Design of the hybrid automated reliability predictorThe design of the Hybrid Automated Reliability Predictor (HARP), now under development at Duke University, is presented. The HARP approach to reliability prediction is characterized by a decomposition of the overall model into fault-occurrence and fault-handling sub-models. The fault-occurrence model is a non-homogeneous Markov chain which is solved analytically, while the fault-handling model is a Petri Net which is simulated. HARP provides automated analysis of sensitivity to uncertainties in the input parameters and in the initial state specifications. It then produces a predicted reliability band as a function of mission time, as well as estimates of the improvement (narrowing of the band) to be gained by a specified amount of reduction in uncertainty.
Document ID
19840043988
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Geist, R. (Duke Univ. Durham, NC, United States)
Trivedi, K. (Duke Univ. Durham, NC, United States)
Dugan, J. B. (Duke Univ. Durham, NC, United States)
Smotherman, M. (Duke University Durham, NC, United States)