How to predict acceleration factor in JMP for multiple stresses

Question

Hi all, I am still familiarizing myself with the reliability and survival platform. I was hoping someone could walk me through how to predict the acceleration factor for a multiple stress system in JMP (ex: Temp, Humidity, and Voltage)? Thanks in advance!

DonMcCormack · Answer

The Temp2 column is a holdover from when the Arrhenius transformation wasn't available in the Fit Model dialog. While it's not necessary to use it, as you've shown, sometimes it makes the model easier to understand.

Jason_Wiggins · Answer

Nice Don! I just recreated your example and it works well. I did something a little different though. I labeled the Arrhenius transformation column (temp 2) to X1. I created a new column for log(:Volts) and labeled it X2. In fit parametric survival, I specified the model as u = b0 + b1X1 +b2X2 + b12X1*X2. For me, this is easier to unravel and show. Creating the quantile profile works well when I choose to expand intermediate formulas. Will be on a call with Clovis shortly and will ping you with any additional questions.

DonMcCormack · Answer

A bit trickier with more than one accelerating factor. Here's what I'd think about doing, be interested to hear what others think. Start by modeling the relationship with Fit Parametric. Save the Quantile Probability formula. The AF is the ratio of this formula at given acceleration conditions to this formula at standard operating conditions. I don't believe the quantile will matter, as the AF should be the same regardless of value. With this info, you can build a Profiler using a custom formula. I'm attaching an example using a data set I use to show how to model two acceleration factors. I've fit Model 2 to the data. I've also saved the AF Profiler I built off the model as one of the table variables.

Tantalum Capacitor.jmp