Optimization of Vaneless Volute and Mixed Flow Impeller for Pulsating Flow

Abstract

Vaneless turbocharger turbines are commonly used for automotive engines due to their low cost and better off-design performance. It consists of a vaneless volute and a radial or mixed flow rotor, where both components are important to the overall device performance. With the pulsating nature of the exhaust flows, most energy is contained at the peak of the pulse. Therefore, during one engine cycle, optimizing the turbine performance for the peak pulse region is more straightforward to improve the cycle-averaged shaft power generation. This study sought to optimize both the volute and rotor simultaneously for the peak point of the pressure pulse (2.4 bar). Thirteen design parameters in total are considered during the optimization process. Six volute design parameters were used to control the aspect ratio, intake area, exit area, and the circumferential distribution of the cross-sectional area. Seven rotor parameters were utilized to modify the cone angle, blade axial location, and the camber-line angle distribution. The optimization was conducted by a novel optimization algorithm based on Kriging surrogate model, and compared with the conventional genetic algorithm. Commercial turbulent viscous CFD solver ANSYS-CFX was used to predict the turbine performance. Full-stage turbine, including ten blade passages, is explicitly modeled for better accuracy. In order to ensure the matching between turbocharger and engine maintained the same as the original turbine, special attention was paid to constraint the swallowing capacity characteristic of the optimized turbine to be similar to the baseline turbine, with a maximum 2.5% difference at the design point. Compared with the baseline turbine, the turbine efficiency was improved by 3 percentage points with the using the genetic algorithm, and an improvement of 3.65 percentage points was achieved by using the Kriging surrogate model based optimization algorithm. Although the optimized turbine has a lower peak efficiency, the optimal velocity ratio of optimized design shifted from the baseline value of 0.71 to 0.61, implying a better performance will be achieved under high loading conditions. The improvement of the turbine performance is attributed to a better blade loading that is achieved in the 0.2–0.4 stream-wise location. The elementary effectiveness has been studied, and the camber-line distribution of the rotor is found to be the most influential factor on the turbine performance.