Study On Sensitivity Test And Application Of Neyer D-Optimal Method

Sensitivity is one of key properties of energetic material such as explosive, powder, pyrotechnic powder and initiating explosive device, and is also an important index of safety and reliability of energetic material. However, the determination of sensitivity is a greatly complex task. In addition, the increasing complication of modern weapons needs better safety and reliability of initiating explosive device, which makes sensitivity test of initiating explosive device few and exact as possible as. It is therefore of great importance to study new sensitivity test method. This dissertation studies Neyer D-optimal sensitivity experimental design methods, sensitivity data analyses and application of Neyer D-optimal method in impact sensitivity test.The developments of sensitivity experimental methods are reviewed and various sensitivity tests are compared respectively. Sensitivity experimental methods are classified as two series of methods. One is the first series of sensitivity experimental methods which determine sensitivity distribution with experiments and calculate point estimation and corresponding interval estimation. The other is the second series of sensitivity experimental methods which assume known sensitivity distribution and calculate point estimation and corresponding interval estimation. the design principle of each series of procedure is analyzed and summarized. The mathematic model of D-optimal design theory is established, from which a train of thought can be provided for the new sensitivity experimental design.Based on the features of sensitivity data different from the treatment of a common random variable, the likelihood function suitable for sensitivity data of Neyer D-optimal method is established by means of statistical theorem. The numerical solution of the maximum likelihood estimation and maximization of determinant of the information matrix is given out by the Newton iteration method. The program is made for sensitivity mean and standard deviation with FORTRAN.The application of Neyer D-optimal method is studied in impact sensitivity experiment. The impact experiments of E level HMX explosive are done at various sample size with explosive probability method, Bruceton method, Langlie method and Neyer D-Optimal method respectively. The sensitivity mean and standard deviation of E level HMX explosive are obtained by treatment of experimental data of the four methods and distribution graphs of sensitivity curves are fitted. Sensitivity mean and standard deviation, sensitivity distribution curve and extrapolated results of sensitivity curves are analyzed and primary conclusions about Neyer D-optimal method are drawn from experiments as follow:(i) Neyer D-optimal method can produce efficient sensitivity mean and standarddeviation estimates of the parameters of the distribution, even with limited prior knowledge.(ii) Compared with other experimental methods, sensitivity mean is measured with few sample size in Neyer D-optimal experiment test, which is more suitable for much expensive products.(iii) The extrapolated results of sensitivity curves of Neyer D-optimal method is not good and is less accurate than those of explosive probability method and Bruceton method.(iv)The calculating results of Neyer D-optimal method depends on the assumption of sensitivity distribution.These research results are of significance to the measurement of sensitivity mean and standard deviation of expensive products, and the workers who study new sensitivity experimental method.