Bulletin of the American Physical Society
APS March Meeting 2022
Volume 67, Number 3
Monday–Friday, March 14–18, 2022; Chicago
Session W31: Learning and analyticals for measurementsRecordings Available
|
Hide Abstracts |
Sponsoring Units: GIMS Chair: Chris Jacobsen, Argonne Lab/Northwestern U Room: McCormick Place W-192A |
Thursday, March 17, 2022 3:00PM - 3:12PM |
W31.00001: Application of Quantum Information Processing Algorithms to Advanced Reactor Informatics Maria Pantopoulou, Madeleine Roberts, Lefteri H Tsoukalas, Alexander Heifetz We are investigating application of quantum information (QI) algorithms to communication and data analytics of advanced nuclear reactors (ARs). The AR's, which are currently under development as part of energy de-carbonization initiative, are intended as highly automated replacements to existing fleet of aging nuclear reactors. Integration of QI into AR control and monitoring systems is one potentially promising approach to achieving AR autonomous operation. The first part of this work explores quantum key distribution (QKD) for secure wireless communications with a nuclear facility. This is investigated through computer simulations with SeQUeNCe (Simulator of QUantum Network Communication) software package developed for modeling of quantum networks. The second part of this work involves investigating performance of Quantum Principal Component Analysis (Q-PCA), using data sets from a thermal hydraulic flow loop, which are relevant to nuclear facility operation. Classical PCA is a frequently used unsupervised machine learning algorithm for data dimensionality reduction and analysis. Performance of Q-PCA, which is implemented using IBM Qiskit software package, is benchmarked relative to that of classical PCA. |
Thursday, March 17, 2022 3:12PM - 3:24PM |
W31.00002: Convolutional Neural Network for Classification of Material Defects in Pulsed Thermal Tomography Images Alexander Heifetz, Victoria Ankel, Wei-Ying Chen We have developed a deep learning convolutional neural network (CNN) to classify material defects in pulsed thermal tomography (PTT) images. PTT is a non-destructive imaging technique that uses heat diffusion to visualize subsurface internal defects in materials. In particular, PTT has been investigated for applications in imaging of subsurface pores in metals produced with laser powder bed fusion (LPBF) additive manufacturing method. PTT method involves delivering a thermal pulse on material surface with a flash lamp, and recording surface temperature transients with a fast frame infrared camera as heat diffuses into the material bulk. Thermal tomography reconstruction algorithm visualizes internal structures by converting time-dependent measurements of surface temperature into thermal effusivity spatial depth profile. However, interpretation of PTT images is not trivial because of blurring of images with increasing depth. We address this by developing a CNN for classification of size and orientation of subsurface defects in PTT images. Performance of CNN was investigated using PTT images created with computer simulations of heat transfer in metallic structures. We trained CNN on a database of simulated PTT images of structures with elliptical subsurface air voids, and demonstrated the ability of CNN to classify radii and angular orientation of voids in test images. In addition, we showed that CNN trained on elliptical defects is capable of classifying irregular-shaped defects. Simulated PIT images of such defects were obtained using shapes of actual defects in scanning electron microscopy (SEM) images of sections of stainless steel specimens printed with LPBF method. |
Thursday, March 17, 2022 3:24PM - 3:36PM |
W31.00003: Improving data acquisition and thermal resolution of ultra-low temperature tuning fork thermometers in liquid 3He using machine learning and other algorithms Alexander Donald, Andrew J Woods, Lucia Steinke We are investigating the use of quartz tuning forks in liquid 3He as small sample thermometers at ultra-low temperatures near 1 mK and in high magnetic fields up to 16 T. The resonant frequencies and damping are dependent on the viscosity of the surrounding 3He, which has a well-known temperature dependence [1]. Such thermometry has been successfully used in relatively large 3He cells. Its adaptation to thermal studies of small solid samples requires engineering miniaturized 3He cells with thermal masses comparable to the samples and improving data acquisition and analysis techniques [2]. The traditional measurement setup uses a frequency sweep to attain the resonant curve which is used to generate one temperature data point. The time required for full frequency sweep of 100 data points is on the order of two minutes, compared to estimated thermalization times around 30 seconds. Using optimized analysis techniques and measurement methods guided by machine learning, we aim to improve both the thermal resolution and data acquisition times of our thermometers. We use machine learning as a powerful tool to select maximally useful frequency point measurements for resonance sweeps while maintaining the accuracy of the curve features. We investigate other optimizations such as resonance tracking to further improve resolution and measurement times. A fast and compact thermometer compatible with ultra-low temperatures and high magnetic fields while having sufficiently fine resolution to measure small samples will enable new calorimetric measurements in this difficult to access regime. |
Thursday, March 17, 2022 3:36PM - 3:48PM |
W31.00004: Detection of Anomalies in Environmental Gamma Radiation Background with Hopfield Artificial Neural Network Alexander Heifetz, Luis A Valdez, Miltos Alamaniotis Environmental screening of gamma radiation consists of detecting weak nuisance and anomaly signal in the presence of strong and highly varying background. In a typical scenario, a mobile detector-spectrometer continuously measures gamma radiation spectra in short, e.g., one-second, signal acquisition intervals. The measurement data is a 2D matrix, where one dimension is gamma ray energy, and the other dimension is the number of measurements or total time. In principle, gamma radiation sources can be detected and identified from the measured data by their unique spectral lines. Detecting sources from data measured in a search scenario is difficult due to the highly varying background because of naturally occurring radioactive material (NORM), and low signal-to-noise ratio (S/N) of spectral signal measured during one-second acquisition intervals. The objective of this work is to investigate performance of a Hopfield Neural Network (HNN) in detection and identification of weak nuisances and anomalies events in the presence of a highly fluctuating background. Performance of HNN algorithm is benchmarked using search data from an environmental screening campaign. One data set contained a 137Cs source, and another dataset contained a 131I source. We also compare performance of HNN, which is a supervised learning method, to those of K-means clustering and Self-Organizing Map (SOM) unsupervised learning methods. |
Thursday, March 17, 2022 3:48PM - 4:00PM |
W31.00005: Machine Learning for Estimation of Fiber Optics Temperature Field Sensing in a Thermal Hydraulic Flow Loop Styliani Pantopoulou, Matthew Weathered, Darius Lisowski, Lefteri H Tsoukalas, Alexander Heifetz Fiber optic distributed temperature sensors are potentially promising options for nuclear reactor thermal-hydraulic system because they provide information about fluid temperature field. Compared to an array of point sensors (e.g. thermocouples), fiber optic sensors based on Rayleigh backscattering offer temperature sensing with higher spatial density and faster response time. Data measured with fiber optic sensor is a 2D matrix, where one dimension is length along the fiber, and another dimension is time of measurements. When exposed to high temperature and ionizing radiation environment in a nuclear reactor, fiber optic silica material degrades over time. The objective is to use machine learning methods to self-monitor or validate distributed measurements to detect early signs of failure. In our approach, we use long short-term memory (LSTM) neural networks trained on prior history of measurements to predict the next data point in real time. The basis set of fiber optic segments is identified by calculating correlation coefficients between nearest neighbors. Performance of this approach is studied using dynamic temperature field data sets obtained with single mode 1550nm optical fiber installed in a water flow loop with a thermal mixing Tee. This study also aims to determine the minimal number of training fiber segments for reliable estimation of measurements made with other segments of the fiber. |
Thursday, March 17, 2022 4:00PM - 4:12PM |
W31.00006: Neural Network Based Solution of Heat Equation for Thermography Modelling Pola Lydia Lagari, Alexander Heifetz, Lefteri H Tsoukalas We introduce a neural network-based solution of the heat conduction partial differential equation (PDE), to interpret the results of thermography, which is a non-destructive method for evaluating material structure. We are simulating a pulsed thermography system, in which transients of heat diffusion, following a heat pulse deposition on the material surface, are used to infer material internal properties. In this work we construct solutions for the PDE, using neural forms with the proper initial and boundary conditions embedded. As a benchmark problem, we solve the heat conduction equation in a one-dimensional rod of finite thickness, having both sides thermally insulated. The rod is initially brought at zero temperature, except for the front surface where a non-zero temperature distribution is specified. Future work will consider the extension of this neural network-based approach higher space dimensions and several different geometries. |
Thursday, March 17, 2022 4:12PM - 4:24PM |
W31.00007: Visualization of acoustic mode conversion and power flow in suspended thin-film phononic device Daehun Lee, Qiyu Liu, Huan Li, Lu Zheng, Xuejian Ma, Shawn I Meyer, Songbin Gong, Ruochen Lu, Mo Li, Keji Lai Acoustic wave plays a critical role in electromechanical devices, which are widely utilized in wireless communication and quantum information systems. Since conventional terminal-to-terminal measurement misses important local features of acoustic fields, a new characterization method with high spatial resolution of the acoustic profile is highly sought after. Using a transmission-mode microwave impedance microscope (T-MIM), we directly visualize the acoustic mode conversion and power flow in suspended thin-film phononic devices. Moreover, with fast Fourier transform (FFT) filtering, the forward and backward propagation can be separately analyzed and the propagation loss for both waves calculated. Our result shows that T-MIM can be used as advanced characterization method for acoustic platforms including filters, resonators, phononic crystals, and acoustic metamaterials. |
Thursday, March 17, 2022 4:24PM - 4:36PM |
W31.00008: Full Physical Modeling of EM Signals in a Photon Nano-Sensor Platform Andrew J Nonaka, Revathi Jambunathan, Prabhat Kumar, Zhi Yao Breakthroughs in photon detection devices will have a profound impact for measurement capabilities across a wide range of fields in imaging and sensing. We present recent developments in circuit layout design in the context of photon nano-sensors on a CMOS platform using full physical EM modeling. Characterizing and optimizing the transmission of signals in electrodes connecting carbon nanotube sensors to integrated circuits is crucial in the design of single or low-count photon detectors. Our approach is to use the ARTEMIS code for microelectronics to perform full physical modeling of signals to help in the design process. ARTEMIS is an extreme-scale, GPU-accelerated software framework for modeling electromagnetic wave propagation in microdevice circuitry. Using ARTEMIS we are able to model a wide range of geometrical configurations and material properties which will in turn inform proper device fabrication specifications. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700