Physics Simulator Response Filtering for Validation-Bounded Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current model validation techniques for physics system simulators do not provide assurance that post-validation calibrated (PVC) changes keep predictions within the predetermined model validation domain (MVD), leading to potential violations of model validation boundaries due to unquantified interactions between parametric uncertainties and modeling errors.
Innovation Solution
Implementing a validation assist filter that ensures calibrated model parameters and responses remain within the MVD by filtering out parameter and response variations that could cause predictions to fall outside the MVD, using methods such as singular value decomposition, project pursuit techniques, or neural networks to maintain model accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If post-validation calibration changes are introduced to improve predictions or respond to changes in modeling conditions, then model accuracy and adaptability are improved, but the assurance that predictions remain within the model validation domain is lost
Solution Approach 1:
The patent implements a feedback mechanism where the validation domain boundary is used to monitor and control post-validation calibration changes. The system continuously checks whether calibrated parameters keep predictions within the pre-determined validation domain, providing feedback that ensures reliability while allowing adaptability through controlled parameter adjustments.
Solution Approach 2:
The patent allows parameter changes through post-validation calibration but constrains these changes by requiring that the calibrated parameters must not cause predictions to fall outside the pre-determined validation domain. This resolves the contradiction by enabling parameter adaptation while maintaining boundary assurance through the validation domain constraint.
2Ease of manufacture
If conservative assumptions are used to model interaction terms between parametric uncertainties and modeling errors, then the model validation is simpler and more straightforward, but the validation domain is restricted and further improvements are difficult
Solution Approach 1:
The patent performs preliminary action by pre-determining the validation domain boundary before post-validation calibration is applied. This boundary serves as a pre-established constraint that guides subsequent calibration efforts, allowing improvements while ensuring predictions remain within validated limits. The validation domain is established in advance using conservative assumptions, then used to enable controlled improvements.
3Reliability
If the physics system simulator is revalidated after post-validation changes to ensure predictions remain within the validation domain, then reliability is maintained, but time and resources are consumed
Solution Approach 1:
The patent performs preliminary action by establishing the validation domain boundary in advance, before post-validation calibration is applied. This pre-determined boundary serves as a reference for evaluating whether calibrated parameters maintain predictions within the validation domain, eliminating the need for time-consuming revalidation while preserving reliability assurance.
Data Source
AI summary
This disclosure relates to systems and methods for post-validation adjustment of a physics system simulator using response-based filtering. A single simulator instance is used to generate simulated responses for both a target application model and a set of scaled-down experimental models. A validation assist response filter removes simulated responses that fall outside the boundaries of a mathematically defined model validation domain. The filter may be constructed using pseudo-runs that compute mutual information between the simulated responses of pseudo target models and associated pseudo experimental models. The filtered experimental responses are used by a response calibration module to compute a posteriori application response predictions without modifying underlying model parameters. The approach improves predictive confidence while requiring only a single simulation implementation.