SEQUENTIAL EXPERIMENTS FOR TEST MINIMIZATION

Characterization and modeling of complex physical processes such as the structural response of a hardened target to detonation of a conventional weapon requires destructuve testing. Due to the expense of such tests, it is essential to keep the number of required destructive tests to a minimum. The technique of Design of Experiments associates each experiment with a term in a polynomial that predicts damage as a function of the inputs. Thus, the terms (but not their coefficients) are known prior to obtaining the results of the experiments. This paper presents an overview of Design of Experiments as a tool for systematic planning and efficient analysis of physical tests. The methodology is used to determine the effect of a single bomb casing fragment on a concrete slab. A sequential experimental design approach decreased the number of required tests from 27 to 17 over more traditional experimental designs. Repeated experiments allowed estimation of natural process variability which increases the statistical validity of the prediction interval of the polynomial. The polynomial can be used directly in a simulation, or if the simulation already has a predictive algorithm, the algorithm can be checked to see if its result falls within the prediction interval of the polynomial.

Language

  • English

Media Info

  • Features: Figures; Tables;
  • Pagination: p. 695-698

Subject/Index Terms

Filing Info

  • Accession Number: 00778318
  • Record Type: Publication
  • ISBN: 0784404216
  • Files: TRIS
  • Created Date: Nov 5 1999 12:00AM