question

Axel Kohonen avatar image
1 Like"
Axel Kohonen asked Sam Stubbs commented

How to get a certain percentage of breakdown when I know the downtime distribution and the states it applies to?

Hi,

I would like to know how to create a model where I have a processor that can only break down in the processing state and the downtime is from 30 seconds to 2 minutes which is approximated in FlexSim by the function:

maxof(0, exponential(30, 20, 0)))

I know that the downtime should be say 5% of the time that the object is in the processing state. How can I calculate what the correct uptime should be so that the downtime becomes 5% of the processing time? Is it even possible when I do not know ahead of time how much the processor will be in the processing state and how much it will be in some other state?

One option in my case would be to run the model without breakdowns and check how much the processor is in the breakdown state and then calculate the uptime from that. But it seems that it is prone to error as the processing time changes if I change something in the model later and then I need to do this again. Is there some better solution to this?

Thank you!

Kind regards,

Axel

downtimeappliedtoasetofstates.fsm

FlexSim 16.2.0
mtbf mttrbreakdowndown time
· 2
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 ·

@Axel Kohonen the function maxof(value1, value2) has no effect because: exp(30,20,0) = exp(0,20,0) + 30. The distribution cannot be lower than 30. There aren't any negative values in the exponential distribution as long as the location parameter isn't negative.

1 Like 1 ·
Axel Kohonen avatar image Axel Kohonen Joerg Vogel commented ·

Hi @Jörg Vogel,

You are right. I guess I used the maxof function just to make sure as I also used normal distributions in which I had to use the maxof function to not get negative values at times.

0 Likes 0 ·
Sam Stubbs avatar image
1 Like"
Sam Stubbs answered Sam Stubbs commented

In your MTBF table, since you already designated that the down time is only affected during the processing state, FlexSim automatically only takes the processing time into consideration for time between down times. And since you want the down time to be roughly "5% of the time that the object is in the processing state," couldn't that just be making the Up Time the same as the downtime *20?

As in:

(maxof(0, exponential(30, 20, 0)))*20

· 2
5 |100000

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

Axel Kohonen avatar image Axel Kohonen commented ·

Hi @Sam Stubbs,

Thanks, that works, but I guess that is is more intuitive to use

(maxof(0, exponential(30, 20, 0)))/0.05 as you can write the percentage directly then instead of 1/percentage.

Axel

0 Likes 0 ·
Sam Stubbs avatar image Sam Stubbs ♦ Axel Kohonen commented ·

Yes I agree, that would be more directly intuitive for using percentage values.

1 Like 1 ·
Jeff Nordgren avatar image
1 Like"
Jeff Nordgren answered Axel Kohonen commented

@Axel Kohonen

Axel,

When you set in the MTBFMTTR under the Breakdowns tab to only breakdown when processing, that is exactly what the breakdown will be based on, ONLY the processing time. So you have already done that. So I'm confused as to why you need to figure out ahead of time what the processing time is going to be? It doesn't matter, the breakdowns will be based ONLY on the time spent processing.

In the attached model, I used your model and added a dashboard that proves that is what is happening. I also added labels in the processor to keep track of the total processing time as well as the total down time (Code in the processor and the MTBFMTTR).

Is there something I'm not understanding about what you are needing?

downtimeappliedtoasetofstates-jn1.fsm


· 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.

Axel Kohonen avatar image Axel Kohonen commented ·

Hi @Jeff Nordgren,

Thank you for the model. You have understood correctly, but the problem was that I was resetting the values in the state_profile nodes of the objects when I wrote a state report every hour. It seems that FlexSim uses the values in the nodes to calculate the uptime and thus I got way too few breakdowns.

Axel

0 Likes 0 ·