The residual axial thrust of a centrifugal pump is the vector resultant of the hydraulic components of impellers and sleeves, the momentum force and other imbalance forces and is bear by the thrust bearing. Among all the components, the hydraulic is by far the most important because it is typically one order of magnitude bigger than the others and the final residual axial thrust itself; but it is also the most difficult to calculate or estimate. This is mainly due a lot of uncertainties in the definition of the pressure distribution in the sidewall gap, dependent on the rotational speed, the leakages through the annular seals, the inlet swirl to the sidewall, the axial alignment of the impeller with the stator, the geometric tolerances and the pump operation.
In this paper an experimental approach to validate and calibrate the formulation to predict the hydraulic component of the impellers is presented. The typical formulation to evaluate the parabolic behavior of the pressure distribution is based on a series of coefficients, coming from literature, to consider all the above-mentioned influences. This formulation can be considered satisfying when dealing with pumps with back-to-back arrangement of the impellers on the rotor, since the hydraulic components are almost balanced. But with in-line configuration, all the hydraulic components of the impellers act against one direction and most of their force can be balanced by means of a balance drum, much more reliable than a balance disk. An experimental test campaign on a high energy diffuser pump was performed. The pump was equipped with load cells and temperature probes on the thrust bearing, and a special balancing line with a regulating bleed-off valve whose aim was to partialize the flowrate routed from the back side of the balancing drum to the suction. By throttling the bleed-off valve, it was possible to measure the hydraulic components of the impellers in very different operating conditions of the balancing drum, at different rotational speed and at different flowrates included zero residual axial thrust capability. As results of this campaign, once calibrated the coefficients on the full-scale pump, it is possible to calculate the residual axial thrust based on specific pressure measurements of the model test of a stage hydraulic; this also allows a thorough optimization of the thrust bearing selection.