Turboexpander constitutes one of the vital components of Claude cycle based helium refrigerators and liquefiers that are gaining increasing technological importance. These turboexpanders which are of radial inflow in configuration are generally high-speed micro turbines, due to the low molecular weight and density of helium. Any improvement in efficiency of these machines requires a detailed understanding of the flow field. Computational Fluid Dynamics analysis (CFD) has emerged as a necessary tool for the determination of the flow fields in cryogenic turboexpanders, which is often not possible through experiments. In the present work three-dimensional transient flow analysis of a cryogenic turboexpander for helium refrigeration and liquefaction cycles were performed using Ansys CFX®, to understand the flow field of a high-speed helium turboexpander, which in turn will help in taking appropriate decisions regarding modifications of established design methodology for improved efficiency of these machines. The turboexpander is designed based on Balje's nsds diagram and the inverse design blade profile generation formalism prescribed by Hasselgruber and Balje. The analyses include the study of several losses, their origins, the increase in entropy due to these losses, quantification of losses and the effects of various geometrical parameters on these losses. Through the flow field analysis it was observed that in the nozzle, flow separation at the nozzle blade suction side and trailing edge vortices resulted in loss generation, which calls for better nozzle blade profile. The turbine wheel flow field analysis revealed that the significant geometrical parameters of the turbine wheel blade like blade inlet angle, blade profile, tip clearance height and trailing edge thickness need to be optimised for improved performance of the turboexpander. The detailed flow field analysis in this paper can be used to improve the mean line design methodology for turboexpanders used in helium refrigeration and liquefaction cycles. © 2017 Elsevier Ltd