ALT Planning Questions
Join Chris and Fred as they discuss accelerated life testing or ALT. ALT is like the name sounds – testing done in an accelerated timeline. What questions do you need to ask? Which questions do you need to answer?
- Where do you start with modeling? We typically think about two things when it comes to ALT. The first thing is the ‘characteristic life model.’ The model which shows how things like temperature or stress will accelerate the failure of the thing you are testing. The second thing is the ‘time to failure probability distribution.’ This is the model that characterizes the natural variation in how quickly your thing fails when you are not increasing things like stress and temperature. Which is what we will talk about now.
- Part #1 – the characteristic life model. There are popular models like the Arhennius model that describes how much faster chemical reactions occur as temperature increases. The idea is that if you increase temperature, you can make things like corrosion or dendritic growth occur faster. Which means you can reduce a 10-year process into a 2-day process – which is great for testing. The model allows us to work out how 2 days at a higher temperature relates to time at a normal operating temperature. So even though we know our thing is failing faster, we know how to ‘convert’ it back to normal use.
- Part #2 – the time to failure probability distribution (model). Even when we test different but seemingly identical things in a particular test environment, they will fail at different times. Why? Because there are material imperfections, variability in manufacturing and lots of other sources of uncertainty. So failure is random. But it doesn’t mean that it is unpredictable.
- … and on part #2 – we are dealing with wear-out failure mechanisms (according to Chris). When we ‘accelerate’ things, we are making things accumulate damage faster. Which implies wear-out. This being the case, the time to failure probability distributions need to have an increasing hazard rate. Think normal, lognormal and Weibull distributions. The shape parameter for the Weibull distribution in this context needs to be greater than 1.
- … and on part #2 – we look for different failure mechanisms beyond wear-out failure (according to Fred). When we ‘accelerate’ things, we increase stress. Which means that we can see where stress exceed strength without accumulating damage.
- What do you think?
- But the most important thing is starting with the decision you are trying to inform. We often forget about this. Because with ALT, we need to know the failure mechanism we are accelerating. For example, if you are trying to find which material you should use, then you might only need to focus on part #1 – the characteristic life model. So – what decision are you trying to help?
Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques to field data analysis approaches.