Your title goes here Application of STAR-CCM+ to Turbocharger Your Performance subtitle goes Prediction: here ICE Workshop Fred Mendonça
Content - STAR-CCM+ for Turbobcharger CFD - Power side turbine - Guide vane incident losses - Conjugate Heat Transfer - Part-to-high load thermal cycle - Air side compressor - Compressor virtual performance prediction - Installation effects - Future perspectives in transient flows - Compressor acoustics - Pulsed flow into turbine 2
Your title goes here APPLICATION OF STAR-CCM+ TO TURBOCHARGER Your MODELING subtitle goes AT here BORGWARNER TURBO SYSTEMS BorgWarner: David Grabowska CD-adapco: Dean Palfreyman, Bob Reynolds
Overview Turbine side - Inlet Guide vanes aerodynamics - Reduce incident angle flow losses - STAR-CCM+ used to determine the correct orientation of the guide vanes - Thermal requires Conjugate Heat Transfer (CHT) - Heat-up and cool-down in part-full load cycles - Non-isothermal housing affects compressor performance - Bearing / shaft oil lubrication 4
Geometry The variable guide vanes sit in the volute just upstream of the turbine leading edge The vanes each have their own pivot but are connected to a ring, which is in turn connected to a hydraulic actuator. This moves according to the operating condition of the engine to provide uniform flow guidance into the turbine
Computational Domain and Mesh The computational domain consisted of the volute inlet, all guide vanes and the complete turbine wheel Polyhedral mesh with Local refinement, Focused, prismatic cells in the wall layer, ~2.7 million cells
32.2-Degree Vane Opening, Nominal Position Mach Number Axial Section Mach Number Vane midspan -Relative frame Vane Pressure Pressure side In-Plane Relative Velocity Vane Midspan Static Pressure-Vane Midspan Vane Pressure Suction side
Different Vane Angles Mach Number Vane midspan -Relative frame (2 GV) Mach Number Vane midspan -Relative frame (10 GV) In STAR-CCM+ the vanes can be rotated and the mesh reconstructed Previous solution is mapped onto the new grid Analysis is continued Mach Number Vane midspan -Relative frame (63.3 GV)
Thermal Modeling of Turbochargers The turbocharger is typically connected directly to the engine and is thus subjected to high temperatures, both from the exhaust flow entering the turbine but also externally from engine mounting points. Thermal simulations at various operating conditions have been performed to investigate the temperature distribution through the turbine and bearing housing. In addition, thermal heat up and cool down transient calculations have performed from part to full load conditions to investigate Hot spots due to changing operating conditions, e.g. waste gate opening Thermal soak back into bearing housing and (bearing) oil Transient stress analysis for thermal cycles to failure
Geometry: Sequential twin turbocharger Low Pressure (LP) stage High Pressure (HP) stage Diverter valve
Geometry Diverter valve
Thermal Modeling: Simulations Performed Steady state thermal calculations were undertaken at part and full load conditions to determine the temperature distribution as well as serve as initial conditions for thermal transient calculations. Part load represented an engine coasting condition, full load is full engine rpm and full engine load. Thermal transient calculations simulating the thermal heat up and cool down of the turbine stage between part and full load.
Volume Mesh The computational mesh was constructed by merge/imprinting components and using split-bysurface topology to identify the different regions and automatically create interfaces Fully conformal mesh 4.5 million polyhedral cells Fluid Volume
Fluid Flow Results Full load (steady) Pressure Choked Wastegate Impingement Velocity near wastegate
Temperature Results Full load (steady)
Temperature results: components HP Housing IGV Back Plate and Guide Vanes LP Housing LP Turbine Wheel and Shaft HP Bearing Housing
Transient Conjugate Heat Transfer Analyses Following steady state CHT calculations, work was extended to transient heat up and cool down simulations Calculations from steady state part and full load analyses used as initial conditions for the thermal cycles Heating Cycle: Fluid @ full load, Solid @ part load Cooling Cycle: Fluid @ part load, Solid @ full load The analyses were run with a varying time step to capture the high temporal temperature gradients early in the analysis and for computational efficiency later in the analysis: 0<time<15sec, time step = 0.5 sec 15<time<80sec, time step = 1.0 sec, 80<time<150sec, time step = 2.0 sec 150<time<(end), time step = 4.0 sec
Heating Cycle Time = 4 secs Time = 18 secs Time = 42 secs Time = 105 secs Time = 355 secs
Heating Cycle
Monitoring Temperature Mon 4 Mon 45 Heating Cooling Mon 24 Mon 39
Cooling Cycle Time = 4 secs Time = 18 secs Time = 42 secs Time = 102 secs Time = 352 secs
Conclusion BorgWarner Turbo Systems have been successfully applying STAR- CCM+ to turbocharger design since v.2.02 in the following areas. Turbine guide vane analysis Thermal conjugate heat transfer analysis» Steady state full and part load conditions» Transient heat up and cool down simulations» Thermal soak back into the bearing housing and oil» Thermal transient stress calculations through data mapping to an FEA grid Compressor performance maps Effects of on engine installation pipe work
Your title goes here Application of STAR-CCM+ to Your Turbocharger subtitle goes Performance here Prediction Ford Motor Company: Onur Baris CD-adapco: Fred Mendonça
Virtual Compressor Map - Compressor performance map > obtained from standard test rig - Deviation from performance map - with an engine air intake system - Non-isothermal operation - Different Air Intake System (AIS) - Hard for testing - CFD is an ideal comparative tool - Benchmarking required - Purpose is to benchmark the method and compare vehicle setup with ideal setup and create an efficient method for compressor performance analysis 24
Compressor Analysis - Scope Benchmarking Ideal setup (i.e. test rig setup) is run at three speeds: 100000 RPM 160000 RPM 200000 RPM Ideal vs Vehicle setup comparison Compressor performance deviation when vehicle ducting installed 160000 RPM speed is chosen having the highest efficiencies on the map Three cases compared Ideal setup Vehicle Installation no IGV Vehicle Installation with IGV 25
Geometry of the Compressor Total pressure and temperature values are taken from inlet (compressor) and outlet (compressor) sections for the calculation of pressure ratio and corrected mass flow rate. Compressor Inlet Compressor Outlet Compressor Volute Compressor Blades 26
Test vs Vehicle Set-up Ideal Set-up Ideal (left) vs vehicle (right) configurations for the compressor - Straight pipes added to the inlet and outlet of the compressor - No changes in cross-section area - Ideal case for the compressor operation 27 Vehicle Setup - The piping is not straight including bends, contractions (changes in cross-sectional area of the pipe) - The pipe may include guide vanes before the compressor inlet - Real case in which the compressor will operate
Vehicle Set-up w or w/o IGV Pre-swirlers added to the inlet of the compressor due to acoustic requirements Significant impact on performance Comparison with ideal and vehicle setup without IGV needed 28
CFD - Volume Mesh View of volume mesh on a cross-section through the blades - Approx 3million cell model -3 hrs on 32 proc. Per operating point -Speed line is fully automated 29 Prism mesh adapted in small clearances
CFD Flow Modeling - Turbulence: SST k-ω model is used as turbulence model - Governing equations: Momentum and energy are solved coupled (high compressibility effects since Ma>0.3) - Motion: Rotationary region is defined for blades and hub, which rotate at the operation speed Other regions are stationary Mixing plane interface is used between stationary and rotationary regions - This interface applies a commonly used treatment which circumferentially averages out the velocities along several radial bands at the outlet of the upstream domain and applies the mean circumferential velocity value to each radial band at the inlet of the downstream domain 30
CFD Boundary conditions - Total conditions (pressure and temperature) are used at the inlet being equal to atmospheric conditions - Downstream boundary condition is defined by static pressure -All y+ wall boundaries are treated as adiabatic -Alternatives; - Mass-flow and pressure - Stagnation pressure and mass-flow Initialization -Start at a low pressure ratio (close to choke) and use the results as the initial condition for the next operatin point at that speedline (higher pressure ratio) - Grid Sequencing initialisation seems to be able to start at all operating points robustly 31
CFD Y+ values Typical y+ plot at 160000 RPM 32
Results Ideal Setup Figure showing the results of ideal setup i.e. test rig setup Results are very close to test Deviation from test results at low mass flows close to surge region 33
Results Ideal Setup 1 2 3 Streamline plots with pressure contour for 1: PR=1.73 2: PR=2.15 3: PR=2.25 At 160000 RPM speed, ideal test setup 34
Results Ideal Setup Residual behaviour for different pressure ratios at 160000 RPM (left: PR=2.15, right PR=2.25) 35
Results Ideal vs Vehicle Figure showing the results of ideal and vehicle (with and without IGV) setups at 160000 RPM A certain deviation occurs from ideal setup when vehicle setup is introduced Further deviation when IGV is included All speedlines for different intake ducting become close as mass flow decreases 36
Results Ideal vs Vehicle A B C Axial velocity distribution for A: Ideal, B: Vehicle no IGV, C: Vehicle with IGV cases 37
Results Ideal vs Vehicle A B C Tangential velocity distribution for A: Ideal, B: Vehicle no IGV, C: Vehicle with IGV cases 38
Results Ideal vs Vehicle A B C Streamline plots with pressure contour for A: Ideal Case B: Vehicle installation no IGV C: Vehicle installation with IGV 39 At 160000 RPM and PR~2.05
Conclusion Good agreement between CFD and test for ideal setup Deviation from the ideal compressor map with vehicle intake ducting Further deviation when IGV included > IGV has high impact Detailed insight information for different inlet ducting configurations using CFD, which cannot be obtained from test - More information available to investigate performance deviation Successful methodology developed for compressor performance analysis - A good match with the compressor map from the test is obtained - Robust process for different intake system geometries established 40
Future Perspectives Compressor transience To investigate the general characteristics in a transient run Investigation of surge, where the flow is highly unsteady. Transient runs are needed to investigate in detail Transient runs are also needed to investigate the acoustic effects of the IGV Prediction shows original design (top) with a predicted broadband noise problem between 4-8kHz Modified design (bottom) prediction shows removal of broadband noise issue while BFP and engine 1 st order tones are dominant 41
Future Perspectives Pulsed Turbine Pulsed inflow from engine Hysteresis in the efficiency during the cycle Experiment Pulse time average performance 42 CFD
Thank YOU any questions?