Upgrade to Pro — share decks privately, control downloads, hide ads and more …

maia's phd summary for @hopperties, march 2013

maia
March 12, 2013

maia's phd summary for @hopperties, march 2013

maia

March 12, 2013
Tweet

More Decks by maia

Other Decks in Research

Transcript

  1. ROBOTS!!
    !
    !
    OR

    HOW I LEARNED TO GIVE UP AND FINISH MY PHD
    maia sauren

    View Slide

  2. View Slide

  3. 1 10 100 1 10 100 1 10 100 1 10 100 1 10 100


    Hz kHz MHz GHz THz
    Radiofrequency

    radiation
    Ionising radiation
    UV
    Infrared

    radiation
    visible light
    50 Hz

    Power

    Lines
    AM

    radio
    FM

    radio
    Mobile

    Phones
    TV
    Satellite X rays Gamma rays
    Lasers

    View Slide

  4. RF  SAFETY  STANDARDS
    • ICNIRP or IEEE recommendations
    • Different standards in different countries (humans aren’t so different)–
    no difference at 900 MHz

    View Slide

  5. RF SAFETY STANDARDS
    • Basic restrictions vs. fundamental limits
    • ICNIRP basic restrictions: mandatory limits on direct modes
    of RF interaction with the body
    • Current density (mA/cm²)
    • Whole Body Averaged SAR (W/kg)
    • Localised SAR in the head and torso (W/kg)
    • Localised SAR in the limbs (W/kg)
    • Specific absorption per pulse (J/kg)
    • Power flux density (W/m²)

    View Slide

  6. THIS WORD, ‘SAFE’...
    “4 W/kg of whole body exposure for ~30 minutes
    produces a temperature increase of 1oC, a level at
    which established biological effects begin to take
    place”
    (Health Phys, ’98)
    • safety factors
    •Based on conservative and crude assumptions
    •Too conservative = uneconomic

    View Slide

  7. REAL- TIME EXPOSURE OF LIVING HUMANS
    • Can’t measure absorption, only exposure
    • Dead tissue? (Surowiec, Stuchley et al 1985)
    • Animals? (Forster 1979, Schwan 1989)
    • ...

    View Slide

  8. Absorption is based on: a) size b)
    shape c) dielectric properties of
    tissues

    • How much do these vary in
    humans?

    • How much effect do these
    variations have on absorption?

    • Do some factors affect absorption
    more than others?

    • How can we test any of this?`
    PROBLEM STATEMENT

    View Slide

  9. PHYSICAL MODELLING - MEET SAM
    specific anthoromorphic mannequin

    View Slide

  10. COMPUTATIONAL MODELS
    • Canonical models:

    • Stuchley 2008, layered homogenous spheres & cubes

    • Watanabi 2006: cubes

    • slab models: Melbourne Uni
    • SAM and his homogeneous friends
    • CT or MRI-based models: Visible Human(s), Korean, Chinese,
    NORMAN

    • Child models:

    • simply scaled adults (Dimbylow 2002, 2005),

    • morphological scaling (Wang & Fujiwara 2006, 2008) (Wiart 2008)

    View Slide

  11. NUMERICAL SOLVING TECHNIQUES
    •Solving Maxwell’s equations

    •Different methods

    • FDTD

    • MoM

    • FEM

    • Hybrids

    View Slide

  12. PROBLEMS WITH COMPLIANCE
    • Assumptions in both kinds of models:
    • Size, thickness, relative position of tissues
    • Dielectric properties of tissue
    • Average human = Caucasian male
    • Women, children = scaled Caucasian males
    • Image-based models are one-offs
    • Contradictions, inconsistencies in the data
    • Testing assumptions is difficult

    View Slide

  13. Relatively easy parametric
    adjustment of:
    • Tissue size
    • Tissue relative location
    • Dielectric properties
    • Morphology
    Skin - 2 mm thick
    Skull - 7 mm thick
    Brain – 80 mm radius
    Eyes – 15 mm radius
    Ears – 2 mm skin
    Nose – average head
    coated by 2 mm skin
    Fat around eyes – 7 mm
    Filler: average head
    Based on anatomic
    measurements taken from
    large sample of adult
    Caucasian males (Farkas,
    1994)
    FEKO, finite element method/method of moments (FEM/MoM)
    Plane wave excitation (10 W/m), sagittal plane
    Dielectric properties as per Gabriel
    GEOMETRY HEAD
    WILL SAVE THE DAY

    View Slide

  14. SAR  &  CRANIAL  THICKNESSES
    TRL Geometry Head VH Norman
    Phanto
    m
    4mm 6mm 11mm
    Whole head
    average SAR
    (W/Kg)
    0.043 0.036 0.036 0.035 0.034 0.043
    10g Ave SAR
    (W/Kg)
    0.27 0.12 0.13 0.13 0.15 0.19
    Tissue ear head
    (near
    head
    (near ear)
    head
    (near
    top of
    brain
    back of
    neck
    Max peak
    SAR (W/Kg)
    0.66 0.90 0.96 0.89 0.59 0.73
    Position ear head
    (behind
    eye)
    skin (at
    ear)
    skin (at
    ear)
    back of
    neck
    ear
    27 literature
    sources

    !
    multiple
    measurements

    !
    3103 skulls
    • compromise  model  
    • point  of  diminishing  returns

    View Slide

  15. SKIN THICKNESS

    Peak 10g SAR distribution
    0.000000
    0.000005
    0.000010
    0.000015
    0.000020
    0.000025
    0.000030
    0.000035
    0.000040
    0.000045
    0 50 100 150 200 250 300 350 400
    Penetration into head (mm)
    SAR (W/Kg)
    1mm
    2mm
    5mm
    • 16 literature sources
    • multiple measurements
    • 1520 cadavers or live
    volunteers
    • Significant effect of skin
    thickness , but not at
    human anatomic ranges
    • Absorption is within
    safety limits
    • Skin in heterogeneous
    models is 1.5 mm or 2
    mm thick
    • Predicted SAR results
    from Geometry Head
    model are comparable

    View Slide

  16. DIELECTRIC  PROPERTIES
    • Original dielectric properties as per Gabriel (1996), varied to
    ±10, ± 20 and ± 30 %

    !
    • Assumptions:

    • Approximately linear association with SAR

    • Variation for SAR of 30 % is an accepted uncertainty in SAR
    measurement standards

    • Skull tissue close to surface - a priori reason that it plays a
    significant role in energy absorption

    View Slide

  17. DIELECTRIC PROPERTIES
    • SAR in skull increases as
    conductivity increases: a more
    lossy layer near surface
    • Increased shielding effect in the
    head – drop in whole head SAR as
    conductivity increases
    • Skin dielectric properties affect
    SAR – as expected – but not
    enough to exceed limits
    • As eye dielectric properties drop,
    SAR increases a little
    •All SAR variations seen are
    within safety limits

    View Slide

  18. Head size scaling factor
    Testing:
    Effect of varying entire head size ± 30%
    !
    Dimensions
    GHead: height = 230 mm width = 176 mm
    GHead 70%: height = 161 mm width= 123 mm
    GHead 130% height = 299 mm width = 229 mm
    VH: height= 292 mm width = 229 mm
    SAM:
    height = 228 mm width = 190 mm
    !
    Previous work (scaled down adult):
    Dimbylow (1993) – higher SAR in eyes
    Dimbylow & Mann (1994) – lower SAR in smaller heads
    Ghandi (1996) – SAR increase in smaller heads
    Schönborn (1998 ) – some differences
    Guy (2002) – no characteristic differences
    Christ 2005 – no correlation between head size & SAR

    View Slide

  19. HEAD  SIZE:  1-­‐GRAM  AND  10-­‐GRAM  
                     70%  -­‐  85%             90%  -­‐  100%  
               105  -­‐  110  %              115-­‐130  %

    View Slide

  20. HEAD SIZE: RESONANCE
    Variables 70% 80% 85% 90% 95% 100% 105% 110% 115% 120% 130%
    Head diameter 123.20 140.80 149.60 158.40 167.20 176.00 184.80 193.60 202.40 211.20 228.80
    (head dia)/λ 0.37 0.43 0.45 0.48 0.51 0.53 0.56 0.59 0.61 0.64 0.69
    Skull diameter 120.40 137.60 146.20 154.80 163.40 172.00 180.60 189.20 197.80 206.40 223.60
    (skull dia)/λ 0.36 0.42 0.44 0.47 0.50 0.52 0.55 0.57 0.60 0.63 0.68
    Brain diameter 112.00 128.00 136.00 144.00 152.00 160.00 168.00 176.00 184.00 192.00 208.00
    (brain dia)/λ 0.34 0.39 0.41 0.44 0.46 0.48 0.51 0.53 0.56 0.58 0.63
    Head height 161.00 184.00 195.50 207.00 218.50 230.00 241.50 253.00 264.50 276.00 299.00
    (head height)/λ 0.49 0.56 0.59 0.63 0.66 0.70 0.73 0.77 0.80 0.84 0.91
    Skull height 158.20 180.80 192.10 203.40 214.70 226.00 237.30 248.60 259.90 271.20 293.80
    (skull height)/λ 0.48 0.55 0.58 0.62 0.65 0.68 0.72 0.75 0.79 0.82 0.89
    Brain cavity height 149.80 171.20 181.90 192.60 203.30 214.00 224.70 235.40 246.10 256.80 278.20
    (brain cavity height)/λ 0.45 0.52 0.55 0.58 0.62 0.65 0.68 0.71 0.75 0.78 0.84
    Freq = 900 MHz, λ = 330mm
    σ = 0.77
    σ = 0.97

    View Slide

  21. VALIDATION PROJECT

    View Slide

  22. View Slide

  23. DISCUSSION
    • Confounding factors: Sources of uncertainty

    • Increased uncertainty in Geometry Head by simplifying

    • Simplification of the tissues into geometrical shapes

    • Computational methods: solver introduces uncertainty – tetrahedral
    meshes as estimate for tissue shapes

    • Dielectric properties – measured or guessed at

    • Decreased uncertainty by including fewer tissues
    • is variability in humans high enough to affect compliance with
    safety standards?
    • is this tool any good?

    View Slide

  24. APPLICATIONS OF GEOMETRY HEAD
    • good tool for exploring the relationship between SAR and
    anatomical features
    • middle ground between homogeneous and highly
    heterogeneous models
    • Provides info on how limits should be set – uncertainty
    within the standard, vs. uncertainty with measurements
    • Can be used for looking at epidemiological population
    variations and how these affect standards – are all
    populations safe? Are kids safe?

    View Slide

  25. HAVE I SOLVED THE PROBLEMS?
    Does the anatomic variability in humans place some people in danger
    of exposure above the safety limits?
    • How much do these factors vary in humans?
    • a fair bit, some factors more than others
    • How much effect do these variations have on absorption?
    • a fair bit, some factors more than others
    • Do some anatomic factors affect absorption more than others?

    • How can we test any of this?
    • using my fabulous model!

    View Slide

  26. NEXT (NOT ME)
    •What happens at extremes?
    •Remove sharp angle in the middle of the head?
    •Other tissue variations, more tissues?
    •Eventually – expand model
    • Mobile handset, not just dipole
    • Hand on mobile
    • Whole body
    • Other frequencies

    View Slide

  27. View Slide