Secondary solution (grad u) is not computed correctly

[VikingScientist]

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>                        
<simulation>                                                                   
  <!-- General - geometry definitions !-->                                     
  <geometry dim="3">                                                           
    <patchfile>PipeFlow3D.g2</patchfile>                                       
    <raiseorder patch="1" u="0" v="0" w="1"/>                                  
    <refine patch="1" u="3" v="3" w="7"/>                                      
    <topologysets>                                                             
      <set name="Inflow" type="face">                                          
        <item patch="1">5</item>                                               
      </set>                                                                   
      <set name="Outflow" type="face">                                         
        <item patch="1">6</item>                                               
      </set>                                                                   
    </topologysets>                                                            
  </geometry>                                                                  
                                                                               
  <!-- General - boundary conditions !-->                                      
  <boundaryconditions>                                                         
    <dirichlet set="Inflow"  comp="1" type="constant"> 3 </dirichlet>          
    <dirichlet set="Outflow" comp="1" type="constant"> 0 </dirichlet>          
  </boundaryconditions>                                                        
                                                                               
  <!-- Problem-specific block !-->                                             
  <darcy>                                                                      
    <permvalues>1|1|1</permvalues>                                             
    <bodyforce>0|0|0</bodyforce>                                               
    <source>0</source>                                                         
  </darcy>                                                                     
                                                                               
  <postprocessing>                                                             
    <vtfformat nviz="2">ascii</vtfformat>                                      
  </postprocessing>                                                            
                                                                               
</simulation>                                                                  

Where PipeFlow3D.g2 is a straight cylinder:

700 1 0 0                                                                      
3 1                                                                            
3 3                                                                            
0.000000 0.000000 0.000000 1.000000 1.000000 1.000000                          
3 3                                                                            
0.000000 0.000000 0.000000 1.000000 1.000000 1.000000                          
2 2                                                                            
0.000000 0.000000 1.000000 1.000000                                            
-0.000000 -1.414214 1.414214 1.000000                                          
0.000000 -2.000000 0.000000 0.707107                                           
0.000000 -1.414214 -1.414214 1.000000                                          
-0.000000 0.000000 2.000000 0.707107                                           
0.000000 0.000000 0.000000 1.000000                                            
0.000000 0.000000 -2.000000 0.707107                                           
-0.000000 1.414214 1.414214 1.000000                                           
0.000000 2.000000 0.000000 0.707107                                            
0.000000 1.414214 -1.414214 1.000000                                           
18.000000 -1.414214 1.414214 1.000000                                          
12.727922 -2.000000 0.000000 0.707107                                          
18.000000 -1.414214 -1.414214 1.000000                                         
12.727922 0.000000 2.000000 0.707107                                           
18.000000 0.000000 0.000000 1.000000                                           
12.727922 0.000000 -2.000000 0.707107                                          
18.000000 1.414214 1.414214 1.000000                                           
12.727922 2.000000 0.000000 0.707107                                           
18.000000 1.414214 -1.414214 1.000000                                          

The pressure should form a linear drop from left to right with a gradient that is constant [a,0,0] for some value a. Instead, this is what I get:

[VikingScientist]

Running this with the flag "-LR" gives the following output

Projecting secondary solution ...
	Greville point projection
 *** ASMu2D::evalSolution: LR B-splines require the same number of evaluation points in u-, v- and w-direction.
 *** SIMbase::project: Failure when projecting patch 1.

It looks like it tries to enter 2D projection methods for this 3D example. Is similar things the reason for the errors in tensor product?

[akva2]

"somebody" was eager on the copy-paste when doing ASMu3D is all.

[VikingScientist]

Hehe, would you look at that. It appears that you are right 😸 Then back to the drawing board. I have no idea what causes this.

[akva2]

debugged a bit further. B-splines = all good. what happens is that gotools give us broken basis function derivatives whenever the evaluation point hit the node of a spline for rationals - which explain why the nastyness is near the boundary. haven't digged further yet, i hope to find a 2D case to ease debugging.