Dear Anita,
According to the documentation, the threshold has indeed to be a float.
However if the threshold you want to define is something like :
output_of_interest < random_threshold, then it may be defined this way :
Event(X, ot.Lower(), 0)
with X = output_of_interest - random_threshold
Hope it will help.
Best regards,
Rodrigue
--
*Rodrigue DÉCATOIRE*
Ingénieur d'Etudes et de Recherche
Centre d'affaires du Zénith - 34 rue de Sarliève - 63800 Cournon d'Auvergne
Tél : + 33 (0) 4 73 28 93 66 | Fax : + 33 (0) 4 73 28 95 78
www.phimeca.com <http://www.phimeca.com> | Viadeo
<http://www.viadeo.com/fr/company/phimeca-engineering-sa> | LinkedIn
<http://www.linkedin.com/company/1012024?trk=prof-0-ovw-prev_pos>
Le 01/08/2016 18:32, Anita Laera a écrit :
Dear OpenTURNS users,
I am working with threshold exceedance approaches and I wanted to know
if it is possible to specify a distribution for the threshold, instead
of a float.
Until now I have always used the constructor:
Event(output_variable_of_interest, comparison_operator, threshold)
where threshold is just a number.
There are other available options, for example:
Event(output_variable_of_interest, domain)
In this case, if I use Interval as domain type, I think it means that
we are considering a uniform interval (and not a distribution).
The other available constructor is with Event(process, domain), but it
is not really clear to me how it works.
My question is: is it possible to treat the limit state function as a
stochastic process instead of just a number?
Thank you for your help!
Kind regards,
Anita
_______________________________________________
OpenTURNS users mailing list
[email protected]
http://openturns.org/mailman/listinfo/users
_______________________________________________
OpenTURNS users mailing list
[email protected]
http://openturns.org/mailman/listinfo/users
