Neutron Spectrum Extraction Using Neural Networks for Light Element Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing neutron detection methods for hazardous substances face challenges due to complex gamma spectra with substantial background noise, the need for costly and complex equipment, and the difficulty in obtaining high-energy photons from conventional sources, making it hard to identify discrete neutron components for light elements like carbon, nitrogen, oxygen, and chlorine.
Innovation Solution
A method using a trained multilayer and multichannel neural network processes neutron spectra from photo-nuclear reactions, enabling the extraction of discrete neutron components from bremsstrahlung or multi-energy photon sources without requiring single-energy sources or specialized electronics, by training the network with simulated neutron spectra from both bremsstrahlung and single-energy photon irradiations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If bremsstrahlung photons are used to induce photo-nuclear reactions, then intense high-energy photons can be obtained, but the photon energy spectrum becomes continuous rather than discrete
Solution Approach 1:
The continuous bremsstrahlung photon spectrum is segmented into discrete energy components through neural network processing. The network analyzes the continuous spectrum and separates it into discrete energy bins, enabling the system to treat each energy component independently as if it originated from a monochromatic source.
Solution Approach 2:
A neural network acts as an intermediary between the bremsstrahlung photon source and the detection system. The network processes the complex continuous spectrum and extracts discrete neutron components, effectively mediating the transition from continuous to discrete energy analysis without requiring physical spectral filtering.
2Measurement precision
If gamma spectrometry is used to measure gamma radiation from photo-nuclear reactions, then light elements can be identified, but the complex gamma spectra with substantial background noise make accurate measurement difficult
Solution Approach 1:
The neural network extracts discrete neutron components from the complex gamma spectrum by identifying and isolating specific spectral features associated with photo-nuclear reactions. This extraction process separates the signal of interest from the background noise and other gamma radiation contributions.
Solution Approach 2:
The physical gamma spectrometry system is replaced with a neural network-based analysis system. Instead of relying on physical detectors and spectral filtering, the invention uses computational processing to analyze gamma spectra and identify light elements, substituting mechanical detection with intelligent algorithmic processing.
3Measurement precision
If the tagged photon technique is used to detect neutrons coincident with bremsstrahlung electrons, then neutron component extraction for each photon energy is possible, but complex equipment including magnetic dipoles and specialized electronics is required
Solution Approach 1:
The complex mechanical detection system (magnetic dipoles, coincidence electronics) is replaced with a neural network-based computational system. The network processes spectral data to achieve the same neutron component extraction functionality without requiring specialized hardware for coincidence detection.
Solution Approach 2:
The functionality of the tagged photon technique is copied through software simulation rather than physical implementation. The neural network replicates the spectral analysis capability of the tagged technique using computational models trained on simulated data, eliminating the need for complex physical coincidence detection apparatus.
4Use of energy by moving object
If conventional radioactive sources are used for photon irradiation, then simple equipment can be used, but sufficient energy levels (greater than 6 MeV) cannot be reached
Solution Approach 1:
The energy parameter of the photon source is changed from the limited range of conventional radioactive sources to the high-energy range of bremsstrahlung photons. By using an electron accelerator to generate bremsstrahlung radiation, the system achieves photon energies greater than 6 MeV necessary for photofission reactions, while the neural network processing maintains analytical simplicity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate identification of light elements like carbon, nitrogen, and chlorine by extracting discrete neutron components from complex neutron spectra, reducing the need for costly equipment and allowing detection with conventional, more intense photon sources, thus enhancing the efficiency and cost-effectiveness of illicit substance detection.
Implementation Method 1
photo-nuclear reactions, the effective cross-sections (probabilities) whereof are well known for most of the nuclei of light elements such as nitrogen, carbon, oxygen, chlorine
Implementation Method 2
bremsstrahlung photons, which are produced using an electron accelerator. In a known manner, in an electron accelerator, an electron beam of sufficient energy (several MeV) strikes a target consisting of a material of high atomic number Z, which gives rise to braking radiation (bremsstrahlung photons)
Data Source
AI summary
The invention relates to a method for extracting discrete neutron components (from photo-nuclear reactions between photons and at least one chemical element to be detected) of a neutron spectrum from photo-nuclear reactions obtained by irradiation, of a material comprising said at least one chemical element to be detected, with a photon source, at least one energy of the photons of the source being greater than the photo-nuclear reaction threshold of the chemical element to be detected, using a multilayer and multichannel neural network having an architecture with a convolution stage and a deconvolution stage. The method comprises a preliminary step of training the neural network by supervised learning and a prediction step, using the trained neural network.