bio-medical rf simulations with cst microwave studio
TRANSCRIPT
www.cst.com | May-10
Bio-Medical RF Simulations
with CST Microwave Studio®
Biological Models
Specific Absorption Rate (SAR)
Bio-Medical Examples
www.cst.com | May-10
Biological Models
The right choice of the biological model is essential for the reliability
of a SAR or EMI simulation.
Visible Human
voxel data
SAM Phantom,
homogeneous
models
other voxel data
CST Voxel Family
www.cst.com | May-10
• New .obj import allows import of biological models, e.g. from
Poser® 8 (http://my.smithmicro.com/win/poser/index.html)
• For most high frequency applications fully sufficient
• Simulate much faster then voxel models
Homogeneous Hand/Body Models
www.cst.com | May-10
SAM - Standard Anthropomorphic Model
www.sam-phantom.com
Frequency dependent material
properties (according to
standard) can be modelled by
dispersive materials via
tabulated input.
Only one simulation run for
broadband results!!
tissuesimulantliquid
(TSL)
plastic shell• Originally created for measurements
• Shape specified in IEEE/CENELEC/IEC
standards
• Filled with homogeneous glycol-containing
tissue-simulant liquid, only two materials
for simulation
• Virtual prototyping through simulation
www.cst.com | May-10
Hand monoblock
Hand fold
Hand narrowdata
Hand PDA
CTIA Hand Models
www.cst.com | May-10
CST Voxel Family
KATJA
(pregnant)
LAURA GUSTAV
BABY
CHILD
EMMADONNA
www.cst.com | May-10
CST Voxel Family
Macros -> Solver -> Calculate
Human Material Properties
www.cst.com | May-10
Available in
different
resolutions
Materials of
interest can be
chosen
Visible Human Project
produced by the National Library of
Medicine (NLM), Maryland
http://www.vr-laboratory.com/
HUGO
www.cst.com | May-10
Cole-Cole-Materials
www.cst.com | May-10
SAR: Overview and Background
SAR – Specific Absorbtion Rate
Unit of SAR: W/kg
P: Power loss density
E: Electric field strength
J: Current density
s: Conductivity
r: DensityTypically averaged over pre-defined mass
A measure for electromagnetic energy absorbed by biological
tissue mass when exposed to radiating device (e.g. mobile phone)
22
22 JEPSAR
www.cst.com | May-10
Averaging Procedure
1. Point of avg. SAR calculation
2. Search for 10 g cube (iteratively)
3. Integrate losses in cube
At boundary treatment depends on
chosen averaging standard:
IEEE C95.3, IEEE 1528.1, CST C95.3
CST legacy
The „constant volume“ assumption uses an averaged cube size:
- Faster (no iterative search for cube with correct mass)
- Only approximative (not according to official SAR standard)
www.cst.com | May-10
• Several guidelines and standards specify SAR safety limits (i.e.
ICNIRP).
• Standards like IEEE 1528 regulate measurement methods for
practical assessment of compliance.
• A simulation standard IEEE 1528.X is in development
• 1528.1 requirements for hexahedral time domain codes (end
of 2010)
• 1528.2 application to cars with passenger/bystander (~2011)
• 1528.3 application to mobile phones near head (~2011)
• 1528.4 requirements for tetrahedral frequency domain codes
• CST participates in standards committee.
• IEEE C95.3 Annex E specifies SAR averaging scheme for
simulation.
• CST MICROWAVE STUDIO® has already been approved by the FCC
(USA) to comply with hex td standard drafts.
SAR Standards under Development
www.cst.com | May-10
Visualization of SAR
2D or 3D plot including information about position of the maximum
www.cst.com | May-10
Visualization of Max. SAR Cube
www.cst.com | May-10
Re(er) Im(er)
0.9 GHz 41.5 17.98 (= 0.9 S/m)
1.8 GHz 40.0 13.98 (= 1.4 S/m)
Dispersive Broadband Simulation
Typical requirement for dual band phones:
Frequency dependent
material definition:
Second order dispersive
fit for tabulated values,
only one simulation run
required
www.cst.com | May-10
0.9 GHz, 1g 1.35 1.31 1.74
0.9 GHz, 10g 0.96 0.93 1.13
1.8 GHz, 1g 0.69 1.32 1.32
1.8 GHz, 10g 0.99 0.83 0.83
Dispersive Broadband Simulation
Compared material settings:
Constant settings for 0.9 GHz
sim. time 45 min.
Constant settings for 1.8 GHz
sim. time 45 min.
Dispersive broadband fit
total sim. time: 57 min.
Dispersive fit agrees very well for S-Parameter and SAR
values in both bands for only 25% extra simulation time
S-Parameter comparison:
SAR value comparison:
www.cst.com | May-10
Measured vs. Computed SAR Distribution
0 dB = 2.8 W/kg
Example: 7T MRI endorectal coil
MeasurementSimulation
Overall:
SAR computed
SAR measured 1.08 – 1.15
Courtesy of Erwin L. Hahn Institute Essen,Germany
www.cst.com | May-10
Magnetic Resonance Imaging (MRI)
Three EM-fields needed for imaging
• STRONG magnetostatic field (human: 1 – 9.4
Tesla, up to 21 T for animals)
Mostly superconducting magnets, aligning the
spinning protons
-> M-Statik Solver
• Gradient field for positioning (in kHz range)
-> Magneto-Quasistatik Solver, LT-Solver
• HF field to excite spinning protons and
receive relaxation signal (60 – 500 MHz)
Rotating B-Field most interesting (B1+)
-> Both T- and F-Solvers are of intererst!
Most interesting for MRI R&D
www.cst.com | May-10
Design Challenge: Increase SNR of image
For 7T MRI -> fres = 297 MHz -> lbody ~ 13 cm
-> It is difficult to obtain homogeneous field
distribution inside body, specialized coils
required
Safety issue: SAR ~ fres2
-> SAR critical for higher fres
-> Alternative: queck directly body
temperature increase, bioheat solver!
SNR ~ static biasing field ~ spin resonance frequency fres
Advantages of CST: Complete Technology, Static, LF, T, F and
bioheat solvers in one frontend, Voxel Family, fast SAR, etc…
www.cst.com | May-10
Courtesy of Erwin L. Hahn Institute Essen,Germany
8 Channel Head Coil
www.cst.com | May-10
8 Channel Head Coil
arg(B1+)
Vs/m²
|B1+| SAR voxel
SAR 10g
[°]
www.cst.com | May-10
8 Channel Head Coil
SAR10g
location of max. SAR10g in
left shoulder for off-centre
position of head
max. perm. power = 23 W
(CW)
location of max. SAR on
left side of the head
most critical aspectSAR10g SAR10g SARhead
max. perm. power25 W 27 W 33 W
www.cst.com | May-10
• loops overlapped and shifted
• 70 cm cable length
• box with TR-switches + pre-amps
Spine Loop Array
z
x
43 cm
20 cm
www.cst.com | May-10
Comparison to Measurement
Measurement
max B1+ = 15.9 µT
Simulation
max B1+ = 13.5 µT
www.cst.com | May-10
SAR Compliance
critical aspect: localized SAR (10g averaged)
www.cst.com | May-10
Microwave Breast Cancer detection
Dr. Maciej Klemm, Electromagnetics Group, Centre for Communications
Research (CCR), University of Bristol, United Kingdom
e-mail: [email protected]
www.cst.com | May-10
Model setup and clinical results
• dipole antennas
• dispersive tissues
• inhomogeneous breast !
• model 30-40M cells
• full imaging (30 simulations)
takes about 10h (hardware
accelerated; 4 GPU cards)
www.cst.com | May-10
Pace Maker Simulation
T-Solver
F-SolverT-Solver
Complete Technology:
www.cst.com | May-10
Results at 400 MHz
Inside biological tissue phantom
Averaging Cube for
max SAR
E-Field
SAR
www.cst.com | May-10
Pacemaker inside Human Body Model
www.cst.com | May-10
Cardiac Pace Maker
Courtesy of Lehrstuhl für Theoretische Elektrotechnik, Bergische Universität Wuppertal, Germany
Frequency dependent field coupling into a Cardiac Pace Maker (CPM)
www.cst.com | May-10
BABY besides Baby-Phone
Stimulated power: 500 mW at 865 MHz
Max. SAR value (averaged over 10g): 0.02 W/kg
(well below accepted maximum of 2 W/kg for public exposure)
Courtesy of Lehrstuhl für Theoretische Elektrotechnik, Bergische Universität Wuppertal, Germany
www.cst.com | May-10
New CST Examples!
Can only be opened by customers who have
-Voxel Import
- BioModel License
-> offer for evaluation license!!
www.cst.com | May-10
CST STUDIO SUITE offers a wide range of tools for
bio-medical simulations (MRI, cancer treatment,
diathermy, implants, etc.)
Both flexible homogeneous and detailed voxel models
are available
„Complete Technology“ allows combined simulations
from static to GHz including circuit simulation
SAR and Bio-Thermal simulations help to improve
performance and safety of medical devices
Summary
www.cst.com | May-10
Appendix