Abstract

This paper presents the development of an additively-manufactured rocket turbopump by Relativity Space. For the past 80 years, centrifugal turbopumps have offered liquid rocket engines well-known benefits in performance and payload to orbit. Often regarded as the “heart of the engine,” rocket turbopumps require careful design attention, as they are typically pushed to the absolute limits of achievable power density when compared to land-based turbomachinery. Unfortunately, traditional methods of turbopump development are constrained by expensive fixed tooling, high part counts, and long-cycle times. Additive manufacturing removes many of these constraints, and allows development to occur at an accelerated pace, thus significantly reducing the time required to develop new engines. Combining modern turbomachinery design and analysis software with additive manufacturing know-how, Relativity was able to complete the development of a prototype in record time, and confidently use the data obtained to drive design upgrades in subsequent iterations. This study shows some aspects of the layout and testing of this innovative design.

Together with the high-level description of the layout and manufacturing of the turbopump, this study will show normalized results for pressure rise and efficiency from an initial test series of the design. The tests were conducted using liquid nitrogen as a stand-in for liquid oxygen, which is the fluid the pump will ultimately use in its final application. This data will be compared to systematic mean-line and CFD analysis. The mean-line analysis consists of a full-stage layout and demonstrates how parameters of the 1D model can to be adjusted to better match this class of turbomachinery design. The expected results for the ultimate fluid intended to be used in the pump (oxygen) are shown using the calibrated data in the mean-line model derived from the test data. CFD results will also be shown and compared to the test data. The CFD analysis was for the primary flow path of the stage and are generated using a full Navier-Stokes solver with a time invariant assumption.
The potential for additively-manufactured turbomachinery for future applications to space flight, as well as other areas where high performance is needed and constrained time scales are common, is discussed. The advent of these new fabrication methods poses special concerns for the designer and will require new methods to design and analyze them. Special accommodations in the modeling for techniques such as meanline, CFD, geometry parametrization, FEA analysis, and optimization that are required for these next generation machines is discussed.