Temp and Flow Control for two pumps

aemcgo0

Member
Join Date
Jul 2014
Location
Upstate NY
Posts
6
Ok, I have a upgrade to our brewhouse that I'm having trouble getting my head around how to program.

I have a hot water tank with a 5hp pump on a VFD, and a cold water tank also with a 5hp pump on VFD. Both of these flow into the mash cooker at the first operation of the brew cycle; so, we are looking on blending these two water streams to an operator set temperature using a 3 way blending valve. No problem if we didn't care about the flow rate upstream of the system, just turn both pumps up to full speed and let the PID loop for the 3 way valve modulate the blend for temperature. We do this all day long on our fermentor glycol system.

However, they also want to control flow rate of the now combined and temperature controlled water flow. This will allow the brewers to synchronize the water flow and the grain flow to end at the same time. Is there a way that you can think of, using the pump speeds as the flow rate changer, that will allow me to control both temperature and flow rate in to the mash vessel without adding an additional control valve after the blend?

I'm concerned that when I go to modulate the temperature and flow together it will cause the system to oscillate between two situations, at temperature and not at flow rate, and not at temperature and at flow rate. I think this will make the whole thing unstable and impossible to tune. Any help or ideas in this would be great.

Thanks,
Adam
 
If the 3-way blending valve is a control valve, I'd first consider using it to control temperature. Control flow with speed to one drive; the other drive will get its setpoint as some fixed ratio of the signal sent to the first drive (maybe 1:1).

Just a thought...
 
Flowrate is the boss, ratio is the slave.
example.
Flowrate setpoint is 500 l/h meaning the pumps running both on 50%
if flow is 400 the pumps should be on 43%
if temp is too low the ration should be higher as for example.
at 50% ratio cold pump runs on 22% hot pump on 21%
if too cold cold pump will be 18% and hot pump on 25%.
So you run with 2 PID loops.

You will need some time as the sensor must be behind the mixer.
No need for three way valve.
 
jamesau:
Yes the 3 way valve is a control valve so hitting the temperature shouldn't be a big deal, as I'll just have a PID in the ControlLogix do the blending between the tanks to hit the temp.

Using a fixed point on one pump could work, I'm concerned that the combined flow rate of the two will be off. But I guess using a PID loop for the other speed to make up the difference should work. I think I would want to set the fixed speed for the cold water as if the temp is a little cold it's not the end of the world as we heat up immediately after we are done filling the tank. The catch is that there is an outside possibility the the brewmaster will want to be transferring wort out of the brew kettle at the same time as mashing in and we use the same cold water supply to cool the beer on its way out. This we time varying the pump speed to hit a temperature target for wort after the heat exchanger. So PID controlled cooling water flow but keep wort flow constant. So the combined flow of the mash water and cooling water will now vary.

shooter:
Yes I could use the speed of the two pumps as the blending but I've found that being less accurate and more difficult to tune in.

These are both great ideas and I'll contemplate them both tonight over a beer.

Thanks,
Adam
 
Using a fixed point on one pump could work,...


Just to be clear, my first recommendation was to have a fixed ratio, not point, on the pump. So the ratioed pump's VFD setting would track (according to the ratio) the VFD setting of the pump used for flow control. (If the ratio was 1:1, each pump's output would be the same. If 1:0.8, one pump's would be 80% of the other...).

If concurrent use of cold water for cooling is called for, the above strategy would compensate to some degree. If the cooling load demands too much water too quickly, you may have to embellish the controls with additional features to satisfy both objectives.

Enjoy your beer tonight:beer:.
 
Is there a way that you can think of, using the pump speeds as the flow rate changer, that will allow me to control both temperature and flow rate in to the mash vessel without adding an additional control valve after the blend?
I would set up two separate PID loops, one for temperature control, and the other for flow rate control.

1. Brewers set the flow rate setpoint. Flow Rate PID calculates speed of VFDs to achieve rate. You will need a flowmeter to feed back the Process Variable to your Flow Rate PID. You may be able to use 1/2 of the calculated speed for both VFD's, but most likely the temperature difference will mean that the cold water pump will need to run slightly faster to get an approximate equal hot-to-cold mixture. So feed the calculated speed to the cold pump, then use a field-tested ratio of hot/cold X PID Speed for the hot pump.

2. Operator sets temperature of the water. The water temperature is measured downstream of the mixing valve and fed back to the Temperature PID Process Variable. The Temperature Control Variable is sent to the 3-way mixing valve. This valve must be set up so that it does not block or change the total flow rate, but only changes the ratio of hot-to-cold water. Changing the ratio will change the flow rate slightly, but the Flow Rate PID should be able to make an automatic adjustment as the hot-to-cold ratio changes.
 
Last edited:
I like the idea of being able to set a ratio as well.

One PID for temperature with the valve.
One PID for both pumps in parallel + the ability to set a multiplier for each. If you see that the temperature control valve is in the first or last quadrant, you could add a multiplier to the according pump. This would put your valve in an approximate 50% position and give you a better temperature control.
 
Pandiani and I worked out a similar problem a long time ago.
One method was linear quadratic control.
The other was the ratio method but controlling the temperature by controlling the relative speed of the two pumps rather than using a mixing valve.
I still have the mathcad work sheet. I will see if I can find it tomorrow.
If I remember right Pandiani used the LQC method and I used the ratio method similar to what Shooter suggested.

It doesn't make much difference which method is used if the temperature and flow remain constant.
 
After thinking over your guys suggestions last night and then talking to one of my old programming mentors this is what I came up with.

First, since I know the temps of both tanks and they both change temps slowly I can calculate the theoretical ratio of the two pumps flow rate that would give me that temperature using this formula.

(Ttarget - Tcold)/(Thot - Tcol) = Percent of flow Hot

This would therefore become for lack of a better way of saying it, fixed at the start of mash in. In reality this could be allowed to change as the tank temps change, but they are such large tanks it will be slow and I probably will not see it then.

Then use a PID loop to scale the now relative speeds of the two pumps to give me the desired flow rate at the flow meter.

Finally use a PID the "fine tune" the temperature at the blending valve. In theory this should keep the 3 way valve towards the middle of the range and not disrupt the flow rate drastically.

This looks to be the way that boneless is suggesting and I can't think of how this wouldn't work. Any other thoughts?

Thanks for all you help,
Adam
 
I like it! :)

Let us know how it works out for you, I might have some applications for this to.
 
(Ttarget - Tcold)/(Thot - Tcol) = Percent of flow Hot

Just to clarify, the right hand side of the above should be the ratio: (hot flowrate)/(total flowrate)
which is basically %hot/100, where the percentage is on a total flow basis.

Do you have a flowmeter after the hot pump that you plan to use? Assuming these are centrifugal pumps, they by their nature create head (pressure) which will vary with speed. What this means is that just setting the relative speeds won't get you relative flowrate, especially when the mix valve output is changing. (If you have a hot flowmeter, by all means employ the above relation).
 
Thanks all,
Yes there is a flow meter there so it should work. The non-linear aspect of the centrifugal pump is the part that was throwing me. So I've relocated a further upstream meter to just after the pump to get the proper relationship. Our brew house is quite the Frankenstein in that it's 20 years old and about a dozen people have changed it through the years. I was brought on to try and bring it up to modern automation standards. But the Brewers themselves can be quite resistant to changes. So it's been a bit of a challenge working with what management wants and what the Brewers want.

Thanks for all your help.
Adam
 

Similar Topics

The objective: Reach about 10’ horizontally into a 1300F furnace, and pick up a 140lb rectangular container, then pull the container back out of...
Replies
15
Views
1,811
Hi Everyone, We have Silo System which has temperature sensor and these sensor data will be communicating through RTU unit to the Dedicated PC...
Replies
2
Views
570
We have a small plastic parts oven (SP at 210F). It used to have an old bimetal temperature regulator. We are upgrading it using two temperature...
Replies
7
Views
1,499
We have a vertical Axis which was overheating while not moving. After further investigation I found that when the axis moves upward to a...
Replies
5
Views
1,166
Needing some help. I have a machine that was put in storage and has ended up in my lap. It had a 5/02 originally but the card was pulled and I...
Replies
3
Views
1,350
Back
Top Bottom