question

arn-k avatar image
0 Likes"
arn-k asked Jason Lightfoot edited

mean time to detect

Hello,

in my simulation, I have a system that has different types of faults/failures with different times to detect and diagnose the issues and finally do recovery actions. I am aware of MTBF and MTTR but I was wondering if there is a way I can add a time MTTD (mean time to detect) for my system as well? Thank you so much.

FlexSim 23.0.1
mtbf mttrdetectfaults
· 1
5 |100000

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

Joerg Vogel avatar image Joerg Vogel commented ·
You can always use a message to create an event in the future. By reacting on receiving a message you can do anything to evaluate this event.
0 Likes 0 ·

1 Answer

·
Jason Lightfoot avatar image
0 Likes"
Jason Lightfoot answered Jason Lightfoot edited

If the stoppage happens only when detected, then for the TBF just enter the sampled time to the failure (TBF) plus the sampled time to detect (TTD). If you only have data on when it was detected, then these are the same thing so you can just use the fitted TTD distribution. If you didn't detect the failure when it happened, how will you know when it happened and what the TBF distribution should be?

· 4
5 |100000

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

arn-k avatar image arn-k commented ·
Got it, thanks

In my scenario, stoppages happen once faults/failures are detected. However, for some amount of time the system is operating faultlessly until a fault happens but it is not detected right away, hence I have TTD.

In that case, is there no way to differentiate between time to detect and time to failure in the dashboards? My main goal here is to show stoppages and the time it takes for detection and recovery actions separately.

Thank you.

0 Likes 0 ·
Jason Lightfoot avatar image Jason Lightfoot ♦ arn-k commented ·
You can set a flag to indicate there's a fault using the MTBF - and change the triggers to not stop the object.

Then you can decide if the time detect is based on time or the number of items processed as part of some QA if needed.

If it's sampled randomly and that's the TTD distribution then you can use another distrubution in to check if the fault label is set.

If it's based on the number of items processed then you can model that explicitly.

If the TTD stacks after the fault occurs then you can send a message to the machine and use a process flow to detect that and initiate a TTD delay before stopping the machine.

There are many possible solutions, but what you've described so far is a little ambiguous to us (hence the options above).

0 Likes 0 ·
arn-k avatar image arn-k Jason Lightfoot ♦ commented ·

how can I set the number of items before detecting the fault? Another question I have is that can I set the delay time in process flow before stopping the machine? thank you so much for your help again

0 Likes 0 ·
Show more comments

Write an Answer

Hint: Notify or tag a user in this post by typing @username.

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.