Reliability, Life Testing, and Shelf Life
Everything will fail, eventually. Reliability is a probability that a system (or a component) will fail no sooner than t time units from the time it begins operating. Reliability has many manifestations, or one might consider reliability as a special case of probabilities that some specific type of event will occur no sooner than t units from some initial reference time. Other manifestations include survival of patients having a particular disease, or the shelf life of drugs. For simplicity, we will refer to reliability in terms of time to failure, with the understanding that this time could be the time from a reference point to the occurrence of some specific type of event (such as death, progression of disease, a drug concentration in a human body drops below some threshold, or potency/reactivity loss). The point is to derive a model by which the probability of interest can be predicted, and to incorporate design parameters into this model. In this way, we hope to help the EAS design a system so that it will have a desired reliability for a particular time-to-event.
KeywordsFailure Rate Shelf Life Failure Time Apply Scientist Life Test
- Conover, W. J. (1999). Practical nonparametric statistics (3rd ed.). New York: Wiley.Google Scholar
- Lee, E. T. (1992). Statistical methods for survival data analysis (2nd ed.). New York: Wiley.Google Scholar
- Mann, N. R., Schafer, R. E., Singpurwalla, N. D., (1974). Methods for Statistical Analysis of Reltability and Life Data. Johw Wiley and Sons, New YorkGoogle Scholar
- Whitten, K. W., Davis, R. E., Peck, M. L., & Stanley, G. G. (2004). General chemistry (7th ed.). Belmont, CA: Brooks/Cole.Google Scholar