Measuring stability of flow - Compounding processes

[davis007]

I am changing one of our compounding processes. Currently we use a K-tron solids feeder to feed a rather tacky material. We have constant issue with bridging of the material and inconsistent material feed rates. The material has a rather low melt point ~60 deg C and is thermally stable. So we have agreed to look at the possibility of pre-melting the material and feeding as a liquid.

My question is what type of testing should I perform on the liquid pump to ensure that it is calibrated correctly and that the equipment provides a consistent flow.

The simplest idea is pump the liquid into a bucket and measure weight vs. time. But I am struggling with the specifics of how long, how often to repeat the test etc.

More specifics so that you can help confirm that I am setting this up correctly. (I hope.)
Require that the feed rate is 7 lbs per hour. This material is compounded into a compound produced at a rate of 100 lbs per hour. The specification of the loading for this material is +- 1% of the compound (6% - 8% of total compound.) I think that I need to confirm that the pump can generate a flow of 7 lbs hour average with a standard deviation of ~1/3 lbs/hr.

To confirm that the pump can accomplish this I plan on setting up the melting unit and the pump. Setting the flow to the desired rate (7lbs/hr) and running into a bucket on a scale. I would record the weight every min. for 1 hour and calculate 30 (2 min.) flows. These flows could be used to calculate a mean and SD.

Does that make sense?
Do you have a different suggestion?

[Marc]

Is this within the field of expertise of anyone here?

[CarolX]

davis007,

Looks like we don't have any experts here to help you. Any luck resolving your issues?

[Jim Wynne]

Originally Posted by CarolX

davis007,

Looks like we don't have any experts here to help you. Any luck resolving your issues?

I think I would try and get in touch with the pump manufacturer's engineering people (not sales engineers, if they can be avoided). This type of study might have already been done.

[davis007]

Jim and CarolX;

Thanks for trying to follow up. I have spoken with the pump manufacturers and the performance of the pump can be determined. What I am trying to ensure is that from a quality stand point the pump performance meets my need. My first pass is that the specs are +/-1 lb/hr with a target of 7 lbs/hr, and I want 6 sigma performance so the pump needs to be centered on the 7 lbs/hr with a standard deviation of 1/3. The main issue is that since this is a continuous process I can measure the flow over any time period I want. As you increase the time period the standard deviation of the flow rate decreases. So what I am really stuck on is how do I determine the proper time period over which to measure the flow. It seems that it should have something to do with my process, which is a continuous mixer, perhaps 1/10 the residence time of the mixer?

Anyways thanks for the help.
davis007