Prognostics Health Management of Hybrid Powertrains4 · Prognostic Health Management of Hybrid Powertrains Dr Suresh Nayagam, ePHM Group, [email protected] February 2016

Prognostic Health Managementof Hybrid Powertrains

Dr Suresh Nayagam, ePHM Group, [email protected]

February 2016

www.cranfield.ac.uk/ivhmwww.cranfield.ac.uk/ivhmIVHM©2016 Cranfield University All Rights Reserved

Why a System’s FunctionalReliability Is Important

Toyota’s Sudden Acceleration Air Asia Crash

https://www.youtube.com/watch?v=cOWdWHSgI-4

www.cranfield.ac.uk/ivhmwww.cranfield.ac.uk/ivhmIVHM©2016 Cranfield University All Rights Reserved www.cranfield.ac.uk/ivhmwww.cranfield.ac.uk/ivhmIVHM©2016 Cranfield University All Rights Reserved

Traditional ReliabilityAssessment

From Field Data

Test in an Environmental Chamber


Product Service System(PSS)

Product Service System:“an integrated product and serviceoffering that delivers value in use”

GoldCare is a comprehensive recurring fleet maintenance &engineering management service which provides flexible solutionsfor Material Management, Engineering, and MaintenanceExecution.


Integrated Vehicle HealthManagement

...the transformation of system data into information tosupport operational decisions that results in:

• Minimised maintenance action/time

• Enhanced operational awareness

• Improved readiness and availability

• Reduced inspections and troubleshooting

• Reduced redundancies & design margins

• Reduced schedule interruptions

• More efficient logistics operations

• Reduced environmental impact

hence creating real business benefit

…multi-sector application:

• Aerospace• Marine• Automotive• Rail

• Energy• Health

• Agriculture• Manufacturing


Integrated Vehicle HealthManagement

• Production,certification & testing

• Total ownership costs• System & life cycle• Requirements• FMECAs• Design models• Failure modes/models• System test data

Design EngineeringManufacturing

• Maintenance Scheduling• Spares’ Supply• Asset Tracking• Maintenance Execution

Vehicle Maturation / New Product

• Operational Demand• Fleet Availability• MR & O leading

• Operational Schedule• Operational Effectiveness

Health Status

ActAcquire

Transfer

Sense

Health Status• Current• Predicted

Analyse

Asset

Data Repository& Ground Processing

Maintenance &Logistics

Operational Control


Unscheduled and ScheduledMaintenance

Embraer – AHEAD system


PowerDistributionUnit

HVBATTERY

ElectricalMotor

NeuralNetwork

Invertor

KalmanFilter

GearBoxWheels

PowerDissipation

CompactElectro-thermal

Rain FlowCountingAlgorithm

RUL

Torque

Ambient Temp

LifetimeModels

Lifetime Estimationof Power ElectronicModule

Cu

rrent

Vo

ltage

Reliability and Health Monitoringof Powertrains


Case Study 1: Power Module IGBT

Repeated heating and cooling leads to repetitive mechanical stress and eventual failure.

Exposure to sustained high temperatures drives diffusion-related mechanisms (creep,intermetallic growth, annealing).

Mismatch in CTE causes fatigue failure(de-bonding) of bond wires.

CTE mismatch causes fatigue failure atsoldered interfaces.

Heatsink

Thermal GreaseCopper baseplate

Substrate SolderSubstrate

DieSolder

Wirebond

Elements of the heat transfer path ofthe power electronic module

Silicon Die


Power Cycling Ageing Test

.

– On state collector emitter voltage (Vce) changes with different powercycling.

– The junction temperature and the collector-emitter are measured andrecorded constantly until the IGBT fails in accelerated ageing experiments.

– The failure mode involves wire bond lifting off and progressively endingbefore reaching the open circuit.


Data Clustering

IGBT Degradation Phase Duration

0 500 1000 1500 2000 2500 3000 3500 4000 4500 50001

2

3

4

5

6

7

8

9

10

Cycles (Times)

Vce

(Volts)


Parameter Optimization

Using analytical maximumlikelihood estimation (MLE)method to estimate best fitof the modelling parameter λ for Poisson distribution

�

��

��

�

� ��

MLE for Poisson Probability Distribution


IGBT failure model develops usingfuzzy nature of failure mechanism

�� is chosen as the precursor parameterfrom the experiment data and this providesthe best degradation indicator incomparison with other measurableparameters.

Junction temperature ( � ) resulting from

power switching is also used with fuzzysets.

ANFIS structure uses fuzzy Sugeno modelreasoning and all parameters are optimizedusing neural networks. Structure of the Proposed ANFIS Model


Case Study 2:Differential of anDrive Generator

• Function of Drive Generator:Transform mechanical power fromthe turbine into electrical AC powerfor the aircraft systems

• Function of the planetarytransmission:Transform a variable input speed(turbine) into a constant outputspeed (generator)

Integrated drive generator Constant Speed Drive


Background: Failure Modes

Metal-metal contact

• Failure mode: Lack of lubrication/overload + low speed

• Consequence: Metal-metal contact at the planetary bearing

• Current detection method: Temperature threshold at the sump


PbM: Metal-Metal ContactThermal Model

Calculate temperature distribution

– Temperature health indicator

– Temperature dependency of friction


Thermal Model - Results

• Temperature profiles for several loadcases (combination of angular velocityand normal load)

• Maximum temperatures

– Oil

– Bearing

– Gear

120˚C

100˚C

160˚C


Test Rigs to Provoke Failures

• Lubrication fully controlled (flow/pressure)

• Operational conditions:

– Input Speed: up to 3,000 rpm

– Crown gear speed: up to 6,000rpm

– Torque: max 276 Nm


ComponentComponent BoardBoard SubsystemSubsystem SystemSystem

Silicon Odometer-EnhancedComponentsSilicon Odometer-EnhancedComponents

Smart Driver & BoardSmart Driver & Board Smart Multi-Agent SystemConnected VehicleSmart Multi-Agent SystemConnected Vehicle

1- Design For Prognostics &Diagnostics2- Self-Healing3- Chip Variation

1- Design For Prognostics &Diagnostics2- Self-Healing3- Chip Variation

1- Integration of tests forlow power with test forhigh power2- Prognostic constraint

1- Integration of tests forlow power with test forhigh power2- Prognostic constraint

1- Design for Similarity2- Inference Engine3- Experience Sharing

1- Design for Similarity2- Inference Engine3- Experience Sharing

1- IVHM SpecificCommunication2- Middleware

1- IVHM SpecificCommunication2- Middleware

Technical LevelTechnical Level

ECUPower ManagementECUPower ManagementD

eliv

erab

leD

eliv

erab

leO

bje

ctiv

eO

bje

ctiv

eIVHM-Enabled AutonomousSystems


IVHM-EnabledAutonomous Systems


Future Vision (Our Challenge)

The car that looks after itself

Self-aware assets monitor current health, reliably predictingremaining useful life and automatically reconfiguring to optimiseand plan future MRO&L actions to minimise cost.


IVHM©2013 Cranfield University All Rights Reserved

Prognostics Health Management of Hybrid Powertrains4 · Prognostic Health Management of Hybrid Powertrains Dr Suresh Nayagam, ePHM Group, [email protected] February 2016

Documents