RO for Dummies - Step by Step RO design help please.

[wmick]

I am so excited to design, build and commission a Hobby Sized RO for next season. I am obviously not alone... because there are lots of folks on here, seemingly in the same boat... or trying to optimize their current projects.
I have been combing through the plethora of RO threads on this site, in hopes of gaining a thorough understanding... A ton of good information and experience. Fun to read.. informative... and now I am more confused than I ever was.
Everyone has different requirements... Throughput requirements, budget requirements, batch, vs continuous, concentrate brix requirements, single-pass vs multi-pass. Intregal recirc, vs separate recirc, vs no recirc.... The list goes on and on... Every time I think I get close , I read a post that seems to contradict my understanding...

I thought maybe I might try to get a new thread started... RO for Dummies... (me being the "dummy") Focusing on design criteria, and theory, step by step... Not focusing on any one real life project, but in hope that myself and others can apply the general knowledge and theory to their specific projects. With any luck it will get some traffic and serve some purpose. If it doesn't, I will know it was a bad idea. A spot for definitive questions, explanations, and informative discussion or debate... Hopefully creating a bunch of new RO experts... or at least one.

I will start it off with a first question.. in some cases I think I know the answer..... and hope that someone cares enough to blow my theories out the door.
I'm really not confident that I know all the "steps", so Steer me straight if I'm off the mark...

STEP 1 - Establish System Requirements - I assume this is where everyone should start, ideally, if starting from scratch.
Using some really round numbers as an example. Requirement for this fictitious system project example is to take 100 Gal of 2% sap to 10% in one hour...

STEP 2 - Calculate Permeate and Concentrate Flows.. (this will aid in establishing a re-circulation requirement and overall system fluid flow-rates)
Formula to establish volume of concentrate would be brix_in / brix_out X sap_vol or 2 / 10 X 100 = 20 Gallons of Concentrate. Volume of permeate is sap_vol - conc_vol or 100 - 20 = 80 gallons of permeate (water removed)
Is this correct ?

If so
STEP 3 - Establish Total System Flow - (Including re-circulation)
As I understand it... "Recovery" (which is the amount of permeate produced, as a percentage of total membrane flow) is recommended to be no more than 15% to prevent membrane fouling and optimize its performance. Is this correct?
If this is the case.... In order to produce our 80 Gallons of permeate per hour.... we would calculate total flow with this formula perm_vol / recovery or 80 / .15 = 533.33 gal/hr total system flow to the membrane(s) (including re-circulation). Re circulation would be achieved by creating a constant loop of concentrate back through the membrane over and over... creating a good constant high volume flow across the concentrate side of the membrane. This could be done by diverting some of the flow from a single high pressure pump,.... or with a second pump dedicated to re-circ. EACH HOUR - 80 Gal goes to the permeate tank... 20 Gal goes to the concentrate tank and 433.3 re-circulates.
Is this Correct?

That's a good start, I think.... Have I missed a step yet? Do my numbers make sense? Thanks in advance, to those that share and participate.

[Mead Maple]

I will be following this thread very close and appreciate you starting it!


Sent from my iPhone using Tapatalk

[Urban Sugarmaker]

Your numbers make sense. There might be some minor total flow differences in real life though. Sap temp plays a role as does starting sugar content but you have the right idea on the theory. You may have trouble getting these specific numbers to work with a single membrane because you'd need 600gph pump. Controlling pressures at that volume could be a challenge with a 4x40 membrane. Two 4x40 membranes with a 330GPH will give you the performance you describe above.

[bowhunter]

Everything you have looks good and is close enough for this kind of project. Step 2 is an approximation which as I said is ok for this kind of project since the mathematically correct calculation isn't worth the extra effort. Your understanding of recovery and circulation is excellent.

I would add some additional thought to the first step. You need to decide whether or not you want to run the RO before boiling or at the same time you're boiling. So how much time do you want to run the RO each day? You also need to estimate how much sap you have to process and how fast you can run the evaporator. I prefer to run them together. I don't store concentrate so I run my RO to match my evaporator capacity. If you do it in two steps i.e. run the RO first then run the evaporator the RO will have to be much larger. I usually like to boil at least 6 - 8 hours. If you only have enough concentrate to run 3-4 hours the evaporator operation is really inefficient. It takes a lot of heat to start up each time for a short run. In fact you will lose a lot of efficiency you gain with the RO with really short evaporator runs.

Overall your approach is correct. Decide what you want to do first and then how to do it.

[wmick]

Thanks Urban. You are correct... Good Point. I should have established a minimum sap temperature in the initial requirements... I assume that temperature compensation will come in to play, when we start looking at pump and membrane specs in subsequent steps, but not sure yet??.. Would 32 F be a good temperature to plan for?..(as a worst case scenario). As soon as I did the re circulation calculation, I figured that this high flow scenario would not likely be a candidate for a single post, single pump system like many are building... (but the design process theory should still apply, I think?)
.....Here's where things get interesting already.... A second pump might be an option to compensate for the high recirc requirements... It would not need to be a high pressure pump.. Just a circulator... Lets leave that one for another time and stick to the single pump systems for now...

You make a point that I would like to explore a bit... You suggest "Two 4x40 membranes with a 330GPH". When I first read this, I thought how can we possibly achieve the 533 gph with a 330 pump??? Without thinking it through, I thought "more membrane would require more flow". But then it dawned on me this is not the case. Adding a second membrane can change things up considerably.. Lets take a look at the parallel membrane configuration 1st.. (as I think it is the simplest math)
I will call this Step 4.

STEP 4 - Consider Different Membrane configurations.
A... Parallel Configuration of 2 Membranes. Each Membrane will produce half the permeate (40Gal/Hr ea) and each membrane will produce half the concentrate (10Gal/Hr ea) In order to abide by the 15% recovery guideline, each membrane will require 267 gph flow to it. (40/0.15) ... Because the total pump flow will be split evenly between the two membranes, total system flow would still need to be 267 X 2 or 533 gph.. This is the same as my original scenario, and no advantage for the 100 GPH requirement. Is this correct?
B... Series Configuration of 2 Membranes.... Membranes in series is a different ball-game. Each membrane will each be creating some of the total permeate, just like the parallel configuration, However... The remaining flow and re-circulation will go through both membranes in series, rather than getting split between them. Therefore, allowing BOTH membranes to take advantage of more of the pump's flow, for the re-circulation/recovery goals. I sketched things up a bit to show my thoughts... in this scenario, we can still achieve the production goals and the 15% recovery goals with much less system flow... 307 gph, rather than the 533. I do understand that this is very idealistic conditions... Flow control and pressure drop compensation will likely need to be considered if we would want these 2 membranes to flow exactly the same amount of permeate. As my sketch demonstrates.... Flow across the 1st membrane would need to be approx 40 gph higher, to compensate for the permeate draw-off and still provide enough flow to the second membrane. Does this sound correct?

This is definitely where some real life experience and expertise could help... I've oversimplified with the 40/40/20 out-flows. I know there will be some pressure and flow differences between the 2 membranes, resulting in uneven permeate flows. Do we know what to expect here? Is it significant enough that we need to plan for it? Hopefully not, but if so, how can we calculate this ??? or is there a reasonable margin of error to plan for? Would a person put a flow control on the #1 permeate line, to try and even them up... or just let it go, and increase overall flow requirement to make up for it? ?? I assume that a 330 pump would likely handle this scenario... but I'd really like to figure out how we can "Know" that.

I sketched up two scenarios, first one with even 40 gph from each membrane,, and a second one, where the first membrane produces more permeate... I think it demonstrates how this can affect the overall flow requirements... at 45/35 rather than 40/40 we would require 345 gph rather than the 307.

Thanks for playing....
series flow.JPG
series flow 2.JPG

[wmick]

Thanks Bowhunter.... Understood... Even before deciding on throughput requirements, one should consider all your good points here... I guess I skipped a step, assuming everyone would have thought this through. but maybe not... It might not be about how much sap you collect per day, but more-so how how much you boil per hour... This is a really good point.!!!

Lets Call this STEP 1-A
STEP 1-A - Establish batch or continuous flow and how much throughput you really need...

If I were going to design a system for myself to feed my evaporator, I would want a system that can produce my desired concentrate at a rate of about 30 gph... (which I think is about max throughput for my arch (as-is)) So... Working backwards but keeping with the 2% to 10%.... 30 gal of concentrate to boil, at 10%, would require 150 gal of 2% raw sap. Sap Vol = Concentrate Vol / (BrixIn / BrixOut) or 30gph / (2 / 10) or 150 gph raw sap in. My RO system would need to be good for 150 GPH.

If I want to RO in batches through the week and store my concentrate till the weekend, it would look something like this... The most sap I intend to collect in one day will be about 400 gallons... and I want to be able to take this from 2% to 10% in 5 hours , so I would need to have a system capable of 80 GPH.

Of course, if I relax on my concentration specs, or any other variables, the numbers would all change..

Does this sound right?
Thanks

[Urban Sugarmaker]

You make a point that I would like to explore a bit... You suggest "Two 4x40 membranes with a 330GPH". When I first read this, I thought how can we possibly achieve the 533 gph with a 330 pump??? Without thinking it through, I thought "more membrane would require more flow". But then it dawned on me this is not the case. Adding a second membrane can change things up considerably.. Lets take a look at the parallel membrane configuration 1st.. (as I think it is the simplest math)
I will call this Step 4.

I was quoting you a performance for a series setup. The recovery rate is not double (30% vs. 15% with one membrane) with 2 in series. It's 27.75% because the second membrane is starting with a higher concentration. That's also why you can achieve approximately 80% water removal with less total flow. Parallel flow requires more pump volume and will produce more permeate faster, but not at a higher concentration. It seems most commercial units are configured in series. That would be my preference.

[wmick]

Gotcha...
This is the stage that I am really stuck... and need some major help.
What I'm really lacking is the understanding of mechanics of specific membranes and how they might react to different pressures and temperatures, brix levels, etc.
All I really know at this point, is how much flow I need in the overall system.....
I don't understand why the recommended recovery rate % would be lower with higher brix.
I don't understand how to choose a membrane or how to read and use the pertinent specifications.
I don't know how much pressure I need or how to figure it out...
I don't understand what the flow and pressure differences might be, between the membranes of a 2 post system.

I've searched for test results, etc, online... and If I were processing Sea-Water, I'd be inundated with information.... but not so much for sap.??

I am really hoping someone might take the time to go through these numbers with me... (Maple Membrane 101)
Knowing what we know... and based on the system we started off with. How would one go about specifying a set of membranes and a pump...
and how to calculate what to expect from those membranes... at various temps, pressure, brix, etc...
.... for a system that needs to be capable of processing 100Gal/hr of 2% sap at 32F into 10%concentrate... on a single pump , double membrane in series system.

Thanks a ton.

[bowhunter]

You are correct the about the recovery and also about the series vs. parallel configuration.

[bowhunter]

I spent about 3 months developing a tool in an EXCEL spreadsheet that I use on the forum to estimate sizes for people wanting to build their own or troubleshoot problems with their existing system. It is pretty rigorous relying on manufacturer's data, physics, physical chemistry and some empirical data but not something I would offer to others to use because it does require some interpretation. I am happy to do a sizing run and send you a copy of the recommended sizing for the membrane(s), pump and flow meters you want to build.

Sizing from scratch pretty complicated to answer so let's start with some basics if you want to try and develop you own sizing calculations.

1. In reverse osmosis we are supplying enough pressure on the high sugar side to force water to permeate through the membrane from the high sugar side to the no sugar side. To do this we have to overcome the osmotic pressure of the high sugar side. During osmosis mother nature is trying to dilute the high sugar side with water from the no sugar side of the membrane and osmotic pressure is the driving force.
2. As the sugar content in the concentrate increases the osmotic pressure we have to overcome increases, so we need to apply more pressure to the concentrate side to get the same water permeation rate.
3.At a given osmotic pressure, permeation rate is directly proportional to the pressure differential between the osmotic pressure and the applied pressure.
4.Osmotic pressure can be fairly complicated to calculate because it includes primarily sugar but also other trace minerals in the sap and the absolute temperature.
5. Temperature is a major factor in permeation rates. Permeation rates drop about 3% for every degree drop in fluid or membrane temperature. For example permeation rates are about 50% lower at 34 F than at the rating temperature of 77 F. This is a result of the viscosity of water increasing as the temperature drops. You can notice this in nature sometimes as the temperatures approach freezing during a rain. You will see big droplets hanging on overhead wires and tree branches as the viscosity of the water increases near the freezing point.
6.Membranes are generally rated by gallons of permeate collected over 24 hours on water at the rating temperature and pressure. They're usually rated at a recovery rate of 15%. The rating is always done at 77 F but, the rating pressures are different. Nano-filtration membranes are generally rated at 70 psi. The XLE membranes are rated at 100 psi and I believe MES rates their membranes at 125 psi.
7. Recovery is the % of the water that is removed per pass through the membrane. Most manufacturers recommend a maximum of 15% recovery per membrane per pass. You can run higher than 15% on sap and many of the once through systems run 25% or so. The only downside is the membrane may foul more quickly and require more frequent soap washing.
8. Each size membrane have minimum and maximum velocities that have to be evaluated. Too high and you can mechanically damage the membrane. Too low and the membrane will foul much more quickly.
9. There are generally three options for membranes. The old workhorse for the industry in North America has been the nano-filtration membranes such as the NF270. It removes a lot of water and so it will generally maximize the capacity of the RO. The downside is that they are prone to pass sugar at higher concentrations much about 10%. They are also more delicate to soap wash because they are not tolerant of higher pH. The FilmTec XLE's are now very popular. They are less expensive than the nano-filtration but they permeate at a lower rate than the nano-filtration and they are much more tolerate of mistakes with higher pH during soap washing. The MES membranes are becoming very popular also. They were developed with sap processing in mind and I believe they are pretty tolerant of soap washing and running high recoveries. They're not as high of capacity as the nan-filtration membranes but they are much less expensive.