An approach to predicting the reliability of engineering products in the absence of, or with incomplete, reliability data for their components and elements is considered. It is shown that the uncertainty associated with reliability prediction can be overcome through a reinterpretation of regulatory documents and established scientific and methodological principles. Design methods for reliability prediction are investigated based on the substantiation and evaluation of engineering solutions adopted during the development and manufacture of highly critical, one-of-a-kind (unique), non-repairable aerospace products. It is established that any design solution can be reduced to the substantiation and assessment of the redundancy between the required product structure and random environmental factors. It is demonstrated that, during the preparation of design documentation, the design engineer is able to select the structural parameters and dimensions required to achieve the specified level of redundancy on the basis of frequency-based, parametric, and logical-probabilistic indicators characterizing the probability of changes in the product state. It is concluded that compliance with the required functional performance can be assessed by predicting product reliability using a reliability calculation method based on the probabilities of the required functions being performed by individual components and elements.

EDN  DTIQGX