The Reliability Study For A Load-Sharing Parallel System Of Two Components

It a common practice to use redundancy techniques to improve complex systems reliability in systems engineering. Parallel system, k-out-of-n system, warm standby system, cold standby system which are typically redundancy system. What kind of strategies can be use to improve redundancy system reliability is the emphasis research in redundant systems.When analysis system reliability indexes, we assume that the lifetime of the working components are independent in the system, but regardless of the working components lifetime is affected by system bearing load in reliability studies of general redundant system. This assumption will generate some deviation in the analysis of load sharing system reliability which does not accurately reflect the real work state of the system. For example, the multi-traction cable suspension bridge, CPU of a computer system can be regarded as load-sharing system. In generally, the distribution of lifetime and higher failure rate will be change due to increase bearing work load of components.Thus, based on load-sharing redundant system, this paper analyzes the reliability for a parallel load-sharing system which is consisted with two components. The parallel system bearing load is a continuous function of load assignment ratio a. Due to the relationship of components lifetime distribution in full load and partial load, we use Markov model obtaining the system state transition matrix and transfer rate matrix with the components can be repair and un-repair respectively. Then, we calculate reliability of the system, momentary availability and steady availability, mean work lifetime and mean repair time ift??. We also give a conclusion that the load-sharing system is better than ordinary parallel system with two components in view of mean lifetime. Secondly, using Bayesian estimation method, we obtain the parameter estimation of component failure rate which the lifetime of the components are independent and identically distributed exponential distribution in the load-sharing parallel system. Finally, we derive the residual lifetime reliability function RX1,X2(x)of the system under the conditions min(Xp1,Xp2)=tby the introduction of the virtual age and take into account for the residual lifetime reliability function RX1,X2(x) when the components are distributed exponential distribution.