Automated Dephlegmator Cooling Water Flow Control

[Pierrot Lunaire 55]

When I did a full defleg setup adding to my air-cooled setup (As discussed here: viewtopic.php?p=7723811), I wanted to find a practical way to control cooling water to the defleg to control reflux flow. The setup is now fully automated with PID control via a PLC and HMI to maintain a consistent vapor temperature at the defleg exit.

Rather than writing the full story out again, it has become the topic of an application "notebook" for AutomationDirect, the company that provided the control hardware. Read it here: https://library.automationdirect.com/au ... gKeIfD_BwE

The author cited did the actual PLC setup and programming. It works very nicely.

There are also other references available online, and a video showing it in more detail is in the works.

Pierrot.

[LWTCS]

What is your affiliation?
Be very clear or find yourself heavily moderated moving forward.

Speak up now.

[Copperhead road]

Your “As discussed here” topic…
Is that a thread on HD, could you please copy and paste the thread here so it can be opened.

[shadylane]

Is this the one Pierrot?


viewtopic.php?p=7727362#top
Scroll all the way down to the last post.

[Pierrot Lunaire 55]

OK, forgive me that I don't understand the mechanics of how to get from thread to thread on this platform. Here's the earlier thread about building the still with its forced air cooling system: viewtopic.php?p=7723811&hilit=Pierrot+L ... 5#p7723811

I'm not sure what 'affiliation' you're interested in. This is my still, and once I added the defleg, it was clear that controlling it would be tricky. My friend, the article's byline author, is a EE and has worked on a variety of similar automation projects, often with AutomationDirect products, and several of which turned into these types of small case studies. Anybody reading the article who is capable of doing a project like this will know that there is at least a dozen companies that make similar equipment.

The point is: defleg control via a PID routine is a textbook example of closed-loop temperature control. Obviously I am not the first to think of this idea as such discussions appear from time to time on this forum. Shadylane has cited an excellent example above. My post at the end of the thread about using a peristaltic pump seemed like a practical way to control water flow, and this is an extension of the control methodology.

[shadylane]

I've never done it, but it seems a peristaltic pump would work for automating dephleg cooling flow.
If going down that route, probably there would be a need to vary the power also.
I think a big problem with automating the process is going to be accurately measuring the vapor temp above the dephleg.

[Pierrot Lunaire 55]

Let's answer those in reverse order.

The defleg has an RTD probe inserted just above the exit tubes. As you can see from the photo, the probe is longer than it needs to be, which could cause some thermal siphoning making the reading slightly lower than the steam actually is. This application doesn't call for high precision since the final measurement is proof at the outlet via the hydrometer. We can change the setpoint to reflect final proof. The photo also shows the RTD measuring the cooling water exit temperature (lower left). There is also the feed water temperature taken at the bucket. Since we know the speed-to-flow measure for the peristaltic pump, we can calculate the water flow as well as temperature differential to determine how much heat is being removed.

750_4998 copy.JPG

The peristaltic pump is a handy approach in that it can provide a very specific flow and is not dependent on supply pressure. We started out with a simple speed control and moved to controlling speed via an Arduino in a closed loop with steam temperature at the defleg exit. This worked but was a bit trickier to control. The PLC system as shown in the write-up is certainly overkill given its complexity. A stepper motor with an appropriate controller is ideal in this case. You could do the same with far less hardware.

[Pierrot Lunaire 55]

BTW, if anybody is interested, I can probably provide more information on the earlier version with the Arduino controller. We used a peristaltic pump bought off Amazon and created an honest-to-goodness PID controller. It worked and had a method to load in the setpoint, but had no user interface of its own. If you have a working data display mechanism, it may be what you need.

[Yummyrum]

Thanks for your feedback on this project PL55.

Reading between the lines , are you saying that the theoretical steam (vapour passing the deflag) temp setpoint in reality needs to be tweaked up or down with the goal being to achieve a desired ABV at the spout . ?

In other words , say the theoretical ethanol vapour temp is 78.3°C .
But in reality you might need to set the setpoint to say 79.5°C or what ever to actually get the desired ABV ( say 95.5)

Some of the fellas have already been using Parasteltic pumps such as the Kamoer DIP1500 which has external speed control via 4-20mA , 0-10v or modbus486.

I’m guessing a much simpler setup with just a PID controller driving the above style of Parasteltic pump would do a similar job .

[Pierrot Lunaire 55]

What I'm suggesting is that many variables enter into the process. If we look at the vapor vs. liquid temperature graph, if the process is perfect, we should be able to predict the final ABV from the vapor temp at the defleg exit. Of course the process is far from perfect.

Distillation_cr.jpg

As mentioned, my probe is too big and likely reads low because too much heat is being transferred out through the probe itself. I should insulate it. Moreover, it takes some time for a change in the defleg temp setpoint to be recognized at the outlet. Let's say I change the defleg exit setpoint lower by two degrees C. The pump has to change speed, running faster, changing the flow through the defleg, bringing the temperature down, probably over a couple of minutes. We want the change to be slow, so we use the D factor in the controller to keep it from overshooting. The now cooler vapor has to get through the defleg and main condenser, and start filling the parrot. It could easily take several minutes for the change to be recognized fullly on the hydrometer. The PID programming can be set with this in mind, as temperature loops are generally pretty slow.

This project, as described in the article linked in the first post, is more elaborate than is necessary for the basic function. It was a demonstration of what is possible given some basic automation technology. The earlier version running on an Arduino was able to control the loop, but it was difficult to tune the PID parameters and it lacked an effective user interface. Someone else approaching the challenge could take a different strategy entirely, but peristaltic pumps in this context seem like a really good idea.

I suppose the ultimate objective is a controller that can compensate for the changing boiler temperature as a run cooks off its initial alcohol content. It will adjust the reflux rate to maintain a desired overall ABV at the outlet.