Stefano Longo Lilantha Samaranayake Daniel Auger
LOOSELY COUPLED HARDWARE IN THE LOOP
TESTING FOR ELECTRIC VEHICLE POWERTRAINS
CRANFIELD UNIVERSITY
BIGGER PICTURE
Developing FUTURE Vehicles Fundamental Understanding of Technologies for Ultra Reduced Emission Vechicles
FUTUREVehicles
Three work packagesThree cross-cutting scientific themes
Cross-cutting Theme 1: Thermal analysis, failure and degradation diagnosis and prognosis
Energy flow
Information flow
Cross-cutting Theme 2: Reduced order models
Cross-cutting Theme 3: Limits of performance by application of advanced control
Work Package 1: chemical energy storage systems plant modelling
Work Package 2: electrical energy
converter systems plant
modelling
Work Package 3: reduced order modelling and
controls
Unified modelling framework
Optimisation and control design
Fuel cells, batteries and
supercapacitors Electrical machinesPower electronics
BIGGER PICTURE
Developing FUTURE Vehicles Fundamental Understanding of Technologies for Ultra Reduced Emission Vechicles
FUTUREVehicles
Three work packagesThree cross-cutting scientific themes
Cross-cutting Theme 1: Thermal analysis, failure and degradation diagnosis and prognosis
Energy flow
Information flow
Cross-cutting Theme 2: Reduced order models
Cross-cutting Theme 3: Limits of performance by application of advanced control
Work Package 1: chemical energy storage systems plant modelling
Work Package 2: electrical energy
converter systems plant
modelling
Work Package 3: reduced order modelling and
controls
Unified modelling framework
Optimisation and control design
Fuel cells, batteries and
supercapacitors Electrical machinesPower electronics
OBJECTIVES
Minimise components degradation via control design
Hardware-in-the-loop testing
Minimise components degradation via control design
Hardware-in-the-loop testing
MINIMISE COMPONENTS DEGRADATION VIA CONTROL DESIGN
Batt. SoC
Sup. Cap SoC
MINIMISE COMPONENTS DEGRADATION VIA CONTROL DESIGN
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
MINIMISE COMPONENTS DEGRADATION VIA CONTROL DESIGN
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
MINIMISE COMPONENTS DEGRADATION VIA CONTROL DESIGN
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
Results
Conclusions
x�Degradation with Optimal control is lower compared to Maximum Torque Per Ampere control
x�Lower overshoot and settling time in Torque transients
x�Both transient and steady state voltage and current demands are lower; which allows
��Smaller battery��Smaller converter��Lower battery degradation��Lower cooling requirements
x�However, high computation power is required to complete real-time control law calculation within the sampling interval (0.5 ms).
x�Output quality depends on model accuracy.
Max. Torque / Ampere (MTPA) controlMethodology
Optimal Control for Minimizing Degradation in Electric Motors
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
Objectives
Use multi objective optimization containing Closed Loop performance (JCL) and Degradation (JDEG) cost functions in deriving the optimal control law for Field Oriented Control of Permanent Magnet Synchronous Motors (PMSM) in Electric Vehicles.
Optimal Controller
IPMSM Electrical Circuit
IPMSM Electromagnetic
Circuit
vd, vq id, iq eWmWrefZVehicle Dynamics
>> Read Motor output>>Calculate Optimal Control Law using �Reduced order Motor Model to Minimize
Cost Function
�Predicted torque profile �i.e., Model Predictive Control
>>Apply First Step of Optimal Control Law>> Repeat
� �1CL DEGJ J JD D � �
oqqe iLZldL mdLR
cRdv
di odi
cdi
pmeOZ
odde iLZlqL mqLR
cRqv
qi oqi
cqi
d-axis
q-axis
d
q
pmO
sO
oqmqiL
odmdiL
si
odi
oqioqmqe iLjZ
odmde iLjZ
sv J
Worst possible transient from maximum positive torque to minimum negative torque transient is presented.
<< Comparison of motor degradation with Optimal Control & MTPA Control
Comparison of input voltages in the d-axis and q-axis as calculated by the Optimal Control & MTPA Control
<< Comparison of motor currents on d-axis and q-axis with Optimal Control and MTPA Control
Motor Data
Optimal Control Algorithm
>>
OBJECTIVES
Minimise components degradation via control design
Hardware-in-the-loop testing
OBJECTIVES
Minimise components degradation via control design
Hardware-in-the-loop testing
HIL TEST RIG ExperimentsTest-Rig
Test Rig for Electric Vehicle Powertrain Optimization
Stefano Longo, Lilantha Samaranayake, Daniel AugerCentre for Automotive Engineering
FUTURE Vehicles
ObjectivesDesign and develop a test-rig which enables experimental verification of control techniques designed for powertrain components such as the motor, battery and supercapacitors in order to have
9�Improved closed loop control performance9�Minimized component degradation
DC Power linesAC Power linesSignal lines
3 Phase Brushless DC
Motor
DC3 Phase Brushless DC GeneratorDC
3 Ph. AC
3 Ph. AC
ACGrid
AC
DC
DC
DC
DC
DC
Cell/Module/Battery under test
Supercapacitors under test
Torque Sensor
Torque & Shaft SpeedRead: Voltage & Current
Write: Torque reference Read: Voltage & CurrentWrite: Load reference
HIL System
Battery State of Charge
Supercapacitor State of Charge
Power DemandPo
wer
/ 10
(W)
1. Degradation cost function evaluation based on power measurements using��Cumulative Work Ratio (CWR)��Cumulative Loss Ratio (CLR)��Cumulative Input Energy Ratio
(CIER).
2. Closed loop speed control loop performance of nonlinear 3 Phase Brushless DC motor plant with ��PID��Type 1 Fuzzy Logic Control
(T1-FLC) ��Type 2 Fuzzy Logic Control
(T2-FLC).
3. Optimal control to minimize battery degradation using battery-supercapacitor hybrid energy storage.
Experiment 1Power Demand
Pow
er /
10 (W
)
-18.1
-38.3
-12.9
-43.2-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0NEDC ARTEMIS
Percentage reduction in total energy demandPercentage reduction in peak power demand
Experiment 2
Experiment 3
(A)
(C)
(B)
(A) Experiment with combined ARTEMIS drive cycleTracking accuracy of CASE 1 > CASE 2 > CASE 3.Integral Absolute Error (IAE) CASE 1 < CASE 2 < CASE 3(B)Degradation cost functions, CWR, CLR & CIER CASE 1 > CASE 2 > CASE 3
(C)Same torque demand in all CASEs
CONCLUSIONS�CWR, CLR & CIER can successfully represent degradation.�They can be used as degradation cost functions in control design.
� �
³
¦³
¸̧¹
·¨̈©
§�
life
ifinal
iinitial
t
ratedm
N
i
t
t ie
dtP
dttPCIER
0 _
1_
11
_
_
K
� � � � � �� �
³
¦³
¸̧¹
·¨̈©
§�
�
life
ifinal
iinitial
t
ratedm
N
i
t
t imimie
dtP
dtttTtPCLR
0 _
1___
11
_
_
K
Z � � � �
³
¦³
life
ifinal
iinitial
t
ratedm
N
i
t
t imim
dtP
dtttTCWR
0 _
1__
_
_
Z
CONCLUSIONS�Peak power demand
T2-FLC < T1-FLC < PID�Total energy demand
T2-FLC < T1-FLC < PID
Battery – Supercapacitor hybrid optimal control gives the least battery degradation
HIL TEST RIG
IN SUMMARY
Development of control algorithms to mitigate
components degradation
IN SUMMARY
Development of control algorithms to mitigate components
degradation
IN SUMMARY
Optimal power flow control impacts on
degradation
Development of control algorithms to mitigate components
degradation
IN SUMMARY
Development of control algorithms to mitigate components
degradation
Optimal power flow control impacts on
degradation
IN SUMMARY
Development of control algorithms to mitigate components
degradation
Optimal power flow control impacts on
degradation
A HIL rig to evaluate, test and optimise electric
powertrains
IN SUMMARY
Development of control algorithms to mitigate components
degradation
A HIL rig to evaluate, test and optimise
electric powertrains
Optimal power flow control impacts on
degradation
IN SUMMARY
Development of control algorithms to mitigate components
degradation
A HIL rig to evaluate, test and optimise
electric powertrains
Optimal power flow control impacts on
degradation
HIL components under test do not need to match
in terms of power
IN SUMMARY
Development of control algorithms to mitigate components
degradation
A HIL rig to evaluate, test and optimise
electric powertrains
Optimal power flow control impacts on
degradation
HIL components under test do not need to match in terms of
power
THANK YOU