| Popis: |
Component reliability utilizing test to failure is usually performed with random samples. Samples are randomly selected to represent the entire population, and are then tested to failure. Weibull analysis is conducted on the failure data. If the calculated reliability meets the requirement, then the component passes the test. Otherwise, the component need to be redesigned. However, sometimes random sampling from a large population is not possible (i.e. during the DV (design verification) stage when tooling is not yet finalized). Only limited parts are made for function and reliability assessment. Since the number of these parts are limited, their dimension, surface treatment, and other properties can be purposely controlled during the prototype manufacturing process. Some of the properties will affect the component's reliability. By considering these properties into reliability testing, the reliability of production parts after tooling kick off can be properly estimated. This paper introduces a new method which integrates the accelerated life testing theory with stress-strength interference method to estimate component reliability using controlled parts. By applying the proposed method for certain components, test time and cost can be reduced and reliability of production parts can be accurately estimated as early as possible. |
