Search Results

Radial flow Variable Nozzle Turbine (VNT) enables better matching between the turbocharger and engine. At partial loading or low-end engine operating points, the nozzle vane opening of the VNT is decreased to achieve higher turbine efficiency and transient response, which is a benefit for engine fuel consumption and emission. However, under certain small nozzle opening conditions (such as nozzle brake and low-end operating points), strong shock waves and strong nozzle clearance flow are generated. Consequently, strong rotor-stator interaction between turbine nozzle and impeller is the key factor of the impeller high cycle fatigue and failure. In present paper, flow visualization experiment is carried out on a linear turbine nozzle. The turbine nozzle is designed to have single-sided clearance, and the Schlieren visualization method is used to describe the formation and development process of clearance flow and shock wave under different clearance and expansion ratio configurations.

Variable nozzle turbine (VNT) technology has become a popular technology for diesel engine application. To pivot the nozzle vane and adjust the turbine operating condition, nozzle clearances are inevitable on both the hub and shroud side of turbine housing. Leakage flow formed inside the nozzle clearance leads to extra flow loss and makes the nozzle exit flow less uniform, thus further affects downstream aerodynamic performance of the rotor. As the leakage mixing with nozzle wake flow, the process is highly unsteady, which increases the fluctuation amplitude of transient load on the rotating turbine wheels. In present paper, firstly steady CFD analysis of a turbocharger turbine was performed at different nozzle openings. Then unsteady simulation of the turbine was carried out to investigate the interaction between the leakage flow through nozzle clearance and the main flow. Nozzle clearance's effect on turbine performance was investigated.

Performances of a centrifugal turbocharger compressor are investigated and validated in this paper. Based on the validation results, numerical optimizations are performed using ANN and CFD methods. Different impeller geometry with free parameters controlling stacking laws, end-wall, blade sectional camber curves and corresponding performances are used as input layer of ANN in the optimization, while adiabatic total-to-total efficiency and total pressure ratio are used as output layer of the optimization cycle. With this method, the performances of the compressor investigated in this paper are improved notably.