The effect of number of nozzle vanes in the turbine stage of a turbocharger is studied using computational fluid dynamics. The nozzle vane unit having 8, 9 and 10 numbers of nozzle vanes configuration is proposed for the radial flow turbine with 30 mm wheel tip diameter. At maximum opening position of the nozzle vanes and for the typical turbine expansion ratio of 2.5, the reduction in mass flow parameter with 10 numbers of nozzle vanes is about 1% lower compared to the 8 numbers of nozzle vanes. The maximum turbine flow range is not affected with higher number of nozzle vanes. The improvement in flow guidance is observed in nozzle vane unit having 10 numbers of nozzle vanes. The improvement in pressure distribution is observed in both the nozzle vane and turbine wheel with increase in number of nozzle vanes. The entropy generation in a turbine stage is found to decrease with increase in the number of nozzle vanes.

Turbocharger has a paramount influence on the performance of an internal combustion. Improved emission requirements have led to complex after treatment systems, which add pressure drop to the air management system. One of the ways to mitigate negative effects of pressure drop is to improve turbocharger efficiency. The scope of performance improvement for a typical turbocharger majorly lies on the modification of compressor wheel, turbine wheel, volutes etc. The major challenges in compressor wheel modification include setting the right major geometrical dimensions, considering compressor operability at different application requirements, design cycle time and the cost of computation. Present study is about evolving an effective optimization methodology, which comprises of parametrization of compressor stage at preliminary design stage and optimization of the chosen parameters through coupling one dimensional flow analysis tool with a robust optimization tool. The parameters were chosen based on their influence on overall efficiency and pressure ratio at different mass flows and varying engine rotational speeds. Surrogate models have been used to choose the optimal designs from the preliminary design space as per requirement and optimized designs were analyzed further for verification. Final validation has been carried out using a 3D RANS code.