Rotodynamic Pump Efficiency Variations Due to Changes of Operating Speed or Temperature

The same pump is often required to operate at speeds and temperatures different from those at which it was tested by the manufacturer. This can be and is frequently done with appropriate design measures taken to meet the most stringent operating conditions dictated by the highest operating speed and temperature. In other cases testing of the unit at its design speed (or temperature) may be impractical and too costly to he justified.

In all cases a prediction of the unit's performance at its full operating conditions is important, particularly for high-power, large or high-speed machines. Theoretical considerations and test evidence show that changes of performance do take place at varying speed and temperature. These changes are due mainly to the variation of the flow Reynolds number. However, a reliable method of quantifying these changes has been and still is a matter of controversy.

A large number of formulae has been published by different authors in an effort to establish a general rule of efficiency ‘scaling’ for all hydraulic machines. Initially all losses of the machine were considered; it was found later that certain types of losses (kinetic) should not be ‘scaled’. However, in all cases, it is still expected that the higher the increment of speed (or temperature) in relation to test, the higher the efficiency improvement for all hydraulic machines.

The present work suggests that each category of energy loss behaves in a different fashion at increased speed or temperature. Therefore, whereas some losses remain constant at all speeds, other losses vary at different rates with Reynolds number. For this reason it follows that, in certain circumstances, an ‘optimum’ speed ratio or temperature ratio may exist at which efficiency increment would become a maximum. Two expressions describing such an optimum speed ratio and temperature ratio in terms of the loss breakdown of the pump are developed and discussed. An analysis of the expressions shows that, in certain circumstances, which may easily occur in practice, a loss of efficiency rather than an increment can occur with increased speed (or temperature).

Moreover, since disc friction and volumetric losses exert a major influence on the magnitude and the position of the optimums, the individual unit breakdown of losses in test conditions is required to make a reliable performance prediction for speeds (or temperatures) different from test speeds (or temperatures).