Reliability with multiple responses and same input variables

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

Hi All,

Thanks for the wonderful uqlab software. Please I need help on the best approach of dealing with the reliability analysis of three limit state functions describing a particular structure with the same input variables (X,Y,Z) but 3 different responses ( g1(x),g2(x),g3(x)) . variables are standard normal.

                                             g1(x) = aX +bY +c Z
                                             g2(x) = fX+hY+kZ
                                             g3(x) = mX+nY+pZ

Should there be solved separately or is there a way of dealing with them together in reliability analysis using uqlab?

2

Dear @aokoro

If you want to solve three independent reliability problems (and get three probabilities of failure), you only need to define a single g function in a .m file with a vector output. Then 3 reliability analyses will be carry out independently, see the UQLab reliability user manual, Section 2.4.2 page 35 for details.

If you have a “system reliability analysis” in mind, you can get inspired by this example.

Good luck
Bruno

3

Thanks Prof Sudret for the information and direction.

4

Hi @bsudret or @styfen.schaer,

As this topic is an old topic, I am not sure how I should “re-open” it and make my question.

Regarding the system reliability example mentioned (https://www.uqlab.com/reliability-parallel-system), I still have a question: I need to run a system reliability analysis but my limit-state functions (g1, g2, and g3) are created from an m-file.

Could you support me on how a full parallel system limit state function can be calculated using the 3 m-files?

Thank you in advance,

Best,