ODIN: A Retrospective on Integrated System Design for EV Dr. Melanie Michon 15th September 2016 Contents • Targets for ODIN • Setting up the ODIN Consortium • Design Challenges • Lessons Learned • Project Update • Summary Slide 2 CONFIDENTIAL © Copyright 2016 Initial Targets for ODIN (2011) • Focus subcompact/compact EV‘s with 65kW peak power (30s) • Integration of e-machine, power electronics and gearbox into one shared housing • Validation of high speed concepts o Not just a “me-too” design • Up to 50% cost reduction compared to reference powertrains • Avoidance of rare earth metals o Switched Reluctance Motor Slide 3 CONFIDENTIAL © Copyright 2016 Romax work on EV’s prior to consortium selection 2007 2008 2009 2010 2011 2012 2013 2014 E-AWD Rear Gearbox Design Support HEV efficiency optimisation HEV APU control optimisation 2015 2016 EV NVH opt. at concept design x2 EV gearbox & motor noise, thermal & efficiency optimisation EV gearbox HEV gearbox EV design support efficiency durability, NVH and HEV gearbox & motor AWD bearing EV FDU gear EV concept selection improvement efficiency opt. analysis whine analysis design and control opt. HEV concept, detail design and system integration PTU for HEV EV gearbox design optimisation HEV EDU design assessment EV gearbox concept design HEV NVH optimisation EV Differential Design and Analysis Off-road e- CVT noise troubleshooting Slide 4 EV gearbox design process EV gearbox design Hybrid Vehicle Gearbox assessment Optimisation EV NVH troubleshooting CONFIDENTIAL © Copyright 2016 Consortium Set Up system development • Kick off September 2012 gear e-components lubricant Slide 5 casting sim mechanic demo car CONFIDENTIAL © Copyright 2016 Design Challenges • What layout is best? How many stages? • What motor maximum speed/overall ratio? • How efficient will the drivetrain be? How do we make it more efficient? • Can we integrate cooling of motor/gearbox? • How do we design for low noise? Slide 6 CONFIDENTIAL © Copyright 2016 Which layout is best? • Assessment qualitative rather than quantitative Slide 7 CONFIDENTIAL © Copyright 2016 Ratio limits Packaging Second stage ratio Overall Ratio? Maximum Speed? Efficiency? ? Innovation ? Overall ratio Smaller, lighter, cheaper electrical machine Higher gear ratio Gearbox efficiency validated at conventional speeds but not high speeds Motor efficiency calculations slow, esp. SRM Design decision: R&D cannot be “me-too”. The project needs to set a challenge, even at the risk of failure Slide 8 CONFIDENTIAL © Copyright 2016 Innovation Highlights • System Simulation o • Power electronics NVH simulation in early development stages E-Motor Cooling & Lubrication o • One cooling and lubrication circuit (oil) Switched Reluctance Motor (SRM) o • High speed (22.5K RPM); NVH optimized design and control Power Electronics o High switching frequency o Highly dynamic current sensing • Main cover Gear Box Gear set o High efficiency despite high speed (max ratio for 2 stages) o NVH optimized layout • Housing and integration o Synergies by high integration Main housing Main focus: optimize efficiency, NVH, driveability, packaging and costs Slide 9 CONFIDENTIAL © Copyright 2016 Development of Integrated Cooling • Thermal Simulation carried out by Romax • System Level: 1D system model study of layouts • Integrated Cooling: Through system heat sharing, enables efficiency gains for the gearbox achieved on cold start performance Non Integrated Slide 10 Integrated Cooling system CONFIDENTIAL © Copyright 2016 Shared cooling circuit development Philosophy of dry sump topology: • The e-Motor cooling requirements dictate the major cooling flow demand • The power electronics require consistency of temperature for best reliability • Gearbox cold running performance can be improved, through use as an alternate heat dump to the radiator Slide 11 CONFIDENTIAL © Copyright 2016 Development of Lubricant 1. High load, low speed Implementation of variable friction coefficient over LoA, Mixed-EHL Model 2. Low load, high speed “Designing the oil for the machine and the machine for the oil” Slide 12 CONFIDENTIAL © Copyright 2016 Lubricant effect on efficiency in RomaxDESIGNER • 2 lubricants with identical viscosities investigated for 2-stage EV gearbox ISO standard (ISO TR14179-DE) suggests identical performance FVA345 method enables differentiation of lubricants Slide 13 CONFIDENTIAL © Copyright 2016 How to “Design for Low Noise” • Predict noise at a time where it can be minimised • Need for rapid design-analyse-assess-redesign iterations: new electrical machine NVH capability No housing, but complete with driveshaft and vehicle definition System Concept Sum total power through bearings System test System Detailing/ Sub-system design Basic housing design Mount stiffnesses from reference data Sub-system test Sub-system Detailing/ Component Selection Component Detailing Component Test Slide 14 CONFIDENTIAL © Copyright 2016 NVH Step 1: Initial concept assessment for NVH performance Unique capability to assess NVH performance for concept design System test System Concept System Detailing/ Subsystem design Sub-system test Identify potential NVH issues early – Right First Time design Sub-system Detailing/ Component Selection Component Detailing Component Test Gearbox (SOURCE) Bearing 2 Bearing j Acoustic power transmitted through bearings Powertrain definition, no housing definition Representative model of generic housing is used Bearing 1 Bearing n Sound pressure at a given location Slide 15 CONFIDENTIAL © Copyright 2016 Method Validation Vibratory force (N) Vibration velocity (m/s) No Housing Acoustic Power Transmitted Through Bearings in 0-5000 Hz band (mW) 21.6 mW Significantly off 4.8 mW Generic Housing Transmitted acoustic powers are very close Full Model 4.5 mW Slide 16 CONFIDENTIAL © Copyright 2016 NVH Step 2: Concept Housing and unit excitations give indication of dominant response frequencies for gear and e-machine excitations System test System Concept System Detailing/ Subsystem design Sub-system test Sub-system Detailing/ Component Selection Component Detailing Component Test Front Mounts Rear Mounts Slide 17 CONFIDENTIAL © Copyright 2016 Identify potential design actions from the simulation: “Breathing” Mode at 13000 RPM • Frequency and noise radiation through motor end plate matches with experience • Design Action: o Design rib patterns to reduce response Slide 18 CONFIDENTIAL © Copyright 2016 NVH Step 3: Detailed Housing and calculated excitations System test System Concept System Detailing/ Subsystem design Sub-system test Sub-system Detailing/ Component Selection Component Detailing Component Test • Housing designed based on guidance from concept simulation • Further guidance based on detailed simulation Slide 19 CONFIDENTIAL © Copyright 2016 6 Mode, 12Cycle – 1400N 6 Mode 16 1 Detailed housing at 4761 Hz (23808RPM) 17 18 2 3 15 14 5 6 4 13 12 11 8 9 7 10 Radial Force- 6 Mode_12cycle Frequency (Hz) 800 1600 2400 3200 4000 4800 0 5.0E-03 Reduction Factor = 0.36=8.81dB 4.0E-03 Velocity (m/s) Radial Forces 3.0E-03 2.0E-03 1.0E-03 0.0E+00 0 0 4000 25 8000 16000 Input Shaft (RPM) 20000 75 100 125 50 Vehicle Speed (km/h) Air-6Lobe_12Cycle Slide 20 12000 24000 150 Old_Air-6Lobe_12Cycle CONFIDENTIAL © Copyright 2016 0 Mode, 36Cycle – 190N 0 Mode 17 18 2 Concept housing at 8236Hz (13728RPM) Radial Force- 0 Mode_36cycle Frequency (Hz) 10 1 3 Detailed housing at 7373Hz (12288RPM) 0 2.5E-02 2400 4800 7200 9600 12000 14400 Reduction Factor = 0.22=13.22dB 2.0E-02 Velocity (m/s) Radial Forces 1.5E-02 1.0E-02 5.0E-03 0.0E+00 0 4000 0 25 8000 Input Shaft (RPM) 16000 20000 75 100 125 50 Vehicle Speed (km/h) Structure-0Lobe_36Cycle Slide 21 12000 24000 150 Old_Structure-0Lobe_36Cycle CONFIDENTIAL © Copyright 2016 Significant reduction in response achieved for gear and e-machine excitations Response to transmission error Response to radial force Slide 22 CONFIDENTIAL © Copyright 2016 Test Results: Impact of Lubricants on Efficiency • Extensive testing of lubricants to confirm the process “Designing the oil for the machine and the machine for the oil” Slide 23 CONFIDENTIAL © Copyright 2016 Test Results: Efficiency • ODIN gearbox showed better efficiency than baseline gearbox (a current EV gearbox) at all speeds and loads Slide 24 CONFIDENTIAL © Copyright 2016 Other test results …. and design decision • Gearbox design for efficiency successful • NVH test run with very different model set up and boundary conditions, so no correlation possible • Switched Reluctance Motor efficiency did not meet design targets, despite novel control strategy development • By the end of Loop 2 the price of rare earth metals less problematic o Economic justification for SRM had reduced • Decision for Loop 2: Go with Induction Motor, similar gearbox to Loop 1 Slide 25 CONFIDENTIAL © Copyright 2016 Lessons Learned • Design of integrated electro-mechanical drivetrains brings new uncertainties • Difficult to develop new design/simulation process and new design at the same time. Check the “technological building blocks” are in place • However, use of up-front simulation (e.g. NVH) can yield great benefits • Close collaboration between motor and gearbox design activities essential • Engineering decisions still subject to economic influences Slide 26 CONFIDENTIAL © Copyright 2016 Project Update • Motor changed to Induction Motor by Bosch • Maximum motor speed changed to 15,500 RPM • Gearbox design process from Loop 1 replicated, made use of successes for efficiency, lubricant design • Methods for design for noise, thermal in an integrated system carried over • Testing started in the last couple of weeks… watch this space! Slide 27 CONFIDENTIAL © Copyright 2016 Summary • Romax has seen at first hand the challenges in a design project, with cross-industry collaboration (motor, gearbox, vehicle) • Much R&D work has been kicked off based on limitations of the current state-of-the-art for design of EV drivetrains • Understanding of the interactions across the electromechanical drivetrain is essential for design success Slide 28 CONFIDENTIAL © Copyright 2016