Designing an atmospheric storage tank recirculation system - pump control strategies? 2

[HaxNobody]

I'm working on a 10,000 gallon storage tank for reverse osmosis purified water that needs to be recirculated through a UV sanitizing reactor. I would like help on the correct strategy for controlling the pumps in this loop. The loop design is as follows:

I have two 10HP centrifugal pumps in parallel, each with its own check valve on the discharge. They can achieve up to 62 PSI of head pressure at zero flow. These pumps feed directly to the UV reactor.
After the UV reactor, I have a 240 GPM flow restrictor valve that is intended to prevent the UV system from exceeding its design flow rate. This restrictor valve could have 15-20 PSI drop when running at its full flow rate.
The line then goes out to the process loop, where various equipment taps in to draw from the loop to feed mix tanks and other processes. The loop pressure should be around 30 PSI when recirculating under "idle" conditions with no draw off the loop.
When the loop returns, it goes to a set of 1 micron absolute rated sediment filters.
After the filters, the loop then goes to a backpressure regulating valve that is set to maintain around 170 GPM of tank return flow when the loop is idle, to achieve 1 tank turnover per hour.
The backpressure regulating valve is set to 10 PSI, and it will maintain around 30 PSI at design flow. If the loop pressure drops because other processes are drawing from it, the backpressure valve will close and create a "dead leg" at the end - to prevent this, the backpressure valve also has a 10 GPM flow restrictor in parallel to ensure some amount of minimum flow is still achieved.

Both pumps are on their own VFDs, and I originally intended to control their speed based on a pressure sensor located just before the backpressure regulating valve at the end of the loop. However, I have identified a problem: If the loop pressure is too low because it is being drawn down by process equipment, this strategy would cause the pumps to keep running faster and faster, even though the 240 GPM flow restrictor is going to prevent them from performing additional work. So I need to figure out how to "reset" the pump speed based on flow across the UV reactor. I'm not sure whether I should try and detect actual flow, or if I should try and put a pressure sensor on the UV reactor and try to calculate the differential to estimate flow rate.

There may be other pitfalls that I have not yet identified that you guys might know about. Looking forward to any insights on this - thanks!

 

[georgeverghese]

Error in my last post : PIC should be Reverse Acting, not Direct Acting.

 

[LittleInch]

Yes, I'd go with that.

For the last valve I would maybe look at a mechanical stop at say 10% open to give you a constant low flow through the filters and back to the tank.

I think the OPs pressure drop seems very high based off that graph. But I'll look later.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

The pressure drops noted seem far too big. I get about 3.5psi per 100ft. Even with the added equivalent length so say 450ft, this is only 15 psi not 57...

I would set the final valve at say 35 psi, but you can play with that. You should be able to get flow transmitters with built in control logic to drive your VFD and same for the valve with pressure signal input.

You would flow back into the tank at 240gpm when no other users are there, but that seems OK as the pump would ramp down. Saves messing about with low select blocks and similar.

How does that work for you?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[HaxNobody]

Think you can do without that 240gpm restrictor which burns a lot of pressure. The scheme I have in mind to load the VFD for the pumps is as follows:
a) Install a flow element and FT downstream of pumps or downstream of UV steriliser. Configure a reverse acting hi set FIC with this - setpoint 240gpm - this will be the override controller to provide the load signal to VFD
b) Retain the current plan to sense pressure just upstream of backpressure regulator and configure a direct acting PIC with this - normal setpoint 25psi - this will load the pump VFD most of the time
c) Configure in your control system an auto signal select to select the FIC output when it reaches SP of 240gpm, thus cutting off the PIC output from going to VFD

See diagram attached. This scheme enables you to operate the pump at the min required supply header without losing developed pressure, so it saves on power. If you dont have PLC, you can just as well configure all this with equivalent pneumatics devices, all mounted locally.

Thank you for this diagram and suggested implementation. I'm not familiar with the PIC and FIC terminology - are they different terms for a PID loop/controller?
It looks like I will be adding a flow meter to the setup instead of relying on passive devices to prevent overrunning the UV.

The pressure drops noted seem far too big. I get about 3.5psi per 100ft. Even with the added equivalent length so say 450ft, this is only 15 psi not 57...

I would be happy if the pressure drop is lower than I found off of that graph. Indeed, I looked around and found a few charts that agree that the friction loss should be closer to 3.5-3.6 PSI per 100ft at 170ish GPM like you say, which is an improvement over the 5.5ish PSI that I was beginning with off of the graph. Now it is still 4.5-6 PSI per 100ft at 200-240 GPM, but this flow rate is only required near the beginning half of the loop while performing bulk tank fill operations, and isn't relevant to the end of the loop at the longest possible distance where the tank is returning.

I did some incorrect math as well while converting between friction loss and head loss and between feet and PSI, so even if it were the worst case of 6 PSI per 100ft that would be a total drop of 25 PSI not 57. This gives me back the pressure I will need across my filters, which start at about 8 PSI when brand new at the 170 GPM flow rate.

I would set the final valve at say 35 psi, but you can play with that. You should be able to get flow transmitters with built in control logic to drive your VFD and same for the valve with pressure signal input.

You would flow back into the tank at 240gpm when no other users are there, but that seems OK as the pump would ramp down. Saves messing about with low select blocks and similar.

How does that work for you?

Much appreciated! The filters are not rated to exceed 200 GPM total (50 each, 4 in parallel), so I can't run the full return flow at idle, but having a flow meter combined with inputs to indicate whether the consumers are using anything from the loop should allow me to select between two different flow setpoints on the fly.

 

[LittleInch]

PIC is pressure indicating controller
FIC is Flow...

I.e. it tells you the pressure and flow as well as sending signals to a device.

Someone this is in the instrument, sometimes in the controller on the valve. But these are simple control loops.

Or fit 5 filters? Makes life easy.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[HaxNobody]

OK, so I'm thinking about the suggestion to put the regulating valve in front of the filters at the end of the loop. As the filters load up, this will increase the backpressure on the valve. With a pilot operated valve, my assumption is that I would sense flow downstream of the filters in order to control the valve, is that correct? I assume this would be a calibrated orifice plate that the valve manufacturer would supply with it.

The reason I was looking at the plastic valve was because I had it on hand. I wonder if I could still achieve control over it by sensing the pressure on the inlet and the outlet and varying the pump speed based on the delta across the valve? Since I have the flow curve, I could set it at like 5 PSI and then run the pump up until I achieve a 25 PSI delta, which gives 170 GPM on the flow curve. Then as the filters load up, this will affect backpressure, which will affect delta across the valve, so I just keep speeding up to compensate for filter loading.

Adding that 25 PSI of valve delta to the piping losses of 15ish PSI gives 40 PSI, and then tack on 8 PSI for the clean filters gives 48 PSI. Based on the pump curve, this would give me an additional 7-8 PSI of filter loading before I drop down below my minimum tank return flow, and that's still just with a single pump. I think I could stage up the second pump to achieve an additional 5 PSI of filter loading. This means I could only achieve around 20 PSI across the filters before I run out of pumping capability, even though the filters can handle up to 40 PSI.

That would allow me to specify and install a proper control valve at a later date while still being able to get started with what I have.

 

[LittleInch]

No it won't.

As the filters load up the pressure drop required to be taken by the valve will reduce for the same flow. You are sensing pressure upstream of the valve because that valve is set for pressure control.

Again no. The purpose of that valve is to maintain pressure upstream of it so that your users get a decent pressure. As the flow to the users increase, you need the flow to the tank to fall to near zero to send most of the flow to your users. That valve is really nothing more than a slightly sprung orifice. It won't work the way you actually need it it. IMHO.

You are free to do what you want, but everyone here is telling you that the valves you have chosen won't do what you want them to or take too much pressure drop (the flow restriction valves). The system you started with will work, but poorly, and will only supply a certain amount to the users when it could supply much more at higher pressure.

They are both essentially orifices which don't really "control" anything and hence can't react the way I think you need them to.

Georges diagram is, for me, just about perfect that allows you to do the two different things that the circuit needs to do - flow to your users with a decent pressure and flow OR flow in recirc mode to the tank when you have no users. Without exceeding the limits on the UV filter or anything else. And everything in-between.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[HaxNobody]

Georges diagram is, for me, just about perfect that allows you to do the two different things that the circuit needs to do - flow to your users with a decent pressure and flow OR flow in recirc mode to the tank when you have no users. Without exceeding the limits on the UV filter or anything else. And everything in-between.

Thanks, that makes sense and I was thinking about it a bit backwards. I'm definitely going to start with George's diagram and use it to update my initial design. I appreciate you guys for letting me bounce ideas off the wall because I'm still working to understand and grok the how and why of the components and concepts that I'm dealing with here.

Related to that understanding of concepts, I did dream up another potentially harebrained idea last night that I want to try and understand better. Coming back to this comment:

These ultrapure water loops are not necessarily constructed for efficiency. There are issues such as maintaining velocities to sweep the microorganisms off the piping walls.

This makes sense, are there some rules or calculations I can use to ensure I have the right velocity through my system?

The idea I dreamed up was to reduce the main loop velocity by separating the tank recirc from the loop completely. So in other words, the end of my loop would no longer maintain the full tank return flow, and would instead return through a much smaller line at a much lower flow rate. But it sounds like if I do that, I risk promoting bio growth because of the reduced pipe velocity? The minimum tank turnover would then be achieved with a much lower head loss by occurring near the pump with a wye or something of that nature.

Again, I'm not trying to knock or disparage any of the excellent suggestions and ideas I have received so far, I'm just wanting to understand the hows and whys better.

 

[LittleInch]

I have no idea tbh.

All I do know is that you are operating in the turbulent region of flow at all times here ( as opposed to laminar flow at very low velocities) which should act to reduce any growth, but there may be a minimum velocity recognised as well.

A quick google say the recommended minimum velocity is 5 ft/sec to prevent biofilm build up and bacteria (Urgh)

you could indeed do this, but then your pipe to the users could become contaminated as a dead leg?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Heaviside1925]

OP,
LittleInch did a good job in summation.

flow to your users with a decent pressure and flow OR flow in recirc mode to the tank when you have no users. Without exceeding the limits on the UV filter or anything else. And everything in-between

I think you said that the pumps were already purchased, if so, do you have the actual rated curves, including the power? I am asking because there may be an opportunity to check-off one of those design requirements through drive settings as opposed to external instrumentation. Full disclosure, it's been many (many) years since I've actually cracked an owner's manual and set up a drive, so I am not greatly familiar with the current "nifty" drive settings, to say, where I am going to go may not work at all. I seem to remember a setting where the drive estimates the HP, that you could use as a high limit, to prevent the excess flow condition. I doubt 240 gpm is a hard and fast number but a manufacture's reasonable not to exceed number. So, if you limit the HP to something like 230 gpm then that should be close enough to prevent an excess flow scenario. If this is a viable solution, it reduces the design requirements to supplying sufficient pressure to downstream users and provide a recirc of 170 gpm. With only these two requirements needing to be met, it could greatly simplify your instrumentation and control.

 

[HaxNobody]

I think you said that the pumps were already purchased, if so, do you have the actual rated curves, including the power? I am asking because there may be an opportunity to check-off one of those design requirements through drive settings as opposed to external instrumentation. Full disclosure, it's been many (many) years since I've actually cracked an owner's manual and set up a drive, so I am not greatly familiar with the current "nifty" drive settings, to say, where I am going to go may not work at all. I seem to remember a setting where the drive estimates the HP, that you could use as a high limit, to prevent the excess flow condition. I doubt 240 gpm is a hard and fast number but a manufacture's reasonable not to exceed number. So, if you limit the HP to something like 230 gpm then that should be close enough to prevent an excess flow scenario. If this is a viable solution, it reduces the design requirements to supplying sufficient pressure to downstream users and provide a recirc of 170 gpm. With only these two requirements needing to be met, it could greatly simplify your instrumentation and control.

I've reached out to see if I can get a power curve, although I don't have much hope for getting one from this manufacturer (AMT).

Regarding the UV reactor, it has 4 inch in/out flanges and a 6 inch body, so a massive amount of flow *could* go though it, but then you lose all your dwell time to actually kill any bacteria. If I remember correctly, it was rated at 178 GPM if we were targeting 40 mj/cm2, and 500 GPM if we only wanted 16 mj/cm2. 30 mj/cm2 is sort of an industry standard, and from what I understand it represents a 3-log reduction in specific bacteria and coliforms like E. Coli etc.

Now to be fair, filling the mix tanks is immediately followed by adding all the ingredients, which includes preservatives and bacteria-killing chemicals, so perhaps it is less of an issue to fill the tanks at a much higher flow rate than the base recirculation rate for the main water storage - so overrunning the UV during tank fill might not be as critical as keeping it within spec for idle recirculation. I suppose this is something I could try if I ever had issues with tank fill rate being "too slow" even if it meets the design criteria.

 

[Heaviside1925]

(not a water treatment person) but I wonder if there is an argument to be made concerning tank turn-over vs passes through the UV at 170 gpm? Depending on rate and batch demands, can the water be considered to have had two passes through the UV? So, a limited "excess flow" condition may not be a concern. May not be viable at all but maybe I'll learn something as well.

 

[pierreick]

Hi,
The weakest point could be the UV system, you may have to consider 2 units in // with all the safety and alarm system in place.
To prevent bacteria growth, it's also good to consider recirculation line to avoid dead legs.

https://www.aiche.org/resources/publications/cep/2016/september/minimize-biofouling-ro-membranes

https://aqua-chem.com/purified-water-piping-guidelines/

Good luck
Pierre

 

[georgeverghese]

I've just realised that flow restriction orifice sitting parallel to the self op pilot backpressure control valve IS the min flow recycle to the tank when there are no users. At this time, the pump would be running at some low speed to maintain 25psig on the header, and the backpressure control valve would be full closed. So do ask the pump vendor what is min flow on these pumps at close to zero flow at 17-18m head at this reduced speed. At this time, the filters will be burning some 7psi, assuming all 4 1micron abs filters are running, min flow through the min flow restriction orifice is perhaps 40gpm(?). So sizing case dp for this min flow RO would possibly be 40gpm of water at a dp of 25-(7+x+y), where x = tank static head at normal level in psig, y = any line frictional loss getting back to the tank from the filters. In summary, when there is no demand, pump flow at low speed may be some 40gpm(?) through the min flow restriction orifice, recycling to tank. In this arrangement, the location of this min flow RO is correct, since it prevents a dead leg at the filters and return line when there is no flow demand at users.
Control systems vendors can offer you, if so require, all pneumatics field located PID controller units for the FIC and PIC , low select relay PY - very reliable if you ask me, as long as you've got good quality dry air for pneumatics. Put them all in a weather proof cabinet if all this is outdoors.

 

[HaxNobody]

The weakest point could be the UV system, you may have to consider 2 units in // with all the safety and alarm system in place.
To prevent bacteria growth, it's also good to consider recirculation line to avoid dead legs.
Pierre

Thanks for your input and for the interesting links. We are currently taking city water through carbon pretreatment to remove chlorine/chloramines and then through water softeners before going to the R.O. units, and did have some fouling after several years. So I wonder how much of a difference it would make to inject a biocide after the water softener in this case?

On the UV, although it is a single unit, it does have dual lamp drivers, UV monitoring, and alarm outputs for various failure modes.

I've just realised that flow restriction orifice sitting parallel to the self op pilot backpressure control valve IS the min flow recycle to the tank when there are no users. At this time, the pump would be running at some low speed to maintain 25psig on the header, and the backpressure control valve would be full closed. So do ask the pump vendor what is min flow on these pumps at close to zero flow at 17-18m head at this reduced speed. At this time, the filters will be burning some 7psi, assuming all 4 1micron abs filters are running, min flow through the min flow restriction orifice is perhaps 40gpm(?). So sizing case dp for this min flow RO would possibly be 40gpm of water at a dp of 25-(7+x+y), where x = tank static head at normal level in psig, y = any line frictional loss getting back to the tank from the filters. In summary, when there is no demand, pump flow at low speed may be some 40gpm(?) through the min flow restriction orifice, recycling to tank. In this arrangement, the location of this min flow RO is correct, since it prevents a dead leg at the filters and return line when there is no flow demand at users.
Control systems vendors can offer you, if so require, all pneumatics field located PID controller units for the FIC and PIC , low select relay PY - very reliable if you ask me, as long as you've got good quality dry air for pneumatics. Put them all in a weather proof cabinet if all this is outdoors.

So that little bypass orifice was selected as a Flomatic CDX10, i.e. 10 GPM regardless of pressure above a certain point. Since I'm now looking to upgrade the backpressure valve to something that is not just a spring operated orifice, this should allow me to dial in the flow rate pretty accurately and be out of the danger zone on the pump flow spec. Unfortunately, while these pumps work as published, they are relatively cheap and don't have any decent support from the manufacturer, including a power curve. This is part of the reason why I have 2 in parallel from the very beginning. Thanks again for your input, it's been very helpful to get me on the right track.

 

[georgeverghese]

Error in my last post: press drop through 4 filters at total min flow of 40gpm ( when at min speed) is less than 1psi, not 7psi.

Took a look at this Flomatic CDX-10 literature on the Flomatic website. No example flow - pressure graphs here to make things clear. You say it maintains 10gpm at all upstream pressures beyond ??psig. Presumably this is then nominally better than the simple RO shown on my sketch - okay. But 10gpm seems low for this min flow at pump min speed - check with pump vendor. Take care with the duty spec for the self - pilot operating backpressure regulator - it only needs to flow the min flow for this pump what the CDX10 cannot at max operating speed, where max op speed is less than 3450rpm.

 

[pierreick]

Hi,
To answer to your question, consider this handbook from Veolia,
link attached:

HTTPs://www.watertechnologies.com/handbook/chapter-09-membrane-systems#MICROBIOLOGICAL FOULING

It is said to keep a residual chlorine content to mitigate the fouling.
If I was in your shoes, I would contact your chemical treatment supplier for advice and follow up.
Good Luck
Pierre

 

[Heaviside1925]

To point to what bimr and pierreick are implying, it what you are doing is not uncommon. There are likely pre-engineered vendor options readily available. From a purely instrument and controls aspect, this thread has provided you with excellent options, especially georgeverghese. That said, there is something to be said for manufactures who have been in the game for decades and decades and the experience they have gained. Not saying you are not capable of designing your own system that functions to the design parameters you have established but there may be unforeseen or unintended consequences to your design that could be avoided by engaging with more use specific folks.

 

[georgeverghese]

At max operating speed for this application, which would be much less than 3450rpm, min flow for pump operating stability is possibly some 40gpm. So if the CDX10 enables 10gpm, the pilot op backpressure control valve should do only the remaining 30gpm at an upstream pressure of 30psig. So line size after all consumers need only be 40mm or 50mm going to PCV / CDX10. I see Flomatics have pilot operated control valves also..
Adding to comments on water, presume the tank is kept under positive cryogenic quality N2 pressure to keep bacteria laden air out.

 

[HaxNobody]

To answer to your question, consider this handbook from Veolia,
It is said to keep a residual chlorine content to mitigate the fouling.
If I was in your shoes, I would contact your chemical treatment supplier for advice and follow up.

We remove the city water chlorination to prevent R.O. membrane damage on the advice of the R.O. manufacturer Suez, a.k.a. Veolia after a rebranding recently. But it looks like there may be other pretreatment options that won't damage the membranes, so I will approach them to discover what options they suggest for that.

To point to what bimr and pierreick are implying, it what you are doing is not uncommon. There are likely pre-engineered vendor options readily available. From a purely instrument and controls aspect, this thread has provided you with excellent options, especially georgeverghese. That said, there is something to be said for manufactures who have been in the game for decades and decades and the experience they have gained. Not saying you are not capable of designing your own system that functions to the design parameters you have established but there may be unforeseen or unintended consequences to your design that could be avoided by engaging with more use specific folks.

Thanks. I came here first because I am able to get a better idea of what to ask of these suppliers now. The rough design of this loop was supplied by the R.O. unit vendor years ago for a smaller system, but I'm finding that some of the design concepts don't scale as well to the larger system we are building now.

At max operating speed for this application, which would be much less than 3450rpm, min flow for pump operating stability is possibly some 40gpm. So if the CDX10 enables 10gpm, the pilot op backpressure control valve should do only the remaining 30gpm at an upstream pressure of 30psig. So line size after all consumers need only be 40mm or 50mm going to PCV / CDX10.

I think something got lost in translation somewhere along the line, since min flow is 1 tank turnover per hour on a 10k gal tank, which is around 170 GPM, not 40. Unless you are exploring the possibility of separating the tank turnover from the loop recirc flow to reduce pumping head? If that is the case, then one of the documents posted previously seemed to indicate that min flow rate could be as low as 3-4 feet per second without causing excess fouling so the loop return could indeed be smaller as you are suggesting.

I see Flomatics have pilot operated control valves also.

I approached them first, but they don't offer an all-stainless option, only the epoxy coated cast iron. Additionally, after talking to a vendor for the Fisher valves, I was quoted a price that was more than double the cost of both of my pumps combined, so it looks like my best option will be to use an electrically actuated valve tied to my controller, rather than a pilot operated self-contained valve.

Adding to comments on water, presume the tank is kept under positive cryogenic quality N2 pressure to keep bacteria laden air out.

This is actually the first time I have heard of this concept, although I immediately understand why it would be a good idea. Right now the UV is the last line of defense, but the product we are making with the water contains preservatives and biocides in the formula so it has been less of a concern. We make a cosmetic product, and we keep samples of the product from many years ago to test ongoing efficacy, and so far product samples from 7 years ago are still testing as good as new.

I drive past a water bottling plant every day and that would probably explain why they have a gigantic dewar of nitrogen outside - cool!