Drawdown in Multilayer well

I'm calibrating to few pumping tests using a multilayer well and realized that the drawdown in the pumping well seems always under-estimated.
So I created a simple one layer model with a well in the middle. I made sure my elements fit the "ideal element size" by moving the nodes and discretizing as needed. I compared the results to the theis solution and FEFLOW is under-estimating the drawdown in the pumping well by as much as 20 m!
Am I missing something?

Thanks.

Hi,

FEFLOW has been tested with several benchmark exercises. One of them is the Theis' solution. You may take a look on the section 9.8.1 "[i][b]Transient Flow to a Well in a Confined and Unconfined Aquifer[/b][/i]" in the FEFLOW Book (pages 429-435).

Regards,

Carlos

Thanks for your quick reply,
I'm not questioning the accuracy or correctness of FEFLOW. I'm rather checking if I missed something. I know the mesh around the well is a factor. From a previous response on this Forum too fine or too coarse gives different drawdown results at the well. But using the "ideal element size" should yield realistic results. This is what I was trying to establish. Does the benchmark you reference address the ideal element size in a multilayer well? Or is it the axisymmetric example from the white papers? Also does the book come with a FEFLOW license or does it need to be purchased separately?

Thanks 

have you considered well losses?How is your match for the obs well data?

Goktug

Hi Goktug,
Thanks for your suggestion. I'm looking into that. But as I mentioned I created a simple theoretical model using the multilayer well and compared it to the Theis solution, so no well loss in this case.

Have you made sure that all the assumptions for Theis' solution to be valid are met, e.g., confined conditions, infinite lateral extent, fully penetrating well etc.?

Regards,
Julia