The variation of failure rate is interpreted in terms of the interaction mechanism between load and strength (product property). It is highlighted that the variation of product failure rate with service time is controlled by two types of failure rate variation trends. One is the decreasing trend dominated by load statistical characteristic, the other is the increasing trend dominated by strength degradation. Under the action of a stationary random load process, the statistical characteristics of load peaks leads to product failure rate decreasing with service time, provided that the product property does not degrade. The reason is simply that, a product will not fail to a impact of load not higher than those to which the product has successfully resisted, while the possibility that a higher impact (higher than all the preceding ones) appears in a unit time interval will decrease with the increase of operation experience. On the other hand, load history dependent product property deterioration leads to failure rate increasing continuously. From the viewpoint that failure is the reflection of load-strength interaction, this paper derives failure rate model for product subjected to random load sequence. As the foundation, discrete time parameter, i.e. number of load actions is used, "discrete failure rate" is defined, and it is clarified that the discrete failure rate at the nth load action is equivalent to the failure probability caused by the (n + 1)th load action, given that the product has survived to the foregoing n times of load actions. Based on the failure rate model, the effects of load uncertainty, component strength uncertainty, and strength degradation as well on failure rate are discussed.
Reliability Evaluation Based on Different Distributions of Random Load
