TMP or Flux as Control

[FiltrationEngineer]

Generally speaking, is it better to control membrane systems (water filtration) on constant TMP, or constant flux? I have seen a lot of systems where the permeate is throttled, so TMP and flux change inversly proportional to each other as the membrane fouls. I am designing a system with some automatic control on the permeate side.

Does anyone have experience controlling the TMP or Flux rate of the system? Is there any difference in performance?

 

[brennanpe]

This really dependent upon the nature of the process. What specific process (MF, UF, NF, RO) and what is it that you would lile to process (water, wastewater, milk, juice, paint, caustic, acid, etc.)? Also, how are you intending to operate (feed & bleed, single pass, dead ended, batch, modified batch, etc.)?

brennans7

 

[FiltrationEngineer]

This is a wastwater UF process pumping from and returning concentrate to a tank. I will pilot the process, and will be removing permeate with a pump because of relatively low pressure inside the tubular membrane; this way I can keep either TMP or flux constant using a PID loop with a VFD on the permeate pump.

I am thinking to keep flux constant and let TMP vary between backpulses.

Is there an argument for keeping TMP constant and letting flux vary?

 

[brennanpe]

If this is a pilot system I would do both. Are you planning on performing pressure and flow excursion exercises as well? What UF tube are you planning on using?
Sounds like you plan on running in a batch configuration. What is the size of the process tank and what will be the membrane configuration (number of tubes in series and number of passes)? Also membrane area? What concentraion factor are you hoping to attain? Is this a supported membrane?

To approach your question more directly, all of the UF wastewater systems I have ever commissioned have kept the the pressures constant over the course of the run. Not many work the other way (constant permeate rate). Having said that, really the determining factor is whether or not this is to be a batch concentration or continuous (no concentration) process. Due to the viscosity changes from concentration and pressure limits on most membranes constant TMP is the way to go. By increasing the driving force for permeation as you go all you really due is increase the fouling rate.

 

[aspearin1]

"Constant Flux" is a bit misleading in this sense. If you operate at constant flow, (which is real easy with a PD pump), and you are blocking membrane pores, you are in actuality reducing your effective filtration area exponentially. That constant flow now has to plug through a smaller window of flow area. Constant pressure (best with a centrifugal pump or compressed gas) will allow for an exponential flow decay equivalent to the rate at which your membrane plugs. In a sense, your constant pressure gives you a constant flux (though I'm not sure it's viewed this way in the industry). UF and RO membranes are high pressure driven systems. I agree with brennans7. TMP sounds most logical.

The industries I serve do operate very much on constant flow principles. The products are, however, disposable, and the membrane can be taxed and put to its functional limit. If you're using so-called clean-in-place systems, try not to push it too hard. Preserve the life and avoid premature replacement.

ChemE, M.E. EIT
"The only constant in life is change." -Bruce Lee

 

[FiltrationEngineer]

This is a continuous process using 5mm tubes. The way the filtration works is propriatary, but the system is operating under low TMP's (around 3 psi)& flux (25 gfd) so I don't feel like I am pushing the system too hard & causing plugging when I allow TMP to increase between backpulses.

The current system design is for constant flux. This makes it easier to sell a X gpm system because the flow rate is constant and prevents some problems on the front end (like larger EQ).

I have the feeling that constant TMP would make the system run at a more steady state. TMP seems to be the main driving force for fouling, and if it were held steady, the fouling trend would be less variable...