International Journal of Reliability, Quality and Safety Engineering

This paper considers optimal designs of step-stress accelerated life tests in which each lognormally-distributed test item is first run at low stress, and if it does not fail for a specified time, then it is run at high stress until a predetermined censoring time. It is assumed that a log-linear relation exists between the lognormal location parameter and stress, and that a cumulative exposure model for the effect of changing stress holds. The optimum stress change point minimizes the asymptotic variance of maximum likelihood estimator of a specified percentile at design stress. For selected values of the design parameters, the optimum plans are tabulated. Designs of high-to-low step-stress accelerated life tests (ALTs) in which each item is first run at high stress and then at low stress, and the optimality criterion of minimizing the generalized asymptotic variance of maximum likelihood estimators of model parameters, are also considered. The effects of the incorrect pre-estimates of the design parameters are investigated.

We propose a systematic approach for design and validation of error detection software. Formally, the semantic of a specification is represented by a transition system. This representation is then used to generate a flowgraph or ddgraph which is used to construct an execution path tree. The information obtained from this algorithm representation is used to aid in the design of software-based fault detection techniques for hardware faults.

Flowgraph and ddgraph representations provide information to predict future program flow. During execution, the current program path is recorded, along with the expected path. Checks are placed to verify that the program path follows the predicted path.

Algorithm-based fault tolerance (ABFT) techniques are used to detect data structure corrupting faults and to improve the fault coverage. Fault coverage provided by this approach for different types of hardware faults has been estimated through experiments with the software-based fault injection tool (SOFIT) and the data is presented to demonstrate the effectiveness of the method.

Component redundancy is used to achieve high reliability and fault tolerance in many ultra-reliable systems. Moreover, redundant components can be checked and repaired at the end of periodic intervals of time, called maintenance check times. In such scheduled maintenance (SM) systems, multiple faults may occur and remain undetected because the effects of the faults on the system are masked by the redundant components. Furthermore, specific maintenance actions may be carried out at the end of each mission to guarantee that the system is not in certain specific or failed states. Accurate reliability computation depends on the appropriate modeling of the system. In this paper, we consider scheduled maintenance checks and conditional maintenance checks, and develop a methodology to analyze systems employing such maintenance strategy. We also describe a tool that we have developed using our algorithm. Examples are presented to prove the aptness of our method. We show that conventional technique vastly overestimates the unreliability of such systems. We also show that for the SM systems, our technique provides a more realistic estimate of reliability.

The common version of Genichi Taguchi's parameter design entails a marginal analysis procedure for determining parameter settings that will maximize or minimize a response. Unless all significant parameter interactions are known a priori and provided for in the orthogonal array of the parameter design experiment, conclusions drawn from marginal analysis will not necessarily be correct. In this paper, the probability that the routine parameter design procedure will actually succeed in realizing the optimization objective is discussed, followed by illustrations based on data from the commonly cited literature.

The current trend in modern production is directed towards shorter and shorter production runs. The two major reasons causing this trend are the rapid spread of the just in time (JIT) philosophy and the constantly increasing multiplicity of customer demands. The short runs of modern production not only constitute a challenge for production management, but they also cause some problems when applying traditional statistical methods, designed to be used for large sets of data. One of these methods is process capability studies. Since theories on how to use process capability studies in short production environments are incomplete, the aim of this paper is to present some ideas which will partly fill this gap.

The theories of process capability studies for short runs presented are based on ideas of focusing on the process, not on the products, and on using data transformation. By using the transformation presented, it is possible to conduct process capability studies in a traditional straightforward manner. A simulation study shows that the suggested transformation technique works satisfactorily in real situations. Finally, the -plot is introduced as a method of interpreting and analyzing the capability of a short run production process. By using the -plot, additional information is obtained concerning the capability of a process, compared to using traditional process capability indices only.

Two-level fractional factorial design is an efficient technique for experiments considering a large number of factors. To evaluate the efficiency and analyze the data for such a design, we need to know the generators for the design, so that, using the generators, we can generate its defining relation and alias structure. Although knowing the generators is important for a two-level fractional factorial design, it is not unusual in actual industrial situations for the generators used in the design to be lost or overlooked while the design is performed. Since Taguchi methods has been widely applied in industry, in this research, an efficient algorithm based on Taguchi orthogonal arrays (OA's) and interaction tables is developed to identify the generators for given designs. Furthermore, with the investigation of the insights of Taguchi OA's and interaction tables, this research may provide ideas for making Taguchi methods a simple tool for developing optimal designs for 2^{k - p} experiments.