The Influence of the Reynolds Number and Roughness on the Efficiency of Axial and Centrifugal Fans – A Physically Based Scaling Method

C2 - Performance Measurements

Acceptance tests on large fans to prove the performance (efficiency and total pressure rise) to the customer are expensive. Hence there is a need to reliably predict the performance of fans not only at the design point but also at part- and overload from measurements on down-scaled test fans. The commonly used scale-up formulas give satisfactorily results only near the design point, where inertia losses are small in comparison with friction losses. At part- and overload the inertia losses are dominant and the scale-up formulas used so far fail. This work develops a new method, which aims to fill this gap. Furthermore, the new method considers the influence of surface roughness and geometric variations on the performance. The validation of the new scale-up formula is performed with test data from two axial fans with a diameter of 1000 mm / 250 mm and two centrifugal fans with 2240 mm / 896 mm diameter. Except Reynolds number, Mach number and relative roughness the two axial fans and the two centrifugal fans are similar to each other.

The method, discussed in this work, consists of two steps: Initially efficiency is scaled. Efficiency scaling is derived analytically from the definition of the total efficiency. With the total differential and one simplification it can be derived, that the change of friction factor is inversely proportional to the change of efficiency.

The second step allows for calculating the shift in flow rate, which can typically be observed in most test data. Prediction of this shift bases on a physical model concerning the displacement thickness of the boundary layer on the blades in a fan.

Experimental results concerning the efficiency of the four fans at different rotational speeds (i. e. Reynolds numbers) are shown. The new correction method is validated and comparad to other methods. The predicted performance characteristics show a good agreement to test data.