OT: Volume Sensing

[leitmotif]

Dan,
Unfortunately to my understanding, spraying too much color, won't change the color of the product, but will significantly change the cost of operation. Also, it is possible to spray too little dye,and have a product that looks good initially, but fades quickly. Not ideal for control.

What may be worth considering also is that the dye goes on the outside ie the surface area. So how do you equate volume to surface area when you have varying sizes.

Interesting problem.

Sorry all I have is questions.

Dan Bentler

 

[Brijm]

What may be worth considering also is that the dye goes on the outside ie the surface area. So how do you equate volume to surface area when you have varying sizes.

Interesting problem.

Sorry all I have is questions.

Dan Bentler

Me too,that's why I'm throwing it out there.

 

[Brijm]

Just to give an idea as to how they where controlling it before the automation. Every 15-30 minutes they would take a scale reading of the color used, get the load count from the loader (x bucket size), and take a 1 minute measure of the water flow. From this they would calculate the per yard values of water and color used, and adjust to targets.

As the kid who is running it say's "It's not rocket science" But there are a lot of variables into getting it to something controllable. I'm currently taking the average time of the last four loads (to calculate yards/hr, and doing the calculations for them. This works fairly well, but is subject to human error, and I'm not comfortable controlling from it.

 

[Lancie1]

A quick review of the process. They use a bucket loader to load wood chips into a grinder hopper. The ground material is fed by belt into the coloring machine hopper. Both the coloring machine hopper, and the grinder can vary in speed. Color and water are added to the wood, and then discharged onto a v-belt, which carries the end product to the correct pile. They are processing up to about 200 yards of material an hour, and from 40-80 gpm of water.

We are currently counting the wood being dumped into the machine from the front end loader, which is not very accurate to control to.

Several good methods have been mentioned (color sensor, weigh belt) but the most economical method is probably to weigh the wood in the grinder hopper. Install 3 or 4 load cells on the hopper supports, run the weigh to a loss-in weigh controller. Use the LIW controller to add proportional water and color to the coloring machine hopper as the wood is moved from the grinder to the colorer.

You would need the load cells, a summing box, a LIW controller, and 1 or two signal-controlled mixing valves. You may can add a fixed amount of water to each batch, and just vary the amount of dye based on wood weight.

The moisture content of the wood will affect results. The wetter the wood the less color it will soak up so will need more coloring dye, but also the wetter it is the heavier it will be, so if you base your color addition on wood weight, your system will be self-compensating for the moisture content. You may have to make a small adjustment at times in the ratio of color-to-wood weight. Eventually you could add a moisture detector on the belt between the two hoppers to automatically adjust the ratio based on moisture of the incoming wood, but for now you could have a PLC input (3-position selector switch maybe) to set 3 levels of wood moisture manually (dry, medium, wet).

A cheaper (but not as effective) method would be to use a ultrasonic or radar level detector in the grinder hopper to determine the height of the wood, after grinding. This height is multiplied times the hopper dimensions to determine the wood volume, then the color amount could be set proportional to the volume. This method would have no built-in self-compensation for the incoming wood moisture content. If the wood chips always come off an indoor dry pile, or a sawing operation for cured wood, then maybe the moisture does not vary significantly.

 

[bce123]

running through an oven first would help with consistency but may drive cost up

 

[Lancie1]

My only thought to detect the volume of material being processed, is to use a laser or other sensor to measure the height on the v-belt exiting the machine. Assuming a constant belt speed, I could somewhat accurately calculate the volume.

About now you need to tell us whether this is a "batch" operation (which I had assumed previously), or a "continuous" operation (which your above comment implies).

If a batch operation where the grinder bin is fully loaded, then emptied into the coloring hopper, then the method of weighing the grinder hopper should work beautifully. The increase in weight due to moisture will actually help compesnsate by adding more color for heavier wetter wood.

On the other hand, if the grinder is continously loaded and never allowed to empty out, then batch-weighing will not work very well, and you are back to weighing or trying to measure the weight/time flow-rate output of the grinder hopper. A weigh belt not only weighs the material crossing the belt, but also measures the speed of the belt and computes the flow rate output in weight per time unit. However a weigh belt would be an expensive item for this operation.

You might simulate the effect of a weigh belt by using a level detector like your said, or by using a rotary airlock feeder or screw feeder on the bottom of the grinder hopper, and adding a prox switch to count the revolutions per minute (assuming that you can keep the feeder full all the time). Then use the bucket count to compute the volume and multiply by the known color/volume ratio. Alternatively, if you use a VFD drive for the feeder, you could simply use the speed setpoint of the VFD to compute the volume output of the feeder (Volume Output Weight per minute = VFD RPM x Feeder Volume per revolution).

 

[maverick33]

as mentioned previously, there are moisture sensors that are commercially available. they are used widely in concrete ready mix plants and are used to get the moisture content of materials and then add the appropriate amount of water which is obviously a critical component in concrete.

the moisture sensor could be used to detect the water content and then you could weigh the material and accurately adjust for the water content. i have not seen everything (very little really) but from what I have seen pushes me to thinking that weighing is the way to go. i have seen some of the laser scanners but have not seen them work very well on high speed belts. however, belt weigh scales are used regularly in many industries and I have seen both good and bad results. having a consistent material that stays on the belt helps tremendously.

another option (now getting to higher $) is nuclear densometers type gauges. they work by using a small amount of radioactive material as a gamma energy source and then detect how the gamma rays are attenuated by the material being conveyed. looks like a small frame that fits around the conveyor with a ball on top.

not sure what would be best but seems like lots of factors go into this.

 

[Brijm]

About now you need to tell us whether this is a "batch" operation (which I had assumed previously), or a "continuous" operation (which your above comment implies).

This is a continuous operation

 

[bce123]

http://www.engineeringtoolbox.com/wood-density-d_40.html http://www.engineeringtoolbox.com/weigt-wood-d_821.html

 

[leitmotif]

Me too,that's why I'm throwing it out there.

OK just an idle curiosity question
Why dye mulch green?

Dan Bentler

 

[Lancie1]

Dan,

A lot of the wood-chip type mulch is dyed dark red. Apparently people think that makes it look better than the raw wood. As for me, here in the South I have seen problems with termites, so I use river gravel for mulch. Let the little buggers try to eat that!

 

[Lancie1]

This is a continuous operation.

Then you are reduced to trying to measure a difficult-to-measure weight or volume on a moving belt. Without a weigh-belt feeder, you probably can guess at the volume about as accurately as you could measure it with any type of level sensor. The problem is that wood chips are not consistently sized, do not fill a belt in any consistent pattern so vary in volume (sometimes there are large clumps that stick up high, other times there will be fine chips that lay in the bottom of the belt), and vary in moisture content. The more consistent variable would be weight, but that will be costly to measure on a moving belt.

 

[Brijm]

Our IFM Efector rep. recently bragged about this sensor:

http://www.ifm.com/ifmus/web/news/pnews_8r3n6b.html

We plan to try them to detect volumes in hoppers on equipment unsuitable for load cells. According to the rep. they're easy to set up...you teach zero to tune out background equipment, and they prove a 4-20mA output for easy interface to a PLC.

quote]

Spoke with IFM and the local rep stopped by yesterday... Going to give this a try. I'll let you all know what I think.

Thanks for the assistance!