Vibrator Speed Control using Load Cell weight

[Cody Anderson]

Hello All,

I have two bags of two different types of peanuts hanging on a load cell above two separate feeders. I am using Studio 5000 and would like to add a speed control to each feeder so I can run one at 10 percent or 20 percent, and run the other at 90 percent or 80 percent. We are trying to blend small and large peanuts together. I am trying to run at about 12500 lbs/hr. Each bag typically weights right around 2200 lbs. Not sure if I need a PID to have my speed constantly updating? Any help would be apricated!

Thank you!

 

[parky]

PID is definitely the way to go, on a continuous process, providing you get the tuning right one sure way to get a quick response is to load pre-set values for the set points to get it started & use bump less transfer to the set point to take over, it will depend on the PID software but most will allow a change from manual to auto SP in bump less mode. This can reduce the time for the PID to stabilise at the given rate. I have done this on a system that combined ingredients for a particular food system & also for a aggregate mix.

 

[Cody Anderson]

I am not really sure how I would begin to set that up, if you could explain a little further?

 

[parky]

The way I have done it is slightly different, it was not loss in weight but on the conveyors, so it was measured in kg/h or in the case of the food product done using flow meters. This was far more simple as it just required the ratio of the different products with a little maths to calculate the flow for each product. Loss in weight is a little more difficult as it is not exactly the same so this is a guess:
Measure at given intervals the loss in weight from the 90% product (this probably does not need a PID i.e. it is fixed (at least in setting it up). Then at sample times take the loss in weight and calculate the loss to give the figure in lbs/tons/kg what ever per hour, divide it by 10 will give you the weight per hour from the 10% figure this would be your set point for the 10% addition.
It could be that the 90% is fixed rather than PID, sample every second or x seconds to get the flow, divide this by 10 or what ever percentage you choose & use a PID for the lower addition.
As for the fixed start of a PID lets say for example a drive would supply a given product at 100kg per hour at 60% of full speed, (assume 60% output of the loop) then you put the PID in manual at start up, with a fixed set point of 60%, wait x seconds then switch to auto control, so you need to determine the kg/h from the loss in weight at intervals that give you good control this may be from every second or maybe even a number of seconds. what you will be trying to do is find the rate that one feeder is giving & calculate the set point of the other feeder. To be honest it is a bit of trial & error to find the best way to do it, if you do use both on PID control it is easy if you find one is better fixed, just set the PID to manual & set the required feed to a constant that way you have an easy way of trying both ways i.e. Both in PID control, One fixed or both fixed.

 

[Cody Anderson]

Thank you for the information. I am planning on keeping the feeder that is running at the higher percent constant. Will be running that feeder close to 100 percent always. So I Will just need to adjust the speed of my second feeder to feed in a certain amount of peanuts depending on what percent I want.

 

[drbitboy]

A PID has three* primary and variable external elements:

1. The Controlled Value (CV). In this case, this would be a signal sent to one of the feeders. What physical control element mechanisms are used to set (adjust, change) the rates of peanuts moved by the feeder the bags? Are they e.g. a motor, or a valve position, or something else? Note that this is not the lb/h of peanuts; these are signals, typically expressed in percent and digital counts (e.g. 0-4095) inside a PLC. Outside the PLC they are a voltage or current or pressure if they are analog, or some other form if some sort of digital control (serial, network) e.g. digital counts.
2. The Process Value (PV); in this case, it would be the measured rate of only one of the types of peanuts, in lb/h. How does this process measure the rate of peanuts from each bag? The OP only mentions one load cell, so it is not clear how the two rates, one of small peanuts and the other of large peanuts, are to be measured independently; how is this done?
3. The Set Point (SP); in this case it would be the desired or target rate of one of the types of peanuts, in lb/h.

A PID will change its CV to drive the error (difference) between the measured PV and the target SP to zero. I.e. for each type of peanut, the measured lb/h of peanuts (PV) will be equal to the target lb/h of (SP). Once the PVs are measured and brought into the PLC, the PID CV connected to a PLC output and that output connected to the physical control element, and the PID is tuned, all that is left is to calculate the SPs based on target operating conditions.

The two SPs, one for each type of peanut, would be set using simple scaling from two parameters; one parameter sets a flow rate and the other expresses the relationship between the two types of peanuts. The particular parameters, e.g. total flow (12500lb/h) and percentage of one type (10-20%), will be specific to the process, but will be convertible to the two PIDs' flow rate SPs, one for each type of peanut.

* there are several other elements to a PID (e.g. tuning parameters, PID equation form, auto/manual and other options, etc.), but PV, SP, and CV are the ones most in play when a PID is controlling some part a process.

 

[parky]

The maths is pretty simple for example assume the main feeder is giving 30g (note we are decimalised here so bear with me), per second i.e. measured every second the loss in weight is 30g, then multiply by 3600 gives you kg per hour, i.e. 108kg/h so you now know that 10% is 10.8 kg/h that is the set point for the 10% feeder needs a setpoint of 10.8.
Although I did not do it we had a system where we were cooking rice in-line, this required a given amount of hot water + steam to be sprayed into the mixer at a ratio of 1 ton of rice to 1.2 tons of water (I think that was the ratio), however, the water sprays were pulsed so had a mark space ratio, i.e. on for a time & off for a time, the raw rice feed could be adjusted to run at any tonnage per hour, to calibrate it, we had a manifold tht connected to the series of spray heads, this fed a vessel external to the main cooker, we would run with all spray nozzles open for x seconds, weigh the water collected, this calculated the mark/space ratio of the sprays, the calculation took into consideration the rice feed in kg/h & adjusted the on/off time to give the required water addition.

 

[Cody Anderson]

I have an HMI where they can adjust the percent that they want the vibrator to run at. Each feeder is ran by a vibrator running at 60 hz using 4-20 analog.

Each bag has its own load cell so I have a weight of each bag coming back to my plc. So there are two load cells.

 

[drbitboy]

Thank you for the information. I am planning on keeping the feeder that is running at the higher percent constant. Will be running that feeder close to 100 percent always. So I Will just need to adjust the speed of my second feeder to feed in a certain amount of peanuts depending on what percent I want.

If one feeder is kept at a constant rate, then

• the process needs to only measure that rate, Fmeas, and
• only one PID will be needed, for the other feeder, and
• a single parameter expressing the ratio of peanut types, in combination with Fmeas, is all that is needed to calculate the SP of that PID.

For example,

• if the parameter is the percentage of the measured-only peanuts, PCTmeas, then SP = Fmeas * (100 - PCTmeas) / PCTmeas
• if the parameter is the fraction (0 to 1) of the PID-controlled peanuts, FRACpid, then SP = Fmeas * FRACpid / (1 - FRACpid)
• if the parameter is the ratio of PID-controlled peanuts to measured-only peanuts, RATIOpid, then SP = Fmeas * RATIOpid
• etc.

N.B. the flow of BOTH types of peanuts will need to be measured.

 

[parky]

DR Your up early this morning lol, yes DR is right. create a sample time to get the loss in weight (make both sampling times the same), calculate the flow rate, one becomes the control, the other is the process variable you need to control, calculate the ratio & that becomes the set point for the PID of the percent addition.

 

[drbitboy]

...i.e. 108kg/h so you now know that 10% is 10.8 kg/h that is the set point for the 10% feeder needs a setpoint of 10.8. ...

Actually it's a setpoint of 12.0, so 12 / (108 + 12) = 10%, but 10.8 or 9%+ is probably close enough. Actually, @parky is right, that is good enough as long as the ratio is small.


Keepin' the streak alive!


The hardest part is going to be getting an accurate difference measurement in real time from load cells near a vibrator; some form of filtering and/or a long time-base (difference over several seconds), with a correspondingly long update time for the PID, might work.

 

[Cody Anderson]

Really appreciate the help so far guys! Thank you for the info, im starting to get a better understanding of what needs done.

So would my Fmeas be the 12500 lb/hr?
And my PTCmeas would be either the 10 percent or 20 percent?

 

[Cody Anderson]

Would my PV in my PID be the rate of my constant feeder or the rate of the feeder im controlling with my PID?

 

[parky]

your SP for the 10% will be the 10% of the PV of the 90% so for example if your main feed is 12,500 lb/h then the SP for the PID to control the 10% feed will be 10% of the 12,500 so in actual fact you will want 11,250 lb/hr on the 90% feed & 1,250 on the 10% feed or what ever the ratio is this then gives you the total of 12,500.
So if on your HMI the operator sets the flow rate at 12,500 lb/hr, & the ratio is 90/10 then divide the 12500 by 100 * 10 sub that from the total that gives two figures of 11,250 for the 90% & 1250 for the 10%. One thing you need to stop the operator from doing is having a mix of more than 100% so I suggest a little logic that adds the two ratio's up & flags an error if it is not 100%. Easy to do as the operator inputs are only copied to internal variables if the two equal 100%.
Or very simply operator enters main percentage & thr PLC calculates the rest to make 100%.

 

[drbitboy]

So would my Fmeas be the 12500 lb/hr?
And my PTCmeas would be either the 10 percent or 20 percent?

Definitely not for PCTmeas.

Maybe not for Fmeas:

Looking at the OP ... it seems the 12500 lb/h is the target total rate of peanuts, from both bags?

Fmeas is the rate of peanuts from one bag, i.e. the flow rate that is controlled, not by PID, but by the operators setting the vibrator speed directly from the HMI, typically to a value around 100% speed, on the vibrator controlling the peanuts that are to have the higher percentage weight in the mix. Fmeas is the rate of peanuts that are only measured and not PID-controlled.

Also, there is no statement yet

• of desired accuracy;
• of scope of the targeted weight ratio of small vs. large peanuts.
  • It could be the ratio is a spec for peanuts being put into small bags, so the moment-to-moment ratio is important,
  • OR
  • It could be that the ratio is a spec for peanuts being put into a large storage bin, comprising many bags being loaded onto the feeder over a few hours, with mixing in the bin.

In the case of the large bin, a moment-to-moment rate calculation is not as important, PID could be overkill, and it may be enough to simply bump the smaller-percentage peanuts' vibrator speed up or down by a fraction of a percent every 10s based on the ratio of accumulated weight changes since the process started. This could be done across several replacements of empty bags with full bags. E.g. presumably there is an HMI interface for "the lower- or higher-percentage peanuts' bag is empty; stop both vibrators while we bring in a new bag" event.