• centrifugal turbopumps
• efficiency
• geometry
• impeller
• liquid-propellant rocket engine
• noncavitating performance
• preliminary design
• pressure side
• space
• splitter blades

In the space field, the seek for higher performance has been always the key of technological and social evolution. The conception of proven means of flight propulsion such as liquid-propellant rockets represents a remarkable outcome of this quest, since it raised humankind to the role of multiplanetary species. Nowadays, high power-density turbopumps represent the most weight-effective solution for liquid propellant feed systems of rocket engines for primary space propulsion. These applications are in fact characterized by the joint requirements for high total and specific impulses, where large amounts of usually cryogenic propellants are to be
injected at high pressures in the engine thrust chamber. Under these conditions turbopump-fed systems are preferable, because the mass reduction allowed by the use of low-pressure tanks far outbalances the additional weight of the machine. Since the design of a turbopump determines one of the major impacts on their final cost, these machines must ensure high reliability, low cost, light weight, stable flow for the required operating range, high efficiency, adequate suction performance and long life. Thus, its importance cannot be underestimated.

In the aim of further extending centrifugal impellers performance, splitter blades are sometimes used as a viable alternative way to increase the head of the pump with acceptable efficiency. Splitter blades are a kind of blades which do not extend the full length of the impeller and are generally placed between two full-length blades. The state of the art research on this topic shows that there is not clear-cut evidence as to what splitter blade position could be the most effective to get advantages. Turbopump positive effects such as head increase, flow separation reduction, suppression of pressure fluctuations at discharge, and a better behaviour
in the presence of cavitation are acknowledged benefits. Concerning efficiency, many
opposite effects can concur to influence it in a good or bad way, typically because of the greater blockage and losses provided by viscous effects.

The present study aims to investigate on the influence of splitter blades on space centrifugal turbopumps performance. This has been possible by suitably modifying a reduced order model developed in recent times by prof. d’Agostino and colleagues at ALTA S.p.A. in Pisa (Italy). The proposed model approach represents a useful and powerful tool capable of rapidly predicting the geometry and performance of centrifugal turbopumps radial impellers to provide indications for their preliminary design. Its application led to the design, optimisation, and experimental testing of the VAMPIRE turbopump at ALTA’s Cavitating Pump Rotordynamic Test Facility (CPRTF). The excellent agreement of theoretical and experimental results successfully validated the above model, therefore paving the way for the present
research study.

The new modelisation had to account for the discontinuity in the flow field introduced at the leading edge of splitter blades and for the additional blockage and viscous losses due to boundary layers developing on the splitter blades walls. Suitable assumptions have been introduced to model both the flow incidence angle and losses at the entrance of splitter blade channels. Moreover, the splitter blades shape has been designed to resume that of the VAMPIRE blades. This choice allowed to retain the mathematical description of the 2D vorticity correction slip flow field to consistently introduce the desired modeling assumptions mentioned above. Ultimately, the set of model input parameters has been completed by adding two nondimensionalised variables for the splitter blades definition, namely the adimensional axial position of the splitter blades leading edge and the azimuthal position relative to the suction and pressure sides of the main blades. The noncavitating performance of the VAMPIRE turbopump with splitter blades have been predicted by making vary these two parameters.

In general, both the final efficiency and total pressure head have been observed to increase above the values of the VAMPIRE turbopump without splitter blades. In particular, efficiency turned out to improve in the best case by +1%. This apparently modest increment is due to the already optimised geometry of the VAMPIRE, whose efficiency was already high (i.e., 83%). Nevertheless, configurations with lower performance have been identified as well. The use of a stall parameter meant to detect regions more or less prone to develop flow separation turned out to be a useful indicator for the smart positioning of the splitter blade inside the
impeller vanes. Actually, the observed gains in performance have occurred by moving the splitter blade closer to the pressure side of the adjacent full-length blade.