Characterizing magnetic soils: State of the art and future needs

1
Characterizing magnetic soilsState of the art
and future needs
Stephen D. Billings Sky Research,
Inc. Leonard R. Pasion Douglas W.
Oldenburg UBC-Geophysical Inversion Facility U.
of British Columbia
2
ITEP test-lanes at Benkovac and Oberjettenberg
3
Why does the soil with lower susceptibility but a
larger change in susceptibility with frequency
cause more problems to a metal detector?
4
Outline

• Theory
• Induced magnetisation
• Eddy currents
• Magnetic viscosity
• Application to test-lanes
• Future directions

5
Soil effects on time-domain metal detectors

• Induced magnetization (Susceptibility c)
• Eddy currents (Susceptibility c and conductivity
  s)
• Magnetic viscosity (frequency dependent
  susceptibility c(w))

6
Soil effects on time domain metal detectors
Induced magnetization
Time-domain
Frequency-domain
7
Soil effects on time-domain metal detectors

• Induced magnetization (Susceptibility c)
• No effect
• Eddy currents (Susceptibility c and conductivity
  s)
• Magnetic viscosity (frequency dependent
  susceptibility c(w))

8
Soil effects on time domain metal detectors
Eddy currents

• When the primary field turns off, eddy currents
  are induced on the surface that oppose the change
  in the field
• Initial field
• These currents then diffuse and decay in a way
  that depends on the conductivity and geometry

9
Soil effects on metal detectorsEddy currents
10
(No Transcript)
11
Soil effects on time domain metal detectors
Eddy currents
Scheibel AN-19
Typical range
12
Soil effects on time-domain metal detectors

• Induced magnetization (Susceptibility c)
• No effect
• Eddy currents (Susceptibility c and conductivity
  s)
• Late time decay in a half-space is
• Magnetic viscosity (frequency dependent
  susceptibility c(w))

13
Soil effects on time-domain metal detectors
Magnetic viscosity
14
Soil effects on time-domain metal detectors
Magnetic viscosity
Primary Field
Magnetization

• Time constant t characterizes time for M to
  rotate to its new orientation
• If time dependence characterized by a single t

15
Soil effects on time-domain metal detectors
Magnetic viscosity

• 1. Magnetic mineral composition
• Magnetic character of the soil dominated by the
  presence of ferrimagnetic minerals
• Maghaemite a Fe2O4
• Magnetite Fe304
• 2. Concentration
• Mass fraction of the dominant magnetic carrier
• 3. Grain size distribution in the soil sample
• Size determines the nature of domains within each
  grain

16
Magnetic viscosity, cont.

• Assume a non-interacting system of magnetic
  grains.
• Integrate over a distribution of time constants
  f(t)
• For a large range of relaxation times distributed
  log-uniformly over their spectrum
• A induction coil measures the time-rate of change
  so

17
Soil effects on time-domain metal detectors

• Induced magnetization (Susceptibility c)
• No effect
• Eddy currents (Susceptibility c and conductivity
  s)
• Late time decay is
• Magnetic viscosity (frequency dependent
  susceptibility c(w))
• Late time decay is

18
Relationship to Frequency Dependent Magnetic
Susceptibility

• The susceptibility of a sample containing a
  collection of particles with a uniform
  distribution of energy barriers (Lee, 1983)

t2-1
t1-1
19
Relationship to Frequency Dependent Magnetic
Susceptibility
20
Outline

• Theory
• Induced magnetisation
• Eddy currents
• Magnetic viscosity
• Application to test-lanes
• Future directions

21
ITEP test-lanes Frequency dependent
susceptibility
22
ITEP test-lanes Conductivity 0.01 S/m
Magnetic viscosity response t-1
Eddy current response t-5/2
23
ITEP test-lanes Conductivity 0.1 S/m
24
ITEP test-lanes Ground reference heights
Amplitude at fixed time channel
25
Changing susceptibility magnitude Effect on
time-decay amplitudes
26
Changing susceptibility magnitude Effect on
amplitude decay with height
Amplitude at fixed time channel
27
Changing susceptibility difference Effect on
time-decay amplitudes
28
Changing susceptibility difference Effect on
amplitude decay with height
Amplitude at fixed time channel
29
Can we predict susceptibility difference from
ground reference height?
Debas Skinner (1996) Dc correlates well
with magnitude of 1/t decay
30
Outline

• Theory
• Induced magnetisation
• Eddy currents
• Magnetic viscosity
• Application to test-lanes
• Future directions

31
More measurements of frequency dependence

• We know that susceptibility difference is the key
  parameter
• Is our assumption of a log-uniform distribution
  correct?
• If not what other distribution is appropriate?

32
We need more soil measurements

• Yoga Das from DRES, Canada has offered to
• Supply instrumentation for in-field measurements
• Make laboratory measurements of soil samples
• SERDP funded proposal to look at the geological
  influence on unexploded ordnance detection
• Colorado School of Mines, University of British
  Columbia, New Mexico Tech., Sky Research
• The US-Army Corps of Engineers are building a
  susceptibility meter for measurements over a wide
  frequency range

33
Better understanding of soil compensation
algorithms

• Time-domain instruments generally use the 1/t
  response to magnetic viscosity (or its equivalent
  for a finite transmitter waveform e.g. Candy,
  1996).
• Frequency-domain rely on a flat quadrature
  response and/or linear change in the in-phase
  response with log frequency.

34
Kahoolawe Geonics EM63 time-channel 1
Channel 1 (mV)
35
Kahoolawe time decays
36
Minimizing false alarms due to geology

• From knowledge of transmitter waveform predict
  magnetic soil response
• Calculate how closely each decay matches the soil
  model
• For example, with a rectangular pulse of duration
  Dt the soil model is
• Determine the value of A that minimises the
  difference between this equation and the actual
  decay.

37
Minimizing false alarms due to geology, cont.
60 mm mortar
Mark 81
Soil
5 HE round
38
Conclusions

• Magnetic viscosity will generally be the dominant
  soil effect for an EM sensor
• Magnetic soil and ground effect responses are
  independent
• Induction coil response to step-off is 1/t
  regardless of spatial distribution of soil

39
Conclusions, cont.

• Can predict response of an arbitrary soil for an
  arbitrary waveform
• Is the assumption of a uniform distribution of
  time constants valid?
• Need more information on the complex
  susceptibility of different soils

40
UXO Detection in Magnetic Environments

• Kahoolawe, Hawaii
• 60 000 anomalies dug
• 32-33 False Alarm Rate
• 70.3 due to non-hazardous metal objects
• 27 due to Geology

41
Kahoolawe Geonics EM63
42
Effect of transmitter on-time
Primary Field
Magnetization
D t

• Each time constant will respond as

43
Effect of transmitter on-time D t 20 ms
Time constants
Induction coil response
Pulse
44
Effect of transmitter on-time D t 200 ms
Time constants
Induction coil response
Pulse
45
Effect of transmitter on-time D t 2000 ms
Time constants
Induction coil response
Pulse
46
Effect of different waveforms
47
Minimizing false alarms due to geology, cont.
Susceptible sphere Misfit 0.72
Susceptible half-space Misfit 0.8
48
Minimizing false alarms due to geology, cont.
Steel sphere Misfit 101 Time constant 42 ms
Aluminium sphere Misfit 439 Time constant
796ms
49
Loop radius 1 m.
Conductivity 0.01 S/m
Conductivity 1 S/m
50
Loop radius 25 m.
Conductivity 0.01 S/m
Conductivity 1 S/m