Statistical Analysis of Mechanical‐Shock Fragility Test Results
Statistical Analysis of Mechanical‐Shock Fragility Test Results
复制标题
机械冲击脆性测试结果的统计分析
DOI：
10.1002/pts.2290
复制
发表时间：
2017
影响因子：
2.6
通讯作者：
Takamasa Nakajima
中科院分区：
文献类型：
--
作者：
Shogo Horiguchi;Takamasa Nakajima
根据日本工业标准Z 0119：2002，用于包装和产品设计的机械冲击脆性试验方法在统计分析方面存在一些问题。一个这样的问题是所施加的冲击水平逐步增加。我们无法确定每个样品的真实脆性值。这就是所谓的interval censored data。如果为每个样本选择最大非破坏性冲击水平，则计算的平均值将过小。另一个问题是，冲击的次数被限制为最大5或6次，以便使累积疲劳的影响最小化。因此，如果我们确定相当高的首次冲击水平，则通过产品在经受计划的首次冲击时的失效获得初始删失数据。如果我们确定了相当低的最终冲击水平，则在产品未因计划的最终冲击而损坏时获得最终删失数据。这些问题使统计分析变得困难。因此，我们提出了一种改进的数据分析方法，包括区间删失，初始删失和最终删失数据，并提供了该方法应用于跌落试验结果的例子。算例结果表明，该方法是可行的，但截尾数据会产生误差，使统计量的精度损失。最后，通过对ISO 7765 - 1：1988阶梯法和JIS Z 0119：2002阶梯法的比较，分析了各自的优缺点。版权所有© 2017约翰威利父子有限公司
The mechanical‐shock fragility testing methods applied to packaging and product designs according to Japan Industrial Standards Z 0119:2002 incur some problems related to statistical analysis. One such problem is that the applied shock levels are increased in steps. We cannot determine the true fragility values for each sample. This is called interval‐censored data. If the maximum level of non‐damaging shock is selected for each sample, the calculated average will be too small. Another problem is that the number of shocks is limited to a maximum of 5 or 6, so as to minimize the effect of the accumulated fatigue. Therefore, if we determine the considerably high first shock level, initial‐censored data are obtained by the failure of the product when subjected to the planned first shock. If we determine the considerably low final shock level, final‐censored data are obtained when the product has not been damaged by the application of the planned final shock. These problems make statistical analysis difficult. We therefore propose an improved analysis method for data including interval‐censored, initial‐censored, and final‐censored data, and provide examples in which the method is applied to the results of a drop test. The results of the examples show that the proposed method is practical and that the censored data yields errors and loss of accuracy for the statistics. Finally, by means of a comparison between the staircase method of ISO 7765‐1:1988 and the method of JIS Z 0119:2002, we analyse the advantages and disadvantages of the respective methods. Copyright © 2017 John Wiley & Sons, Ltd.