Determining Mixing Time in Storage Tanks 1

[mp08]

We currently have several large (105m3) final product tanks. Since the plant I work at is a batch operation, we typically have material in the storage tank and once another batch in the reactor of the product is done, we pump it into the storage tank. We then mix the the product in tank simply through re-circulation with the pump. Also, there is no dipleg in the tank.

We have to take final samples of the storage tank to determine what the viscosity is before shipping to the customer. The issue we're seeing is that we don't know how long it takes to actually mix the tank so that it is homogeneous. The operators take a sample once it's been mixing and if it is within specification, are okay with it. If it's not within spec, then they take a resample. We then have issues where a customer will come back saying their results are different and when we take another sample from the storage tank, we see that our results are very different than what was put on the COA. We're trying to figure out a way that we can estimate how long it will take a tank to mix so that we can get an accurate result of what exactly we are shipping.

Does anyone have suggestions? I could only think of just continuing to take samples until the values stabilize, but to do that continuously takes a lot of effort.

 

[Latexman]

You need to research "jet mixing" or "jet mixers". Perry's should lead you to THE key reference by Fossett. There's another reference mentioned in my companies manual by Fossett and Prosser.

Could you characterize the fluids so we can help more?

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Latexman]

Also, check out a Harnby, Edwards, and Nienow reference.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[mp08]

We make polyesters at the plant. The tanks are heated to keep the material molten.

 

[bimr]

An eductor is probably the least expensive method to mix. You need the viscosity and the mixing time. You can lower the required mixing energy to match the time between batch discharges.

http://www.transvac.co.uk/brochures/Jet Mixers.pdf

 

[Latexman]

That doesn't really tell us much. Please characterize the fluids in terms of the parameters that are important to mixing technology. Viscosity is probably #1. Then maybe solids and density. For example, mix a 50,000 lb batch at 10,000 cP, 80% solids, and 8.0 lbs/gal with a 50,000 lb batch at 20,000 cP, 80% solids, and 8.0 lbs/gal. But that's me speculating. You tell us what you need the mixing to do.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[MortenA]

The point is - there are loads and loads of info and theory on this subject. Most process engineering text boos would comment on the subject. If you dont have the people in-house (a chemical engineer should be able to dig out a text book and brush up on the subject) then you should consider getting external help.

Best regards, Morten

 

[mp08]

Our products hav[sub]e a huge range of viscosities (up to 35,000cps) and the size of the batch could be anything ( 0 - 150,000 lbs) that need mixing. There are essentially no solids in the tank.

I couldn't find anything in my text books (I'm currently the only chemical engineer on our plant site). I plan on looking for the references above and already have looked up jet mixing online and found a few helpful papers / equations.

 

[Latexman]

No wonder the results seem random. 35,000 cp! Jet mixing works best if the jet is turbulent. It is worse if the jet is transitional. It is even worse if the jet is laminar. You did not mention having a jet mixing nozzle. I assume you have none. I use a standard pipe reducer about 1 foot off the bottom of the tank and aimed at the high liquid level diagonally across the tank, to maximize the jet length. It works good for us with cP viscosities in the 100's. It's not as good with cP viscosities in the 1000's. I doubt you are getting any significant mixing at all. Thus, the randon results. Unmixed!

Study up on this technology! Let us know what you find. I suspect you will need some mechanical agitators for your tanks.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[LittleInch]

Latexman,

Fully agree - Once mp08 came out with 35,000 cP I'm fairly astounded that this stuff flows at all, never mind mixes well in a tank(!)

I suspect you would need to re-circulate at a rate of 3-400m3/hr per tank to get any sort of decent mixing and possibly use two or three different sets of nozzles at different levels inside the tank and it probably needs at least 2 complete turnovers of the tank to get close to a homgeneos mix. Mechanical agitators sounds like the only sensible solution at present.

I don't think you'll get a theoretical answer to this.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[geesamand]

With high viscosity fluids the right mixing setup can make incredible differences in performance and mix quality. There are certain impellers, applied correctly, that will turn thick stuff with ease and minimal power. Getting it right is often a combination of temperature regulation, mixing wetted parts design, and tank configuration.

With high viscosity product you need an impeller design that can occupy a large amount of the vessel and impel the fluid without relying on large "return" areas in the flow. Which is to say that high flow agitator impellers, propellers, and jet mixers are highly unsuitable for this kind of material. Slow and steady torque makes this work.

I cannot offer a specific commercial pitch in this forum but I will say that among the current mixing providers, only one is still serious about research and advancements in this field. More than one company have solutions that are more or less capable of good results but only one has enough lab work to really make it work well. This has been applied with excellent success to very high consistency paper pulp, polymers, paints and coatings, chocolate production, etc.

David

 

[IFRs]

Asphalt is sometimes mixed with full tank height paddles rpms turned with a high torque low speed motor. This might work for you also.

 

[rkg1]

Jet mixing is not going to work. The viscosity is too high and Reynolds number too low.

The question is a common one for producers of viscous fluids (paint is a very common example). How long to wait before you can take a sample that is truly representative of the contents of the whole vessel?

Mixing in the Process Industries (edited by Harnby, Nienow & Edwards) is OK but the Handbook of Industrial Mixing (edited by Paul, Atiemo-Obeng and Kresta) is good. There is a chapter on estimation of blend time which includes worked examples of how to use the appropriate correlations. It also includes information on blending of non-Newtonian fluids - which I suspect may be relevant in this case - and design of jet mixers.

The sizing of the agitator you need will depend on the shape of the vessel (what is the ratio of diameter to liquid depth) and the time you are prepared to wait before taking a sample. You also need to be able to specify the allowable deviation from the mean; what is acceptable to the customer? +/- 5 %? +/- 1 %? The tighter the spec, the longer you will need to mix at a given power input.

While the viscosity of the fluid is high (35000 cP) it is the Reynolds number should be your focus. Mixing this fluid in a vessel 1 m in diameter will be laminar. In a vessel ~5 m in diameter (a 100 m3 tank with liquid depth equal to diameter) the Reynolds number will be transitional. As we scale-up Reynolds number increases because it is proportional to the impeller diameter squared.

I think that there are calculations that can be carried out that will enable you to assess the performance of your system, i.e. there is a theoretical answer out there.