Authors:

Abhishek Raj

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Accurate, cost-effective, and near real-time detection and identification of neutron vs. alpha radiation is critical for many applications, including national security, nuclear medicine, energy, and search for dark matter. Conventional ionization and excitation-based neutron/alpha detectors are also sensitive to gamma and beta particles and can become saturated, making them ineffective for capturing and distinguishing neutron and alpha radiation. The centrifugally tensioned metastable fluid detectors (CTMFDs) enable gamma-beta blind detection of neutrons and alpha particles. However, deciphering between the interactions presents a hurdle. To overcome this limitation, this research aims to develop an AI-based solution to discern neutron and alpha particles, by analysing acoustic shock spectra resulting from femto-scale radiation interactions. In CTMFDs, the femto-scale interaction of neutron and alpha radiation with tensioned metastable state atomic nuclei of sensing fluids produces rapid-growing bubbles. These shock spectra are not amenable to standard signal analysis. However, machine learning methods provide confidence in identifying these particles with greater than 90% accuracy. The acoustic shock signals accompanying the experimentation with neutron and alpha particles were recorded separately and converted into spectrograms. Various convolutional neural network architectures have been investigated to accurately classify the acoustic signatures of radiation particles, allowing us to record and classify neutron vs. alpha interactions with any given detection event in near real- time. A convolutional neural network had been designed, achieving ~99.2% accuracy. Further work involves studies to discriminate energetically diverse neutron sources and understand the underlying physics of interactions that give rise to unique acoustic signatures on the formation of transient cavitation events. Keywords: Neutron Detection; Alpha Radiation; Centrifugally Tensioned Metastable Fluid Detectors; Acoustic Shock Spectra; Convolutional Neural Network

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Abhishek Raj