The released report contains the details of 22 individual blind numerical tests conducted by DTRA in concordance with ASI. AEM is the mathematical solver on which ASI’s Extreme Loading for Structures (ELS) software is based.
In comparison with the established but confined Finite Element Method (FEM), AEM is relatively new to the scene of blast and progressive collapse analysis.
Developed only 17 years ago through the calculated combination of FEM and DEM techniques at the University of Tokyo by Dr. Hatem Tagel-Din, AEM has had to wait for an opportunity to show the engineering world its full potential. With the release of the comprehensive DTRA report, however, that day has come.
The initial numerical tests, which required that AEM’s results be recorded first to ensure blind study integrity, involved nine blast walls with differing boundary conditions, thicknesses, and distances to the blast’s source. These first tests took place from an entirely virtual standpoint, pitting FEM directly against AEM, like a grizzled veteran that’s been called out by a young challenger. The tests complete, the DTRA contractor wrote that in all nine numerical tests, AEM was in excellent agreement with FEM in regard to load displacement and failure pattern. A draw can be a victory when someone’s proving a point.
Simulated blast column in Three Viewing Modes
The next two tests introduced a new variable with which to compare AEM — a physical experiment. Both tests applied a blast load to a building column modeled in ELS and compared it with live testing results. Because of complicated column behavior and failure localized at the column ends, researchers needed a fully nonlinear dynamic analysis in order to predict accurate column behavior. Essentially, they needed AEM to perform. And it did. When comparing AEM’s virtual results to the live experiments’ results, the contractor concluded that “AEM was able to capture the mode of deformation, damage locations at the top and bottom, overall deformed shape, and the effects of column damage to the structure in reasonable time and reliable accuracy.” The reasonable time in which AEM operates is best illustrated by comparisons with other methods. AEM took 12 hours to complete its analysis, which according to the DTRA report author is “a relatively short time compared to other analysis methods.” Such unmatched efficiency is another testament to how AEM and ELS are revolutionizing the world of structural engineering software.
The third and final slew of tests researchers threw at AEM and ELS were the most challenging, composed of 11 blind numerical case studies meant to compare AEM and FEM for progressive collapse applications. A complicated model with reinforcement details, nonlinear behavior of concrete and steel, separation between elements, and a potential for collision when failure takes place gave each of the tests a desired level of realism expected from an actual collapse by such a structure. A 5 story reinforced concrete building was modeled by Extreme Loading for Structures and an undisclosed FEM proprietary software. Modeling in the AEM software was “significantly easier and faster than FEM,” according to the report. Both methods matched each other closely for accuracy before collapse, however, once collapse occurred FEM based software fell short. The study found that not only did AEM not have the problem of hour glassing frequently associated with FEM, but it also did not show collapse as FEM did when no collapse was expected to occur.
The conclusion of the researchers and, specifically, the DTRA contractor, was that AEM, being a “well-established method for structural analysis,” is simply more practical than FEM in the use of 3-d brick elements during the demolition of complicated structures.
To view the full DTRA report, please click here: DTRA Report