IRIS_2012 OECD/NEA/CSNI benchmark: Numerical simulations of structural impact

Abstract

A benchmark of numerical simulations related to the missile impact on reinforced concrete (RC) slabs has been launched in the frame of OECD/NEA/CSNI research program “Improving Robustness Assessment Methodologies for Structures Impacted by Missiles”, under the acronym IRIS. The goal of the research program is to simulate RC structural, flexural and punching, behavior under deformable and rigid missile impact. The first phase called IRIS_2010 was a blind prediction of the tests performed at VTT facility in Espoo, Finland. The two simulations were performed related to two series of tests: (1) two tests on the impact of a deformable missile exhibiting damage mainly by flexural (so-called “flexural tests”) or global response and (2) three tests on the impact of a rigid missile exhibiting damage mainly by punching response (so-called “punching tests”) or local response. The simulation results showed significant scatter (coefficient of variation up to 132%) for both flexural and punching cases. The IRIS_2012 is the second, post-test, phase of the benchmark with the goal to improve simulations and reduce the scatter of the results. Based on the IRIS_2010 recommendations and to better calibrate concrete constitutive models, a series of tri-axial tests as well as Brazilian tests were performed as a part of the IRIS_2012 benchmark. 25 teams from 11 countries took part in this exercise. Majority of participants were part of the IRIS_2010 benchmark. Participants showed significant improvement in reducing epistemic uncertainties in impact simulations. Several teams presented both finite element (FE) and simplified analysis as per recommendations of the IRIS_2010. The improvements were at the level of simulation results but also at the level of understanding of impact phenomena and its modeling. Due to the complexity of the physical phenomena and its simulation (high geometric and material non-linear behavior) and inherent epistemic and aleatory uncertainties, the scatter still remains important especially comparing to conventional, linear structural engineering analysis. However, the majority of teams achieved the results within ±40% range, comparing to the test results, which is considered acceptable for these types of simulations. The paper provides guidance on how to perform impact simulation in order to achieve acceptable results.