Dean Price
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Dean Price has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Brendan Kochunas
University of Michigan
10 shared publications - Mohammed I. Radaideh
Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, USA
6 shared publications - Krishna Garikipati
Department of Mechanical Engineering University of Michigan United States
3 shared publications Nathan Roskoff
Westinghouse Electric Company LLC, 1000 Westinghouse Drive, Cranberry Township, Pennsylvania 16066
3 shared publications- Paul Seurin
Massachusetts Institute of Technology 77 Massachusetts Avenue 02139 Cambridge MA
3 shared publications - Thomas Folk
University of Michigan
3 shared publications
Latest Publications
- Open Access
- 4 November 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 460
Techno-economic optimization of a heat-pipe microreactor, part II: Multi-objective optimization analysis
- 4 November 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 460
Open AccessTechno-economic optimization of a heat-pipe microreactor, part II: Multi-objective optimization analysis
- 29 October 2026
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 114 (21) , 1257
Physics-informed SINDy with application to nuclear reactor dynamics
- 29 October 2026
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 114 (21) , 1257
Physics-informed SINDy with application to nuclear reactor dynamics
- Open Access
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Sequential Optimal Experimental Design for Efficient Training of Reduced Order Models of Reactor Pin Power Distributions
High-fidelity reactor simulations are expensive, and that cost limits how much training data can be realistically generated to build reduced-order models (ROMs). This work develops a sequential optimal experimental design (sOED) algorithm for constructing Proper Orthogonal Decomposition (POD)-based ROMs while keeping the number of full-order model (FOM) evaluations very low. We demonstrate the approach on a heat-pipe microreactor, predicting the 3D pin power distribution as a function of control drum position. POD provides a compact representation of the pin powers, Gaussian Process Regression (GPR) approximates the POD coefficients, and Bayesian Optimization–guided by an approximate error model and an expected-improvement acquisition function–sequentially chooses new inputs at which to generate training data to update and improve the ROM.We compare the performance of several ROMs built from different training data set samplingstrategies: including a random sampling, latin hypercube sampling, manually optimization based on knowledge of the true model error, and the proposed sOED approach, which operates without access to this information. For the problem considered, sOED selects points reasonably close to those obtained through manual selection, while using only seven FOM evaluations to achieve reconstruction accuracy comparable to the reference ROM constructed from the full dataset of 181 simulations. Overall, the results show that a usefully accurate POD-based ROM can be built from fewer than ten expensive Monte Carlo runs, offering a practical, non-intrusive path to lowering ROM training costs–and setting up a natural next step of testing the method on higher-dimensional sets of reactor operation parameters.
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSequential Optimal Experimental Design for Efficient Training of Reduced Order Models of Reactor Pin Power Distributions
High-fidelity reactor simulations are expensive, and that cost limits how much training data can be realistically generated to build reduced-order models (ROMs). This work develops a sequential optimal experimental design (sOED) algorithm for constructing Proper Orthogonal Decomposition (POD)-based ROMs while keeping the number of full-order model (FOM) evaluations very low. We demonstrate the approach on a heat-pipe microreactor, predicting the 3D pin power distribution as a function of control drum position. POD provides a compact representation of the pin powers, Gaussian Process Regression (GPR) approximates the POD coefficients, and Bayesian Optimization–guided by an approximate error model and an expected-improvement acquisition function–sequentially chooses new inputs at which to generate training data to update and improve the ROM.We compare the performance of several ROMs built from different training data set samplingstrategies: including a random sampling, latin hypercube sampling, manually optimization based on knowledge of the true model error, and the proposed sOED approach, which operates without access to this information. For the problem considered, sOED selects points reasonably close to those obtained through manual selection, while using only seven FOM evaluations to achieve reconstruction accuracy comparable to the reference ROM constructed from the full dataset of 181 simulations. Overall, the results show that a usefully accurate POD-based ROM can be built from fewer than ten expensive Monte Carlo runs, offering a practical, non-intrusive path to lowering ROM training costs–and setting up a natural next step of testing the method on higher-dimensional sets of reactor operation parameters.
- Open Access
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Sequential Optimal Experimental Design for Efficient Training of Reduced Order Models of Reactor Pin Power Distributions
High-fidelity reactor simulations are expensive, and that cost limits how much training data can be realistically generated to build reduced-order models (ROMs). This work develops a sequential optimal experimental design (sOED) algorithm for constructing Proper Orthogonal Decomposition (POD)-based ROMs while keeping the number of full-order model (FOM) evaluations very low. We demonstrate the approach on a heat-pipe microreactor, predicting the 3D pin power distribution as a function of control drum position. POD provides a compact representation of the pin powers, Gaussian Process Regression (GPR) approximates the POD coefficients, and Bayesian Optimization–guided by an approximate error model and an expected-improvement acquisition function–sequentially chooses new inputs at which to generate training data to update and improve the ROM.We compare the performance of several ROMs built from different training data set samplingstrategies: including a random sampling, latin hypercube sampling, manually optimization based on knowledge of the true model error, and the proposed sOED approach, which operates without access to this information. For the problem considered, sOED selects points reasonably close to those obtained through manual selection, while using only seven FOM evaluations to achieve reconstruction accuracy comparable to the reference ROM constructed from the full dataset of 181 simulations. Overall, the results show that a usefully accurate POD-based ROM can be built from fewer than ten expensive Monte Carlo runs, offering a practical, non-intrusive path to lowering ROM training costs–and setting up a natural next step of testing the method on higher-dimensional sets of reactor operation parameters.
- 25 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSequential Optimal Experimental Design for Efficient Training of Reduced Order Models of Reactor Pin Power Distributions
High-fidelity reactor simulations are expensive, and that cost limits how much training data can be realistically generated to build reduced-order models (ROMs). This work develops a sequential optimal experimental design (sOED) algorithm for constructing Proper Orthogonal Decomposition (POD)-based ROMs while keeping the number of full-order model (FOM) evaluations very low. We demonstrate the approach on a heat-pipe microreactor, predicting the 3D pin power distribution as a function of control drum position. POD provides a compact representation of the pin powers, Gaussian Process Regression (GPR) approximates the POD coefficients, and Bayesian Optimization–guided by an approximate error model and an expected-improvement acquisition function–sequentially chooses new inputs at which to generate training data to update and improve the ROM.We compare the performance of several ROMs built from different training data set samplingstrategies: including a random sampling, latin hypercube sampling, manually optimization based on knowledge of the true model error, and the proposed sOED approach, which operates without access to this information. For the problem considered, sOED selects points reasonably close to those obtained through manual selection, while using only seven FOM evaluations to achieve reconstruction accuracy comparable to the reference ROM constructed from the full dataset of 181 simulations. Overall, the results show that a usefully accurate POD-based ROM can be built from fewer than ten expensive Monte Carlo runs, offering a practical, non-intrusive path to lowering ROM training costs–and setting up a natural next step of testing the method on higher-dimensional sets of reactor operation parameters.
- Open Access
- 17 August 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 458
MAEO: Scalable Multiobjective Animorphic Ensemble Optimization for large-scale optimization applications
- 17 August 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 458
Open AccessMAEO: Scalable Multiobjective Animorphic Ensemble Optimization for large-scale optimization applications
- 26 May 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 456
Techno-economic optimization of a heat-pipe microreactor, Part I: Theory and cost optimization
- 26 May 2026
- Published by Elsevier in Nuclear Engineering and Design
- Vol. 456
Techno-economic optimization of a heat-pipe microreactor, Part I: Theory and cost optimization
- Open Access
- 30 July 2025
- Published by Elsevier in Nuclear Engineering and Technology
- Vol. 57 (12)
Uncertainty quantification and sensitivity analysis for SPERT III E-core reactivity measurement benchmarking
- 30 July 2025
- Published by Elsevier in Nuclear Engineering and Technology
- Vol. 57 (12)
Open AccessUncertainty quantification and sensitivity analysis for SPERT III E-core reactivity measurement benchmarking
- Open Access
- 6 June 2025
- Published by Elsevier in International Journal of Heat and Mass Transfer
- Vol. 251
Additional considerations in analytical solution for time-dependent heat conduction in a three-dimensional multilayer sphere
- 6 June 2025
- Published by Elsevier in International Journal of Heat and Mass Transfer
- Vol. 251
Open AccessAdditional considerations in analytical solution for time-dependent heat conduction in a three-dimensional multilayer sphere
- 4 November 2024
- Published by Taylor & Francis in Nuclear Science and Engineering
- p. 1-20
Multiphysics Modeling of Heat Pipe Microreactor with Critical Control Drum Position Search
- 4 November 2024
- Published by Taylor & Francis in Nuclear Science and Engineering
- p. 1-20
Multiphysics Modeling of Heat Pipe Microreactor with Critical Control Drum Position Search
- 22 July 2024
- Published by Taylor & Francis in Nuclear Technology
- Vol. 211 (5) , 1014-1043
Calculation of Isotope-Specific Fission Rates for Neutrino Source Calculation in PWR Fuel Diversion Safeguarding
- 22 July 2024
- Published by Taylor & Francis in Nuclear Technology
- Vol. 211 (5) , 1014-1043
Calculation of Isotope-Specific Fission Rates for Neutrino Source Calculation in PWR Fuel Diversion Safeguarding