Neutron detection using Poisson distribution comparison independent of count rate based on correlation signals
View Patent ↗Embodiments are directed to comparison-based methods of conditionally assessing the excess in correlation of an unknown neutron count measurement compared to the correlation present in a data defined as background, and to providing a technical definition of excess correlation intended to properly handle the measured excess correlation. The degree of correlation between an unknown source and background can be used to prevent masking of neutron count data for the source by background radiation.
1. A computer-implemented method of analyzing radiation of a target source relative to background by conditionally assessing an excess in correlation of neutron emission from the target source compared to background neutron emission, comprising:
measuring, for a first number of cycles, first neutrons detected by a detector not placed in proximity to the target source to derive a background multiplicity count distribution of different multiplet sizes for a defined observation time; normalizing the background multiplicity count distribution by dividing a number of first neutron detections for each multiplet size by the first number of cycles to generate a first set of normalized multiplets (b 0 , b 1 , b 2 , b 3 );
calculating, from the first set of normalized multiplets, a lambda value (Λ n ) for each multiplet size to determine if the detected neutrons represent a non-Poisson distribution based on non-zero values for at least second and third lamba (Λ 2 , Λ 3 ) values;
defining the detected neutrons as background if at least the second and third lambda values are non-zero;
measuring, for a second number of cycles, second neutrons detected by the detector placed in proximity to the target source to derive a source multiplicity count distribution of different multiplet sizes for the defined observation time;
normalizing the source multiplicity count distribution by dividing a number of second neutron detections for each multiplet size by the second number of cycles to generate a second set of normalized multiplets;
calculating, from the second set of normalized multiplets, corresponding lambda values for each multiplet size; dividing, for the first neutron detection, the second lambda value by the first lambda value (Λ 2 /Λ 1 ) to derive a normalized correlation indicator;
using the normalized correlation indicator to scale all lambda values for the first set of normalized multiplets to match a count rate of the second neutron detections;
comparing the scaled lambda values to each of the corresponding lambda values for the second set of normalized multiplets and characterizing the target source as fissile material if each normalized source multiplet exceeds each respective normalized background multiplet by a threshold defined by the normalized correlation indicator.
2. The method of claim 1
wherein the first set of normalized multiplets comprises a first integer number n of multiplets, and wherein the second set of normalized multiplets comprises a second integer number m of multiplets.
3. The method of claim 2 wherein the scaling eliminates an effect of the count rate in the respective count distributions.
4. The method of claim 2 wherein the different multiplet sizes ranges from zero to six neutrons, and the defined observation time comprises 512 milliseconds.
5. The method of claim 1 , wherein calculating the lambda value (Λ n ) for first set of normalized multiplets is performed by:
Λ 1 =b 1 /b 0 ;
Λ 2 =(2b 1 /−Λ 1 b 1 )/2b 0 ; and
Λ 3 =(3b 3 /−Λ 1 b 1− 2Λ 2 b 1 /3b 0 .