ABSTRACT
Transportation systems are characterized by high demands with respect to reliability and availability corresponding to only a small number of unintended events (failures) per period that must be credibly demonstrated in quite limited periods (e.g., 3 years). Often questions arise such as, “If we expect more or less constant MTTF of 100,000 hours, what failure times will we actually see in operation? Is it good at all to require MTTFs?” “If we observe more failures than expected on what grounds, can we accept/dismiss if measurements cannot be repeated?” “In transportation systems, many diverse components fail with each different down times/restoration times, so what is the resulting availability? How can we specify and measure this?” The chapter intends to respond to issues such as those mentioned earlier. Those of the stochastic/statistical basics that are implied today by typical specifications (e.g., for constant failure rates) are described; potential misconceptions are highlighted; and advices on reasonable specifications and practical acceptance schemes are introduced.
