International Journal of Reliability, Quality and Safety Engineering

A new periodic damage model is proposed in this paper, where the damage produced by shocks is only measured at periodic times, without considering the stochastic shocks and the cumulative process of damage in discrete times. We model five replacement policies for the periodic damage models using three approaches of whichever triggering event occurs first, whichever triggering event occurs last, and replacing over a planned measure, which are named as replacement first, replacement last, over-damage replacement, over-damage replacement first, and over-damage replacement last, respectively. In analytical ways, we obtain their optimum replacement policies to minimize the expected cost rates and compare them to find which model is more economical. Finally, numerical examples are given to illustrate the validity of the presented models.

A general method is proposed for establishing the fatigue test specification of torsion beam rear axle (TBRA) in the car. The main characteristic parameters of entire vehicle are analyzed by using ADAMS simulation, to study which vehicle parameter is associated with the load, and figured out the relationship between the testing load and the key vehicle parameter based on the existing test specifications. An illustrative comparison in theoretical fatigue lifetime of the TBRA is presented to demonstrate the feasibility and effectiveness of the proposed fatigue test specification. The proposed method of establishing fatigue test specification can provide a reference for durability evaluation of chassis components and help suppliers achieve the synchronous development with OEMs.

The inherent variability of major infrastructure can be associated with structural properties such as member size and geometry, elastic constants, density, strength characteristics or external load types. These variables and factors may give rise to risk, safety and uncertainty for general structures. In this paper, a comprehensive reliability evaluation framework is presented for a laminate composite plate under hydrostatic pressure. An establishment and verification of a response surface, the determination of performance function in terms of input and output random variables, and the comparative application of combined algorithms such as Monte Carlo simulation, artificial neural network and fuzzy theory are conducted.

A gear transmission system such as a gear train is a series system in sense of reliability evaluation, while it is different from a traditional series system such as a chain. Therefore, the conventional series system reliability model is not appropriate for gear transmission system. Based on the time-dependent configuration and the time-dependent loading path of a gear train, system specific reliability modeling technique is presented, by which a gear train is taken as a time-dependent series system, consisting of different components (gear teeth) in different time intervals. To illustrate the application of the reliability analysis method, several time-dependent series system reliability models are developed for different types of gear trains, the reliability estimation results are compared with those obtained by traditional series system reliability model.

It has been commonly known in a Database Management System that constant plans of incremental/differential backup schemes are set up and their full backups should be done to balance the costs of data backup and failure recovery. In this paper, we consider a database system that is running for 24/7 and has several random busy states of updates, and we suppose that incremental/differential backups are implemented after each update and full backups are triggered at constant time $T$ and random updates $N$. That is, models with decision variables of $T$ and $N$ for full backups with random incremental/differential backups are considered. We obtain the expected costs of data backup and failure recovery and the expected cost rates for full backups. Optimum solutions to minimize the expected cost rates are discussed in analytical ways. Furthermore, full backups that are made at time $T$ and updates $N$ are considered as constant and random modes respectively and compared analytically. For the discussions, numerical examples in exponential cases are given.