MBR block

Hi,

I would be really pleased if you could help me a little with a MBR system that I am not able to make work properly.

Basically the main problem is with the flow balance. The value of the “influent  flow rate” is equal  to the value of “Q_Out_membrane” as a result of the estimation of the value of the flux as “Q_in/Surface membrane area” and considering the relaxation time.

Q_In= 113,00 m3/d
Q_Out= 112,97 m3/d
Q_Out_Membrane= 112,97 m3/d
Q_Concentrate = 100,00 m3/d
J1= 0,4035 m/d
Q waste= 10 m3/d
Q recirculation = 90 m3/d
Membrane Surface area: 280 m2
Volume of the tank: 50 m3
Volume minimum: 50 m3
T1 (relaxation period)=0,11666667 min
T2 (filtration period)=7,4999995 min

When simulating, the initial volume of the MBR decreases, and when it arrives to 50 m3 (the desired one), the simulation shows strange values and stop (please, see the attached image). With this information, could you have any idea of where is my mistake?

I have tried to synthetize the main information in order you can find a solution, but sometimes is more difficult to explain it than showing it in the project.

Thanks a lot for your time.

Hello Hector,
indeed, it may not be easy to pin-point specific issues that may have got to do with parameter settings.
However, in your current set up, I believe the main problem is with the volume ('min' and 'tank'): it is not advisable that they have the same value as it would make a test like IF ((V > Vmin) & (V < Vtank)) THEN fail. And, as a consequence, the solver to become unstable.

Rather use the Vmin as some lower bound (not zero, but one order of magnitude smaller that the maximum should do); and Vtank as the actual, desired tank volume.
See how the simulation outputs improve.

Also, carefully set the integrator and integrator settings, as you have phases in the MBR model.

I cannot judge from the snapshot, but I see a negative quantity (flow rate or volume): what is it? why does it happen?

Hi again Enrico, and first of all thanks for your explanations.

Asking your last questions,  all represented parameters are in the left vertical axe, except the volume, which is in the right vertical axe. The negative value appears when the volume decreases until the value of 50 m3 (in that simulation, corresponds at the same time with the “V_tank” and “V_min”).

Unfortunately, following your instructions, after decreasing the value of the “V_min” to 40 or even 10 m3, the situation that I find is more or less the same, despite I have proved with different integrators models and diminished the “tolerance” in order to go slower but safer.

The only integrator model that does not show negative values in the plot of flows and volume is “VODE”, but in return, the quality of the effluent fluctuates without sense (see the attached images, with V_min=10 m3 and VODE integrator)…something happens with the integrator I am afraid.. Any more advice, please?

Thanks a lot for your dedication.

I already see a significant improvement: there's no negative volume or flow rate.
Anyway, I guess the only way to gain a better view on your set up is that you share 1) your [b]WEST layout[/b]; and 2) the settings for the following [b]parameters[/b]:
[list]
[li]average QIn to your system and desired QWaste[/li]
[li]filtration and backwashing periods (relaxation are less relevant, can be zero)[/li]
[li]filtration and backwashing fluxes[/li]
[/list]

Alternatively - it may be faster, you can send your project to 'mike' then we can post the working settings here.

Hello Hector,
ok I see: you used the [b]MBRIdeal[/b] model (I thought we were dealing with fouling and backwashing too): so your layout is correct (cf image "MBR_Layout"). Interesting application by the way: [b]retrofitting an existing facility[/b] with the addition of an MBR unit.

I see only one issue: you're starting at an incorrect initial volume (at time 0), 1000 m3. This is the library default value which, I will agree with you, is kind of hidden ...
If you reset it to 50 (or 40) m3, the initial transient is reduced to a bare minimum and the simulation runs smoothly.
To do so - and this is the 'hidden' part, you need to [b][u]reset the initial mass of water[/u][/b] (M[H2O] which is a derived state variable of the MBR).

Then it's all correct - except that you cannot use an adaptive step size integrator (like RK4ASC) nor VODE, essentially because you have a couple of timer blocks in your layout and the MBR has also got phases.
These are a couple of runs I've done:
[list]
[li]RK4, step size 1E-5; communication interval for plots 0.01 -> 5 days in SS take 88"[/li]
[li]RK4, step size 1E-4; communication interval for plots 0.01 -> 5 days in SS take 17"[/li]
[li]RK4, step size 1E-4; communication interval for plots 0.005 -> 5 days in SS take 21"[/li]
[li][u][b]RK4[/b][/u], [u][b]step size 1E-4[/b][/u]; [u][b]communication interval for plots 0.005[/b][/u] -> 100 days in SS take 2'14"[/li]
[/list]
Steady-state is reached after approximately 30 days (cf image "MBR_Simul").

Hi Enrico,

You are absolutely right and now it runs perfectly. Indeed I tried, before your last advice, changing the “initial mass of water”, but I pressed the button of “play” in Dinamic State, and not in Steady State.

Thanks a lot for your useful explanations.