thermal transient simulation with ansys software · pdf filethermal transient simulation with...
Post on 12-Feb-2018
389 Views
Preview:
TRANSCRIPT
© 2011 ANSYS, Inc. June 26, 2014 1
Thermal Transient Simulation with ANSYS Software – Case Studies
Sandeep Sovani, Ph.D.
Manager, Global Automotive Strategy
ANSYS Inc, Ann Arbor, MI, USA
Xiao Hu, Ph.D.
Principal Engineer
ANSYS Inc, Ann Arbor, MI, USA
© 2011 ANSYS, Inc. June 26, 2014 2
New Paradigm 1 – Multidomain Simulation
Phenomena at one level affect those at other levels and need to be simulated in simultaneous co-simulation
Electrode Level
•Electrode layout •Manufacturing process development •Life
Molecular Level
•Material innovation •Material selection
Cell Level •Charging, dischar-ging profiles •Heating •Safety under abuse •Swelling, deformation
Pack Level •Thermal Mgmt •BMS Logic •Safety •Durability •NVH •EMI/EMC
Powertrain and Vehicle Level
•System Integration
Smal
l Sca
le
Larg
e S
cale
© 2011 ANSYS, Inc. June 26, 2014 3
A Total Li-Ion Battery Simulation Solution
© 2011 ANSYS, Inc. June 26, 2014 4
Newman’s 1D Electrochemistry Model Electrode Level
Manufacture Material selection Electrochemistry
Starting from the electrode level . . .
3D Electrochemistry Simulation
Multidomain Simulation
© 2011 ANSYS, Inc. June 26, 2014 5
Multidomain Simulation – Electrode Level
Li concentration in electrodes during discharge
ANSYS Case Study – 3D Electrochemistry Modeling
© 2011 ANSYS, Inc. June 26, 2014 6
ANSYS Simplorer’s Results
x=0
x=Lp+Ls+
Ln
x=Lp+Ls
x=Lp
1/10 C
1/2 C
1 C
2 C
4 C
6 C
8 C
10 C
White et al’s Results
Multidomain Simulation – Electrode Level
ANSYS Case Study – Pseudo 2D Electrochemistry Newman Model – Quantitative Validation
Long Cai, Ralph E. White, Journal of Electrochem. Soc., 156 (3), A154-A161 (2009)
Xiao Hu, Shaohua Lin, and Scott Stanton, SAE 2012-01-0665
© 2011 ANSYS, Inc. June 26, 2014 7
ANSYS Simplorer’s Results White et al’s Results
Multidomain Simulation – Electrode Level ANSYS Case Study – Pseudo 2D Electrochemistry Newman Model – Quantitative Validation
Long Cai, Ralph E. White, Journal of Electrochem. Soc., 156 (3), A154-A161 (2009)
Xiao Hu, Shaohua Lin, and Scott Stanton, SAE 2012-01-0665
© 2011 ANSYS, Inc. June 26, 2014 8
Newman P2D Electrochemistry Model in Fluent (CAEBAT project)
0 Li
s j
r
cr
rr
D
t
c sss 2
2
0ln Li
eDe jckk
Li
eeee j
F
tcD
t
c
1)(
RT
F
RT
Fij ca
a
Li expexp0
Domains
• negative electrode
• separator
• positive electrode
• spherical particles
A system of DAEs are solved for every CFD cells:
- Total number of equations: (n_Cs+2) *n_NE + 2n_SP+(N_Cs+2) n_PE
- Memory requirements: N_cell*{(n_Cs+1) *n_NE+ n_SP +(N_Cs+1) n_PE)}
L. Cai and R.E. White, “Reduction of Model Order Based on Proper Orthogonal Decomposition for
Lithium-Ion Battery Simulations” J. of Electrochemical. Soc. 156(3) A154-A161 (2009).
© 2011 ANSYS, Inc. June 26, 2014 9
Temperature Results Based on Electrochemistry
-110
-90
-70
-50
-30
-10
10
30
50
0 200 400 600 800 1000 1200 1400
Cu
rre
nt
(A)
Time (s)
© 2011 ANSYS, Inc. June 26, 2014 10
. . . incorporating electrode level effects into the cell level . . .
Cell Thermal Model with Electrochemistry Cell Level Electrical Model Heat Generation Abuse Nail Penetration Crush
Multidomain Simulation
© 2011 ANSYS, Inc. June 26, 2014 11
Cell Level – Equivalent Circuit Model (ECM)
Battery Equivalent Circuit Model (ECM)
Simulation Testing
X. Hu, L. Collins, S. Stanton, S. Jiang, "A Model Parameter Identification Method for Battery Applications", SAE 2013-
01-1529.
© 2011 ANSYS, Inc. June 26, 2014 12
• 28 cells are connected in a module
• 4 modules are connected to the final configuration
Module/Pack Level – ECM
X. Hu, L. Collins, S. Stanton, S. Jiang, "A Model Parameter Identification Method for Battery Applications", SAE 2013-
01-1529.
Battery Pack ECM
Simulation Results
© 2011 ANSYS, Inc. June 26, 2014 13 X. Hu, L. Collins, S. Stanton, S. Jiang, "A Model Parameter Identification Method for Battery Applications", SAE 2013-
01-1529.
Cell Level – ECM Extraction Tool
ECM Extraction Toolkit
ECM Model Workflow
© 2011 ANSYS, Inc. June 26, 2014 14
Module Level – CFD Thermal
Temperature Distribution
Temperature and Pathlines Temperature Distribution
Temperature Distribution
© 2011 ANSYS, Inc. June 26, 2014 15
Battery Thermal Behavior via ROM
Type 1: Thermal Network
• Careful calculation and calibration needed
• Accuracy compromises
Type 2: LTI – state space
• Can be as accurate as CFD
• No calibration
LTI t t Step
Input
Step
Respons
e
1
© 2011 ANSYS, Inc. June 26, 2014 16
1. Create the CFD model
2. Generate step responses Icepak has specialized tools
Other CFD codes are fine
4. Simulate inside Simplorer
3. Extract LTI ROM Use Simplorer
LTI ROM for Battery Cooling ANSYS CFD
© 2011 ANSYS, Inc. June 26, 2014 17
LTI ROM gives the same results as CFD.
LTI ROM runs in less than 2 seconds while
the CFD runs 2 hours on one single CPU.
X. Hu, S. Lin, S. Stanton, W. Lian, “A Foster Network Thermal Model for HEV/EV Battery Modeling,” IEEE
TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011
X. Hu, S. Lin, S. Stanton, W. Lian, “A State Space Thermal Model for HEV/EV Battery Modeling", SAE 2011-01-1364
Heat Profile
LTI ROM for GM Battery Module
© 2011 ANSYS, Inc. June 26, 2014 18
Module Level – ROM for Thermal
ROM vs CFD ROM for the Battery Module
© 2011 ANSYS, Inc. June 26, 2014 19
SVD+LTI technology allows for quick temperature distribution calculation in addition to average temperature calculation.
Using a heat source from GM, SVD+LTI ROM is applied to the GM 16 cell case.
SVD+LTI ROM : GM 16 Cell Test Case
Heat source used
© 2011 ANSYS, Inc. June 26, 2014 20
1. Create the Fluent CFD model
2. Generate step responses
3. Extract SVD+LTI ROM • Use Simplorer
4. Simulate inside Simplorer
SVD+LTI ROM for Battery Cooling
ANSYS CFD
© 2011 ANSYS, Inc. June 26, 2014 21
SVD+LTI ROM Validation: GM 16 Cell Test Case
CFD (200 sec)
SVD+LTI ROM (200 sec)
CFD (400 sec) CFD (600 sec) CFD (800 sec)
SVD+LTI ROM (400 sec)
SVD+LTI ROM (600 sec)
SVD+LTI ROM (800 sec)
© 2011 ANSYS, Inc. June 26, 2014 22
SVD+LTI ROM Statistics for GM 16 Cell Test Case
CFD mesh size 3,025,067
Number of steps in step response run 180
Size of matrix A 3,025,067×180
Number of snap shots used for SVD calculation 180
SVD calculation time 4 min 50 sec
SVD calculation memory usage 8.1 G
SVD+LTI ROM extraction time 5 min
CFD validation simulation time on single CPU 5 hr
SVD+LTI ROM simulation time on the same
single CPU
a few seconds
© 2011 ANSYS, Inc. June 26, 2014 23
SVD+LTI ROM Validation
CFD (10 sec)
SVD+LTI ROM (10 sec)
CFD (20 sec) CFD (30 sec) CFD (40 sec)
SVD+LTI ROM (20 sec)
SVD+LTI ROM (30 sec)
SVD+LTI ROM (40 sec)
© 2011 ANSYS, Inc. June 26, 2014 24
SVD+LTI ROM Statistics for the Test Case
CFD mesh size 12,209,486
Number of steps in step response run 120
Size of matrix A 12,209,486×120
Number of snap shots used for SVD calculation 60
SVD calculation time 3 min
SVD calculation memory usage 18 G
SVD+LTI ROM extraction time 5 min
CFD validation simulation time on six CPUs 20 hr
SVD+LTI ROM simulation time single CPU a few seconds
© 2011 ANSYS, Inc. June 26, 2014 25
ECM calculates heat source and sends it to the two ROMs.
LTI ROM calculates average temperature and sends it to ECM.
SVD+LTI ROM calculates temperature distribution.
GM Battery Module – ECM Coupled with ROMs
© 2011 ANSYS, Inc. June 26, 2014 26
ROM Results
LTI ROM calculates average temperature.
SVD+LTI ROM calculates temperature field.
Average Temperature Heat Source for One Cell
© 2011 ANSYS, Inc. June 26, 2014 27
Module/Pack Level – Bus Bar Model
Electromagnetic FEA Analysis for Busbar RLC Network Extraction
© 2011 ANSYS, Inc. June 26, 2014 28
Full Battery Simulation
Voc = -1.031*exp(-35*(abs(IBatt.V/Vinit))) + 3.685 +
0.2156*(abs(IBatt.V/Vinit)) - 0.1178*(abs(IBatt.V/Vinit))^2 +
0.3201*(abs(IBatt.V/Vinit))^3 +
0.3/30.0*(U1.Temp_block_1-273)
© 2011 ANSYS, Inc. June 26, 2014 29
Full Battery Simulation
© 2011 ANSYS, Inc. June 26, 2014 30
Conclusions
Battery is a multi-scale multi-physics application.
ANSYS provides tools for all aspects of battery simulation.
Furthermore, ANSYS integrates different models into battery system level simulation through model order reduction.
© 2011 ANSYS, Inc. June 26, 2014 31
© 2011 ANSYS, Inc. June 26, 2014 32
© 2011 ANSYS, Inc. June 26, 2014 33
© 2011 ANSYS, Inc. June 26, 2014 34
© 2011 ANSYS, Inc. June 26, 2014 35
© 2011 ANSYS, Inc. June 26, 2014 36
© 2011 ANSYS, Inc. June 26, 2014 37
© 2011 ANSYS, Inc. June 26, 2014 38
© 2011 ANSYS, Inc. June 26, 2014 39
© 2011 ANSYS, Inc. June 26, 2014 40
© 2011 ANSYS, Inc. June 26, 2014 41
© 2011 ANSYS, Inc. June 26, 2014 42
© 2011 ANSYS, Inc. June 26, 2014 43
© 2011 ANSYS, Inc. June 26, 2014 44
© 2011 ANSYS, Inc. June 26, 2014 45
© 2011 ANSYS, Inc. June 26, 2014 46
top related