| 摘要 |
A method is disclosed to determine whether a product, a component or a system (hereafter, we use "group" to simplify such wide scope though "product" is used for clarity in some places; hence, the first is refereed to as "group 1") in reliability life testing has longer lifetime than the other (i.e., "group 2"). This method is free from the assumption on the form of distributions (i.e., it is a nonparametric method from statistics standpoint) and the proposed data analysis approach can be applied to all kinds of data and distributions. This method provides a more accurate solution than parametric methods, whose results may have certain errors from the goodness-of-fit at distribution fitting & parameter estimations. After the pre-check on bimodal, early failures, and the failure mechanisms, our invention employs numerical solutions with good accuracy by nonparametric approaches; the data under consideration can be censored, interval or bimodal, not limited to simple case of the complete type. The method may be used to determine multiplicities of reliability tests at all types (e.g., wafer or package level) and levels (e.g., on products, components, subsystems, or systems). The invented method makes the determination based on a comparability index that is derived by integrating the weighted difference between the reliability functions of the two groups under comparison. Several indices are proposed for the effectiveness on reliability comparability. Other indices are proposed to make our invention more flexible.
|
