Three-Chamber Still

DMdistilling

I'm currently involved in an exciting project where we're constructing what may be Australia's first three-chamber still for producing single malt whisky. We're aiming to simulate the process using Aspen Plus or Aspen HYSYS to understand the temperature and pressure profiles throughout the system.

The still design includes three chambers, each 900mm in height and 1300mm in diameter. To manage pressure, we've incorporated a 2" pressure relief pipe from each chamber extending to the top. However, I'm encountering difficulties with the simulation, particularly with the semi-continuous nature of the process.

I'm reaching out to see if anyone has experience or data on temperature and pressure profiles that could inform our simulation. Specifically, we're looking for insights on:

• Typical temperature ranges that should be maintained in each chamber for optimal distillation of single malt whisky.
• Pressure profiles that are expected in a three-chamber still system.
• Any specific considerations for modeling a semi-continuous distillation process in Aspen software.
• Recommendations for setting up the relief system in the simulation to accurately reflect the physical setup.

Any guidance, references, or examples of similar simulations would be immensely helpful. If you require further details to provide assistance, please let me know, and I'll be happy to provide them.

Bolverk

My recipe development one won't be complete for a bit so can't help... sorry.

I think you're going to be hard pressed to find any hard data. You could reach out to Todd Leopold... he's the only one that I've heard of running a three chamber.... I have no idea how open he'll be with that info though.

DonMateo

Look up the boston apothacry's documents on three chamber still operation. There is enough information there to get started. With some engineering knowledge and temperature pressure calculations you should be able to figure it out. As for the first in Oz there is a distillery in perth that had a vendome three chamber still on order.

grim

I'd consider the fact that it can not be modeled.

Bolverk

@grim said: I'd consider the fact that it can not be modeled.

Funny you say this, I've been emailing with Haggy from HD (dude is far smarter than I am) I was trying to get an idea what to expect with certain design elements that I wanted to implement in mine using his double thumper calculator and he keeps finding variables that I didn't even think of. Best we were able to come up with is that you run each chamber for however long it takes for the preheater charge to point where it would start vaporizing then that's when you drop the charges. Once you establish a few baselines with your first run then everything else will become clear.

grim

Wager a bet that the theoretical plate count is less than 1 - (less separation of volatiles than simple batch distillation with no plates, however abv increases due to water condensing out in the "column").

Smear, baby, smear.

I'd once heard this called "fractional condensation" and not "fractional distillation", and there is something satisfying about that.

grim

Said another way, maybe clearer, maybe not.

At the start of each cycle, the upper stages initially function as dephlegmators, as the temperature rises, they switch to functioning as highly inefficient pot stills.

Maybe another analogy is a capacitor - at the start of each cycle, they accumulate charge, and when they hit the tipping point that charge is released.

Bolverk

My understanding and I'm 100% prepared to be wrong about this...

If you're using it as a finishing still the doubler/thumper have got to count for more somewhere around 1 plate, otherwise how do we get to barreling strength in a single pass?

Smaug

@Bolverk said: My understanding and I'm 100% prepared to be wrong about this...

If you're using it as a finishing still the doubler/thumper have got to count for more somewhere around 1 plate, otherwise how do we get to barreling strength in a single pass?

Driving slower.

Bolverk

I'm missing something, can you elaborate?

DonMateo

Hey @DMdistilling mate. Your not from Tasmania are you ? Where I am from in Oz if people attempt to help you, you normally say thanks mate.

Smaug

@Bolverk said: I'm missing something, can you elaborate?

Sure.
When thinking of batch distillation as an example, low and slow will provide a few extra "proof points".

I wish I could find the passage of how Mike Mc Coy explains ( and I'm paraphrasing) how applying just enough heat to optimize separation. But my search kung-fu is weak.

Bolverk

That makes sense, thanks man

punkin

@DonMateo said: Hey DMdistilling mate. Your not from Tasmania are you ? Where I am from in Oz if people attempt to help you, you normally say thanks mate.

Is that needed? Are you making the place better with that comment?

They haven't responded at all, let alone just to say thanks to you.

DonMateo

You know mate I was wondering about this and I have to ask the question why would you use a three chamber still to make a single malt whiskey unless the only marketing selling factor you want is to say you made it in a three chamber still.

If you look at the historical information in the pincock and holt report from 1908, in the US at that time malted whiskey and bourbon were made either in column stills or in pot stills. Three chamber stills were almost only used for rye whiskeys which had the grain bill over 40% UNMALTED, rye. In the malting process you lose most of the vegetable oils that a three chamber still concentrates. These guys are not going to get the product they think they will get out of a three chamber still. Which is why Leopold brothers uses their three chamber for Rye and their other stills for non rye whiskey.

Bolverk

Interestingly i just saw a post on Instagram that Leopold is going to run a pear brandy through their 3 chamber

DMdistilling

Hey everyone, I want to thank you all for sharing your opinions with me. Currently, I am working on modelling the process using Aspen Plus and HYSYS. However, since it's not a continuous process, it's taking me more time to work on it as a semicontinuous process. @DonMateo , you are right that I could get ferulic acid from rye, which is a phenolic acid that gives more flavour to the whisky. However, we are building a continuous still and three-chamber still exclusively for whisky production. so I'm just trying to get a simulation for chamber still which give me enough data about process. 105 celsius for bottom chamber

DonMateo

Hey mate. Everytime ibdid the calculations incame up with 1.06 bar and a temp of 106 deg thats with a liquid column height of 600mm. The main source of doubt that i have is wash does not behave like water due to the ethanol content. So the temperature effect is probably higher. As well i live at 900masl. So my boiling point for water and wash is 3 degrees lower. There is only one way to find out. Build it and they will come.

Haggy

@DMdistilling,

I just saw your posts and even though this is late, I can tell you about my simulation of the 3 Chamber Still process. If you have done a simulation, please compare this post with your model. If you have built and run a 3 Chamber Still, how does this post compare with your data.

The basis for this simulation is my thumper calculations ( see the HD Calculator Section ) and given in my HD posts on " Hail the Mighty Thumper " and " What Two Thumpers Can Do". Essentially, the 3 chambers are like thumpers with vapor injected in each chamber coming from the chamber below. I will simulate Chamber 2 going into Chamber 1 and then over to the independent Thumper. I do not include a vapor pressure calculation for each chamber. Maybe later.

So, compared to a pot and 2 thumpers setup, this 3 Chamber Still simulation has Chamber 2 like the Pot, Chamber 1 is like the first thumper and the independent Thumper is like the second thumper.

To start, input data is 210 gal of 8% abv beer in the top preheat chamber being heated by the vapor from the below Chamber 1. A 2"coil in the preheat chamber of about 33.6 ft^2 area is estimated. I have a calculation ( not show here ) for the heat transfer through the coils and the results show a 90 minute heat up time to about 62.8C ( 145F ) and about 0.26 gal/min of vapor condensed.

Chamber 1 will start with 210 gal at 8% abv, Chamber 2 will start with 220 gal of 4% abv and the Thumper will start with 10 gal at 6% abv ( from the previous cycle ). It was seen that a steam injection of 48 kw gave a 90 minute cycle time for the abv drop from 8% to 4% in Chamber 1.

At the end of the 90 minute cycle, each chamber drops to the chamber below, beer is added to the preheat chamber, and the cycle is repeated.

So, for each cycle, Chamber 1 is heated by steam/alcohol from Chamber 2 and the abv goes from 8% down to about 4% in 90 minutes. Likewise, Chamber 2 starts at 4% abv and drops to about 0.9% in 90 minutes. Not calculated is the higher pressure and hotter bottom Chamber 3 where oils are extracted, flavors enhanced and sent upward by the injected steam and where the liquid abv drops to almost zero.

The simulation calculates the vapor abv, liquid abv, amount distilled vs time (vapor rate), temperature for each chamber and for the Thumper. A maximum fill of the Thumper of 20 gal is presumed and the overflow goes back to Chamber 2.

As in my posts in HD - a series of graphs of the key variables vs cycle time show the results. The calcs go on out past the 90 minute mark, we will stop the cycle at 90 minutes.

First, the expected liquid abv in the Chambers and Thumper:

Then, the expected vapor abv exiting upward from the chambers and the distillate from the Thumper.

Next, the amount of vapor / distillate from the chambers and the Thumper.

Finally, what you want, the temperature variation of Chamber 1 and the Thumper vs cycle time.

So, those are the simulation results using a heat loss of 30-35% in each chamber and Thumper. The abv will vary somewhat depending on the heat losses. You see that at the end of the 90 minute cycle, the Thumper distillate product was about 60% abv and about 27 gal were made.

DM, are these results close to what you have calculated or seen? If given some real data, I can adjust the chamber and thumper heat losses and modify some calcs to improve the simulation.

And, a KEY POINT,
Look at the temperature chart. You see that at 90 minutes, the Chamber 1 temperature is about 98 C. You can use that temperature indication to stop the cycle with 4% abv left to dump to Chamber 2.

So, I hope you and other stillers can reply to this post. That will let me/us see if this simulation fits real data. Another objective is that I hope this helps others better understand the 3 Chamber Still process and maybe want to build or buy one. I did not go into the product flavoring advantages of the 3 Chamber Still. See other posts for that information.

Haggy

Smaug

Hi @Haggy,

Thank you for the input on this thread.
I hope OP responds.

But also, will try and see if I can get the hobby chamber still data info as well as the 3 Howls data as soon as it becomes available. The 3 Howls system is a full sized, commercial production still.

Fingers crossed, as I have thus far had limited response to existing systems in service :(