引信解除保险距离兰利法试验模拟研究

在测定引信解除保险距离的感度用数理统计试验方法中，兰利法由于理论上可使用较少的试验样本量而应用日渐广泛。利用蒙特卡罗法对引信解除保险距离的兰利法试验进行了计算机模拟计算，结果表明：虽然使用小样本即可实现对解除保险距离50%响应点的准确估计，但对解除保险距离方差的估计却有较大的误差和散布；经模拟计算虽然可以针对统计值得到方差估计值的修正系数，但这样的修正系数对于单次试验却无任何意义；这就意味着用兰利法试验测估引信解除保险距离所得结果会存在较大的误差，并且如果不加修正，则会使测定的解除保险距离上限偏短而下限偏长。

Simulation for Langlie Method Test of Fuze Arming Distance

In mathematically statistical test methods that were used to get arming distance of fuze, Langlie method was applied more and more wider because it only uses a few samples. Langlie method test of fuze arming distance was simulated and calculated with Mote-Carlo method. Its result shows that the estimate of 50 % response level of stimulus for arming distance can be got through using a little sample, but the error and scatter to estimate of variance for arming distance is large; the amending co?efficient to estimate of variance can be got through simulating and calculating, but it is useless to the certain single test; this means that the estimated arming distance of fuze by the langlie method test has large error, and the maximum of arming distance becomes shorter and the minimum of that becomes longer without amending.