Flow Calculation on Conveyor with Loadcell and Encoder

Hello everybody.

Can I calculate flow of dust to be pelletized with a configuration shown in the picture? We will take the loadcell signal with a Siwarex and calculate the speed of conveyor via encoder. I thought of a formula like,

Flow = Mass X Speed / Coefficent [kg X m / s]

Have you ever calculated the flow like this before and what should I check with this system? Since we calculated the overall moment of the conveyor what will be the effect of the material which havent arrived at the adge of the conveyor where the loadcell will be attached? To make my calculation more accurate which way should I go?

In the end I am trying to simulate a mass flowmeter with loadcell and a encoder. What will be the accuracy of the calculation?

Thanks...

I have seen scales from milltronics and they seem to use the same principle, so I guess that you can do the same. The problem is they have the knowledge and you have to start from bottom up.

The basics involve knowing the belt speed, belt travel distance over the weigh cell, distance response of the weigh cell, and the instantaneous weight.
No coefficient is necessary, as this is a direct measurement.

If the weigh cell responds to material over 0.1m for example, try sampling once every 0.1m of belt travel, and totalize that.

Essentially, you reduce the equation to:
MassFlow = m/min * kg/m, which reduces down to kg/min.

A bigger problem with these systems is buildup on the belt, idlers, and mostly the design of the weigh cell.

Buildup and idler-tension problems are usually corrected by a second weigh cell on the clean side of the belt, which acts as a continuous tare.

Seriously, the mechanics involved in doing this right are involved, have you considered a self-contained commercial, working solution?

Why dont you just buy the Siwarex FTC module
This is what it is made to do

JRW said:

Why dont you just buy the Siwarex FTC module
This is what it is made to do

Click to expand...

I know Siwarex FTC is used for this application but the machinery company wants to make the project as cheap as possible. We offered them to use Siwarex FTC, but they say it's very expensive and they offered to go down this way.

Thanks...

Thanks very much for your answer, rdrast.

rdrast said:

The basics involve knowing the belt speed, belt travel distance over the weigh cell, distance response of the weigh cell, and the instantaneous weight.
No coefficient is necessary, as this is a direct measurement.

Click to expand...

i)Belt speed
will be mesaured by encoder so no problem here

ii)Belt travel distance over the weigh cell.
As shown in the picture the loadcell will be hanged to the conveyor so the distance over loadcell = 0 ???
Or this is the lenght of the conveyor???

iii)Distance response of the weigh cell
???Please explain if you can

iv)Instantaneous weight.
will be measure by load-cell no problem here

rdrast said:

If the weigh cell responds to material over 0.1m for example, try sampling once every 0.1m of belt travel, and totalize that.

Essentially, you reduce the equation to:
MassFlow = m/min * kg/m, which reduces down to kg/min.

Click to expand...

How can I calculate or measure this weigh cell respronse? If I can calculate kg/m then everything will be very easy. Can you please explain how to find out kg/m measurement or calculation for this process.

rdrast said:

A bigger problem with these systems is buildup on the belt, idlers, and mostly the design of the weigh cell.

Buildup and idler-tension problems are usually corrected by a second weigh cell on the clean side of the belt, which acts as a continuous tare.

Click to expand...

This is also a good idea. They didnt mention any other load-cell for continous tare but maybe I can offer them but since it is extra cost for them I think they wont like this offer

rdrast said:

Seriously, the mechanics involved in doing this right are involved, have you considered a self-contained commercial, working solution?

Click to expand...

Mechanics is not done by the company I work. There is another machinery company I mentioned which we are contaractors of this machinery company as automating their machines. This is the first time that I and my company will try this application if we can manage this than I think there are more to come.

erdemsvri said:

I know Siwarex FTC is used for this application but the machinery company wants to make the project as cheap as possible. We offered them to use Siwarex FTC, but they say it's very expensive and they offered to go down this way.

Thanks...

Click to expand...

So your engineering time is free?
I dug into this a couple of years ago, I looked into the K-tron etc.
Its more complicated than you think.

JRW said:

So your engineering time is free?
I dug into this a couple of years ago, I looked into the K-tron etc.
Its more complicated than you think.

Click to expand...

These are company politics so I dont have much idea. This is what I get as an answer when I offered Siwarex FTC module to my manager...

This is the exact reason I created a post about this. I can imagine it is complicated then I thougt but I have no other choice at this point, I have to implent this way. Many automaiton engineer may went this way before me like you so I think there will be people in this forum which can help me regarding their experience about this stuff and learning from them, thanks to them I can make this calculation better.

Thanks.

erdemsvri said:

These are company politics so I dont have much idea. This is what I get as an answer when I offered Siwarex FTC module to my manager...

This is the exact reason I created a post about this. I can imagine it is complicated then I thougt but I have no other choice at this point, I have to implent this way. Many automaiton engineer may went this way before me like you so I think there will be people in this forum which can help me regarding their experience about this stuff and learning from them, thanks to them I can make this calculation better.

Thanks.

Click to expand...

*******************************


Point taken

What I'm calling "Distance Response" is actually how much material is actually measured by the weigh cell in an instant.

Ideally, you would only weigh the line of pellets directly above the cell. In reality, you are probably getting a weight of pellets directly above, 10cm after, and 2cm before the weigh cell, so your sampling should take that into account.

Weighing every cm in the above instance will yield wildly inaccurate throughput values, ideally, you might keep a running average of the instantaneous weight every cm, average it over (in the example above) 12cm, and add that to your total every 12cm of belt travel.