question

John B10 avatar image
0 Likes"
John B10 asked John B10 edited

General question about CV in FlexSim

Hi I have a general question. I have simulated in FlexSim a production line with four processes, each of the process has this process time: normal(1.5, 3, 0). Infront of each process is a Queue. I have simluated the production line once with a WIP of 10 then with a WIP of 50 and at least with a WIP of 100. Now I have exported the process Time of each flowitem (process time of each product), of Process 1 and Process three. For process 1 and process 3 I have calculated the coefficient of variation. Now I have seen that the coefficient of variation is almoust equal for a WIP of 10, 50 and 100. Does anybody know why is this the case? It is a conwip system. Is this the case because the process time of each machine is the same?

FlexSim 22.0.0
conwip
5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

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

1 Answer

Jordan Johnson avatar image
0 Likes"
Jordan Johnson answered John B10 commented

Without a model, it's hard to say. I have a couple guesses, based on your description, as to where the issue could be:

  • If you are measuring the duration from when an item enters the processor to when it exits, then you results make sense, no matter how many items you process. You are sampling a distribution with a given mean and standard deviation, which are the values you use to calculate the coefficient of variation. So sampling more data from that distribution shouldn't change your answer.
  • Even if you are measuring the time from entering the upstream queue to exiting the downstream processor, if there is only one item at a time in the queue, then additional items won't change your calculation either.
  • It may just be a mistake in where you are gathering the data. Double check that you are measuring what you think you are measuring.
· 3
5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

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

John B10 avatar image John B10 commented ·

Thanks for your answer.

I want to implement 3 scenarios with a WIP of10, Wip of 50 and a WIP of 100.

I want then calculate the coefficient of variation of each scenario and check

if the coefficient of variation is changing from machine 1 to the other scenarios.


Furthermore I want also check if this is also the case on machine 3 for the three scenarios.

And I would also check if the coefficient of variation differ from machine 1 to machine3.

And at least I want to check if the implemented process times (means and std) corresponds with the simulated results.


I am very new in FlexSim, maybe you can have a look at my flexSim modell to check if I have set the labels right for doing my simulations. Production_line.fsm

0 Likes 0 ·
production-line.fsm (2.3 MiB)
Jordan Johnson avatar image Jordan Johnson ♦♦ John B10 commented ·

In your model, you measure the duration spent in the Punching machine, as well as the duration spend in the Rimming machine. I also see you create your WIP at time 0, and then replace each item when an item gets to the sink. This all seems fine to me.

You are using a process time of

normal(/*mean*/1.5, /*std dev*/3, /*stream number*/0)

for your processors. This seems odd, because you can frequently generate negative numbers with that distribution, and negative numbers don't make sense for process times. FlexSim will set those to 0, which is why you get zeros in the Prozessor1 table. If you are seeing a mismatch between the input distribution and the actual times, that is why.

But even with all that, I don't see how changing the WIP of the model would affect the process time of either processor, and consequently, I don't see how that would change the coefficient of variation for those two processors. If all you want to do is verify that it doesn't change, then you are on the right track.

0 Likes 0 ·
John B10 avatar image John B10 commented ·

Ok thank you very much for you help!

0 Likes 0 ·