The Effect of Machining Marks on the Stress Distribution and Fatigue Durability of Turbocharger Compressor Wheels

Turbocharger compressor wheels are traditionally manufactured using the casting process. However, with the demand for smaller and more durable compressor wheels and the improvement of multi-axial machining technology, machined compressor wheels are becoming popular among turbo-machinery manufacturers. The solid-billet materials used for five-axis machining have negligible internal defects, hence, the durability and fatigue life over equivalent cast components is improved. Nonetheless, machining a complex compressor wheel shape from a solid billet results in tool marks being left on the component surface. In this work, repeatedly running two types of multiple wheels to 10% beyond the design speed limit resulted in fatigue failure initiating from the machining marks in around 83% of cases. The ‘as machined’ geometry of the wheels was determined using both CT scanning and optical surface measurement techniques. The data from these measurements were used to generate solid CAD models suitable for finite element analysis to determine the stress distribution in the reverse-engineered wheels.

The maximum principal stresses predicted were 20% to 36% higher than those obtained from the nominal CAD models of the two wheels. In order to model the measured geometry efficiently, a novel technique was used to enforce cyclic symmetry on geometry that is not precisely cyclically symmetric.
Numerical analysis of dog-bone shaped specimens with cusps showed the effects of cusp depth, tool size, cusp direction and surface curvature on the Stress Concentration Factor (SCF) generation due to machining cusps. Regression-based mathematical models were developed to estimate the contribution to SCF due to each of these factors.

A full factorial experimental analysis was performed to identify the effects of spindle speed, feed-rate, and tool approach angle on surface roughness within the cusps generated by a ball-nose milling tool. The results showed that the spindle speed of 7,500 rpm, feed rate of 250mm/min and tool approach angle of 25° generated the best surface finish. These machining parameters were used to machine dog-bone shaped fatigue specimens with cusps.

A set of fatigue tests of dog-bone specimens with cusps was conducted. The fatigue data acquired was compared with data extracted from standard fatigue data. The results showed a partial correlation. Good correlation was observed over the range of cusp sizes found in BorgWarner’s compressor wheels.

SCF due to cusp depth and tool size generated from the regression model correlated with SCF calculated from the FEA results of nominal and reverse-engineered compressor wheels. The developed regression model was used to optimise the cusps’ specification of five Type B compressor wheels. Fatigue tests of the optimised wheels showed a 34% increase in fatigue life.