I am currently trying to programmatically generate FTA models in Medini using JS and have a couple of questions now as a result: * What is the difference between an FTA event and an FTA event node? * How can we programmatically create FTA Event Nodes without having to create FTA Events as they seem to duplicate every event that I want to create? * Using JS, how can we create an FTA model and then link it to the different SysML models that we have in the same project? * How can we programmatically create new diagrams in an FTA model in order to nest models within models?

nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Programatically derving FTA models from the SysML models (Functional/Software/Hardware)

medini_1063646795

I am currently trying to programmatically generate FTA models in Medini using JS and have a couple of questions now as a result:

What is the difference between an FTA event and an FTA event node?
How can we programmatically create FTA Event Nodes without having to create FTA Events as they seem to duplicate every event that I want to create?
Using JS, how can we create an FTA model and then link it to the different SysML models that we have in the same project?
How can we programmatically create new diagrams in an FTA model in order to nest models within models?

Find more posts tagged with

medini

medini legacy

Accepted answers

All comments

medini_1063646795

Also when I try to create connections between FTA Events I get this error:

medini_1059007410

Hi Thomas,

Events, Event Nodes and how they fit together are quite well explained in the metamodel guide of medini, which is available in the installation package (zip) of medini - if you don't have it, you should be able to get it from your internal medini power user (or let us know that we can send it to you).

In short: The Event defines all the common attributes of an event (name, probability etc.), and the Event Nodes are the usages of this event within the FT structure. An event can be used several times in a fault tree, then being a MOE (Multiple Occurring Event), and each usage is represented by an Event Node.

Programmatically you have to take care which kind of information about an event is stored where - e.g. probability at the event, connections to parent gates at the event node.

As an example, I've attached an FTA importer I've recently written (for Aralia FTA XML files). Ignore all the XML stuff, but have a look at the creation of the FT structure in there.

Linking your FTA with the SysML model can be done via traces or (recommended) via the "represents" property of the events, which can refer to the SysML elements.

Creating diagrams programmatically is not really possible, as you may be able to modify some diagram properties, but not put elements onto the diagram itself. With regards to "nest models within models", this is not possible (is it necessary?). FTA models can have multiple top events with hierarchical structure of the FTA.

Happy Scripting!

Nils

medini_1063646795

I see, so before being able to create an event node the even must first be defined. That's why we can't build connections to the events themselves and only to the event nodes. I'm guessing this was done intentionally then. I'm a little frustrated by this because when I run the script it makes it look like there are duplicates of an even all over the model with a random floating event lying around. If this is a feature then I'll figure out a way to work around it.

Based on how the installation process works for the company I'm at I don't have direct access to the metamodel, would you be able to share it with me? I think it would be very helpful for the continuing development that I will have to do.

For the nesting, it's more of a model management strategy that we were hoping to be able to leverage because it's possible with the system diagrams that we can create in Medini so we were hoping we would be able to do the same with the FTA to make them easier to consume and understand.

Thank you for your help and the example file you sent, I really appreciate it!

medini_1059007410

Hi Thomas,

sorry for the late reply, I was off for some time.

The event/event node concept is indeed a "feature", which is absolutely necessary for the handling of single/MOE events. Just keep in mind that events hold the main data, and event nodes are representation of where they are used in the tree (allowing access to most event data, but not owning it).

I've sent the latest metamodel guide via e-mail to you.

On the nesting, the concept of Component Fault Trees (CFTs) might be very suitable for you. We will provide a technology demonstrator for that with 2022 R2 (~July 2022), and if you are ok with that we will contact you then for further discussion on CFTs.

Regards,

Nils

medini_1064630327

Hi Nils,

where to get the xml.js for library? it is not available at the latest lib scripts

medini_1059007410

Hi Manohar,

Please find attached the latest xml.js.

Are you planning to do a specific XML import (which source), or is it more for a learning exercise?

Happy Scripting,

Nils

medini_1064630327

Hello Nils, thanks for the reply. I will compare xml conversion.

Also I will check with other colleagues as you suggested.

medini_1040753747

I have a few errors when trying to "Calculate Probability". When it says, 'Event node XXX has multiple outputs', is this related to the event id (<id>) not being defined?

When I create the TLE, the event id (E122) is automatically assigned, but this does not seem to be the case for all the following events. Do I need to set the event id from my side? Or is it supposed to be automatically assigned?

medini_1023909254

Hi Johan,

the event id is not considered for any evaluations, so this would not be an issue resulting in a sanity problem. Nevertheless there seems to be a corruption in the internal data structure for your FTA. How did you create the FTA - with and import, using a script or directly manually in medini? Did you also run a validation on the FTA?

In order to investigate this deeper, please use "Contact Support" from the Help menu to provide us with all the log-files (and if possible also give us access to the project - we can send a Secure Transfer link for that purpose)

Eckhardt

medini_1040753747

This FTA is generated using a script I made (based on Nils example) and accessing the data already in my project within the tool. Since id is generated for TLE I would assume id would also be generated for connected events. I ran validation and got this error:

When I double click or right click and 'Go to' it says "The object could not be selected"

medini_1023909254

Hi Johan,

it seems the script is having an issue - FTA's should never have any loops. When using the UI to create an FTA this is automatically prevented, however an incorrect script may bypass this check. ID's should usually not play a role but if your script is looking things up based on the ID it may be different ...

Can you share your script -either here or via a support ticket?

Eckhardt

medini_1040753747

I fixed the script so that there are no longer any validation errors.

But when I run "Calculate Probability", it seems to bring up old errors that are no longer relevant:

It went away once when I deleted the corrupted fault trees and I was able to calculate probability successfully, but they are back again now. Is there a way to clear this error log cache?

medini_1023909254

Hi Johan,

good you solved the validation issue. Nevrtheless the screenshot from "Calculat6e Probability" indicates that there are still more issues created by your script (and no, just deleting the error log will not solve them ;-) ).

Just from the screenshot it is hard to see exactly what goes wrong - but "Node Logical Gate has multiple inputs to..." indicates that you have created (duplicated) connections in the FTA - each gate/event should be connected upwards to just one gate/event. So please check in your script the place where you create the connections and check before creating a new connection, whether already a connection exists ... look for the places with

Factory.createRelation(nodeToConnectFrom, nodeToConnectTo, Metamodel.FTA.Connection)

Just a little guessing: The reason may be related to transfer gates in your input file which do not point to an event but to a gate. Another reason could be some illegal creation/connection of MOEs. Again, Although the screenshot gives not reveal the full information, this looks suspicious:

Eckhardt

medini_1040753747

Eckhardt,

I did not create any transfer gates. I'm not sure what is going on, but I do believe the errors are being incorrectly saved and not cleared out. When I delete all the script generated fault trees and just create a simple manual fault tree, the same dialog box with all the old errors still pops up when I run Calculate Probability. The good news is probability is generated correctly even though medini is showing the old errors (the screenshot pops up no matter the scenario).

Also, if I create a similar manual FTA in another project and calculate probability there are no error messages.

medini_1040753747

Bad news is I cannot run "Evaluate Fault Tree..."

medini_1023909254

Hi Johan,

that is really strange - can you provide me with a small sample model where this issue occurs (and if possible also with the script itself).

Eckhardt

medini_1040753747

I deleted the project and regenerated through svn checkout. Ran my FTA script again and the errors are gone. So I can generate cut set and probability without any false messages!

medini_1023909254

Hi Johan, good to hear everything is working again. Anyway, at some point in time it may be good to find the root cause. Are you using the build-in SVN client (which is in fact deprecated/disabled by default) or the integration of medini into an external SVN client (like Tortoise).

In case the issue happens again please use the COntact Support from Help menu to provide us with all necessary log files.

Eckhardt