A revised lower confidence limit procedure for the reliability of complex quasi-coherent systems

Abstract

This thesis describes a procedure for computing a lower confidence limit on the reliability of a quasi-coherent complex system using test data on its components. the failure times of the components are assumed to have either exponential or Weibull distributions with unknown parameters. The accuracy of this procedure is evaluated using computer simulation for various system structures and sets of parameter values for the assumed distributions. This thesis is an extension of a thesis by Kah Chee Yee in that it uses a different equation for the estimate of the shape parameter in the Weibull distribution than Yee used, and it evaluates the procedure for a larger collection of system structures.