skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Validating neural-network refinements of nuclear mass models

Publication Date:
Sponsoring Org.:
USDOE Office of Science (SC), Nuclear Physics (NP) (SC-26)
OSTI Identifier:
Grant/Contract Number:
Resource Type:
Journal Article: Publisher's Accepted Manuscript
Journal Name:
Physical Review C
Additional Journal Information:
Journal Volume: 97; Journal Issue: 1; Related Information: CHORUS Timestamp: 2018-01-16 10:17:45; Journal ID: ISSN 2469-9985
American Physical Society
Country of Publication:
United States

Citation Formats

Utama, R., and Piekarewicz, J.. Validating neural-network refinements of nuclear mass models. United States: N. p., 2018. Web. doi:10.1103/PhysRevC.97.014306.
Utama, R., & Piekarewicz, J.. Validating neural-network refinements of nuclear mass models. United States. doi:10.1103/PhysRevC.97.014306.
Utama, R., and Piekarewicz, J.. 2018. "Validating neural-network refinements of nuclear mass models". United States. doi:10.1103/PhysRevC.97.014306.
title = {Validating neural-network refinements of nuclear mass models},
author = {Utama, R. and Piekarewicz, J.},
abstractNote = {},
doi = {10.1103/PhysRevC.97.014306},
journal = {Physical Review C},
number = 1,
volume = 97,
place = {United States},
year = 2018,
month = 1

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on January 16, 2019
Publisher's Accepted Manuscript

Save / Share:
  • A diagonal recurrent neural network (DRNN) and a fuzzy logic controller (FLC) were designed for optimal temperature control. The DRNN and FLC utilize the response of a fifth-order, one delayed neutron group linear model to an optimal controller that was previously designed. In this paper, the results of testing a real-time implementation of the DRNN and FLC are presented. The method of testing that was utilized is called hardware in loop simulation (HILS). Here, a distributed simulation of a highly non-linear model of the reactor on a network of workstations is used in the final step of testing prior tomore » experimentation on the reactor. Also, in this setup actual control hardware is used to validate the controller realizations.« less
  • Cited by 15
  • Relativistic mean-field models of nuclear structure have been enormously successful at reproducing ground-state properties of finite nuclei throughout the periodic table using a handful of accurately calibrated parameters. In this contribution, we use powerful theoretical, experimental, and observational constraints on the equation of state of asymmetric nuclear matter--not employed in the calibration procedure--to validate two such models: NL3 and FSUGold. It is observed that FSUGold is consistent with all these constraints, except perhaps for a high-density equation of state that appears mildly softer than required by astronomical observations. It is argued that incorporating such constraints goes a long way inmore » removing much of the ambiguity left over from the standard calibrating procedure.« less
  • A combination of secondary ion mass spectrometry, optical profilometry and a statistically-driven algorithm was used to develop a non-contact volume analysis method to validate the useful yields of nuclear materials. The volume analysis methodology was applied to ion sputter craters created in silicon and uranium substrates sputtered by 18.5 keV O- and 6.0 keV Ar+ ions. Sputter yield measurements were determined from the volume calculations and were shown to be comparable to Monte Carlo calculations and previously reported experimental observations. Additionally, the volume calculations were used to determine the useful yields of Si+, SiO+ and SiO2+ ions from the siliconmore » substrate and U+, UO+ and UO2+ ions from the uranium substrate under 18.5 keV O- and 6.0 keV Ar+ ion bombardment. This work represents the first steps toward validating the interlaboratory and cross-platform performance of mass spectrometry for the analysis of nuclear materials.« less
  • This paper describes an artificial neural network response surface methodology (ANNRSM) for process modeling and optimization. The process chosen is that of chemical vapor deposition (CVD) of silicon in a barrel reactor. A desired performance requirement of the barrel CVD reactor is that the deposited layers be uniform in thickness. For modeling this process, experiments are first planned and conducted following the design of experiments (DOE) methodology. The resulting experimental data are mapped with an artificial neural network (ANN). ANNs with different configurations are systematically trained in a simple to complex order by a back-propagation training procedure. Another set ofmore » designed experimental data is used to test the predictive accuracy of the ANNs and to identify the network with optimum configuration of the networks. The selected model, ANN response surface, in conjunction with a gradient search scheme is used to locate the optimum settings. The results of using this methodology in identifying optimal settings in the presence of noise are also presented. Experiments performed on a mock-up CVD reactor support the optimum settings obtained using the ANNRSM. A comparison between ANNRSM and regression RSM shows that ANNRSM is able to build an accurate global model and find the optimum using fewer data especially when the data are noisy.« less