tmweber asked 15 years ago
Hi,
This is my first time using 217Plus so bear with me. The system reliability model uses the Environmental factor πE to calculate the whole system failure rate. I am looking at the reliability of a unit over a range of temperatures, assuming the same dormant temperature. The πE value is based on the range between the operating and dormant temps. But this means the process grade factor swings between about 1 and 0.5 based purely on a 20-30 degree drop in operating temperature. This seems unrealistic since the base failure rate will change significantly as well, resulting in the 30deg C prediction having a quarter of the failure rate of a 55 deg prediction. I should note that the lower temp predictions are going into unheard of territory for some our LRUs, based solely on the process grade.
So my questions are:
1. Why does 217Plus system model depend so heavily on the temperature swing, when that is also taken into account at the component level?
2. Are the results of the base failure rate valid on their own, not taking into account the process grade? These values seem more realistic with our field data for older but not necessarily predecessor units.
smorris answered 15 years ago
We are looking into your first question.
Regarding the second question, yes, the results you get from 217Plus are valid without going through the process grade evaluation. You do not have to go through the process grading process in order to make a reliability prediction. The process grades default to 1.0, representing “typical” practices if you do not answer the questions.
smorris answered 15 years ago
I had to check with Bill Denson, the “father” of the 217Plus modeling approach for some additional insight
The Pi E in the process grade factors attempts to model the “system”
acceleration as a function of temperature cycling and vibration, excluding
the component acceleration with temperature. It’s based on the screening
effectiveness relationships in the Rome Laboratory technical report on ESS effectiveness.
The inquiry asks why the Pi E is so large when changing from 20 to 30C, but
if it’s only the Pi E that doubles, the overall system multiplier will be
smaller since Pi E is only one of several factors in an additive equation. Even if it does, it doesn’t seem that unrealistic to me since thermal cycling is always
a major accelerating factor.