Magnetics discrimination that will satisfy regulators

1
Magnetics discrimination that will satisfy
regulators?
Stephen Billings, UBC-Geophysical Inversion
Facility sbillings_at_eos.ubc.ca
2
Acknowledgements

• Dr John Stanley
• G-Tek Australia
• Dr Clif Youmans
• Montana Army National Guard

3
Organisation

• Existing approaches to UXO remediation.
• Background on magnetometry.
• A discrimination procedure to satisfy regulators?

4
Detection depth versus anomaly amplitude
5
Detection threshold
6
Guthrie Road Montana
Ordnance present 58 live items 25 seeded items
for QA 83 items in total
G-Tek Australia 8 element magnetometer array 0.5
m sensor separation 804 anomalies
7
Limestone Hills Montana
G-Tek Australia 4 element magnetometer array 0.5
m sensor separation 360 anomalies
Ordnance present 26 live UXO 42 seeded items for
QA 68 items in total
8
Limestone Hills and Guthrie Road Montana

• Guthrie Road
• 804 anomalies, 83 UXO
• FAR 8.7 (i.e. 8.7 false alarms per UXO)
• Limestone Hills
• 360 anomalies, 68 UXO
• FAR 4.3

9
Organisation

• Existing approaches to UXO remediation.
• Background on magnetometry.
• A discrimination procedure to satisfy regulators?

10
Sources of magnetization

• Almost all UXO are steel and therefore
  ferromagnetic.
• Remanent magnetization was acquired by the UXO
  during its history and is constant.
• Induced magnetization arises from exposure to the
  Earths magnetic field its direction and
  magnitude change with orientation.

11
Direction and strength of magnetization varies
with orientation (moment magnetisation x
volume)
12
Dipole moment varies with orientation

• Dipole moment magnitude and directionMagnitude
  at maximum when aligned with earths field.

13
Induced dipole moments for an ordnance library

• By simply changing orientation, the induced
  dipole moment of a UXO can change significantly.

14
Organisation

• Existing approaches to UXO remediation.
• Background on magnetometry.
• A discrimination procedure to satisfy regulators?

15
FIRST STEPGeophysical inversion returns dipole
parameters

• Invert the data about each anomaly for
• position
• horizontal and vertical
• dipole moment
• magnitude and orientation

16
Geophysical inversion returns dipole parameters
17
Second step work out the minimum remanent
magnetization to match an item in the library
ITEM 1
ITEM 2
ITEM 2 WITH LOWER REMANANT MAGNETISATION IS MORE
LIKELY A UXO THAN ITEM 1
18
Dipole moments at Guthrie Road
CLOSE-UP
19
Discrimination results at Guthrie Road
Less than 50 remanent magnetisation
Greater than 50 remanent magnetisation
20
Discrimination results at Guthrie
Road(including QA items)

• Initial dig list (up to 50 remanent mag)
• 224 holes
• Last live UXO
• 216-th hole
• 50 more holes for QC
• 65 anomalies which dont pass minimum fit
  criteria
• Final Dig list
• 331 holes (compared to 804)
• Digging 41 of holes yields 100 of the UXO
• plus all the QA items (FAR 3.0)

71 of shrapnel, metallic debris and geology left
in the ground
21
Dipole moments at Limestone Hills
CLOSE-UP
22
Discrimination results at Limestone Hills
Unmodeled QA projectile
Less than 50 remanant magnetisation
Greater than 50 remanant magnetisation
23
Discrimination results at Limestone Hills
(including QA items)

• Initial dig list (up to 50 remanent mag)
• 122 holes
• Last live UXO
• 110-th hole
• 50 more holes for QC
• 17 anomalies which dont pass minimum fit
  criteria
• Dig list
• 177 holes (compared to 360)
• Digging 50 of holes yields 100 of the UXO
• plus all but 4 of the QA items (FAR 1.8)

62 of shrapnel, metallic debris and geology left
in the ground
24
ROC curves for all UXO live and QA items
Original Limestone Hills
New Limestone Hills
New Guthrie Rd
25
ROC curves for all UXO live and QA items On a
more typical scale
26
(No Transcript)