ARL Penn State COMPUTATIONAL MECHANICS Computational Evaluation of

  • Slides: 17
Download presentation
ARL Penn State COMPUTATIONAL MECHANICS Computational Evaluation of the Cavitating Flow through Automotive Torque

ARL Penn State COMPUTATIONAL MECHANICS Computational Evaluation of the Cavitating Flow through Automotive Torque Converters Acknowledgement: This work is supported by the General Motors Corporation 15 August 2012 J. W. Lindau F. J. Zajaczkowski M. F. Shanks R. F. Kunz 1

ARL Penn State COMPUTATIONAL MECHANICS CONTENTS • Introduction • Computational Methods • Results •

ARL Penn State COMPUTATIONAL MECHANICS CONTENTS • Introduction • Computational Methods • Results • Summary 2

Introduction: Automotive Torque Converter COMPUTATIONAL MECHANICS ARL Penn State Torque Converter from Wikipedia CAVITATION

Introduction: Automotive Torque Converter COMPUTATIONAL MECHANICS ARL Penn State Torque Converter from Wikipedia CAVITATION IN A TORQUE CONVERTER: • Working fluid is ATF, heat, extreme pressures • Torque converters have historically not suffered negative effects from cavitation • However, the trend is to smaller, lighter, etc • Minimum pressure/stator region may cavitate at high torque, low turbine speed • Concerns are performance, vibration, noise 3

ARL Penn State Methodology COMPUTATIONAL MECHANICS • First principals model of… – – –

ARL Penn State Methodology COMPUTATIONAL MECHANICS • First principals model of… – – – – • Mixture of gases and liquids Gas-liquid interfaces Large scale gas cavities Incompressible to compressible: disparate sound speeds Shocks Significant inherent unsteadiness (even in steady, planing configuration) Energetic propulsion Chemistry and phase change – Liquid/vapor mass transfer (stiff) – Chemical reactions (stiff) • • • Control Surfaces--6 DOF: fully coupled to flow Preconditioning (addresses stiff physical eigensystem) Turbulence modeled and (where feasible, required) simulated Numerical model: fully-conservative, unsteady, implicit, multiphase, preconditioned finite volume form Unsteady simulations with many millions of degrees of freedom are feasible/required 4

ARL Penn State DIFFERENTIAL MODEL COMPUTATIONAL MECHANICS • Computational tool— • • • •

ARL Penn State DIFFERENTIAL MODEL COMPUTATIONAL MECHANICS • Computational tool— • • • • n-liquid+n-gas preconditioned all-Mach number compressible total energy conservation any 2 -variable eos/species body forces/propulsors Numerical solution of mixture: mass, mass-transfer=phase change and chemistry momentum, energy, additional phases, shock-capturing species, and turbulence models on moving or static, overset computational level-set—free-surface or cavity interface meshes. multibody-control surfaces-6 DOF overset 2 -eq RANS/DES/transition 5

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Cavitator Lift and Drag

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Cavitator Lift and Drag

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Lift and drag values and comparison of

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Lift and drag values and comparison of experimental and computational geometries and computed cavities (with gas streamlines) from experiments of Waid and Kermeen (1957).

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Cavity Size vs. Ventilation Rate

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Cavity Size vs. Ventilation Rate

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Mesh showing flowpath, rotor, and stator in

ARL Penn State VALIDATION HIGHLIGHTS COMPUTATIONAL MECHANICS Mesh showing flowpath, rotor, and stator in NSWC-CD Tunnel Normalized Power Normalized Head Rise Normalized inlet total pressure 9

ARL Penn State COMPUTATIONAL MECHANICS Torque Converters: Computational Mesh turbine pump COMPUTATIONAL GEOMETRY REPEATED

ARL Penn State COMPUTATIONAL MECHANICS Torque Converters: Computational Mesh turbine pump COMPUTATIONAL GEOMETRY REPEATED OVER FULL 360 deg turbine pump stator Thin Torus: Converter Approximating Current Designs Trends stator Round Torus: Research Converter BOTH A MIXING PLANE AND A BODY FORCE BASED COUPLING APPROACH ARE APPLIED 10

ARL Penn State Round Torus CFD Results COMPUTATIONAL MECHANICS a) Cavitating CFD current effort

ARL Penn State Round Torus CFD Results COMPUTATIONAL MECHANICS a) Cavitating CFD current effort MPa 0. 8 0. 0 Single Phase CFD current effort b) CFD and test results on research converter. K-factor (RPM/[torque]1/2) and torque ratio. 11

ARL Penn State Thin Torus CFD Mesh COMPUTATIONAL MECHANICS Through-flow Stator Pump Turbine Grids

ARL Penn State Thin Torus CFD Mesh COMPUTATIONAL MECHANICS Through-flow Stator Pump Turbine Grids for body force based method. Computational meshes, thin-torus torque converter (coarse). Solid surfaces are illustrated with black mesh. Periodic boundaries are illustrated with green mesh. 12

ARL Penn State Thin Torus CFD Results COMPUTATIONAL MECHANICS CFD results (red) diamond: cavitating

ARL Penn State Thin Torus CFD Results COMPUTATIONAL MECHANICS CFD results (red) diamond: cavitating square: 1 -phase Dyno-250 N-m K-factor/100 Dyno-135 N-m Torque Ratio Speed Ratio • • • Computation and testing of Thin Torus TC. Single-phase and cavitating. Plot of K-factor/100 and torque ratio versus speed ratio. Dynomometer: black marks with black lines. CFD: red diamonds and dashed==single-phase, and CFD: red squares and dashed ==cavitating 13

ARL Penn State Thin Torus CFD Results COMPUTATIONAL MECHANICS MPa 1. 1 MPa 0.

ARL Penn State Thin Torus CFD Results COMPUTATIONAL MECHANICS MPa 1. 1 MPa 0. 5 0. 0 suction side pump and stator pressure-side pump and stator Single-phase solution, pump at 3000 RPM, turbine stationary, thin-torus torque converter. Cavitating CFD solution, thin-torus unit. Elements repeated periodically for visual effect. Surfaces made translucent to better visualize stator and cavity. All surfaces colored by pressure. Isosurface of vapor volume fraction at 0. 5. Pump at 3000 RPM, stall condition

ARL Penn State SUMMARY COMPUTATIONAL MECHANICS • CFD methodology validated for ventilated and natural

ARL Penn State SUMMARY COMPUTATIONAL MECHANICS • CFD methodology validated for ventilated and natural cavitation, supercavitation, and turbomachinery • Torque converters modeled usingle blade passage, multi-blade row (steady, periodic assumption) CFD • Mixing plane and body force coupling • Both approaches are problematic • Cavitation effects on pump torque captured • For high torque/large cavities and impact on noise/vibration, a full 360 deg unsteady analysis may be needed 15

ARL Penn State Preconditioner COMPUTATIONAL MECHANICS Derivation simplified working in terms of mass fraction

ARL Penn State Preconditioner COMPUTATIONAL MECHANICS Derivation simplified working in terms of mass fraction

ARL Penn State Preconditioner COMPUTATIONAL MECHANICS • • We choose: c’=min[ max( Vcut-off ,

ARL Penn State Preconditioner COMPUTATIONAL MECHANICS • • We choose: c’=min[ max( Vcut-off , |V|ijk ), cijk ] (c’=cijk yields the unconditioned result) Introduces artificial sound speeds yielding good convergence/accuracy regardless of Mach number/density ratio