Scale-up of biomass flue-gas condensation: Experimental validation from laboratory to industrial scale

Flue-gas condensation (FGC) recovers substantial latent heat in biomass district-heating plants, yet the performance of industrial tubular condensers predicted by laboratory-validated diffusion-layer models has not been verified against full-scale operating data. This work quantifies the laboratory-to-industry transition through a dual-scale campaign: 70 steady states on two laboratory vertical-tube condensers (tube inner diameter 20 and 27 mm, transitional flow) and 8 states on an industrial condenser with a 49.1 m 2 area downstream of a biomass boiler (turbulent flow), all firing wood pellets and chips. Both scales are evaluated on a strict condensation basis, enabling direct comparison. Confronting the experiment with a diffusion-layer model defines a dimensionless, unit-specific scale-up factor Cscale, the ratio of the predicted to the measured condensation coefficient. The model over-predicts industrial performance by 2.7–3.8 with flue-gas quenching and 4.3–6.4 without it, rising with Reynolds number (overall MAPE 334%). At laboratory scale the same model under-predicts (ratio 0.29–0.70); the prediction-to-experiment ratio thus rises monotonically with tube diameter; one possible explanation, proposed here as a hypothesis only, is a superheated vapour-rich core that bypasses the cold wall. Comparing the two laboratory geometries shows that the smaller tube raises the condensation Nusselt number by 1.5–1.7 at matched Reynolds number, quench level and cooling-water temperature. Quenched operation showed a 106% higher condensation coefficient than unquenched operation, consistent with the reduction of the dew-point superheat from 74 to 41 °C. For this single industrial unit and dataset, the laboratory-validated model's prediction needed dividing by roughly 3 (quenched) to 6.5 (unquenched); broader use as a design rule awaits independent validation.