Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations
This article presents a new approach for the efficient calculation of sensitivities in radiation dose estimates, subject to imprecisely known nuclear material cross-section data. The method is a combined application of adjoint-based models to perform, simultaneously, both the sensitivity calculation...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/15/14/5102 |
| _version_ | 1797406561830699008 |
|---|---|
| author | Andrew G. Buchan Dan G. Cacuci Steven Dargaville Christopher C. Pain |
| author_facet | Andrew G. Buchan Dan G. Cacuci Steven Dargaville Christopher C. Pain |
| author_sort | Andrew G. Buchan |
| collection | DOAJ |
| description | This article presents a new approach for the efficient calculation of sensitivities in radiation dose estimates, subject to imprecisely known nuclear material cross-section data. The method is a combined application of adjoint-based models to perform, simultaneously, both the sensitivity calculation together with optimal adaptive mesh refinement. Adjoint-based sensitivity methods are known for their efficiency since they enable sensitivities of all parameters to be formed through only two solutions to the problem. However, the efficient solutions can also be obtained by their computation on optimal meshes, here guided by goal-based adjoint approaches. It is shown that both mesh adaptivity and sensitivity can be computed by the same adjoint solution, meaning both can be performed without additional costs. A simple demonstration is presented based on the Maynard fixed-source problem where uncertainties, with respect to material cross-sections, of doses received in local regions are examined. It is shown that the method is able to calculate sensitivities with reduced computational costs in terms of memory and potentially computational time through reduced mesh size when using adaptive resolution in comparison to uniform resolution. In particular, it is the local spatial contributions to the sensitivity that are resolved more effectively due to the adaptive meshes concentrating resolution in those areas contributing the most to its value. |
| first_indexed | 2024-03-09T03:28:12Z |
| format | Article |
| id | doaj.art-39806592b437498b9771f743a6f0a5d3 |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-09T03:28:12Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-39806592b437498b9771f743a6f0a5d32023-12-03T14:58:58ZengMDPI AGEnergies1996-10732022-07-011514510210.3390/en15145102Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response CalculationsAndrew G. Buchan0Dan G. Cacuci1Steven Dargaville2Christopher C. Pain3School of Engineering and Materials Science, Queen Mary University of London, Mile End Road, Bethnal Green, London E1 4NS, UKMechanical Engineering College of Engineering and Computing, University of South Carolina, Columbia, SC 29208, USAEarth Science and Engineering, Imperial College London, Exhibition Road, South Kensington, London SW7 2BX, UKEarth Science and Engineering, Imperial College London, Exhibition Road, South Kensington, London SW7 2BX, UKThis article presents a new approach for the efficient calculation of sensitivities in radiation dose estimates, subject to imprecisely known nuclear material cross-section data. The method is a combined application of adjoint-based models to perform, simultaneously, both the sensitivity calculation together with optimal adaptive mesh refinement. Adjoint-based sensitivity methods are known for their efficiency since they enable sensitivities of all parameters to be formed through only two solutions to the problem. However, the efficient solutions can also be obtained by their computation on optimal meshes, here guided by goal-based adjoint approaches. It is shown that both mesh adaptivity and sensitivity can be computed by the same adjoint solution, meaning both can be performed without additional costs. A simple demonstration is presented based on the Maynard fixed-source problem where uncertainties, with respect to material cross-sections, of doses received in local regions are examined. It is shown that the method is able to calculate sensitivities with reduced computational costs in terms of memory and potentially computational time through reduced mesh size when using adaptive resolution in comparison to uniform resolution. In particular, it is the local spatial contributions to the sensitivity that are resolved more effectively due to the adaptive meshes concentrating resolution in those areas contributing the most to its value.https://www.mdpi.com/1996-1073/15/14/5102Boltzmann transport equationadjoint sensitivity analysisgoal-based mesh optimisation |
| spellingShingle | Andrew G. Buchan Dan G. Cacuci Steven Dargaville Christopher C. Pain Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations Energies Boltzmann transport equation adjoint sensitivity analysis goal-based mesh optimisation |
| title | Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations |
| title_full | Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations |
| title_fullStr | Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations |
| title_full_unstemmed | Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations |
| title_short | Optimised Adjoint Sensitivity Analysis Using Adjoint Guided Mesh Adaptivity Applied to Neutron Detector Response Calculations |
| title_sort | optimised adjoint sensitivity analysis using adjoint guided mesh adaptivity applied to neutron detector response calculations |
| topic | Boltzmann transport equation adjoint sensitivity analysis goal-based mesh optimisation |
| url | https://www.mdpi.com/1996-1073/15/14/5102 |
| work_keys_str_mv | AT andrewgbuchan optimisedadjointsensitivityanalysisusingadjointguidedmeshadaptivityappliedtoneutrondetectorresponsecalculations AT dangcacuci optimisedadjointsensitivityanalysisusingadjointguidedmeshadaptivityappliedtoneutrondetectorresponsecalculations AT stevendargaville optimisedadjointsensitivityanalysisusingadjointguidedmeshadaptivityappliedtoneutrondetectorresponsecalculations AT christophercpain optimisedadjointsensitivityanalysisusingadjointguidedmeshadaptivityappliedtoneutrondetectorresponsecalculations |