Designing better gearboxes Titelmasterformat durch Klicken bearbeiten

Transcription

1 Designing better gearboxes Titelmasterformat durch Klicken bearbeiten A new method for taking the housing stiffness into account Urs Fasel - Suter Racing Technology AG; Markus Dutly - CADFEM (Suisse) AG In collaboration with:

2 Outline Introduction: Engineering and economical motivation A high performance motorbike gearbox Gearbox modeling procedure with ANSYS and KISSsys Procedure verification Experimental measurements with GOM Precise Optical Metrology Model validation Conclusions 2

3 Project Team Giampaolo Franzoso, CADFEM (Suisse) He was born in 1978 in Verona (Italy). In 2004 received his Master Degree in Biomedical Engineering at Politecnico di Milano, where in 2008 he also got his PhD in Structural, Seismic and Geotechnical Engineering. Since 2008 he worked at Zimmer GmbH as Project Manager and at Marenco Swisshelicopter AG as Lead Engineer. Stefano Meneghello, CADFEM (Suisse) In 1992 was born in Camposampiero (PD, Italy). In 2014 he received his Bachelor Degree in Mechanical Engineering at University of Padua (Italy). Currently he is studying at the ETH Zürich, Master in Mechanical Engineering Structures. Urs Fasel, Suter Racing Technology Born 1974 in Wetzikon (ZH), received his Master in Mechanical Engineering at the ETH Zürich in Working afterwards at ETH, his research work was focusing on design optimizations through evolutionary algorithms applying them also on motorcycle frames. Since 2004 he is part of Suter Racing Technology as project leader and member of the management. Jürg Langhart, KISSsoft Born 1973 in Ottikon (ZH), was studying mechanical engineering at HSR, Rapperswil. Afterwards working at Curtiss Wright and then at Klingelnberg AG, as an application engineer for spiral bevel gears. Joined KISSsoft AG in 2008 as technical sales and project engineer. Ernst Müller, GOM Born 1951 in Weiningen (ZH), Master and PhD in Mechanical Engineering at the ETH Zürich. Established GOM International in 1998 in Switzerland and built up the Distributor network. 3

4 Titelmasterformat High Performance durch Gearbox Klicken bearbeiten Requirements & considerations designing a modern racing motorcycle gearbox In collaboration with:

5 Motivation for a thorough study of factors influencing gearbox (housing) design Performance - Targeting less friction thus optimizing torque transmission. - Reducing power loss inside gearbox. - Optimizing lubrication, being able to use less oil (splash lubrication). - Predicting better mass & stiffness inertia influences by gearbox. Weight considerations (Inherit the needed stiffness & stability whilst reducing considerably the needed weight). - Weight of the engine in this project: 26 kg - Weight of the gearbox/housing: 7 kg (27 %) Reliability The rear wheel is directly connected to the output shaft of the gearbox. For motorcycle applications, a failure in gearbox almost always results in a heavy crash of man and machine. Costs - Service intervals (also limited by regulations of championship organization) - Using less engines, in a season, results in considerable cost savings of teams. 5

6 Some critical aspects / design goals for the developer of an engine gearbox Gearbox housing is connected / one part with the engine housing. Bigger design changes almost always influence the overall engine design. Stiffness of overall engine / housing in motorcycle application has big impact on the overall vehicle stiffness (frame stiffening). Design targets - Bearing zones should be as stiff as possible (higher critical bending speed, optimal gearing contacts thus an optimal torque transmission). - Each gear has different gearing ratio options. - Gearbox design is also always a compromise between different options. - Each gear has its own life cycle. The overall gearbox must be designed targeting an overall best possible life span. 6

7 Geometric aspects of the gearbox which was used Motorcycle Type / Position of gearbox Moto3 TM Motorcycle. Gearbox specification Sequential 6-speed constant mesh, spur gears, splash lubrication. Packaging constrain of gearbox Positioning of the output shaft and crankshaft are given by vehicle requirements. Primary shaft and shifting actuator are constrained by packaging. 7

8 How gearbox design has been done up to now w/out KISSsys Bearing (housing) dimensioning Bearings and resp. housings in a motorcycle gearbox are designed to last and hence are most likely overdimensioned. They are dimensioned using simplified 2D geometric analysis inputing the resulting forces afterwards into an FEM of the gearbox housing. Mass inertia & stiffness considerations When designing an engine for a motorcycle, the mass inertia and stiffness properties of its components from crank shaft until rear wheel have to be considered in order to build up a frequency analysis and detect eigenfrequencies in the system with different gearing. Up to now, the gearbox shafts have been simplified, assumed stiff. KISSsys can lead to a more realistic prediction of frequencies in future. 8

9 Titelmasterformat Gearbox modeling durch procedure Klicken with bearbeiten ANSYS and KISSsys In collaboration with: 9

10 The new approach: Workflow Housing Stiffness extraction Iterative solution within KISSsys Gearbox design and Bearing Loads extraction EXPERIMENTAL VALIDATION Structural Analysis of the housing 10

11 ANSYS Model Non-linear model, incl. frictional contact and bolt pretension. Remote points definition in correspondence of bearing centres. Superelement analysis for stiffness matrix extraction with perturbation (linearization). Import of the stiffness Matrix in KISSsys. 11

12 KISSsys model: Shifted transmission of Moto3 TM Motorcycle Kinematic of transmission Sequential 6-speed constant mesh, spur gears Conditions for laboratory measurements Gears in 1 st speed Torque output (at primary shaft) blocked by SRT Coupling Torque input: T = 205 Nm at the secondary shaft Calculation of bearing offsets Import of stiffness matrix Calculation of bearing offsets Export of forces and offsets to double check with FE Torque output Torque input 12

13 Results for bearing offsets and shaft deflection Bearing outer race offsets offsets are calculated for - radial and axial offsets - radial tilting radial offsets (ux, uz) are evaluated Shaft bending Due to the forces and moments acting on the shafts, the bending is calculated Gear misalignment The gears are misaligned due to the shaft bending and offset Meshing gears Hence, the gear misalignments are now modelled most realistically! Primary shaft Secondary shaft 13

14 Results for gear meshing and contact analysis Gear meshing the face load factor K Hβ describes the load distribution over the face width considering the housing stiffness, the K Hβ changes from 1.62 to 1.53 Gear contact analysis The gear contact analysis is a simulation of the meshing Considering also bending of teeth Gear flank optimization Strength optimization by tooth trace modification noise optimization by profile modification 14

15 Static Structural Analysis Bearing loads application (calculated iteratively in KISSsys using the stiffness matrix). Static Structural analysis allows to evaluate the deformation of the housing and the level of stress. 15

16 Verification of the numerical procedure Comparison between KISSsoft and ANSYS results for Remote Point displacement: good agreement (within small errors due to non linearity). Non-Linear FE Model vs. Linear Model Remote Point 1 D.O.F. X Y Z Rx Ry Rz KISSsys input F [N] / M [Nmm] KISSsys output u [mm] / r [rad] -1.63E E E E E-05 Δ FE output u [mm] / r [rad] -1.64E E E E E E-05 diff % Δ FE/KISSsys Difference lower than 6%. Consider also that values are fractions of a micron. Good for a non linear model % 5.56% 0.13% % -0.30% diff % Δ FE/KISSsys -0.32% 5.56% 0.13% % -0.30% Similar results for all other points 16

17 Precise Industrial 3D Metrology and Evaluation Software for 3D Measurement Data with GOM Titelmasterformat durch Klicken bearbeiten From images to system understanding In collaboration with: Ernst Müller, 17

18 Modeling the Experimental Setup In the model for the experimental validation, boundary conditions were set according to the experimental setup and the loads obtained in KISSsys. The bending effect due to the weight that was adopted to apply a stable torque is considered in KISSsys. Initial preload of 7.5 Nm after bolt pretensioning was set as reference configuration for deformation evaluation according to laboratory loading conditions. BOLT PRETENSION PRELOAD T=7.5Nm LOAD T=205Nm 18

19 GOM deformation measurements - Single camera, capturing form quasistatic load - Deformation vectors, at preload (7.5 Nm), 180 Nm and 205 Nm load. Ernst Müller,

20 Titelmasterformat Model validation and durch Conclusions Klicken bearbeiten In collaboration with: 20

21 Validation of FE model through Experimental Results Same intensity of deformation vector Same direction of deformation Diff% = 7% Diff% = 7% Diff% = 10% Diff% = 10% Diff% = 12% Diff% = 5% Diff% = 5% Diff% = 3% Diff% = 5% GOM MEASUREMENTS ANSYS FEM 21

22 Conclusions I Interaction of KISSsys and ANSYS leads to: Simple and straightforward procedure: Time saving, less errors. Verified and validated procedure. Interaction between gearbox elements captured properly: Better results compared to separated analyses of housing and inner parts. Design improvement with reduced effort: Increased gearbox performance (in terms of weight, stiffness, safety). Reliability increase: Better evaluation of shaft misalignment. More accurate damage forecast (maintenance costs reduction) and life-cycle estimation. Iterative solution within KISSsys Structural Analysis of the housing Housing Stiffness extraction Gearbox design and Bearing Loads extraction Validation 22

23 Conclusions II Simulate 100 x Test (max.) 1 x Test to validate the Simulation Model Future now: No Tests needed for validated Simulations Simulation and Test should always work hand in hand 23

24 Simulation ist mehr als Software 24