Fire Computational Analysis

Modeling and simulation is the mathematical expression of physical theory. In conjunction with experiments, it represents the range of capabilities normally employed for scientific discovery and operational interrogation. To facilitate the modern application of the scientific method through the integrated use of modeling and simulation with experimentation, fire science researchers at SNL have developed unique facilities, methods, models, and measurement tools.

Who We Are

We have a variety of staff who have experience with simulation analysis, model development, and computational theory who employ a suite of computational tools to support mission critical activities.  The computational analysis team works directly with experimentalists to design and interpret fire tests. We draw on an extensive set of simulation tools to support anything from small benchtop parameter studies to large fuel fire predictions. We have the added benefit of being able to work directly with our in-house SIERRA developers to implement desired features in a code. The cycle of work potentially involves models informing experiments, leading to experiments validating models.

Due to some DOE related missions involving the need for quantitative accuracy in model predictions, our in-house codes are designed and maintained to DOE order 414.1D.  The code team works to a software quality plan, and the software is regularly audited for compliance to quality control standards.  SIERRA/Fuego is export controlled, helping assure the advanced capabilities in the software are not disseminated to people who would use the capability to harm us or our interests.

At the core of most of our analysis work is safety.  We are primarily concerned with fire as an operational hazard.  We are focused on the risks associated with fire in high hazard operations.  We solve high consequence fire problems.

Our Capabilities

Validation

  • Validation of our modeling and simulation capabilities provides confidence in the accuracy of the simulations. Sandia has invested in appropriate datasets to quantify accuracies in model simulations for relevant large-scale scenarios.  Our test facilities were designed to enable validation experiments, and the modeling team is experienced in performing validation assessments.
  • We endeavor to maintain current evidence regarding the simulation accuracy to the relevant datasets.

Model Sensitivity

  • We provide more than just predictions, we use models to inform what parameters have high sensitivities. Analysis tools like Dakota can be leveraged to interrogate the physics involved in a computational project and identify physical and model sensitivities. 

High Performance Computing

  • Access to Sandia’s resource of high-performance computing (HPC) machines enables high-fidelity simulations to be run on complex geometries with reasonable turn-around time
  • Full scale CFD models can be run on our unique facilities (FLAME, XTF) with combustion chemistry using the SIERRA/Thermal Fluids tool Fuego.
thermal comp
Simulation of a contaminated gasoline fire with entrainment and dispersion of the liquid contaminant.

What Makes Us Unique

  • The expertise and experience of the scientists and technologists
  • Full-scale to subcomponent level simulation capability
  • A focus on DOE, DOD, and related problems of national interest
  • Outreach to V&V and UQ groups for quantitative accuracy assessments
  • In-house Sierra codes allow for direct model support and development
  • Coupled codes providing multi-physics solutions
  • Ability to securely handle sensitive (classified or proprietary) models
A low-wind conjugate Fuego/Aria simulated 9.1 m square pool fire engulfing a test object.
A low-wind conjugate Fuego/Aria simulated 9.1 m square pool fire engulfing a test object.

Thermal-Fluid Modeling And Related Items

Thermal-fluid modeling takes advantage of many modeling options, with a focus on the Sierra/Thermal-Fluid code suite. This finite-element based, multi-physics computational capability is built upon the Sierra Toolkit framework and takes advantage of Trilinos solver capabilities to enable highly scalable, massively parallel simulation.