Rate Budget - Imbalance problem (transient model)

Hello,

I’m trying to run a transient model using the Phreatic method in order to calibrate it (It’s an open pit mine dewatering), but I have noticed very large flow imbalance erros in each time step (whether Varying or Automatic time-step control). The particularity of this model is that I represented the elements out of my area of interest as inactive elements, as you can see in the Picture.

I tryed to delete the inactive elements (because I think the Imbalance errors are related to this inactive elements), and I ran it at the same way. It really has improved the imbalance, but the problem is that it is not possible to assign the Sy value (Drain-/fillable porosity – “Specific Yield”) when I use the model with this deleted elements, and I need this parameter in order to achieve better results.

Do you have any tip? Is there any other way to improve the imbalance?

Enclosed you have an image showing the imbalance results using the model with the inactive elements and the model without them (deleted elements).

Thank you in advance!!

Your picture indicates on oscillating imbalances.

I recommend to use the automatic predictor-corrector time-stepping scheme. Apart from that, please check the following options:
[list]
[li]In most of the cases, oscillations are triggered by a mesh which is not fine enough to account for strong gradients and/or sharp contrats in material properties. Of course, bad shaped mesh elements may also cause oscillations.[/li]
[li]Decreasing the error tolerance and/or switching from the default euclidian L2 integeral (RMS) norm to the maximum error norm may decrease the imbalance [/li]
[li]You may try to use the fully implicit FE/BE predictor corrector scheme[/li]
[li]You may try to use the PARDISO equation solver[/li]

[/list]