Dynas+

Case studies

Reducing uncertainties in concrete impact simulations

2 min read
Industries
DefenseNuclear
Disciplines
Impact & blastOptimizationMaterials
Technologies
LS-DYNA · LS-OPT · LS-PrePost
In partnership with
IRSN & CNSC (OCDE)
Reducing uncertainties in concrete impact simulations

The case

As part of the international IRIS 2010 and later IRIS 2012 benchmarks, organized by IRSN and the CNSC under the aegis of the OECD, Dynas+ developed, at IRSN's request, a methodology to improve the reliability of impact simulations on reinforced concrete structures performed with LS-DYNA. The first benchmark reproduced a situation very close to industrial projects: only a few global concrete properties were known. Teams had to produce a "blind" prediction before the experimental results were published, on tests deliberately placed near the ballistic limit between arrest and perforation. The IRIS 2012 benchmark went further by providing elementary concrete characterization tests in order to evaluate the contribution of a genuine identification of material parameters prior to the impact simulations.

The solution deployed

Dynas+ developed a progressive methodology combining the capabilities of LS-DYNA and LS-OPT in order to adapt the strategy to the level of information available on the material and the test. When only a few characteristics were known:

  • automatic generation of the parameters for the LS-DYNA concrete material laws from the CEB-FIP recommendations;
  • comparison of the main available material laws in order to identify those best suited to impacts with perforation (MAT_72R3, MAT_159, etc.);
  • Monte Carlo sensitivity analyses to identify the truly influential parameters;
  • robustness studies to evaluate the scatter in results due to material uncertainties;
  • recalibration of the influential parameters through LS-OPT optimization as soon as the first experimental results became available.

When material tests were provided (IRIS 2012), the approach was enriched with automatic identification of the concrete law parameters directly from unconfined and confined compression tests, and their direct integration into the impact models.

Reducing uncertainties in concrete impact simulations, the solution deployed

The results achieved

The work showed that the reliability of a concrete impact simulation depends as much on the quality of the input data as on the ability to evaluate the robustness of the predictions. When the concrete properties are poorly known, sensitivity and robustness analyses make it possible to identify the most influential parameters and quantify the uncertainties. Conversely, when characterization tests are available, parameter identification with LS-OPT significantly improves the predictive capability of the models.

This work also highlights that concrete properties are only one of the sources of uncertainty: reinforcement, impact conditions, structural geometry, and aggregate size also influence the structure's response. A reliable simulation therefore cannot be reduced to providing a single value: it must make it possible to quantify the level of confidence associated with the prediction.

This work was the subject of three scientific papers presented at LS-DYNA conferences (Strasbourg, Detroit, Manchester) on the modeling and identification of concrete parameters under impact.

Reducing uncertainties in concrete impact simulations, the results achieved