# Bad orientation

**URL:** <https://community.freefem.org/t/bad-orientation/411>\
**Category:** General Discussion\
**Created:** [May 14, 2020, 10:12am UTC](https://community.freefem.org/t/bad-orientation/411 "2020-05-14T10:12:01Z")\
**Posts on this page:** 16\
**Page:** 1

<div class="post-metadata">

**Author:** ![JAlonso](https://avatars.discourse-cdn.com/v4/letter/j/b5e925/32.png) [@JAlonso](https://community.freefem.org/u/JAlonso)\
**Post date:** [May 14, 2020, 10:12am UTC](https://community.freefem.org/t/bad-orientation/411/1 "2020-05-14T10:12:01Z")

</div>

Dear colleages,

I am trying to solve a 3D problem on a thick slice with a pentagonal hole.

I have used gmsh to build up the mesh with the following code

cl1=0.5;  
Mesh.Format=30;

Point(1)={-1,0,0,cl1};  
Point(2)={1,0,0,cl1};  
Point(3)={-1.8,-1,0,cl1};  
Point(4)={1.8,-1,0,cl1};  
Point(5)={0,-2,0,cl1};

Point(6)={-1,0,-5,cl1};  
Point(7)={1,0,-5,cl1};  
Point(8)={-1.8,-1,-5,cl1};  
Point(9)={1.8,-1,-5,cl1};  
Point(10)={0,-2,-5,cl1};

Point(11)={-5,3,0,cl1};  
Point(12)={5,3,0,cl1};  
Point(13)={-5,-15,0,cl1};  
Point(14)={5,-15,0,cl1};

Point(15)={-5,3,-5,cl1};  
Point(16)={5,3,-5,cl1};  
Point(17)={-5,-15,-5,cl1};  
Point(18)={5,-15,-5,cl1};

Line(1) = {3, 1};  
Line(2) = {1, 2};  
Line(3) = {2, 4};  
Line(4) = {4, 5};  
Line(5) = {5, 3};  
Line(6) = {10, 9};  
Line(7) = {9, 7};  
Line(8) = {7, 6};  
Line(9) = {6, 8};  
Line(10) = {8, 10};  
Line(11) = {10, 5};  
Line(12) = {9, 4};  
Line(13) = {7, 2};  
Line(14) = {6, 1};  
Line(15) = {8, 3};  
Line(16) = {11, 12};  
Line(17) = {12, 14};  
Line(18) = {14, 13};  
Line(19) = {13, 11};  
Line(20) = {13, 17};  
Line(21) = {17, 18};  
Line(22) = {18, 16};  
Line(23) = {16, 15};  
Line(24) = {15, 17};  
Line(25) = {18, 14};  
Line(26) = {12, 16};  
Line(27) = {15, 11};  
Line Loop(1) = {5, -15, 10, 11};  
Plane Surface(1) = {1};  
Line Loop(2) = {11, -4, -12, -6};  
Plane Surface(2) = {2};  
Line Loop(3) = {12, -3, -13, -7};  
Plane Surface(3) = {3};  
Line Loop(4) = {13, -2, -14, -8};  
Plane Surface(4) = {4};  
Line Loop(5) = {14, -1, -15, -9};  
Plane Surface(5) = {5};  
Line Loop(10) = {23, 27, 16, 26};  
Plane Surface(10) = {10};  
Line Loop(11) = {22, -26, 17, -25};  
Plane Surface(11) = {11};  
Line Loop(12) = {21, 25, 18, 20};  
Plane Surface(12) = {12};  
Line Loop(13) = {27, -19, 20, -24};  
Plane Surface(13) = {13};  
Line Loop(14) = {16, 17, 18, 19};  
Line Loop(15) = {2, 3, 4, 5, 1};  
Plane Surface(14) = {14, 15};  
Line Loop(16) = {22, 23, 24, 21};  
Line Loop(17) = {6, 7, 8, 9, 10};  
Plane Surface(15) = {16, 17};  
Surface Loop(1) = {14, 10, 15, 11, 12, 13, 1, 5, 4, 3, 2};  
Volume(1) = {1};

all look fine but when read the mesh with freefem

load “msh3”;  
mesh3 Th(“pentagono.mesh”);

the error is "read mesh ok 0Mesh3, num Tetra:= 41384, num Vertice:= 8636 num boundary Triangles:= 7536 Mesh3::meshS, num Triangles:= 7536, num Vertice:= 3768 num boundary Edges:= 354  
read mesh ok 0Mesh3, num Tetra:= 41384, num Vertice:= 8636 num boundary Triangles:= 7536  
Mesh3::meshS, num Triangles:= 7536, num Vertice:= 3768 num boundary Edges:= 354  
Bad orientation: The adj border element defined by [16 246] is oriented in the same direction in element 3922 and in the element 446 \*\*\*\*\*\* bug in mesh construction? orientation parameter?  
"

I do not know if freefem has the capability of handling with this.

Does someone deal with a problem like this?

Thanks in advance

Jose

---

<div class="post-metadata">

**Author:** ![JAlonso](https://avatars.discourse-cdn.com/v4/letter/j/b5e925/32.png) [@JAlonso](https://community.freefem.org/u/JAlonso)\
**Post date:** [May 17, 2020, 6:36am UTC](https://community.freefem.org/t/bad-orientation/411/2 "2020-05-17T06:36:10Z")

</div>

Dear colleages,

I am still working on generating with GMSH and reading the mesh with FreeFem for a slice with a pentagonal hole:  
cl1=0.5;  
Mesh.Format=1;  
// Select older mesh version, for compatibility of FF++  
Mesh.MshFileVersion = 2.2;

Point(1)={-1,0,0,cl1};  
Point(2)={1,0,0,cl1};  
Point(3)={-1.8,-1,0,cl1};  
Point(4)={1.8,-1,0,cl1};  
Point(5)={0,-2,0,cl1};

Point(6)={-1,0,-5,cl1};  
Point(7)={1,0,-5,cl1};  
Point(8)={-1.8,-1,-5,cl1};  
Point(9)={1.8,-1,-5,cl1};  
Point(10)={0,-2,-5,cl1};

Point(11)={-5,3,0,cl1};  
Point(12)={5,3,0,cl1};  
Point(13)={-5,-15,0,cl1};  
Point(14)={5,-15,0,cl1};

Point(15)={-5,3,-5,cl1};  
Point(16)={5,3,-5,cl1};  
Point(17)={-5,-15,-5,cl1};  
Point(18)={5,-15,-5,cl1};

Line(1) = {3, 1};  
Line(2) = {1, 2};  
Line(3) = {2, 4};  
Line(4) = {4, 5};  
Line(5) = {5, 3};  
Line(6) = {10, 9};  
Line(7) = {9, 7};  
Line(8) = {7, 6};  
Line(9) = {6, 8};  
Line(10) = {8, 10};  
Line(11) = {10, 5};  
Line(12) = {9, 4};  
Line(13) = {7, 2};  
Line(14) = {6, 1};  
Line(15) = {8, 3};  
Line(16) = {11, 12};  
Line(17) = {12, 14};  
Line(18) = {14, 13};  
Line(19) = {13, 11};  
Line(20) = {13, 17};  
Line(21) = {17, 18};  
Line(22) = {18, 16};  
Line(23) = {16, 15};  
Line(24) = {15, 17};  
Line(25) = {18, 14};  
Line(26) = {12, 16};  
Line(27) = {15, 11};  
Line Loop(1) = {5, -15, 10, 11};  
Plane Surface(1) = {1};  
Line Loop(2) = {11, -4, -12, -6};  
Plane Surface(2) = {2};  
Line Loop(3) = {12, -3, -13, -7};  
Plane Surface(3) = {3};  
Line Loop(4) = {13, -2, -14, -8};  
Plane Surface(4) = {4};  
Line Loop(5) = {14, -1, -15, -9};  
Plane Surface(5) = {5};  
Line Loop(10) = {23, 27, 16, 26};  
Plane Surface(10) = {10};  
Line Loop(11) = {22, -26, 17, -25};  
Plane Surface(11) = {11};  
Line Loop(12) = {21, 25, 18, 20};  
Plane Surface(12) = {12};  
Line Loop(13) = {27, -19, 20, -24};  
Plane Surface(13) = {13};  
Line Loop(14) = {16, 17, 18, 19};  
Line Loop(15) = {2, 3, 4, 5, 1};  
Plane Surface(14) = {14, 15};  
Line Loop(16) = {22, 23, 24, 21};  
Line Loop(17) = {6, 7, 8, 9, 10};  
Plane Surface(15) = {16, 17};  
Surface Loop(1) = {14, 10, 15, 11, 12, 13, 1, 5, 4, 3, 2};  
Volume(1) = {1};

when I tried to read the resulting mesh with

// Level of output  
verbosity = 10;  
//  
// Loading of libraries  
//  
load “msh3”  
load “gmsh”;  
load “iovtk”;  
//load “medit”;  
//  
// Loading the mesh. It is hardly needed specifying the procedence with gmshload3  
//  
mesh3 Th3I=gmshload3(“pentagonV2.msh”);

the error message is

5665 vertices  
closing file 24009 6078  
24009 tetrahedrons  
6078 triangles  
30459 numElements  
– GMesh3 , n V: 5665 , n Elm: 24009 , n B Elm: 6078mes 877 674.198 , bb: (-5 -15 -5) , (5 3 0)  
– BuildAdj:nva= 3 4 6078  
– warning true boundary element 102 is no in correct orientation  
– warning true boundary element 103 is no in correct orientation  
– warning true boundary element 104 is no in correct orientation  
– warning true boundary element 105 is no in correct orientation  
– warning true boundary element 106 is no in correct orientation  
– warning true boundary element 107 is no in correct orientation  
– warning true boundary element 108 is no in correct orientation  
– warning true boundary element 109 is no in correct orientation  
– warning true boundary element 110 is no in correct orientation  
Warning change orientation of 1938 faces  
Warning error in boundary oriention 1938 faces

These are my last efforts after consulting to the GMSH community, but I must confess I am quite lost.

If you have any suggestion, or suspect this could be a bug, any idea will be welcomed

Thank you in advance,

Jose

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [May 17, 2020, 7:27am UTC](https://community.freefem.org/t/bad-orientation/411/3 "2020-05-17T07:27:32Z")

</div>

Here is how you can have FreeFEM reorient the mesh for you.

```auto
load "msh3";
lockOrientation = false;
mesh3 Th("pentagono.mesh");
lockOrientation = true;
mesh3 renumbered = Th;
fespace Vh(renumbered, P1);
Vh u = x + y + z;
load "medit"
medit("Th", Th, u);

```

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 11, 2020, 7:49pm UTC](https://community.freefem.org/t/bad-orientation/411/4 "2020-09-11T19:49:44Z")

</div>

Hi

as I got the following error message:

Fatal Error in mesh : RFQ.mesh number badly oriented element 532

I tried your proposal for getting Freefem reorienting the mesh. Unfortunaltely, the code does not fix it. Any idea?

thank you for your valuable help,

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [September 11, 2020, 7:54pm UTC](https://community.freefem.org/t/bad-orientation/411/5 "2020-09-11T19:54:51Z")

</div>

Without anything to reproduce the problem: no.

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 11, 2020, 10:22pm UTC](https://community.freefem.org/t/bad-orientation/411/6 "2020-09-11T22:22:41Z")

</div>

OK. I thought that maybe the error message would be enough to ring a bell.

Here is the case. I am trying to build a 3D mesh with gmsh made of a volume with multiple plane surfaces as a start in order to solve Poisson equation with Freefem. I have by the way no problem to do that with the kernel “built-in” in Gmsh and exporting a INRIA .mesh. Also I need at some point to use “OpenCascade” mesher in Gmsh as the geometry will get more complex.

Here is the Gmsh code:

SetFactory(“OpenCASCADE”);  
//---------------------------  
// Parameters of the geometry  
//---------------------------

Mesh.CharacteristicLengthMin = .5;  
Mesh.CharacteristicLengthMax = 10;

DIM = 3;  
L = 3.;  
R0 = 1.;  
rho = 0.85;  
res = 0.1;  
Mesh.Optimize = 1;  
p1[] += newp ; Point(newp) = { 0, 0, 0, res} ;  
p1[] += newp ; Point(newp) = { R0, 0, 0, res} ;  
p1[] += newp ; Point(newp) = { 0, R0, 0, res} ;  
p1[] += newp ; Point(newp) = { rho, 3_R0, 0, res} ;  
p1[] += newp ; Point(newp) = { 3_R0, 3_R0, 0, res} ;  
p1[] += newp ; Point(newp) = { 3_R0, rho, 0, res} ;  
p1[] += newp ; Point(newp) = { rho, R0+rho, 0, res} ;  
p1[] += newp ; Point(newp) = { R0+rho, rho, 0, res} ;  
p1[] += newp ; Point(newp) = { R0+rho, 0, 0, res} ;  
p1[] += newp ; Point(newp) = { 0., R0+rho, 0, res} ;

p2[] += newp ; Point(newp) = { 0, 0, L, res} ;  
p2[] += newp ; Point(newp) = { R0, 0, L, res} ;  
p2[] += newp ; Point(newp) = { 0, R0, L, res} ;  
p2[] += newp ; Point(newp) = { rho, 3_R0, L, res} ;  
p2[] += newp ; Point(newp) = { 3_R0, 3_R0, L, res} ;  
p2[] += newp ; Point(newp) = { 3_R0, rho, L, res} ;  
p2[] += newp ; Point(newp) = { rho, R0+rho, L, res} ;  
p2[] += newp ; Point(newp) = { R0+rho, rho, L, res} ;  
p2[] += newp ; Point(newp) = { R0+rho, 0, L, res} ;  
p2[] += newp ; Point(newp) = { 0., R0+rho, L, res} ;

//+  
Line(1) = {11, 12};  
//+  
Line(2) = {11, 13};  
//+  
Line(3) = {14, 15};  
//+  
Line(4) = {15, 16};  
//+  
Line(5) = {16, 18};  
//+  
Line(6) = {14, 17};  
//+  
Line(7) = {5, 15};  
//+  
Line(8) = {4, 5};  
//+  
Line(9) = {4, 14};  
//+  
Line(10) = {4, 7};  
//+  
Line(11) = {7, 17};  
//+  
Line(12) = {2, 1};  
//+  
Line(13) = {3, 1};  
//+  
Line(14) = {8, 6};  
//+  
Line(15) = {6, 16};  
//+  
Line(16) = {6, 5};  
//+  
Line(17) = {1, 11};  
//+  
Line(18) = {13, 3};  
//+  
Line(19) = {3, 7};  
//+  
Line(20) = {17, 13};  
//+  
Line(21) = {2, 8};  
//+  
Line(22) = {18, 12};  
//+  
Line(23) = {18, 8};  
//+  
Line(24) = {12, 2};  
//+

Line(25) = {11, 1};  
//+  
Line(26) = {1, 2};  
//+  
Line(27) = {2, 12};  
//+  
Line(28) = {12, 11};  
//+  
Line(29) = {7, 3};  
//+  
Line(30) = {17, 7};  
//+  
Line(31) = {13, 17};  
//+  
Line(32) = {3, 13};  
//+  
Line(33) = {8, 18};  
//+  
Line(34) = {16, 6};  
//+  
Line(35) = {15, 5};  
//+  
Line(36) = {5, 4};  
//+  
Line(37) = {12, 11};  
//+  
Line(38) = {17, 14};  
//+  
Line(39) = {13, 17};  
//+  
Line(40) = {13, 11};  
//+  
Line(41) = {15, 14};  
//+  
Line(42) = {16, 15};  
//+  
Line(43) = {18, 16};  
//+  
Line(44) = {12, 18};  
//+  
Line(45) = {5, 6};  
//+  
Line(46) = {7, 4};  
//+  
Line(47) = {1, 3};  
//+  
Line(48) = {8, 2};  
//+  
Line(49) = {6 ,8};  
//+  
Line(50) = {14 ,4};

//+  
Curve Loop(1) = {13, 17, 2, 18};  
//+  
Plane Surface(1) = {1};  
//+  
Curve Loop(2) = {25, 26, 27, 28};  
//+  
Plane Surface(2) = {2};  
//+  
Curve Loop(3) = {29, 30, 31, 32};  
//+  
Plane Surface(3) = {3};  
//+  
Curve Loop(40) = {21, 33, 22, 24};  
//+  
Plane Surface(4) = {40};  
//+  
Curve Loop(5) = {5, 15, 14, 23};  
//+  
Plane Surface(5) = {5};  
//+  
Curve Loop(6) = {34, 4, 7, 16};  
//+  
Plane Surface(6) = {6};  
//+  
Curve Loop(7) = {35, 3, 9, 36};  
//+  
Plane Surface(7) = {7};  
//+  
Curve Loop(80) = {1, 40, 20, 6, 41, 42, 43, 44};  
//+  
Plane Surface(8) = {80};  
//+  
Curve Loop(9) = {45,8,46,19,47,12,48,49};  
//+  
Plane Surface(9) = {9};  
//+  
Curve Loop(10) = {38, 50, 10, 11};  
//+  
Plane Surface(10) = {10};

//+  
Surface Loop(1) = {1, 6, 7, 2, 10, 8, 3, 4, 9, 5} Using Sewing;  
//+  
Volume(1) = {1};

//+  
Show “\*”;

for building the geometry to be meshed with Gmsh. No error message is sent by Gmsh and I can export a file RFQ.mesh (INRIA) mesh as I did with the kernel “built-in”. The FreeFem code is at the moment just

load “msh3”  
load “medit”  
mesh3 Th(“RFQ.mesh”);

and I got first

read mesh ok 0Mesh3, num Tetra:= 1061, num Vertice:= 694 num boundary Triangles:= 1146  
Mesh3::meshS, num Triangles:= 1146, num Vertice:= 637 num boundary Edges:= 305  
Err Border element not in mesh  
0 : 69 70 251  
1 : 73 74 252  
2 : 70 251 254  
3 : 70 71 254  
4 : 76 77 253  
5 : 69 251 253  
6 : 74 252 254  
7 : 76 251 253  
8 : 71 252 254

I understood this was due to the presence of 4 points anticipated to be the centers for constructing circles later. These points are not in the mesh. I deleted these 4 vertices in RFQ.mesh directly (pt 9, 10, 19 and 20).

This solved the issue and I get now

the tet 0 is badly orienated tet -0.696641  
the tet 2 is badly orienated tet -1.41456  
the tet 4 is badly orienated tet -0.106155  
the tet 6 is badly orienated tet -0.642099  
the tet 8 is badly orienated tet -0.197834  
the tet 11 is badly orienated tet -0.0435579  
the tet 13 is badly orienated tet -0.0429028  
the tet 15 is badly orienated tet -0.714966  
the tet 16 is badly orienated tet -0.197144  
the tet 20 is badly orienated tet -0.0885662  
Fatal Error in mesh : RFQ.mesh number badly oriented element 661  
nt 1292 / nbe 1146 / mes 2.72958 mesb 951.542  
current line = 8  
Assertion fail : (0)  
line :988, in file …/femlib/Mesh3dn.cpp  
Assertion fail : (0)  
line :988, in file …/femlib/Mesh3dn.cpp  
err code 6 , mpirank 0

Also, I did check the orientation of the surfaces with Gmsh. It sounds good as all surface normals point out:

 ![RFQ_TEST-Line_OC](https://canada1.discourse-cdn.com/flex030/uploads/freefem/original/1X/c58a3fb0c418afd67a07606a67e9f07b3426a40f.png)

The same orientation went well when I used the kernel “built-in” in Gmsh.

Basically, you have all the story.

Thx

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [September 12, 2020, 6:11am UTC](https://community.freefem.org/t/bad-orientation/411/7 "2020-09-12T06:11:40Z")

</div>

Surround your code between three single brackets

```auto
like this

```

otherwise we can’t copy/paste it easily, please.

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 3:56pm UTC](https://community.freefem.org/t/bad-orientation/411/8 "2020-09-12T15:56:15Z")

</div>

Thx for the tip. Here are the 2 codes.

Gmsh code:

```
SetFactory(“OpenCASCADE”);
//---------------------------
// Parameters of the geometry
//---------------------------

Mesh.CharacteristicLengthMin = .5;
Mesh.CharacteristicLengthMax = 10;

DIM = 3;
L = 3.;
R0 = 1.;
rho = 0.85;
res = 0.1;
Mesh.Optimize = 1;
p1[] += newp ; Point(newp) = { 0, 0, 0, res} ;
p1[] += newp ; Point(newp) = { R0, 0, 0, res} ;
p1[] += newp ; Point(newp) = { 0, R0, 0, res} ;
p1[] += newp ; Point(newp) = { rho, 3 <em>R0, 0, res} ;
p1[] += newp ; Point(newp) = { 3</em> R0, 3 <em>R0, 0, res} ;
p1[] += newp ; Point(newp) = { 3</em> R0, rho, 0, res} ;
p1[] += newp ; Point(newp) = { rho, R0+rho, 0, res} ;
p1[] += newp ; Point(newp) = { R0+rho, rho, 0, res} ;
p1[] += newp ; Point(newp) = { R0+rho, 0, 0, res} ;
p1[] += newp ; Point(newp) = { 0., R0+rho, 0, res} ;

p2[] += newp ; Point(newp) = { 0, 0, L, res} ;
p2[] += newp ; Point(newp) = { R0, 0, L, res} ;
p2[] += newp ; Point(newp) = { 0, R0, L, res} ;
p2[] += newp ; Point(newp) = { rho, 3 <em>R0, L, res} ;
p2[] += newp ; Point(newp) = { 3</em> R0, 3 <em>R0, L, res} ;
p2[] += newp ; Point(newp) = { 3</em> R0, rho, L, res} ;
p2[] += newp ; Point(newp) = { rho, R0+rho, L, res} ;
p2[] += newp ; Point(newp) = { R0+rho, rho, L, res} ;
p2[] += newp ; Point(newp) = { R0+rho, 0, L, res} ;
p2[] += newp ; Point(newp) = { 0., R0+rho, L, res} ;

//+
Line(1) = {11, 12};
//+
Line(2) = {11, 13};
//+
Line(3) = {14, 15};
//+
Line(4) = {15, 16};
//+
Line(5) = {16, 18};
//+
Line(6) = {14, 17};
//+
Line(7) = {5, 15};
//+
Line(8) = {4, 5};
//+
Line(9) = {4, 14};
//+
Line(10) = {4, 7};
//+
Line(11) = {7, 17};
//+
Line(12) = {2, 1};
//+
Line(13) = {3, 1};
//+
Line(14) = {8, 6};
//+
Line(15) = {6, 16};
//+
Line(16) = {6, 5};
//+
Line(17) = {1, 11};
//+
Line(18) = {13, 3};
//+
Line(19) = {3, 7};
//+
Line(20) = {17, 13};
//+
Line(21) = {2, 8};
//+
Line(22) = {18, 12};
//+
Line(23) = {18, 8};
//+
Line(24) = {12, 2};
//+

Line(25) = {11, 1};
//+
Line(26) = {1, 2};
//+
Line(27) = {2, 12};
//+
Line(28) = {12, 11};
//+
Line(29) = {7, 3};
//+
Line(30) = {17, 7};
//+
Line(31) = {13, 17};
//+
Line(32) = {3, 13};
//+
Line(33) = {8, 18};
//+
Line(34) = {16, 6};
//+
Line(35) = {15, 5};
//+
Line(36) = {5, 4};
//+
Line(37) = {12, 11};
//+
Line(38) = {17, 14};
//+
Line(39) = {13, 17};
//+
Line(40) = {13, 11};
//+
Line(41) = {15, 14};
//+
Line(42) = {16, 15};
//+
Line(43) = {18, 16};
//+
Line(44) = {12, 18};
//+
Line(45) = {5, 6};
//+
Line(46) = {7, 4};
//+
Line(47) = {1, 3};
//+
Line(48) = {8, 2};
//+
Line(49) = {6 ,8};
//+
Line(50) = {14 ,4};

//+
Curve Loop(1) = {13, 17, 2, 18};
//+
Plane Surface(1) = {1};
//+
Curve Loop(2) = {25, 26, 27, 28};
//+
Plane Surface(2) = {2};
//+
Curve Loop(3) = {29, 30, 31, 32};
//+
Plane Surface(3) = {3};
//+
Curve Loop(40) = {21, 33, 22, 24};
//+
Plane Surface(4) = {40};
//+
Curve Loop(5) = {5, 15, 14, 23};
//+
Plane Surface(5) = {5};
//+
Curve Loop(6) = {34, 4, 7, 16};
//+
Plane Surface(6) = {6};
//+
Curve Loop(7) = {35, 3, 9, 36};
//+
Plane Surface(7) = {7};
//+
Curve Loop(80) = {1, 40, 20, 6, 41, 42, 43, 44};
//+
Plane Surface(8) = {80};
//+
Curve Loop(9) = {45,8,46,19,47,12,48,49};
//+
Plane Surface(9) = {9};
//+
Curve Loop(10) = {38, 50, 10, 11};
//+
Plane Surface(10) = {10};

//+
Surface Loop(1) = {1, 6, 7, 2, 10, 8, 3, 4, 9, 5} Using Sewing;
//+
Volume(1) = {1};

//+
Show “*”;

```

and here is the Freefem code for just reading the INRIA mesh:

```
load “msh3”
load “medit”
mesh3 Th(“RFQ.mesh”);

```

I also tried:

```
load “msh3”
load “medit”

lockOrientation = false;
mesh3 Th("RFQ.mesh");
lockOrientation = true;
mesh3 renumbered = Th;
fespace Vh(renumbered, P1);
Vh u = x + y + z;
medit("Th", Th, u);

```

but got the same result.

Thx

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [September 12, 2020, 6:27pm UTC](https://community.freefem.org/t/bad-orientation/411/9 "2020-09-12T18:27:42Z")

</div>

Can you try the code you copy/paste, please? I still can’t parse your `.geo`.  
At least the first line should be:

```auto
SetFactory("OpenCASCADE");

```

and not:

```auto
SetFactory(“OpenCASCADE”);

```

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 6:50pm UTC](https://community.freefem.org/t/bad-orientation/411/10 "2020-09-12T18:50:13Z")

</div>

The bad quoting is coming the forum editor as I did not first use the code insertion properly as you pointed out. In my .geo file it is properly quoted.

I have no apparent problem with .geo file at least with Gmsh that reads it and meshes it in 3D without error message.

Thx

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 6:52pm UTC](https://community.freefem.org/t/bad-orientation/411/11 "2020-09-12T18:52:37Z")

</div>

Just in case I copy/paste again the .geo code:

```
SetFactory("OpenCASCADE");

//---------------------------
// Parameters of the geometry
//---------------------------

// ERROR: The mesh file doesn't contain vertices 
// => forgot to mesh

// Bad orientation: The adj border element defined by [89 91] 
// car hérité tel que la version bien orienté avec BI
// Correction de l'orientation par définition de lignes supplém. OpenCascade ne reconnait pas les -
// pour changer l'orientation. Les lignes supplémentaires sont des inverses de premières pour palier.
// Dans ce cas, les faces sont bien orientées mais le maillage 3D dans GMSH plante => PLC error.
// PLC Error: A segment and a facet intersect at point
// 

Mesh.CharacteristicLengthMin = .5;
Mesh.CharacteristicLengthMax = 10;

DIM = 3; 

L = 3.; 
R0 = 1.; 
rho = 0.85; 
res = 0.1;

Mesh.Optimize = 1;

p1[] += newp ; Point(newp) = { 0, 0, 0, res} ;
p1[] += newp ; Point(newp) = { R0, 0, 0, res} ;
p1[] += newp ; Point(newp) = { 0, R0, 0, res} ;
p1[] += newp ; Point(newp) = { rho, 3*R0, 0, res} ;
p1[] += newp ; Point(newp) = { 3*R0, 3*R0, 0, res} ;
p1[] += newp ; Point(newp) = { 3*R0, rho, 0, res} ;
p1[] += newp ; Point(newp) = { rho, R0+rho, 0, res} ;
p1[] += newp ; Point(newp) = { R0+rho, rho, 0, res} ;
p1[] += newp ; Point(newp) = { R0+rho, 0, 0, res} ;
p1[] += newp ; Point(newp) = { 0., R0+rho, 0, res} ;

p2[] += newp ; Point(newp) = { 0, 0, L, res} ;
p2[] += newp ; Point(newp) = { R0, 0, L, res} ;
p2[] += newp ; Point(newp) = { 0, R0, L, res} ;
p2[] += newp ; Point(newp) = { rho, 3*R0, L, res} ;
p2[] += newp ; Point(newp) = { 3*R0, 3*R0, L, res} ;
p2[] += newp ; Point(newp) = { 3*R0, rho, L, res} ;
p2[] += newp ; Point(newp) = { rho, R0+rho, L, res} ;
p2[] += newp ; Point(newp) = { R0+rho, rho, L, res} ;
p2[] += newp ; Point(newp) = { R0+rho, 0, L, res} ;
p2[] += newp ; Point(newp) = { 0., R0+rho, L, res} ;

//+
Line(1) = {11, 12};
//+
Line(2) = {11, 13};
//+
Line(3) = {14, 15};
//+
Line(4) = {15, 16};
//+
Line(5) = {16, 18};
//+
Line(6) = {14, 17};
//+
Line(7) = {5, 15};
//+
Line(8) = {4, 5};
//+
Line(9) = {4, 14};
//+
Line(10) = {4, 7};
//+
Line(11) = {7, 17};
//+
Line(12) = {2, 1};
//+
Line(13) = {3, 1};
//+
Line(14) = {8, 6};
//+
Line(15) = {6, 16};
//+
Line(16) = {6, 5};
//+
Line(17) = {1, 11};
//+
Line(18) = {13, 3};
//+
Line(19) = {3, 7};
//+
Line(20) = {17, 13};
//+
Line(21) = {2, 8};
//+
Line(22) = {18, 12};
//+
Line(23) = {18, 8};
//+
Line(24) = {12, 2};
//+

Line(25) = {11, 1};
//+
Line(26) = {1, 2};
//+
Line(27) = {2, 12};
//+
Line(28) = {12, 11};
//+
Line(29) = {7, 3};
//+
Line(30) = {17, 7};
//+
Line(31) = {13, 17};
//+
Line(32) = {3, 13};
//+
Line(33) = {8, 18};
//+
Line(34) = {16, 6};
//+
Line(35) = {15, 5};
//+
Line(36) = {5, 4};
//+
Line(37) = {12, 11};
//+
Line(38) = {17, 14};
//+
Line(39) = {13, 17};
//+
Line(40) = {13, 11};
//+
Line(41) = {15, 14};
//+
Line(42) = {16, 15};
//+
Line(43) = {18, 16};
//+
Line(44) = {12, 18};
//+
Line(45) = {5, 6};
//+
Line(46) = {7, 4};
//+
Line(47) = {1, 3};
//+
Line(48) = {8, 2};
//+
Line(49) = {6 ,8};
//+
Line(50) = {14 ,4};

//+
Curve Loop(1) = {13, 17, 2, 18};
//+
Plane Surface(1) = {1};
//+
Curve Loop(2) = {25, 26, 27, 28};
//+
Plane Surface(2) = {2};
//+
Curve Loop(3) = {29, 30, 31, 32};
//+
Plane Surface(3) = {3};
//+
Curve Loop(40) = {21, 33, 22, 24};
//+
Plane Surface(4) = {40};
//+
Curve Loop(5) = {5, 15, 14, 23};
//+
Plane Surface(5) = {5};
//+
Curve Loop(6) = {34, 4, 7, 16};
//+
Plane Surface(6) = {6};
//+
Curve Loop(7) = {35, 3, 9, 36};
//+
Plane Surface(7) = {7};
//+
Curve Loop(80) = {1, 40, 20, 6, 41, 42, 43, 44};
//+
Plane Surface(8) = {80};

//+
Curve Loop(9) = {45,8,46,19,47,12,48,49};
//+
Plane Surface(9) = {9};
//+
Curve Loop(10) = {38, 50, 10, 11};
//+
Plane Surface(10) = {10};

//+
Surface Loop(1) = {1, 6, 7, 2, 10, 8, 3, 4, 9, 5} Using Sewing;
//+
Volume(1) = {1};

//+
Show "*";
```

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [September 12, 2020, 6:52pm UTC](https://community.freefem.org/t/bad-orientation/411/12 "2020-09-12T18:52:55Z")

</div>

~~Copy the code you have in your `.geo`, not the one from the first post above, which is wrong.~~

Thanks, it’s OK now.

---

<div class="post-metadata">

**Author:** ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)\
**Post date:** [September 12, 2020, 7:38pm UTC](https://community.freefem.org/t/bad-orientation/411/13 "2020-09-12T19:38:00Z")

</div>

Naive question, but wouldn’t it be best to fix the `.geo` when _not_ using OpenCASCADE? What the FreeFEM error says is that there are triangles in the 3D mesh which are not part of the boundary, which usually happens when there is an issue during mesh generation.

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 8:04pm UTC](https://community.freefem.org/t/bad-orientation/411/14 "2020-09-12T20:04:14Z")

</div>

When I use the kernel “Built-in” instead of “OpenCASCADE”, I do not meet any problem. Gmsh meshes the geometry and the INRI .mesh that I export is read without problem by FreeFEM. I can solve the Poisson equation and get coherent results then.

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 8:18pm UTC](https://community.freefem.org/t/bad-orientation/411/15 "2020-09-12T20:18:12Z")

</div>

OK. To be more accurate, if you just use the code I sent but use “built-in” kernel instead, Freefem still sees an orientation problem. If I mentioned to not have a problem it was with the older file where a good orientation was obtained with +/- sign in the command “Curve loop” in the .geo file. In this case, Freefem accepted the exported mesh.  
It is when I tried to use OpenCASCADE instead that I noticed the following:

- The obtained orientation with “Built-in” was changed by OpenCASCADE. OpenCASCADE in Gmsh apparently ignores orientation ajustment with +/-,

- the only way I found to get the orientation I want was to create new lines with opposite directions. Then OpenCASCADE was able to see the same orientation. This is the file I sent. But this one is rejected by Freefem with an orientation issue even when I plot the normals, they are similar to the older case the Freefem accepted.

Later, I discovered the OpenCASCADE option “Using sewing” that creates a shell even if geometrical elements do not have the required topology. This helped me obtaining the mesh in Gmsh. What I have not tried yet is too use this option with the older file (without additional lines with opposite direction). I try this asap.

I hope it helps understanding.

Thx

---

<div class="post-metadata">

**Author:** ![Romuald](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Romuald](https://community.freefem.org/u/Romuald)\
**Post date:** [September 12, 2020, 8:25pm UTC](https://community.freefem.org/t/bad-orientation/411/16 "2020-09-12T20:25:19Z")

</div>

OK. I just did the test. I still have an orientation issue:

`Bad orientation: The adj border element defined by [89 91] is oriented in the same direction in element 300 and in the element 124 ****** bug in mesh construction? orientation parameter?`