Manufacturers of our time have begun realizing the importance of developing products that can deliver reliable 
performance throughout the product’s useful life cycle. Any product that delivers inferior performance is bound to drive customers away, while also damaging the reputation of a company or brand. While there is a lot more focus today on obtaining reliable performance almost every time, a number of challenges still remain. A few of those challenges are:
- Narrow timelines for developing reliable products
- Issues in recording product failure events and their causes due to limited testing timeframes
- Samples available in much smaller sizes for reliability tests
ALT (Accelerated Life Testing) methods may be put to use for overcoming several challenges related to standard reliability testing. Quantitative Accelerated Life Testing is done to identify stress conditions which can accelerate the failure mediums, so that all the events amounting to failure can be recorded within a much shorter time period. Some of the factors which can lead to stress are temperature, voltage, pressure, thermal cycling, speed, load, etc.
Importance of Accelerated Testing
Accelerated tests are run to identify probably modes of failure, or for satisfying design verification prerequisites. However, these tests are usually not designed for estimating reliability under normal use conditions. Tests that result in quantitative reliability estimates need well-thought planning, covering determination of proper sample sizes, stress levels, as well as unit allocation to different stress levels.
Reliability predictions which use standard methods mostly look towards utilizing a probability distribution for deducing the failure time of a material, component or even a system. Once all of the adequate distributions have been determined, reliability can be estimated as a time function. ALT helps in developing a model that describes how time-to-failure distribution is transferred under different stress conditions. This helps in making predictions under normal use conditions, by modeling failure time distribution changes with the stress level changes.
Things to remember
At the same time, it is also necessary to exercise caution while using ALT. This applies to the selection of stress conditions for accelerating failure modes which will eventually occur under normal use conditions, rather than introducing new failure modes that won’t occur under normal use conditions. It is also essential to quantify prediction uncertainty through confidence intervals. Proper upfront planning with ideal sample sizes can aid in ensuring that utmost precision-based estimates will come out by using the test data.
All the predictions that are made by deploying ALT models must be correlated with the results obtained from testing under normal use conditions. Additionally, interlinking the predictions by ALT models with the actual data facilitates the model to be refined for any future ALT application.
For more on using accelerated life testing methods to overcome challenges in standard reliability testing, join expert speaker Steven Wachs on Friday, January 20, 2017, in a live webinar, titled ‘Estimating Reliability Performance with Accelerated Life Tests’. Take part today, for a better understanding of various strategies for estimating reliability performance using accelerated life testing.
