Fault sample size quadratic allocation method of sequential verification based on Sobol sequence

Aiming at the problem that the current testability sequential verification cannot guarantee the certainty, randomness and coverage requirements of the extracted fault mode, which leads to the insufficient credibility of the testability index evaluation, a method of quadratic allocation of sequential verification fault sample size based on Sobol sequence is proposed. Firstly, based on the hierarchical Bayes network model, the allocation factors of nodes and fault modes are determined by FMECA information. Then, by comparing the homogeneity of different sequences, the sampling sequence is determined, and the sequential verification method of quadratic allocation of fault sample size is further proposed. Finally, the paper makes a comparative analysis through the actual cases. The results show that, compared with pseudo-random sequence sampling, the method can effectively reduce the random error introduced by sampling unbalance, ensure the representativeness and coverage of fault modes extracted by sequential verification testability, and have stronger engineering applicability.