Recognition). Selecting constant power per volume preserves turbulent energy dissipation relevant to mixing‑limited reactions, while constant tip speed preserves stress on shear‑sensitive materials (Action). The chosen correlation determines whether lab performance is reproduced at scale (Consequence).

Misapplication

Assuming geometric similarity alone guarantees identical process performance without checking which dimensionless groups control the phenomenon; the semantic error is conflating geometric similarity with dynamic similarity and neglecting dominant transport or kinetic dimensionless numbers.

Consequence

Incorrect scale‑up correlations can produce poor product quality, inadequate heat removal, unexpected flow regimes, excessive shear, unanticipated fouling, or unsafe conditions; they increase time and cost in commissioning and may require retrofits or process redesign.

Reversal

For some processes dominated by linear phenomena (e.g., purely conductive heat transfer in simple geometries with scale‑independent boundary conditions) or when employing modular replication of identical units, simple proportional scaling or unit replication can be valid; conversely, multiphase, turbulent or reaction‑limited systems frequently require detailed similarity analysis or pilot testing.

Boundary

Clearly within: scaling stirred reactors, heat exchangers, absorbers where transport and mixing dominate. Boundary case: microreactors or processes with surface‑to‑volume effects where scaling introduces new dominant physics. Clearly outside: phenomena governed by molecular‑scale physics (quantum, single‑molecule effects) or systems where upscaling is replaced by parallelization of unchanged small units.

Semantic Tension

Empirical simplicity versus mechanistic fidelity: empirical correlations are fast and based on data but often limited to tested conditions; mechanistic dimensionless‑group approaches are more general but require deeper understanding and may demand data for closure models—practitioners must balance risk, cost and available knowledge.

Synthesis

Effective scale‑up begins with dominant‑phenomenon analysis to identify controlling dimensionless groups, followed by selection of scaling rules (preserve group(s) or use modular replication) and validated pilot studies or targeted experiments to quantify departures from similarity and provide correction factors.