Errors: Integrator (VODE)

I am facing two major problems when I run the simulation on VODE (steady-state):

1. It takes tremendously long time to simulate since the steps are extremely little (1E-7 ish).
2. The whole simulation process crashes by the error "Numerical error: CVODE: Integration error: -6" 

Any ideas why it is so and how I can fix them?

Most likely you have a discontinuity where the integrator gets stuck while trying to reduce the step size - and cannot converge.
How to fix this and/or what is the cause is impossible to say without:
[list]
[li]taking a look at the actual log book; and/or[/li]
[li]analysing the layout and models you have used[/li]
[/list]
I suggest you take a look yourself to the logging (F6): you may find more details about the error than what you get from the error message that is popped up.
Do you have timers and/or on-off controllers in your layout? These introduce instantaneous state changes that may result in the instability ...
If you cannot get through, please send an email to mikebydhi@dhigroup.com - attach your project and input file(s).

Dear Enrico,

My investigations show that the biofilm processes cause the problem. The model works fine if I remove both the Trickling Filter and the MBBRs. I have sent the file and the input files to the email address you mentioned above.

I will be very greatful if you could have a look to the model and help me with the problems.

My regards,
/Salar

Yes, the biofilm model is [b]very "delicate"[/b]: it is essential that one starts from [b]good initial conditions[/b] and uses an [b]appropriate integrator[/b]. And even then, it may take a couple of short runs and re-initialisations of the model, to actually get it to run smoothly.

1- I changed the integrator to "Alg" and I do not get the previous errors.
2- Surprisingly all return flows are "0.00" after the loop-breakers and the hydraulic mass balance of the model does not match. (Return flow are correct at in the flow splitters but then turn to zero after loop-breakers, i.e. return flows are zero in the following flow combiners.
3- Moreover, all the particulate COD (X_??) disappear in the beginning of the process.
4- I am feeling not so good since I am the only one asking questions!
Thanks for taking your time,
/Salar

1) No, I would not use [b]Alg [/b] - I am actually surprised it does not issue an error message. Rather [b]RK4[/b], which will be extremely slow but may be good for the initialization (and only for a limited time horizon); then [b]VODE[/b].
2) This does not make any sense - unless it's a consequence of choosing Alg as integrator.
3) What do you mean by "all particulate COD"? Only the particulate fractions that contribute to COD? or all the particulate fractions, as opposed to the soluble ones?
And do you mean: at the beginning of the simulation - which may be, again, due to the integrator? or somewhere at an early stage in your layout?
4) Nothing wrong in sharing issues and thoughts: may be useful for others too

[i]Anyway, I will take a look at your project and get back to you[/i]

Dear Enrico,

I am trying to model steady-state conditions via dynamic simulation i.e. by introducing same flow characteristic everyday during 365 days. I expected to get a straight steady-state condition in the operation so that I could then introduce instantaneous vaiations in the inflow at certain time points.

The problem is that I do not get steady-state results despite I have introduced exactly the same inflow for all the simulation days. Results are still very dynamic with strong fluctuations. 

Do you have any idea how I could solve this problem?

Regards

Very difficult to say ...
I would first make sure that the experiment is actually using the correct input file - not sure how you manipulated the influent file, it may be tricky.
Also: are there other models or data inputs that may interfere and cause the fluctuations you observe?
What kind of model or system are you using?

[i]If these very generic hints don't help, please send your project to[/i] [b]mikebydhi@dhigroup.com[/b]