Defense Intelligence Reference Document Inertial Electrostatic Confinement Fusion

UNCLASSIFIED//F8A 8FFHil.l1k 'JE'lii &Ilk¥
Detector Array and Analysis System
The final imaging quality is determined by neutron source yield integrated over the area
of coverage by the number of detectors, the detector configuration, the collimated
beam size, the acquisition time, and other scan-related parameters. Scintillators such
as NaI, Csl, and BGO as well as Plastic Nal (Tl) are candidates for gamma and x-ray
detectors. Due to their small size, they can be used in a large array along with TOF
spectrometry to provide detailed special information. Recently, UIUC staff has studied
use of an advanced thick Hgb detector which appears to be particularly advantageous
for this application. Hgb is a room temperature semiconductor material with an
excellent stopping power and a relatively large photo-fraction for detecting high energy
gamma rays. Assuming successful development, Hgb detectors would provide superior
array performance.
With the combined sources, the detector array receives a vast amount of information in
a short scan time. To handle this flow of data most efficiently, a fuzzy logic system is
employed. The development of this advanced fuzzy logic system is patterned after a
methodology developed for the diagnosis of abnormal situations in nuclear reactor
safety analysis. The knowledge box employs a goal tree (and/or decision tree) for
representation. This system will use learning-based rules that evolve from extensive
simulation tests to provide training about the simultaneous occurrence of specific
elements and embedded object shapes. It will also learn to differentiate between
positive and negative data by creating rules from these test runs. Materials listed early
in the simulation study discussion would be employed. The advantage of this system is
twofold. The broad area scan analysis is optimized for fast through-put of objects while
the multiple neutron/x-ray identification reduces the probability of false identification.
The fuzzy logic system will pass the integrated item or sound an alert (suspicious object
in the container) automatically without human involvement. Humans will only be
involved when an alert is sounded. An alert occurs for a range of positive identification
of materials on to suspected identification. The multiple source detector array concept
is designed to minimize items that fall into the category of "in-between" positive
identification and clearance.
The overall concept described here is illustrated for an airplane luggage scan in Figure
5.3. Data from the detection system (TNA, PFNA and x-ray) is processed through fuzzy
logic system. If the item undergoing inspection passes this test, it continues on its way
to baggage claim. If an alert is sounded, the item can either be passed through the
machine for a second scan or be manually inspected, depending on the confidence level
of the alert.
Since the integrated system is modular, the basic components can easily be assembled
in a variety of configurations for use at different facilities. An example is the extension
to inspection of container ships being unloaded in port as illustrated in Figure 5.4.
52
UNCLASSIFIED/ ,'F8"1 8FFU!lit.l! l!l!il! 8111!¥