IMPLICIT ESTIMATES FOR A PLAN WITH LIMITED TEST TIME AND PRODUCT RECOVERY IN THE EVENT OF A FAILURE 

Mikhailov Viktor Sergeevich, lead engineer, Central Research Institute of Chemistry and Mechanics named after D. I. Mendeleev (115487, 16а Nagatinskaya street, Moscow, Russia), E-mail: Mvs1956@list.ru 

618.2.08:531.3 

10.21685/2307-4205-2019-2-5 

Background. The purpose of this paper is to construct an estimate that is implicitly defined and close in its properties to an effective estimate based on an integral approach for tests conducted in accordance with a plan of type NB τ .
Methods. To find an effective estimate, we used the integral numerical characteristics of the accuracy of the estimate, namely, the total square of the displacement (deviation) of the expected realization of a certain valuation variant from all possible values of the estimated characteristic from the different values of the parameter of the Poisson crt characterizing the failure flow of the set of products under test.
Results and conclusions. 1. The integral approach has shown its effectiveness in identifying the properties of implicitly given estimates. 2. An implicitly given estimate of the mean time to failure T (formula (6)) is effective among the proposed biased estimates (Table 1). For fault-free testing, the T estimate can be applied to both the NВ τ type plan and the NB τ type plan. 3. As an estimate of the probability of failure-free operation, a traditional unbiased estimate (formula (11)) should always be used, except for fail-safe tests. In this case, a biased, effective and implicitly defined probability of failure-free operation(formula (9)) should be used. For fault-free testing, the evaluation of¨P can be applied to both the NB τ type plan and the NБ τ -type plan. 

Poisson distribution law; exponential distribution; test plan; point estimation