Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12367408-B2
Publication Date
2025-07-22
Expiration Date
Abstract
This invention relates to a parameter or response assist filter that ensures that the predictions of a post-validation calibrated physics system simulator will remain within boundaries of a predetermined model validation domain. Embodiments utilize one or more filters to ensure calibrated model parameters {acute over (P)} and/or calibrated responses {tilde over (ϕ)} cause physics simulator model predictions to remain within the boundaries of the model validation domain MVD for a target application. The filters can be constructed prior to use or automatically inferred, or otherwise determined, from available measurements and other renditions of the physics system simulator during operation.
Core Innovation
The invention addresses limitations of conventional expert-guided calibration for a validated physics system simulator by providing post-validation adjustment for a post-validation calibrated (PVC) physics system simulator. The approach uses a model validation domain (MVD) whose boundaries are evaluated mathematically based on deterministic or stochastic multi-variate functions of target application model responses, scaled-down experimental-models' responses, experimental measurements, derivatives, and an uncertainty estimator.
The method predicts first experimental responses by modeling the physical system using scaled-down experimental models based on model parameters and corresponding parameter variations. A validation assist parameter filter evaluates parameter variations by filtering those that correspond to variations in responses for experimental models that cause predictions of a second physics system simulator, denoted as the PVC physics system simulator, for target application responses to fall outside of the MVD of the first physics system simulator.
After updating, a parameter calibration module adjusts the physical parameters based on the updated scaled-down experimental responses, the corresponding sets of experimental measurements, and the filtered parameter variations. The method finally predicts an a posteriori application response of the physical system by modeling the physical system using the application model based on the adjusted physical parameters. The MVD boundary evaluation and the filtration can further involve uncertainty estimators and multi-variate boundary mathematical functions, as well as selection or removal using entropy or information criteria comparing outputs from first and second simulator implementations and pseudo target/application and pseudo sets of scaled-down experimental models.
Claims Coverage
The independent claim provided is clm-00001. It includes four main inventive feature groups within one method for post-validation adjustment: model validation domain (MVD) boundary evaluation, predicting experimental responses with a first implementation, filtering parameter variations using an MVD boundary filter operator, and updating, calibrating parameters, and predicting an a posteriori application response. The dependent claims in the partial content further specify filtering criteria, fidelity selection, and parameter-feature selection techniques.
Model validation domain boundaries for post-validation adjustment
A method in which the physics system simulator is validated for a target application model (MA), as described by a model validation domain (MVD) the boundaries of which are evaluated mathematically based on deterministic or stochastic multi-variate functions of the target application model's responses, a set of scaled-down experimental-models' responses, corresponding sets of experimental measurements (φE1, φE2, ...), derivatives thereof, and the model parameters (P), the corresponding parameter variations (ΔP), and an uncertainty estimator.
Filtering parameter variations by preventing out-of-domain target predictions
Filtering, by a validation assist parameter filter having an MVD boundary filter operator, parameter variations (ΔP) corresponding to variations in responses for experimental models (φE) that cause the predictions of a second physics system simulator, denoted by post-validation calibrated (PVC) physics system simulator, for the target application responses to fall outside of the MVD of the first physics system simulator.
Updating experimental responses and calibrating parameters using filtered variations
Updating, by the first implementation of the physics system simulator, the first scaled-down experimental responses by modeling the physical system using the scaled-down experimental models based on the physical parameters (P) and their corresponding filtered parameter variations (fΔP); and adjusting, with a parameter calibration module, the physical parameters (P) based on the updated first scaled-down experimental responses, the corresponding sets of experimental measurements, and the filtered parameter variations (fΔP).
Posterior target application prediction using adjusted parameters
Predicting, by the first implementation of the physics system simulator, a posteriori application response (Φ̃A) of the physical system by modeling the physical system using the application model (MA) based on the adjusted physical parameters (P̃).
Mutual-information threshold filtering rule
The method where the filtering is performed by selecting results where mutual information increases beyond a threshold set by comparison of scaled-down experimental responses from two separate physics system simulator instances.
High-fidelity and low-fidelity simulator implementations
The method where a first implementation of a physics system simulator is high-fidelity and a second implementation of the same simulator is low-fidelity.
Parameter feature selection using singular value decomposition, project pursuit, and neural networks
The method where parameter features are selected using singular value decomposition, project pursuit techniques, and/or neural networks.
Across the provided independent claim and described dependent refinements, the core claim coverage centers on validating an application through an MVD whose boundaries are mathematically evaluated from target and experimental responses plus uncertainty estimation, then performing post-validation adjustment by filtering parameter variations to keep PVC-target predictions inside the MVD, followed by updating experimental responses, calibrating physical parameters using the filtered variations and experimental measurements, and predicting an a posteriori application response with adjusted parameters. Additional inventive features in the partial content specify an information-theoretic mutual-information threshold filtering rule, fidelity differences between implementations, and parameter feature selection via singular value decomposition, project pursuit, and/or neural networks.
Stated Advantages
Prevents the predictions of the post-validation calibrated (PVC) physics system simulator for target application responses from falling outside of the model validation domain (MVD).
Provides an MVD boundary filter operator and post-validation adjustment procedure that uses filtered parameter variations to enable posterior prediction using adjusted physical parameters.
Documented Applications
Validating advanced-fuel designs for nuclear power plants.
Fuel-testing and reactor design.
Burn-up credit.
Irradiated fuel inventory assessment.
Anomaly detection / condition monitoring.
Materials transport.
Weapons-aging code validation.
Detonation modeling of high explosives / aging nuclear weapons.
Interested in licensing this patent?