PSI - Issue 19

Xavier Hermite et al. / Procedia Structural Integrity 19 (2019) 130–139 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

133

4

The range of each influence factor should involve representative extreme service conditions to avoid totally irrelevant failures: burning, explosion, burst, etc. (failure mode out of reliability scope, although it could be interesting to highlight the conditions, which lead to those effects for applications other than reliability). The experimental set-up is highly dependent of the kind of information to look for. Avoiding the full factorial DoE is often a good way to reduce the test plan, but know-how in test results analysis and interpretation is then crucial, and the risk to lose unexpected things is higher. The design of experiment should therefore be flexible, starting from the minimum set-up and being wisely completed through results follow-up (see Figure 2).

Fig. 2. DoE experimental steps example During the tests, unexpected phenomenon might occur, like failure of component other than the studied one, due to severe test conditions (see Figure 3). The test should then be stopped to change the failed component, and restart in the same conditions until the desired failure.

Fig. 3. HALT – HASS experimental discrimination