Far-field Monitoring of Rogue Nuclear Activity with an Array of Antineutrino Detectors Neutrino Sciences 2005 University of Hawaii, Manoa December 14-16, 2005 Eugene H. Guillian University of Hawaii, Manoa Neutrino Sciences 2005 University of Hawaii, Manoa December 14-16, 2005 Eugene H. Guillian University of Hawaii, Manoa
Far-field Monitoring. of. Rogue Nuclear Activity. with. an Array of Antineutrino Detectors. Neutrino Sciences 2005. Neutrino Sciences 2005. University of Hawaii, Manoa. University of Hawaii, Manoa. December 14-16, 2005. December 14-16, 2005. Eugene H. Guillian - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Want sensitivity to anywhere on Earth Want sensitivity to a well-defined region
Can’t optimize module positioning
Module positions can be optimized because of prior knowledge of likely
locations
Larger Modules Required• 10 Megatons• 1 year exposure
Smaller Modules Will Do• 1 Megatons• 1 year exposure
Global Array 15º 5º Array
Total of 1596 modules
Global Array 2Equidistant Array
Total of 623 modules
Minimum nearest-neighbor distance ≈
600 km
Global Array 3Coast-hugging Array
Total of 1482 modules
Minimum nearest-neighbor distance ≈
100 km
Modules removed from coast line by ≈
100 km
Regional ArrayNorth Korea
€
log10 S / S + B
Choose locations based on sensitivity map
(red dots are candidate module positions)
• 250 MWth fission reactor deep inside of
North Korea• Background from commercial nuclear
reactors
Rogue Activity Detection Strategy
(1) Assume that no rogue activity is taking place
(2) If this assumption is incorrect AND if the rogue activity is sufficiently large, there would be a discrepancy between observation & expectation
(3) Use a statistical technique (minimum log-likelihood) to estimate the position & power of the rogue activity
Illustration of the Detection Strategy
If no rogue activity takes place, module 1, 2, & 3 detects B1,
B2, and B3 events
With rogue activity, module 1, 2, and 3 sees an extra S1, S2,
and S3 events
The size of the excess goes as:
Power / Distance2
Seeing the Rogue Activity Above Random Fluctuations
ObservedNumber
ofEvents
Backgroundonly
ObservedNumber
ofEvents
Small Signal + Background
RandomStatistical
Fluctuation
Large Signal + Background
B = # background events
S = # signal events
Signal Strength
€
B = statistical uncertainty
Signal Strength
€
B
SS
€
B
Map of Signal Strength
Rogue Activity
2000 MWth
Equidistant Detector Array Configuration10 Megaton per module
1 year exposure
Detectors with Signal Strength > 3
Detectors with Signal Strength > 2
Detectors with Signal Strength > 1
Remarks on Rogue Activity Detection
Rogue Activity:1. Has sufficiently large power2. Is sufficiently close to detector modules
Cluster of nearby detector modules with significant excess
Pin-PointingRogue activity location
given roughly by the position of the cluster
Cutting on Signal Strength
• Tight cut low background noise, but
loss of signal • Loose cut more
signal, but more background noise
Measuring PowerUse log-likelihood to obtain the most likely
power
P99: Benchmark for Array Performance
Log-likelihood Function• A statistical tool used for hypothesis testing
Hypothesis• No rogue activity is taking place
Information Used in Log-likelihood Function
• Expected number of background events in each detector (from commercial nuclear reactors)• Observed number of events in each detector
• The log-likelihood value is not defined a priori because of random fluctuations in the measurement• Its distribution, however, is defined a priori
99% of measurements give log-likelihood above the
alarm threshold1% of measurements sets
off false alarm
1% False Positive
Rogue Reactor Exists
Hypothesis Incorrect
Log-likelihood function is biased to lower values
Rogue Reactor Power is Weak
Large overlap between observed vs. expected distributions
Can’t reliably detect rogue activity
Unacceptably Large Frequency of False Negatives
Definition of P99
P99 = Rogue reactor power which gives 1% chance of false negative
Global Array Performance
• For each array configuration, make a map of P99
• Procedure for making map:1. Vary the rogue reactor position
2. At each location, determine P99
P99 Map: 5º 5ºMWth
P99 Map: EquidistantScaled to 1596 Modules MWth
P99 Map: Coast-huggingScaled to 1596 Modules MWth
5º 5º
Equidistant
Coast-Hugging
P99 Summary
In Water < 100 MWth
W/in several 100 km of
coast
Several 100 MWth
Deep in continent
Up to 2000 MWth
Regional Monitoring
Example:• A rogue reactor in North Korea
Signal
Background
Signal StrengthAbout the Plots
Signal
• Rogue power = 250 MWth
• Detector mass = 1 Megaton
• Exposure = 1 year
Background
• Commercial nuclear reactors
• 1 Megaton
• 1 year
€
log10 S
€
log10 B
€
log10 S / S + B
Detector Locations
€
log10 S / S + B
23 candidate locations based on map of sensitivity
Performance of Various Array Configurations
Consider configurations with 2, 3, and 4 detector modules
For each configuration, determine:• P99
• Estimated area that contains rogue reactor
Two Modules
95% Confidence
99% Confidence
P99 = 250 MWth
Two Modules
95% Confidence
99% Confidence
P99 = 120 MWth
Three Modules
95% Confidence
99% Confidence
P99 = 626 MWth
Four Modules
95% Confidence
99% Confidence
P99 = 336 MWth
Four Modules
95% Confidence
99% Confidence
P99 = 502 MWth
What if a Georeactor Exists?The Georeactor Hypothesis:
• Unorthodox, but surprising things can happen….• If it does exist, its power is likely to be 1~10 TWth
Total commercial nuclear activity ≈ 1 TWth
If a terawatt-level georeactor does exist, the background level for rogue activity
monitoring increases significantly!
log10 BackgroundNo Georeactor
log10 Background3 TWth Georeactor
Ratio3 TWth / No Georeactor
Squeezing More Information from the Data
Fission Bomb Monitoring
ConclusionGlobal Monitoring
≈ 1000 10 Megaton modules
10 Gigaton-year
P99 Water < ≈ 100 MWth
Several 100 km from
coast
Several 100 MWth
Deep in continent
< ≈ 2000 MWth
Regional Monitoring
Several Megaton-year
P99 100 ~ several 100 MWth
Location(95%
Confidence Level)
Best Worst
< ~ 100 kmBand stretching over
several 100 km
• One module costs several hundred million dollars multiply this by number of modules in array
• A terawatt-level georeactor increases the background level by a factor of several in most locations around the world• A regional monitoring regime seems not-too-outlandish
Appendix
Antineutrino Detection Rate for H2O + GdCl3 Detector
€
3040 Events ⋅T
1 year
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
V
106 m3
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
100 km
D
⎛
⎝ ⎜
⎞
⎠ ⎟2
⋅P
100 MWth
⎛
⎝ ⎜
⎞
⎠ ⎟
€
2.25 events ⋅V
106 m3
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
100 km
D
⎛
⎝ ⎜
⎞
⎠ ⎟2
⋅Y
1 kiloton
⎛
⎝ ⎜
⎞
⎠ ⎟
Reactor• Assume 100% detection efficiency for Eν > 1.8 MeV
Fission Bomb• Assume 100% detection efficiency for Eν > 1.8 MeV• Integrated over 10 sec. burst time
Antineutrino Detection Rate for H2O + GdCl3 Detectors
€
832 Events ⋅T
1 day
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
V
109 m3
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
1000 km
D
⎛
⎝ ⎜
⎞
⎠ ⎟2
⋅P
1 GWth
⎛
⎝ ⎜
⎞
⎠ ⎟
€
22.5 events ⋅V
109 m3
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
1000 km
D
⎛
⎝ ⎜
⎞
⎠ ⎟2
⋅Y
1 kiloton
⎛
⎝ ⎜
⎞
⎠ ⎟
Reactor• Assume 100% detection efficiency for Eν > 1.8 MeV
Fission Bomb• Assume 100% detection efficiency for Eν > 1.8 MeV• Integrated over 10 sec. burst time
Background Processes
Antineutrinos from sources other than the rogue reactor