Wide Bandgap for Aerospace Applications Dr Suresh Perinpanayagam IVHM COE
Outline Overview of Cranfield Power Electronics Capabilities Towards All-Electric Aircraft SiC MOSFET Case Study Developing failure models for Wide Bandgap Prognostics development for Wide Bandgap 2
Welcome to Cranfield We are an exclusively postgraduate university that is a global leader for education and transformational research in technology and management. We provide: Impact and influence Premier learning Transformational research. 3
Impact and influence We work with 750+ businesses and governments around the world Aerospace: we have strategic relationships with global companies such as Airbus, BAE Systems, Boeing and Rolls-Royce Defence and Security: we are one of the world s largest providers of postgraduate defence and security education Environment: working with all UK water utility companies and advising major government departments such as Defra Leadership and Management: we work with major international businesses such as Jaguar Land Rover, L'Oréal and Shell, developing high performance leaders across the world Manufacturing: leading two and key partners in an additional four EPSRC Centres for Innovative Manufacturing. 4
Global reach 5
Cranfield s Centres Wide Bandgap Capabilitie s Power Electronics Group Cranfield Nano The Institute's world-class facilities include clean rooms, laboratories and test/fabrication services through to prototype component manufacture, with extensive analysis, modelling, synthesis and characterisation capability. 6
Towards All-Electric Aircraft Novel architecture for generation and distribution of loads Thermal management and thermal exchange Flight-proven electrical equipment systems, including environmental conditioning and protection 7
Electrical Distribution Architecture for Regional Aircrafts 8
Comparison between a Si IGBT and a SiC MOSFET A comparison of the overall losses between a Si IGBT and a SiC MOSFET with similar nominal ratings 9
10 Inverter for Ground Power Units 45 KW INVERTER DESIGN PARAMETERS Parameter Value Unit DC Link Voltage 500 VDC Output Voltage - Line to Line 232 VAC Output Frequency 400 Hz Power Factor 0.85 Nominal Output Power 45 kw 400% Overload value 180 kw Switching Frequency at Nominal Load 16 khz Switching Frequency at Overload 4 khz Heatsink Thermal Resistance 0.017 K/W Maximum Junction Temperature 150 ᵒC Ambient Temperature 40 ᵒC
Effect of Junction Temperature Junction temperatures of switching devices under nominal load and overload conditions. Note the reduction in temperature which corresponds with the reduction in switching frequency. 11
12 Comparison of Losses 67% reduction in losses can be expected. 1.6% increase in efficiency if the IGBT4 module with SiC diodes is used. 2.3% increase if the SiC MOSFET module is used.
Junction Temperature at 125% Overload Standard IGBT4 technology reaches its limit at a 55kW nominal load. SiC-based devices would be able to increase the nominal rating of the power assembly to 95kW, which represents a 72% increase in power density. SiC-based devices have the potential to double the power density 13
Power Electronics Reliability Studies 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 (debonding) of bond wires and soldered interfaces. 14
Physics-of-Failure Applied in Design Mission Profile I(t) (1)Thermal Profile Generator (2) Rain-Flow Analysis T interface (t) (3) Damage Profile Generator and Lifetime Prediction ΔT interface vs No. of Cycles Electrical systems: Trains, Planes & Automobiles Renewable power sources Power generation & distribution Industrial processes Power-electronics process Switching strategy Device electrical models Device thermal models Interfaces: Wire-bonds Chip solder Substrate solder Failure models POF/Empirical Statistical models Weibull 15
Electro-Thermal and Thermo Mechanical Models load current1 LOAD CURRENT Sine Wave1 0.5 Dutycycle L oad Current da Out D1 Rate Transition (1KHz)1 Power to Lower IGBT_ A (G2) LOW _SIDE IGBT _A (G2) Junction Temperature HIGH_SIDE Diode_A (D1) Junction Temperature G2 D1 LOW _SIDE IGBT _A (G2) SOLDER_2 Temperature Sine Wave2 Sine Wave2 original Sine Wave3 1 Constant1 temp D1 1-duty cycle dc temp D3 temp D4 Out D2 Out D3 Out D4 Out G1 Power to Upper DIODE_ A (D1) Power to Lower DIODE_ A (D2) LOW _SIDE DIODE _A (D2) Junction Temperature1 HIGH _SIDE IGBT_C (G3) Junction Temperature LOW _SIDE DIODE _C (D4) Junction Temperature2 HIGH_SIDE Diode_C (D3) Junction Temperature HIGH _SIDE IGBT_A (G1) Junction Temperature2 G2-SOLDER2 (Leg A LEFT) D2 G3 D4 D3 G1 temp D2 HIGH _SIDE DIODE _A (D1) SOLDER_2 Temperature Sine Wave4 tempg1 Out G2 Power to Upper IGBT_ A (G1) HIGH _SIDE IGBT _A (G1) SOLDER_2 Temperature LOW _SIDE DIODE _A (D2) Solder_2 Temperature G1-SOLDER2 leg A D1-SOLDER2 leg A tempg2 Out G3 Power to Upper IGBT_ C (G3) UPPR _SIDE_3a SOLDER_1 Temperature D2-SOLDER2 leg A UPPER SOLDER1-a LegA temp G3 LOWER_SIDE_3b SOLDER_1 Temperature Out G4 temp G4 BASE PLATE-A Temperature LOWER SOLDER1-b LegA AVERAGE INPUT POWER (W) PROCESS 2 OK Power to Upper DIODE_ C (D3) LOW _SIDE IGBT _C (G4) Junction Temperature Base plate- Leg A HIGH _SIDE IGBT _C (G3) SOLDER_2 Temperature LOW _SIDE IGBT _C (G4) SOLDER_2 Temperature1 G4 G3-SOLDER2 Leg B In3 HIGH _SIDE DIODE _C (D3) SOLDER_2 Temperature G4-SOLDER2 Leg C RIGHT Repeating Sequence Interpolated LOW _SIDE DIODE _C (D4) Solder_2 Temperature D3-SOLDER2 Leg B Power to Lower IGBT_ C (G4) UPPR _SIDE_3a SOLDER_1 (C) Temperature D4-SOLDER2 Leg B UPPER SOLDER1-a Leg C LOWER_SIDE_3b SOLDER_1 (C) Temperature BASE PLATE (C) Temperature LOWER SOLDER1-b Leg B Power to Lower DIODE_ C (D4) BASE PLATE-A IR Temperature 1 Base plate- IR Leg A Base plate- Leg B BASE PLATE (C) -IR Temperature 1 Determined Temparature Base plate-ir Leg B Rate Transition (5KHz) 2 Rate Transition (5KHz) Rate Transition (5KHz) 1 Thermo-Mechanical Model Electro-Thermal Model 16
Life-Time Models Wire bond wear-out models N f (1.4*10 11 ) T 3.597 N f is the material number of cycles to failure ΔT is temperature variation. Reliability life-time models for IGBT bond wire interconnect. 17
IGBT Reliability Diode (D1) IGBT (G2) Infra-Red Measurements CEDIP Titanium high frame rate camera Comparisons between Junction Temperature Estimates and Measurements 18
Integrated Vehicle Health Management 19 Vehicle Maturation/New Product Sense Design Engineering Manufacturing Maintenance & Logistics Production, certification & testing Total ownership costs System & life cycle Requirements FMECAs Design models Failure modes/models System test data Operational Demand Fleet Availability MR & O leading Maintenance Scheduling Spares Supply Asset Tracking Maintenance Execution Operational Control Operational Schedule Operational Effectiveness Health Status Act Asset Health Status Current Predicted Acquire Transfer Analyse Data Repository & Ground Processing
Power Module IGBT Repeated heating and cooling leads to repetitive mechanical stress and eventual failure. Exposure to sustained high temperatures drives diffusionrelated mechanisms (creep, intermetallic growth, annealing). Mismatch in CTE causes fatigue failure (de-bonding) of bond wires. CTE mismatch causes fatigue failure at soldered interfaces. Thermal Grease Heatsink Die Solder Silicon Die Wire bond Substrate Substrate Solder Copper base plate Elements of the heat transfer path of the power electronic module 20
Power Cycling Ageing Test The power cycling ageing test provides monitoring and measurements of temperature and electrics. On-state collector emitter voltage (Vce) changes with different power cyclings. The junction temperature and the collector-emitter are measured and recorded constantly until the IGBT fails in accelerated ageing experiments. The failure mode involves wire bond lifting off and progressively ending before reaching the open circuit. The Vce (on-state) parameter indicates any increases in a non-monotone fashion and shows discrete steps with noise invasion until the IGBT fails. The Vce (on-state) voltage precursor indicates a sudden fall at the end of the ageing process when the IGBT fails after more than 4,500 time units. 21
Vce (Volts) Noise Filtering 2.4 2.3 IGBT 1 IGBT 2 IGBT 3 IGBT 4 2.2 2.1 2 1.9 500 1000 1500 2000 2500 3000 3500 4000 4500 Cycles (Times) All IGBT run-to-failure data sets after filtering First IGBT data set after filtering 22
Vce (Volts) ata Clustering TABLE III: IGBT Degradation Phase Duration 10 Classtring data 9 8 7 6 5 4 3 2 1 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Cycles (Times) 23
Parameter Optimization Using analytical maximum likelihood estimation (MLE) method to estimate best fit of the modelling parameter λ for Poisson distribution MLE for Poisson Probability Distribution P(x i λ) = λx i. e λ x i! λmle = 1 n, x i 0 N n=1 x n 24
Results from RUL Estimation Using Weibull Model RUL of IGBT IGBT 14000 12000 10000 Real RUL Estimated Mean RUL 8000 6000 4000 2000 0 0 1000 2000 3000 4000 5000 Life of IGBT Constructed Markov model & Monte Carlo simulation for RUL prediction 12000 10000 8000 Real RUL Estimated Mean RUL 25
Conclusions Reliability and maintenance-free power solutions are important for wide adaption of Wide Bandgap technology. Cranfield could assist industry to develop reliability models for Wide Bandgap applications. Cranfield could develop prognostics capabilities for Wide Bandgap applications. This will enable safety-conscious industry, such as electric aircraft, to adapt Wide Bandgap. 26