基于二项分布的测试性试验方案优化设计与评估

Optimal Design and Evaluation of Testability Test Scheme Based on Binomial Distribution

测试性验证试验的目的是验证产品是否满足设计指标要求,以故障注入为手段,模拟相应的故障模式,通过规定的故障注入方法进行实际测试,并对产生的结果进行判断是否满足设计指标的要求。为提高测试性试验的一次通过率,降低测试性试验经费和试验时间,提出了基于二项分布的测试性试验方案优化设计与评估方法。实现了已知置信度、单侧置信下限、不能检测的故障个数条件下,求解测试性试验中的“最小样本量”的问题。可以对测试性试验进行提前预判,解决了一定约束条件下测试性试验最小样本量的求解及单侧置信下限评估问题。通过GUIDE人机对话框,不仅交互界面直观,而且极大提高了测试性验证评估的效率和准确性。

The purpose of testability verification test is to verify whether the product meets the design index requirements.Fault injection is used as a means to simulate the corresponding fault mode,conduct actual testing through the specified fault injection method,and judge whether the generated results meet the design index requirements.In order to improve the one-time pass rate of testability test and reduce test cost and test time,a method of optimal design and evaluation of testability test scheme based on binomial distribution was proposed.The problem of“smallest sample size”in testability test is solved under the condition of confidence degree,unilateral confidence lower limit and number of undetectable faults.It can predict the testability test in advance,and solve the problem of solving the smallest sample size of testability test and evaluation the confidence lower limit of single test under certain constraints.Through the GUIDE human-machine dialog box,the interactive interface is not only intuitive,but also greatly improves the efficiency and accuracy of testability verification and evaluation.