Optimization Techniques in Conjunction with Complex ATD FE Models Using LS-DYNA

This paper discusses the optimization methods used by Denton to create FE ATD models. Anthropomorphic Test Devices (ATDs also known as Dummies) are manufactured with a variety of hyperelastic and viscoelastic materials. The production processes used to manufacture ATD’s result in variations between dummies that can have a significant influence on dummy performance. In addition the load cases present during testing with ATDs are very complex and of short time duration. A typical frontal crash test may last 100ms but the load cycle on an ATD body region may last only 10ms. The end result is a complex series of issues that can be very difficult to solve in an FE model, using conventional techniques. The number of potential variables (sample list) associated with a specific performance metric (time history, acceleration, etc) can be very large. The settings for a given variable and the effect of that variable setting are difficult to determine given the complexity and short duration of the simulated event. In particular, the use of the hybrid adaptive SHERPA algorithm in the HEEDS software in conjunction with LS-DYNA for determining parameter values is discussed.