Structural and Impact Analysis of Large Composite Container
Analyze the complete loading sequence of a large composite container from static pressurization loads to PSD transporation loading to rail impact to transient dynamic drop testing.
Composite Analysis from Static to Impact to Transient Dynamic with LS-DYNA
LS-DYNA is well known in the automotive world for its ability to simulate extreme nonlinear transient events such as car crashes. In the aerospace world, it is the standard for turbine blade-out or aircraft seat crash analysis or bird strike. A common theme to its usage is the simulation of short duration events or one might say millisecond events. These types of simulations are solved using an explicit numerical technique. Although well-suited for these dynamic applications it is a tough fit for applications where the loading is static in nature and an equilibrium solution is required. For this type of analysis, an implicit approach is much better suited. With recent advantages over the last several years, LS-DYNA is now fully capable of both implicit and explicit type analyses using the same FEA model. In the example shown in Figure 1, our engineering team ran a complete suite of load cases from static pressurization, PSD spectrum analysis for aircraft transport, quasi-state rail impact at 4g and then multiple transient drop test cases from edge to flat bottom. A video walkthrough of this work is available by clicking on Figure 1. More details on this work can be found in our LS-DYNA Conference Paper “Broad-Spectrum Stress and Vibration Analysis of Large Composite Container”.
We should back up and mention that this was a FEA consulting services project using LS-DYNA. The reason we started with LS-DYNA is that the several of the load cases required drop and impact analysis. To avoid building two separate models, i.e., one for implicit (statics) and one for explicit (dynamic), we went out of the gate with LS-DYNA. The mesh was highly structured to obtain the highest possible explicit time step given the material mix of steel, aluminum, fiber-glass reinforced plastic (Glass-FRP or GFRP).
Although FEA composite material data can be constructed from manufacturer’s data and or base properties using a rule-of-mixtures (e.g., Chamis Model) for glass or carbon plus the plastic, it is much nicer to have testing done and then blend the data together to create the material card. The LS-DYNA card was configured for progressive failure of the plies by specifying strength and failure strains under tension and compression.
The FEA model used a wide range of material models: 6061-T6, Divinycell H100 foam core; shock isolation mounts (nonlinear springs), steel cables and roughly 8 laminate schedules. Also, the complete range of element types were employed from cables, beams, shells, nonlinear springs, solids and CNRB's.
The analysis required the application of bolt preload, closure latch preload (the container is split into two sections) and gravity prior to the drop test loading. To accomplish this load sequence in one analysis, a static implicit analysis was first performed and then automatically switched to explicit for the transient dynamic drop test. Failure or “no-pass” criterion was met by element erosion due to progressive failure of the composite shell.
To cover transportation induced vibration loads, a PSD analysis was performed. In this analysis sequence, the container was initialized into its transport condition (i.e., bolt preloads, latch preloads and gravity) and then solved for up to 500 modes prior to the PSD calculation. For this 2.4 million DOF problem, PSD solution times were on the order of 80 minutes using 16 CPU-cores (LS-DYNA MPP Double-Precision).
At the end of the LS-DYNA consulting services project we were able to make accurate (validated to test) predictions as to the absolute failure load of the composite container. The progressive failure capabilities of LS-DYNA were impressive since the load distribution from failed plies onto the rest of the structure allowed insight into the load line behavior of the structure under overload conditions. Moreover, this provided keen insight into how to optimize the structure since we never had to dig into the stress or strain states between plies – it either held together or failed providing a clean digital solution. The container is now in production.