Model Engineering R&D
Advanced Core Concepts, and its subsidiary ACTA LLC, provides engineering services designed to create and refine physics-based models of events and threats for the Department of Defense and other federal customers. We have extensive modeling capabilities and a history of valuable data generation across a multitude of engineering and modeling activities.
ACTA uses and develops first principle High Fidelity Physics Based (HFPB) numerical procedures and codes, which include: LS-Dyna, CTH, MSC-Nastran, and CartaBlanca. Over many years, we added new capabilities to the Los Alamos National Laboratory developed multi-physics, multi-material, and multi-phase code CartaBlanca. Applications include:
- Collateral damage assessment for buried munitions
- Cased munition penetrating and exploding inside an urban wall
- Various metal objects penetrating metal sheets and plates
- Breakup of a concrete munitions storage building due to an internal explosion
- Hypervelocity impact and breakup of a missile
- Explosion of hypergolic liquids and container breakup
Many applications are unable to utilize HFPB models directly or at all for reasons such as prohibitive cost and time constraints. In other cases, only empirical data are available to infer underlying behavior. ACTA has extensive experience in developing a wide-range of reduced order models (ROM) to simulate complex physical phenomena whenever HFPB models cannot be used. We have ROM modeling technology to simulate essentially any complex modeling problem for which real or virtual data is available.
A key specialty is training and production of very fast running models (FRMs) parameterized by a large number of input and output variables, which can be developed with a limited set of real or virtual data. Another key specialty of our reduced-order models is formal uncertainty quantification of all model output.
Model Verification and Validation
Model verification determines whether a model is properly coded and designed to operate as intended. Validation involves comparing a calibrated model with data not used in its calibration. The quantification of model and experimental data uncertainties is a necessary part of model validation, so that the degree of “closeness” in the comparison may be evaluated statistically, or at least probabilistically, for comparison with pre-established validation criteria. A model that is properly verified and validated may be used with confidence.
ACTA is a pioneer in developing methods and tools for V&V and actively participated in developing relevant standards such as the ASME V&V Guide. We developed predictive accuracy and quantified uncertainty of fast running models (FRMs) provided to our clients.
A risk assessment often involves processing large amounts of data from multiple sources. We have unparalleled experience in addressing all aspects of the most complex forecasting problem.
Some of our capabilities include Bayesian modeling tools, Markov-Chain Monte Carlo-based tools, environment and event-sensitive path-integral methods, proportional hazard-based prediction tools, and principal components and specialized neural network models. We also develop knowledge-based frameworks that integrate our prognostics tools into modular “enterprise” level solutions.