UTILIZING COMPUTATIONAL PROBABILISTIC METHODS TO DERIVE SHOCK SPECIFICATIONS IN A NONDETERMINISTIC ENVIRONMENT

摘要

Certain classes of engineering systems must be designed to survive operational shock environments which are difficult or impossible to describe in a comprehensive fashion. The shock response spectrum (SRS), a common measure of the severity of a shock signal, is frequently used in this situation as a standard for system qualification. Typically, the SRS is calculated by performing a series of extreme-level shock tests that are assumed to envelope the operating environment of the system. If the system can survive these tests, the SRS from each test are combined in some conservative manner to derive a single spectrum. Often, a deterministic safety factor is then applied to this composite SRS, resulting in a reference specification for design qualification. It has been shown, however, that use of the SRS in this manner can lead to an imbalance in marginal reliability. Nevertheless, this technique remains the industry's standard means of assessing system response in a shock environment. In this article, techniques utilizing computational probabilistic methods are proposed to derive an analytically-based reference SRS that prescribes a balanced level of marginal reliability. In addition, it provides a technically sound procedure for constructing design specifications in a nondeterministic shock environment. For illustrative purposes, the method is applied to a soil penetration system consisting of a nonlinear transient dynamics calculation performed on a complex finite element structural mechanics model, coupled with a spherical cavity expansion representation of the soil medium.