资料-基于动态失效率模型的在线可靠度预测-On-Line解析.PDF

452 IEEE TRANSACTIONS ON RELIABILITY, VOL. 57, NO. 3, SEPTEMBER 2008 On-Line Reliability Prediction via Dynamic Failure Rate Model Rosario Toscano and Patrick Lyonnet Abstract— This paper presents a new dynamic reliability ap- I. INTRODUCTION proach which is able to take into account the various operating conditions under which the considered system has evolved (past and present). To this end, we propose a differential model of failure RADITIONAL reliability methods are usually based on rate, combined with the state space representation of the process, T analysis of life-time data. These data can then be used for as well as a parameterization procedure. A numerical example the parameterization of models reflecting population character- shows the practical applicability of the proposed approach. istics, under the same conditions as those constituted in the data Index Terms—Dynamic reliability models, operating functions, [2], [3], [5], [6], [9]. These models have proved their useful- reliability prediction, state-space representation. ness during design phases, but are of limited interest in opera- ACRONYM1 tional phases, because, in this case, operational conditions are constantly changing. Indeed, a real system evolves under var- DFRM dynamic failure rate model ious operating conditions, and this has a direct impact on its re- LTD lifetime data