Abstract:System reliability evaluation under small sample condition is one of the most difficult problems in system reliability engineering. In many case, there may be only one product (complex systems), such as some ship. So the statistical sample is very limited. The commonly used methods are classical method, Bayes method and moment conversion method. Using the samples of lower items for system reliability evaluation, the pyramid method is used to expand the samples. In this way, the more the system layers, the lower the evaluation accuracy. These methods have some shortage for evaluating the reliability of large complex system under small sample condition. By using the Equivalent Fault Tree, this paper discussed a new method in expanding the sample number. This method makes all complex system have only two equivalent layers. Then the evaluation accuracy can be insured. When using this method, the analyzer should only build a model with fault tree method, collect data in a proper way and use these data in calculation. The result can be gathered. At the end of the paper, a comparing calculating and an example of evaluation of the reliability of an electrical power system have been proposed by this way. The method has been used in evaluation of the reliability of a ship power system.

Reliability evaluation of complex system based on equivalent fault tree