Company: Southwest Research Institute (SwRI)
Contracting Agent: Wright Laboratories, Eglin AFB (WL/MNSA)

Purpose: To develop a probabilistic tunnel vulnerability assessment tool that accounts for multiple tunnel failure modes and that is suitable for screening new weapons concepts.

Background: A new methodology for investigating probabilistic tunnel vulnerability was conceived as part of a major program for the Defense Special Weapons Agency (DSWA) entitled Probabilistic Structural Analysis of Deep Tunnels (PSADT). This methodology was successfully demonstrated to yield new and enlightening insights into the effect of uncertainties on the damage and failure of underground tunnels. The particular damage mode was blocked ingress/egress due to rock fall into the tunnel as a result of close-in detonation. This project was aimed at extending the original single failure-mode methodology to handle multiple failure modes and to demonstrate its use with at least one complex numerical damage model.

Summary: A probabilistic analysis methodology was developed in which failure models for shallow tunnels in a sloping hillside were described and linked together to yield a system failure model. The physical phenomenon considered include earth penetration through layers of soil and rock, tunnel ingress/egress blockage due to rock fall from a close-in explosion, airblast damage to equipment stored in the tunnel, weapon fragmentation damage to equipment stored in the tunnel, and high-velocity debris throw damage to control center lifelines attached to the tunnel roof. The system model integrates these component models into a flexible framework that accounts for correlations between the component models due to common random variables. Several different state-of-the-art probabilistic algorithms were employed to match efficiency with accuracy; thus, the resulting model is suitable for rapid targeting assessments to more computationally intensive concept screening. A method for computing probabilistic sensitivities from simulation samples was also developed and published. Typical outputs from the system model include the tunnel system vulnerability (probability of kill), identification of the probabilistic optimum weapon aimpoint, and probabilistic sensitivity factors, which identify the model parameters contributing the most (and the least) to the tunnel vulnerability.

Key Publications:

Thacker, B.H., and C.J. Oswald, Y.T. Wu, B.C. Patterson, P.E. Senseny, D.S. Riha, "A Probabilistic Multi-Mode Damage Model for Tunnel Vulnerability Assessment, 8th International Symposium on Interaction of the Effects of Munitions with Structures, McLean, Virginia, 1998.

Thacker, B. H. and Y.-T. Wu, "Sampling-Based Probabilistic Sensitivity Analysis," Proceedings of the 24th Midwestern Mechanics Conference, Developments in Mechanics, T. J. Rudolphi and L. W. Zachary (eds.), Iowa State University, Ames, Iowa, Vol. 18, 1995.