Simulation for Development of Chassis Mechatronic Systems

Simulation for Development of Chassis Mechatronic Systems

Dr. Matthijs Klomp, Technical Specialist – Vehicle Dynamics CAE May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

1

agenda • Vehicle Dynamics CAE @ Volvo Cars – Who are we and what do we do? • The software and electronics revolution in cars – what does it mean and how do we meet these new challenges? (Spoiler #1: Simulation!) • Simulation – Implementation examples • Conclusions – (Spoiler #2: Simulation is necessary!)

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

2

product development Vehicle Level

Voice of Customer

Market Acceptance

Subj. Vehicle Requirements

Subjective Vehicle Testing

Obj. Vehicle Requirements System Level

Component Level

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Objective Vehicle Testing

Sub-system Requirements

Sub-system Testing

Component Requirements

Component Testing

Component Design 3

Our reality: Transformation in progress • Trends: Safe, Green & Convenient • Common statement: ”More than 80 percent of automotive innovations are driven by electronics, and amongst them, 90 percent are supported by means of software.” Functions

• What we see clearly is:

• More vehicle variants in shorter time • Functional growth (exponential?) – e.g. autonomous drive

• How do we meet these challenges?

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Software

Electronics

4

Complexity – network topology

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

5

Simulation as an enabler • • • • •

Enables testing of solutions before hardware is built Enables fully automatic repeatable testing of real driving situations Supports continuous integration of SW functional increments Leverages computing power (n*24/7 versus 1*8/5) Measure the immeasurable – all signals are available

Pinpoint Decisions, Inc. May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

6

Vehicle Dynamics @ Volvo Cars • Vision “Fully virtual vehicle dynamics development” • Predict handling, steering & ride comfort

• Virtual tuning of tires • Virtual tuning of suspension (springs, dampers, stabilizer bars, bushings, hard-points and structural stiffness of chassis & body) • Virtual tuning of steering control (boost, damping, active return, path control, etc.) • Virtual tuning of brake control (yaw, side-slip and roll control) • Focus on event based testing (1:1 equivalence to physical testing) May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

7

System Analysis: suspension K&C

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

8

Vehicle manuever: Parking Sweep 100°/s 10 km/h

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

9

ISO 19634: Steady-State Constant Radius

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

10

Virtual vehicle architecture • Virtual Vehicle Architecture

• Platform and infrastructure to virtually develop networked mechatronic system functions • Purpose driven fidelity of simulation models (plant and controller) • Purpose driven integration of simulation models (configuration)

• Unified models

• Tool vendor independent models • Functional Mock-up Interface

• Purpose driven fidelity

• Real-time online • Error controlled offline • Configuration in PLM system Gimbergsson & Törmänen, 2013

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

11

Consistent Use of Simulation Models Integral process for functions, performance, functional safety and quality testing

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Courtesy: 91015/Edo Drenth

12

Methods – test strategy • Structured (questions pre-defined) • Standard load-cases • Highly automated • Pre-defined variations

• Semi-structured (variations of questions) • Monte Carlo / DOE simulations • Stress test (automatically find error states) • Sensitivity studies

New/updated requirements and/or test methods

Find & report issues

• Unstructured (ad-hoc questions) • • • •

May 3, 2016

Explorative testing (driving simulator, 24/7 off-line driving) Random variations / fault injection Visual inspection of data Data mining (new questions on existing data) VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Propose solution

13

Vehicle dynamics domain • Chassis Nodes • • • •

VDDM/BCM: Vehicle Dynamics Domain Master / Brake Control Module PSCM: Power Steering Control Module SUM: Suspension Control Module SAS: Steering Angle Sensor Propulsion CAN HS

Brake System May 3, 2016

Steering System

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Suspension System 14

Vehicle dynamics domain testing • In XC90 development: • • • • •

Fault injection to test correct error handling and diagnostics Test that functions are available and possible to activate Test of software download and other production related features Test of communication with other nodes in the car System safety testing including end-to-end protected signaling

• Today and onwards: • • • •

Moving Base Driving Simulator

Dynamic testing of vehicle dynamics functions Dynamic long-time testing Integration with driving simulator Integration with steering system test-rig Steering System Rig

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

15

Off-line steering system optimisation Electric Power Assist Steering System Parameterization and Optimisation Employing Computer-Aided Engineering Marcus Ljungberg (VCG), Mikael Nybacka (KTH), Gaspar Gil Gómez (VCG) and Diomidis Katzourakis (VCG) SAE 2015 World Congress & Exhibition. 2015-01-1500, Steering and Suspension Technology Symposium

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

16

On-line steering system optimisation DoE / Optimization (ModeFrontier)

Process automation via COM

Performance metrics

Remote GUI Control via TCP/IP

Calibration (INCA / CANape)

Vehicle & Rest-bus Simulation (CarMaker on DS1006) StWhlAg

XCP via CAN

TieRodPos

VehSpd

Result file

Post-processing (Sympathy for Data)

TieRodFrc StWhlTrq Steering System Verification Using Hardware-in-the-Loop Linköping University, Vehicular Systems Bjelevac, Salko Karlsson, Peter ISRN: LiTH-ISY-EX--15/4831--SE

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

17

Virtual homologation of ESC systems • ECE/TRANS/WP.29/GRRF/2014/12: “Where a vehicle has been physically tested in accordance with paragraph 8., the compliance of versions or variants of that same vehicle type may be demonstrated by a computer simulation, which respects the test conditions of paragraph 8. and the test procedure of paragraph 9.9. The use of the simulator is defined in Annex 1 to this Regulation.”

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

Source: www.cvel.clemson.edu

18

Virtual release of brake and steering systems • S90 brake internal SW road release without physical prototypes • Virtual homologation & steering in development

Diomidis Katzourakis, Stavros Angelis, Matthijs Klomp, Albin Johnsson, Robert Hansson Virtual Brake Software Release 24th International Symposium on Dynamics of Vehicles on Roads and Tracks, August 17-21, 2015, Graz / Austria May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

19

Validation of simulation models • ECE regulation: “The simulated vehicle behaviour and operation of he vehicle stability function is comparable with that seen in practical vehicle tests.” • Objective metrics on the validation the simulation model has been missing so far • Proposed standards:

• ISO/DIS 19364: Passenger cars -- Vehicle dynamic simulation and validation -- Steady-state circular driving behavior • “If all [physical] tests lie within the boundaries [specified in the standard], then the simulation is considered valid for determining steady-state behaviour up to the limits in lateral acceleration covered in the testing.

• ISO/DIS 19365: Passenger cars -- Vehicle dynamic simulation and validation -- Sine with dwell stability control testing • “If the number of the first test in each series in which ESC intervention occurs for physical testing matches the corresponding number for the first test in the simulated series ±1 as specified in 9.2.2, and the conditions specified in 9.3.3 are met for the three tests used for comparison (see 9.2.3), the simulation tool is considered valid for representing vehicle behaviour in the sine with dwell test.”

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

20

Model validity – accuracy vs usefulness • Albert Einstein: “As simple as possible, but not simpler” • George Box: ”All models are wrong; the practical question is how wrong do they have to be to not be useful” • Claude Rouelle: ”Make it useful before you make it complicated”

Robert G. Sargent – “Verification and validation of simulation models” Proceedings of the 2011 Winter Simulation Conference May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

21

Example for steady-state circular driving • Boundaries are based on Adams/Car data from ISO/DIS 19364

May 3, 2016

22

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS 22

Vehicle dynamics driving simulator • 0.3 – 30 Hz motion frequency • Motion envelope • • • •

May 3, 2016

X: ± 0.8 m Y: ± 0.75 m Z: ± 0.15 m Rotations: ± 20°

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

23

Example 1: Suspension control & Driving Simulator Rest-bus & I/O Simulation (CANoe on VT System)

Chassis CAN

4xWhlPos [V] 3xVertAcc [V]

FMI Coupling (FDX via Ethernet)

Cockpit & Motion Platform Variables (EtherCat)

4xDmprCrnt (A)

XCP via Measurement CAN

Vehicle Simulation (Concurrent Linux64)

Measurement & Calibration (CANape)

Driving Simulator

Suspension Control Module May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

24

Example 2: Brake control & Driving Simulator Brkpedpos (analog)

controller

Servo

StWhlTq

CarRealTime

StWhlAg BrkPedPos

Concurrent Vehicle Model

Booster

Vacuum pump

CAN (Chassis & Propuslion)

MCAL (INCA, CANape) Master cylinder

XCP on Ethernet

VDDM/BCM

VCM May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

4x WhlSpd

Diagnostics Ethernet

VT System

ethernet (UDP) SWA, VehSpd,brkpedpos

Brake clamping force (4x)

Rest Bus & I/O

FlexRay (Backbone) J1962 Diagnostic Connector

25

Outlook – Move from HIL to SIL Testing • Current focus for release testing is vehicle simulation and ECU-in-the-loop (HIL) • Future focus is on increased SIL testing (virtual ”ECU”) • Why: • Scalable: theoretically only limited to the number of computer nodes • Component black/grey-box testing possible

• How: • • • • •

Component (SWC:s) source code (as for target) built for Windows (desktop) and Linux (HPC cluster). Composition models with emulation and/or stubs of microcontroller abstraction layer (MCAL) Components and compositions as FMU’s according to the FMI standard Support for calibration using same variables/signals as for target ECU via XCP on Ethernet. SIL models (component and compositions) are automatically built at each CI loop.

• Challenges

• Tool chain (standardised?) for creating SIL models as described above • Tool chain for integration and simulation of (100+) SIL models (CANoe?)

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

26

conclusions • Virtual development methods are required to meet the challenges of the future • Unified / scalable simulation architecture in all phases of development (MIL, SIL, HIL) • Leverage strengths of virtual development

May 3, 2016

VECTOR TESTING SYMPOSIUM, STUTTGART, DR. MATTHIJS KLOMP, VOLVO CARS

27