View in EDS

Research on PGNAA online multi-element quantitative method based on SE-Auto-MBR neural network.

Saved in:
Bibliographic Details
Title: Research on PGNAA online multi-element quantitative method based on SE-Auto-MBR neural network.
Authors: Zhang, You-Jian1,2,3 (AUTHOR), Zhang, Yan1,2 (AUTHOR) yanzhang@ecut.edu.cn, Zhang, Hao-ran1,2 (AUTHOR), Hu, Xuan-di1,2 (AUTHOR), Liu, Shi-Liang1,2 (AUTHOR), Luo, Fang-zheng1,2,3 (AUTHOR), Wang, Ren-Bo1,2,3,4 (AUTHOR)
Source: Journal of Radioanalytical & Nuclear Chemistry. Mar2026, Vol. 335 Issue 3, p1929-1943. 15p.
Subject Terms: *Elemental analysis, *Nuclear activation analysis, *Artificial neural networks, *Measurement uncertainty (Statistics), *Deep learning, *Autoencoders
Abstract: Prompt Gamma Neutron Activation Analysis (PGNAA) enables real-time multi-element analysis of bulk materials but suffers from severe peak overlap, high noise, and complex backgrounds. To address these challenges, we propose SE-Autoencoder-MBR Net, an end-to-end deep learning model integrating Squeeze-and-Excitation (SE) attention, a symmetric autoencoder, and multi-branch heteroscedastic regression. The SE module recalibrates spectral features to enhance characteristic peaks of Ca, Si, Fe, and Al, while the autoencoder enforces spectral morphology preservation and implicit denoising. Crucially, the multi-branch regression outputs element concentrations alongside input-dependent uncertainty estimates, enabling adaptive task weighting via heteroscedastic negative log-likelihood. Experiments on a Monte Carlo–simulated cement dataset—verified for spectral fidelity against experimental measurements—demonstrate that the method outperforms CNN and PLS baselines, achieving an average R2 > 0.97 and RMSE < 0.47. Ablation studies confirm the contribution of each module, highlighting the framework's potential for accurate, robust, and uncertainty-aware industrial monitoring. [ABSTRACT FROM AUTHOR]
Database: Energy & Power Source
FullText Text:
  Availability: 0
Header DbId: enr
DbLabel: Energy & Power Source
An: 192873816
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Research on PGNAA online multi-element quantitative method based on SE-Auto-MBR neural network.
– Name: Author
  Label: Authors
  Group: Au
  Data: &lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Zhang%2C+You-Jian%22&quot;&gt;Zhang, You-Jian&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Zhang%2C+Yan%22&quot;&gt;Zhang, Yan&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;i&gt; yanzhang@ecut.edu.cn&lt;/i&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Zhang%2C+Hao-ran%22&quot;&gt;Zhang, Hao-ran&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Hu%2C+Xuan-di%22&quot;&gt;Hu, Xuan-di&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Liu%2C+Shi-Liang%22&quot;&gt;Liu, Shi-Liang&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Luo%2C+Fang-zheng%22&quot;&gt;Luo, Fang-zheng&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Wang%2C+Ren-Bo%22&quot;&gt;Wang, Ren-Bo&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,3,4&lt;/relatesTo&gt; (AUTHOR)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: &lt;searchLink fieldCode=&quot;JN&quot; term=&quot;%22Journal+of+Radioanalytical+%26+Nuclear+Chemistry%22&quot;&gt;Journal of Radioanalytical &amp; Nuclear Chemistry&lt;/searchLink&gt;. Mar2026, Vol. 335 Issue 3, p1929-1943. 15p.
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: *&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Elemental+analysis%22&quot;&gt;Elemental analysis&lt;/searchLink&gt;&lt;br /&gt;*&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Nuclear+activation+analysis%22&quot;&gt;Nuclear activation analysis&lt;/searchLink&gt;&lt;br /&gt;*&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Artificial+neural+networks%22&quot;&gt;Artificial neural networks&lt;/searchLink&gt;&lt;br /&gt;*&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Measurement+uncertainty+%28Statistics%29%22&quot;&gt;Measurement uncertainty (Statistics)&lt;/searchLink&gt;&lt;br /&gt;*&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Deep+learning%22&quot;&gt;Deep learning&lt;/searchLink&gt;&lt;br /&gt;*&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Autoencoders%22&quot;&gt;Autoencoders&lt;/searchLink&gt;
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Prompt Gamma Neutron Activation Analysis (PGNAA) enables real-time multi-element analysis of bulk materials but suffers from severe peak overlap, high noise, and complex backgrounds. To address these challenges, we propose SE-Autoencoder-MBR Net, an end-to-end deep learning model integrating Squeeze-and-Excitation (SE) attention, a symmetric autoencoder, and multi-branch heteroscedastic regression. The SE module recalibrates spectral features to enhance characteristic peaks of Ca, Si, Fe, and Al, while the autoencoder enforces spectral morphology preservation and implicit denoising. Crucially, the multi-branch regression outputs element concentrations alongside input-dependent uncertainty estimates, enabling adaptive task weighting via heteroscedastic negative log-likelihood. Experiments on a Monte Carlo–simulated cement dataset—verified for spectral fidelity against experimental measurements—demonstrate that the method outperforms CNN and PLS baselines, achieving an average R2 &gt; 0.97 and RMSE &lt; 0.47. Ablation studies confirm the contribution of each module, highlighting the framework&#39;s potential for accurate, robust, and uncertainty-aware industrial monitoring. [ABSTRACT FROM AUTHOR]
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=enr&AN=192873816
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s10967-026-10745-y
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 15
        StartPage: 1929
    Subjects:
      – SubjectFull: Elemental analysis
        Type: general
      – SubjectFull: Nuclear activation analysis
        Type: general
      – SubjectFull: Artificial neural networks
        Type: general
      – SubjectFull: Measurement uncertainty (Statistics)
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Autoencoders
        Type: general
    Titles:
      – TitleFull: Research on PGNAA online multi-element quantitative method based on SE-Auto-MBR neural network.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Zhang, You-Jian
      – PersonEntity:
          Name:
            NameFull: Zhang, Yan
      – PersonEntity:
          Name:
            NameFull: Zhang, Hao-ran
      – PersonEntity:
          Name:
            NameFull: Hu, Xuan-di
      – PersonEntity:
          Name:
            NameFull: Liu, Shi-Liang
      – PersonEntity:
          Name:
            NameFull: Luo, Fang-zheng
      – PersonEntity:
          Name:
            NameFull: Wang, Ren-Bo
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 03
              Text: Mar2026
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-print
              Value: 02365731
          Numbering:
            – Type: volume
              Value: 335
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Journal of Radioanalytical & Nuclear Chemistry
              Type: main
ResultId 1