simulator to study "cross" of packages and bottlenecks

[reggio]

hi to everyone, every now and then I'm reproposed the problem of finding the best solution of handling of various necks, random online insertions and from multiple stations, intersections, augments, accelerations and decelerations.

in this case I would like to "similare" (even not graphically) a transport line in such a situation:

fixed path length=18mt
only load at the beginning
only exhaust at the end
insertion of max 22 packages
insertion time of each neck: variable. . .
insert cadence: variable. . .
permanence time on the conveyor of each neck: variable from 45 to 80min.

purpose: to respect the time of stay as much as possible (considering also the times of insertion when, the transport is still but the time is counted), adjusting the speed of transport, possibly remaining more time but never less.

I imagine that the 1st of the row will give the speed of cruise and every time the bike stops to allow an insertion, such speed will have to increase to recover the time lost in the stop for insertion.
In this way, however, the 2nd of the row will travel at the same speed as the 1st that will be too fast, so once you download the 1st in the correct time, you will have to slow down to "recover" the higher speed "subita" ....

I was wondering if there is a software or a "simple & practical" system that can help me in these simulations. . .

what would you suggest?

ps with excel I tried but I miss.... :

 

[meccanicamg]

I think it's complicated, maybe I'll have to read it again. I think you can implement it in excel and get the graphs of positions/permanence even without using a specific program.

 

[myface]

hi to everyone, every now and then I'm reproposed the problem of finding the best solution of handling of various necks, random online insertions and from multiple stations, intersections, augments, accelerations and decelerations.

in this case I would like to "similare" (even not graphically) a transport line in such a situation:

fixed path length=18mt
only load at the beginning
only exhaust at the end
insertion of max 22 packages
insertion time of each neck: variable. . .
insert cadence: variable. . .
permanence time on the conveyor of each neck: variable from 45 to 80min.

purpose: to respect the time of stay as much as possible (considering also the times of insertion when, the transport is still but the time is counted), adjusting the speed of transport, possibly remaining more time but never less.

I imagine that the 1st of the row will give the speed of cruise and every time the bike stops to allow an insertion, such speed will have to increase to recover the time lost in the stop for insertion.
In this way, however, the 2nd of the row will travel at the same speed as the 1st that will be too fast, so once you download the 1st in the correct time, you will have to slow down to "recover" the higher speed "subita" ....

I was wondering if there is a software or a "simple & practical" system that can help me in these simulations. . .

what would you suggest?

ps with excel I tried but I miss.... :

If you send me extremes I can make you a quote.. :biggrin:

 

[reggio]

I think it's complicated, maybe I'll have to read it again. I think you can implement it in excel and get the graphs of positions/permanence even without using a specific program.

...to me with excel did not come the decisive idea. . .

the problem is "simple" to understand, I try to explain it like this:
the transporter is like a tunnel,
The hills are like cars

all cars must stay in the tunnel at least a variable time (e.g. all 60min, and the day after all 45min).

every time you enter or exit a car from the tunnel, all cars stop for "safety", this time of stop is variable and is different for each entry and exit (say 2 to 5min).

all cars travel at the same speed as the car ahead of all (among those still inside the tunnel - those exits do not count anymore), such speed will change (I assume more and faster until the car comes out later) after each stop for entry or exit.

Here... but like the scheme?
...

 

[reggio]

If you send me extremes I can make you a quote.. :biggrin:

Hi!
quote of what? Software?
perhaps, here only pass the card;p

 

[meccanicamg]

I see that the ideas are many and well confused.

then you have an example of 10 hills in the tunnel and every one has to be 60 minutes in the tunnel.
the first time will be given by the base speed of the tape to cross the tunnel.
the second time will be given by the basic speed of the tape - waiting time to stop (then as soon as the first comes out, the second must travel at higher speed)

at each step you have to do the excitement and eventually changing the path space and having a basic speed you will know how much to increase at each step the speed.

It's a little messed up, but you better take paper and pen, draw the axis of time, draw 3 packages, and at every moment do all the accounts. Then after an hour you beat our heads you find it is much easier than you think it is now.

then you have to check whether it is feasible at a practical level

 

[reggio]

(then as soon as the first comes out, the second must travel at higher speed)

...no... the 2nd must slow down

simplify him

22mt
stay 22min

enter 1, ready via necessary speed=1mt/min.
after 11min is in half he misses 11mt and 11min.
now however enters the 2 and stops the bike of all for 3min (in these 3min, 1 spends it cmq inside)
then 1, after 11+3=14min will have to accelerate to perform another 11mt in (22-14min=)8min i.e. it will go to vel=1,375mt/min.
between 8min, 1 will be out and say that stop for 3min, at this point 2 will have gone (8minx1,375mt/min)=11mt, will have to make another 11mt in (8+3min)=11 then will have to slow down to return from vel=1,375mt/min to vel=1,0mt/min.

this obviously if in the meantime there are no other inserts (which instead will be there at irregular times) ...
For me, it's complicated by hand...

 

[stefanobruno]

simplify him
:

stepping up in a handling system is often complicated.
in this case, without any additional carriers that "slash" the time of entry and exit, I see it particularly hard.

Bye.

 

[Fulvio Romano]

I'm begging that I haven't read the problem carefully, but I think it's a classic code problem. If you want to get your headache there are so many code theory books that apply to the most various cases. from the loading/unloading of hoffmann furnaces, to the multi-receiving assembly (from car to mobile), passing through the galvanization tanks.

In your case, you could schematize with a chain of markov, even if I go a little bit by heart, I don't use this kind of analysis... let's say from the university.

It's not a difficult problem, but with excel I see it hard. the tool to study these things, after paper and pen, is matlab/simulink, but it is not an investment recently. alternatively there is the opensource version, scilab.

otherwise, in visualbasic or c# you do, but you have to cuff us a little bit.

 

[meccanicamg]

...no... the 2nd must slow down

I'm sorry. You have to slow down.

actually fewer pieces you have in the tunnel and the easier it becomes the thing and especially you have n speed variations. if you can have constant time intervals even better.

 

[meccanicamg]

the tool to study these things, after paper and pen, is matlab/simulink, but it is not an investment recently. alternatively there is the opensource version, scilab.

I hate matrices. And I don't use matlab anymore since the time of the unit... but there will be a semi-automatic way to do it with excel... Or not?

 

[Fulvio Romano]

I hate matrices. And I don't use matlab anymore since the time of the unit... but there will be a semi-automatic way to do it with excel... Or not?

No, wait, it's the chains of markov that I don't use from college. ..matlab I stopped using it a couple of minutes ago and maybe I take it back in the morning! :tongue:

so in mind I don't get an easy schematization with excel, unless you solve formulas, but at that point you do it with the calculator... so don't exploit the power of excel.
But maybe I'm wrong.

 

[Fulvio Romano]

Maybe there are...the problem can be schematized like this:

each piece has a load and an exhaust. at each piece you can associate a speed function that will be equal to an average speed plus a speed variation to each load of a new piece. By integrating this function over time, the discharge is the instant t where the integral reaches the length of the tape.

In this way you should be able to vary all parameters and see what happens. If you want to optimize, you should clarify what parameters you can do and what you want to optimize.

in simulink you should be able to do it fast enough, unless you manage the antiusality of loads in a smart way. the moment of loading in fact influences the speed function from moments before the phenomenon itself, I seem to have understood. in any case, just use a "furbo" offset and you do the same.

Maybe in excel you can, but certainly not with the electronic sheet. It's about writing macro... if I have some time tomorrow, maybe I try... but I don't guarantee.

 

[reggio]

Maybe there are...the problem can be schematized like this:

each piece has a load and an exhaust. at each piece you can associate a speed function that will be equal to an average speed plus a speed variation to each load of a new piece. By integrating this function over time, the discharge is the instant t where the integral reaches the length of the tape.

In this way you should be able to vary all parameters and see what happens. If you want to optimize, you should clarify what parameters you can do and what you want to optimize.

in simulink you should be able to do it fast enough, unless you manage the antiusality of loads in a smart way. the moment of loading in fact influences the speed function from moments before the phenomenon itself, I seem to have understood. in any case, just use a "furbo" offset and you do the same.

Maybe in excel you can, but certainly not with the electronic sheet. It's about writing macro... if I have some time tomorrow, maybe I try... but I don't guarantee.

Hello and thank you,

1) I read the previous answers and I saw that softw free there is only scilab, I tried to understand what scilab could give me in terms of mathematical and/or graphic result but honestly I couldn't understand it... I found this: http://www.simul8.it/sample_assembly.phpCould scilab make me create something like this?
2) I also read your last, and ahimè my understanding stopped at "integrale" (I tried to read qlc on wiki but I wasn't able...).
3) I finally tried to inform me about "Markov chains" "matrices" "hoffmann furnaces" ... but also there I found little and understood less...

But this subject remains topical and useful for other cases, ... :redface: would you have the holy patience to tutor me through this "dark alley" trying to solve this problem, explaining to me the steps? :redface:

 

[myface]

Hi!
quote of what? Software?
perhaps, here only pass the card;p

No, I was ironing about the complexity of the topic:smile:

 

[Fulvio Romano]

Hello and thank you,

1) I read the previous answers and I saw that softw free there is only scilab, I tried to understand what scilab could give me in terms of mathematical and/or graphic result but honestly I couldn't understand it... I found this: http://www.simul8.it/sample_assembly.phpCould scilab make me create something like this?
2) I also read your last, and ahimè my understanding stopped at "integrale" (I tried to read qlc on wiki but I wasn't able...).
3) I finally tried to inform me about "Markov chains" "matrices" "hoffmann furnaces" ... but also there I found little and understood less...

But this subject remains topical and useful for other cases, ... :redface: would you have the holy patience to tutor me through this "dark alley" trying to solve this problem, explaining to me the steps? :redface:

Okay, you need notions that you probably don't have.
Maybe with excel you can do even without macro... I'm working on it. . .

 

[Fulvio Romano]

I do not guarantee the accuracy of the result, nor its usability, but at least I tried :wink:

I explain to me how this excel sheet works, then if there is any doubt I will answer the questions.

I hypothesized only five pieces, the sheet is scalable, although not exactly immediate.
in sheet1 there are five pieces and the loading time of each piece must be inserted.
the exhaust columns and duration are calculated automatically and represent the moment when the piece arrived at the end of the tape and the time spent between the load and the discharge.
vel nom is the nominal speed at which the tape travels
vel min is the minimum speed to which the tape takes when you have to load a new piece.
the transient represents the number of seconds in which the braking and the new acceleration of the tape lasts. in practice if for example the transient is 10s, it means that the tape that is at a nominal speed, begins to brake with a descending ramp, after five seconds it is at a minimum speed, after another five is returned at a nominal speed.
for simplicity I have considered triangular speed profiles.
the length of the tape... well, is the length of the tape.

sheet 3 accounts. calculates the initial instants in which the tape begins to brake and the final moments in which the tape returns to nominal speed

in sheet 2 there is the heart of the calculation. I broke them in many columns, otherwise they were very long formulons, difficult to understand and above all to debug!

in the speed columns, there is the speed of the individual pieces. this is 0 in the early moments the load, nominal in unperturbed and variable conditions when there is some piece that is loaded.

the delta column represents only variations in speed. in practice the speed is calculated as a nominal speed less delta.

the delta is the sum of the individual deltax relative to the five pieces.

each deltax is broken into deltaxa for deceleration ramp and deltaxb for acceleration ramp.

summing up all speeds (doing therefore an integral) you get the space taken from the five pieces, in the columns spacex.

on the right of each spacex is again reported time, this for a problem of the search.vert function of excel that does not work if the columns are not adjacent.

the unloadingx is calculated by looking in the column spacex the value equal to the length of the tape and returning the corresponding time.

I hope I have been helpful :finger:

 

[PierArg]

Bye to all,
I introduce myself to report anylogic software (it is free).

It's a little program that simulates any kind of process, from the assembly line to an industry to the sorting of passengers at an airport. from the mathematical point of view you can do everything and you can extrapolate any kind of data.
Perhaps It could come in handy.

 

[PierArg]

the type of use is similar to simulink

 

[Fulvio Romano]

but does not affect the problem anymore?
I'd like to know what you think of my excel. .