| Objectives | Background | Approach |
This project will incorporate and demonstrate probabilistic
simulation within the virtual qualification process. Further,
the process for examining multiple failure sites and the ability
to generate iso-time-to-failure curves will be developed for
individual failure mechanisms.
The approach to qualification of electronic products must include an assessment of variation expected in product life cycle. However, testing design variations is costly and time consuming. Hence, probabilistic simulation of electronic product life distribution for virtual qualification is a valuable tool to evaluate design options. To address this production variability problem, we need to be able to define the variability upfront and allow the simulation process to determine the variability in product durability. Monte Carlo simulation provides an attractive solution to quantify the variation in product durability.
In addition to quantifying the distribution in expected failures, designers and product managers need methods for quickly assessing the sensitivity of product reliability to process parameters. At present, the approach is to select a parameter and define a variation in terms of percentage from its nominal value. This process can be used to generate sensitivity plots that are useful in evaluating a particular design but do not provide the all of the desired information. In many cases, we would like to selected a time to failure for a specific environment and determine how tightly we should control material and manufacturing variability. A useful plot in this case presents lines of constant (iso-) times to failure (or damage). A plot of families of "iso-time to failure" curves provides with a method for source selection and understanding the effect of manufacturing parameter variability.
This project will develop and demonstrate an approach for conducting virtual qualification process using probabilistic simulation (see Virtual Qualification and Reliability Prediction flowchart). In this approach, the current virtual qualification process will be used as a starting point for quantifying the distribution on expected field and test life. The purpose of the extension to the virtual qualification process is to determine the distribution related to the identified failures. This will be accomplished by taking the top N failures, identify and assign distributions to the material, geometry, and loading parameters that are required to quantify failure. With this information, a probabilistic simulation can be conducted to calculate the expected distribution of the selected failures. This information can then be used to determine the probability of failure for the specified life cycle condition. As a second task, the software will be developed to evaluate multiple failure sites. As a second task, the project will develop and approach for generating "iso-time to failure" plots and examine their usefulness in product selection and qualification.