Reliability Active Learning - Borehole function

UQ with UQLab Community Q&A and How To
,
1

Dear All,
Thanks for the wonderful software. I had a go at Active Learning using the Borehole function but received the following feedback. Seems enrichment was truncated after a short time then I see NaN and not so sure what went wrong . I have attached my attempt using the Active Learning module.
**boreholefunction2.m (1.5 KB) Any useful thoughts, corrections will be appreciated @moustapha @damarginal
Feedback from run
Active learning reliability 

 Pf               1.0000e-16       **

**Beta 8.2221 **
**CoV NaN **
**ModelEvaluations 32 **
*PfCI [NaN NaN ]
BetaCI [NaN NaN ]
PfMinus/Plus [1.0000e-16 1.0000e-16]
-------------------------------------------strong text

2

Dear Aokoro,

Your problem is purely mathematical. If the total number of failures is zero, then so is the estimated probability of failure. Consequently, the estimated variance is zero, which leads to 0/0 for the CoV, which gives you the NaN.

Best regards,
Styfen

3

Thank you so much Styfen for your timely response.