ENTWICKLUNG DIESELMOTOREN BMW Steyr Diesel Engine Development Center MULTIBODY AND STRUCTURAL DYNAMIC SIMULATIONS IN THE DEVELOPMENT OF NEW BMW 3- AND 4-CYLINDER DIESEL ENGINES Dr. Stefan Reichl, Dr. Martin Kuchler, Mario Prandstötter, Günther Pessl 3 rd International Conference: Dynamic Simulation in Vehicle Engineering 2014 Engineering Center Steyr, St. Valentin, May 22 nd -23 rd
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Project Examples in the Low Frequency Domain 4 Project Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 2
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Project Examples in the Low Frequency Domain 4 Project Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 3
INTRODUCTION DYNAMIC SIMULATION IN 3-, 4-, 6-CYL. ENGINES Modular design for 3-, 4- and 6-cylinder inline diesel engines (38% carry-over, 59% synergy parts) 70kW 290kW / 220Nm 760Nm Structural Dynamics& Acoustics Vibration / durability of auxiliary components Vibration / durability of the exhaust system Vibration / sound radiation of ETU Transfer functions, transmission paths Vibration/ acoustic of air duct parts Multi Body Dynamics Cranktrain Dynamics Durability of Crankshaft Belt Drive Dynamics Chain Drive Dynamics Excitation ETU Chassis Seite 4
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Process Examples in the Low Frequency Domain 4 Process Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 5
MODELING PROCESS AND APPLICATIONS (1) ENGINE GEARBOX ASSEMBLY MODELS The structural dynamic conformable meshingof the relevant engine parts is performed with ANSA by means of the appropriate CAD models. For each engine component an average temperature is determined based on measurements or thermomechanical calculations. The modulus of elasticity, the mass density and the poisson sratio of the engine part material at this average temperature are assigned to the corresponding FEM model. The material parameters are classified in a database which is based on correlative experiments. The build-up of sub-assembly FEM models is accomplished by modelling internal parts such as crankshafts, balancing shafts, flywheels etc. with ANSA and MEDINA. Modulus of elasticity (MPa) Aluminium alloys Temperature ( C) Crankcase CAD model (left), crankcase FEM model (middle) and modulus of elasticity (right). Seite 6
MODELING PROCESS AND APPLICATIONS (2) ENGINE GEARBOX ASSEMBLY MODELS The engine gearbox assembly models reach up to 60 million degrees of freedom and are assembled with ANSA by means of specific connector models and a once defined assembly specification. 4-cylinder engine gearbox assembly FEM model. Seite 7
MODELING PROCESS AND APPLICATIONS (3) COMPARISON OF CALCULATED AND MEASURED DATA Global vertical bending natural vibration: Validation of the global vertical bending natural vibration of an engine gearbox assembly model on the basis of acceleration measurements at the corresponding right gearbox bearing on a roller dynamometer with full load. Acceleration measurements in the local z-direction Acceleration at the right gearbox bearing / local z-direction High Engine speed range Frequency Engine gearbox assembly model (left) and Campbell diagram of the measured acceleration (right). Frequency range that can be associated with the global vertical bending natural vibration Seite 8 Low
MODELING PROCESS AND APPLICATIONS (4) COMPARISON OF CALCULATED AND MEASURED DATA Global vertical bending natural vibration: Comparison of the calculated global vertical bending natural frequency with the corresponding evaluated frequency range from the acceleration measurements. Top bound Global vertical bending natural vibration occurrence Calculated global vertical bending natural frequency 8.4 Hz 3.1 Hz Lower bound Computed inertance(left), global vertical bending mode shape (middle) and frequency range (right). Seite 9
MODELING PROCESS AND APPLICATIONS (5) MBS MODEL Flexible Bodies: implementation via Craig-Bampton method (MSC.NASTRAN / MSC.ADAMS interface) ETU Crankshaft TVD hub Primary part with flex plates of dual mass flywheel or Torque converter Balancing Shafts Rigid Bodies: Pistons Conrods TVD: inertia ring, belt drive pulley Secondary part of dual mass flywheel Stiffness Curves: belt drive reduced onto the decoupled pulley Flywheel, TVD, engine mounts, gears, 350-500 DOFs (depending on EHD resolution) cylinder pressure dual mass flywheel with nonlinear torsional stiffness engine / gear box mounts modeled with static stiffness curves Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 10
MODELING PROCESS AND APPLICATIONS (6) EHD SUBROUTINE EHD subroutine implemented in MSC.ADAMS Numerical solution of Reynolds equations Consideration of crankshaft tilting and deformations of the bearing shells Load application via pressure distributions (MFORCEs) at crankshaft and bearing shells EHD- Parameter Diameter Supporting width Clearance Oil viscosity Convexity Resolution (number ofload shape functions) MFORCE (one load shape function) MFORCE: superposition of several load shape functions Only compressive forces are applied (in contrast to RBE-elements Reynolds equation (nonlinear PDE): H(φ,z) Lubrication gap geometry with elastic deformation of the bearing shell p oil pressure φ angle D nominal bearing diameter B bearing width Z relative bearing width η dynamic oil viscosity ψ clearance ω B angular velocity Seite 11
MODELING PROCESS AND APPLICATIONS (7) INPUTS / OUTPUTS OF A FULL LOAD ENGINE RUN-UP Inputs: Ignition data: measured cylinder pressures Speed profile from test rig Outputs: Engine speed (rpm) Combustion chamber roof forces Piston side forces Forces & torques in the main bearings Radial forces in the balancing shaft bearings Gear teeth forces of balancing shafts Radial forces at the gearbox input shaft Rotational irregulatory of crankshaft Engine torque Engine and gearbox mounting forces Time (s) Seite 12
MODELING PROCESS AND APPLICATIONS (8) LOAD GENERATION Calculated loads are further used for FEA / fatigue analyses of several engine parts and acoustic simulations high Fatigue analysis of crankshaft FEA of normal velocity levels (2600 2800Hz) low High high low Fatigue analysis of crankcase Acoustic camera (2600 2800Hz) Seite 13 Low
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Project Examples in the Low Frequency Domain 4 Project Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 14
PROJECT EXAMPLES IN THE LOW FREQUENCY DOMAIN (1) MBS-MODEL FOR START ANALYSES Rigid body model with reduced belt drive Engine and gear box mounts: nonlinear stiffnesses(static vs. dynamic) Cylinder pressures from measurements starter Seite 15
PROJECT EXAMPLES IN THE LOW FREQUENCY DOMAIN (2) STARTER Implementation of starter, planetary gear, sprockets, free wheel Starter: differential equations of a permanently excited DC machine Free wheel: user-written subroutine rotor planetary gear Parameter Description sprocket U 0,Batt R i,batt R 30 R ges U B L rot C mot M R I Anker i Batteryvoltage Internal resistancebattery Lead resistance Resistance starter Carbon brushesvoltage Rotor inductivity Motor constant Friction torque starter Mass of inertia rotor Gear ratios planetary gear, starter ring, sprocket VTORQUE: starter VTORQUE: free wheel Starter with planetary gear, sprocket and free wheel Equivalent circuit diagram of the permanently excited DC machine Seite 16
PROJECT EXAMPLES IN THE LOW FREQUENCY DOMAIN (3) RIGID BODY MOTION DURING START Comparison between simulation and measurement data Simulation Scale Factor: 5 Measurement Seite 17
PROJECT EXAMPLES IN THE LOW FREQUENCY DOMAIN (4) OPTIMIZATION OF ENGINE MOUNT POSITIONS Objective function: kinetic energy of ETU MIN (reduction of the shaking during start) x-, y-and z-positions of left and right engine mount are varied in a possible range ADAMS-optimizer OPTDES-SQP Kinetic Energy DZ (mm) Basis Optimized variant global minimum Trajectory of the center of mass of the ETU Objective function Lateral displacement of ETU is reduced by 70%, roll angle by 8%! Lateral forces in engine mounts are decreased by 82%! Maxima of lateral forces in engine / gear box mounts Seite 18
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Project Examples in the Low Frequency Domain 4 Project Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 19
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (1) COMPARISON OF 2 CRANKSHAFTS Crankshaft V1 Crankshaft V2 Benefits of Crankshaft V2 Mass crankshaft -2.2% Moment of Inertia (crankshaft) -3.7% Moment of inertia (cranktrain) -4.0% CO2 emission (NEDC) -0.4% Stiffness Reducion of Crankshaft V2 Bending stiffness (single throw) -13.8% Torsional stiffness (single throw) -17.5% Torsional stiffness(entire crankshaft) -14.2% Seite 20
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (2) COMPARISON OF 2 CRANKSHAFTS Shift in Eigenfrequencies of crankshaft V2 1 st vertical bending mode: -23Hz 1 st lateral bending mode: -43Hz 1 st torsionalmode: -89Hz Seite 21
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (3) MAIN BEARING FORCES (CRANKSHAFT V1) Bearing 1 Bearing 2 Bearing 3 Bearing 4 Bearing 5 6000 rpm Vertical Force 5000 rpm 4000 rpm Lateral Force 3000 rpm Lateral Torque 2000 rpm Vertical Torque 1000 rpm Seite 22
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (4) LATERAL TORQUE (CRANKSHAFT V1 / V2) Comparison of the Lateral Torque (referred to ETU) Lateral torque gives information about the vertical bending of the crankshaft Stiffness reduction causes higher forces and torques => stresses in bearing supports increase Seite 23
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (5) DEFORMATION OF THE CRANKSHAFTS One Cycle @ 4000rpm (point of nominal engine power) Crankshaft V2 Crankshaft V1 Seite 24
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (6) DEFORMATION OF THE CRANKSHAFTS Deformation @ ignition cylinder 4 Crankshaft V1 Crankshaft V2 Bending Line for Crankshafts and Bearing Shells Seite 25
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (7) PRESSURE DISTRIBUTION IN BEARING SHELLS Asperity pressure @ 4000rpm #1 #2 #3 #4 #5 #1 #2 #3 #4 #5 #1 #2 #3 #4 #5 Bottom Top low Crankshaft V1 high Crankshaft V2 Bearing Shells (Crankshaft V1) Maximum of asperity pressure: information about failure mode pitting Average of asperity pressure: information about failure mode wear Seite 26
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (8) TORSIONAL VIBRATIONS Campbell diagrams of crankshaft twist angle f 0 f 0 low high low high Seite 27
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (9) TORSIONAL VIBRATIONS Order analysis of crankshaft twist angle limit, 100 C limit, 100 C TVD is adapted to the torsionaleigenfrequencyof the system => a peak below and a peak above this point occur Torsional vibrations of crankshaft V2 are above the allowed limit Seite 28
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (10) TORSIONAL VIBRATIONS Order analysis of crankshaft twist angle limit, 100 C limit, 100 C Another type of TVD is used (broadband operating) in order to reduce the torsional vibrations Seite 29
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (11) FATIGUE COMPUTATIONS FEMFAT Channel MAX is used for fatigue computations Material models are derived from test rig results Safety factors are calculated in main- and crank bearing fillets, oil drillings, -24.7% Seite 30
PROJECT EXAMPLES IN THE HIGH FREQUENCY DOMAIN (13) FATIGUE COMPUTATIONS Distribution of safety factors high high low low Crankshaft V1 Crankshaft V2 Crack produced ontest Rig Calculated Critical Area and Crack Initiation Site from the Test Rig correspond Seite 31
OVERVIEW 1 Introduction 2 Modeling Process and Applications 3 Project Examples in the Low Frequency Domain 4 Project Examples in the High Frequency Domain 5 Summary Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr Seite 32
SUMMARY The introduced dynamic CAE processes are effectively applicable in an industrial development environment. The simulation results show good accordance to corresponding measurements. The CAE methods are continuously enhanced in the framework of the project work to further increase the quality of the computed results. Rigid body models with measured ignition data and electric components from starter etc. can perfectly be used for start analyses of the engine. Such models are applied for optimizations of the starting system, sensitivity analyses of belt drive and flywheel and optimizations of the engine mounts. If new designs of crankshafts or crankcase are evaluated, flexible multi body systems are used to calculate a full load engine run up. The interaction between crankshaft and crankcase plays an important role. A change in the stiffness of the crankshaft effects the main bearing forces and torques, the torsional and bending vibrations, the pressure distributions in the bearing shells and the safety factors of specific parts. Seite 33
SUMMARY Innovations High End Methods BMW Efficient Development Accuracy Reliability Innovations Simulation Speed Innovations Seite 34
BMW Motoren GmbH Diesel Engine Development Center Simulation/CAE Conference Dynamic Simulation in Vehicle Engineering, May 22 nd 23 rd 2014 Stefan Reichl, BMW Steyr