Effects of Component Interaction

In previous chapters the overall reliability of hypothetical complex systems has been derived on the assumption of complete independence of component failure rates. In situations where the system comprises mechanical, thermal, hydraulic, chemical, etc., components, it is found that although these methods are mathematically correct, they do not yield the actual reliability observed in practice. Intuitively, it might appear that this poor correlation is because the model is not a good functional representation of the real system. Although this may, on occasion, be a possible reason, further analysis may show that the model is correct; but that the assumption of independence of components was unjustified. It can readily be shown that in systems where components are reduced by wear, chemical reaction, environmental attack, etc., or where component subsystems share a component medium, interaction will invariantly exist. This often results in a change of component failure distribution and may also greatly affect optimum maintenance and replacement scheduling. Because of this, the failure distribution observed in the single component life test is not applicable and must be modified.