# Negative Jacobian

 Register Blogs Members List Search Today's Posts Mark Forums Read

 July 27, 2006, 16:31 Negative Jacobian #1 PC Guest   Posts: n/a Hi, I wonder if you can help me here. I have created a body-fitted mesh inside a circular nozzle where I have some negative jacobians, I dont undertand how I can remove these negative jacobian. These happens if I increase the mesh points. Thanks in advance

 July 27, 2006, 16:42 Re: Negative Jacobian #2 Renato. Guest   Posts: n/a You can work out it by changing the nodal connectivity of the negative elements, e.g.: A triangle with nodes i, j, k can be described by: i-j-k or i-k-j one will have positive Jacobian while the other will have negative. You only need to set the nodal mapping accordingly. In finite element method this ordering is usually counter-clockwise for positive Jacobians. ps: It can also be due distorted/invalid elements Cheers Renato.

 July 28, 2006, 01:19 Re: Negative Jacobian #3 Dr.Nick Guest   Posts: n/a Renato's right. it depends on the vector product and normal, i.e. say Normal=(V1xV2), angle(Normal,i)=+R1. Hence, Normal=(V2xV1), angle(Normal,i)=-R1; where i={1,0,0}

 July 28, 2006, 08:59 Re: Negative Jacobian #4 Arambatzis George Guest   Posts: n/a I am also writing a 3D code using tetrahedrals. My problem is that some elements have negative Jacobian. Have you any ideas to solve this prolem? Or any books or lecture notes or links?

 July 28, 2006, 10:16 Re: Negative Jacobian #5 Renato. Guest   Posts: n/a Your problem can be solved in the same way described previously. For a tetrahedra formed by i-j-k-l nodes you might have the following conectivities: (assuming FEM standard mapping) I-J-K-L = positive Jacobian I-K-J-L = negative Jacobian You can use the following piece of Fortran code to compute the element volume (x6) V6 = x21*(z31*y14-y31*z14) + x31*(z12*y14-y12*z14) + x41*(z12*y31-y12*z31) where: xij = xi - xj (regarding the nodal coordinates) Regards Renato.

 July 28, 2006, 10:31 Re: Negative Jacobian (NOTE) #6 Renato. Guest   Posts: n/a I should have also emphasized that negative Jacobians can be due to distorted and invalid elements!!! In this cases, you can't do anything else remeshing your model or fixing the bad elements locally (usually made by hand, picking and moving nodes). Another possibility can be some float point problem related to nodal coordinates. You can always be facing the range the machine number representation -- usually IEEE for PCs (*). In this case, changing the scale of your model can helpful. Mesh problems usually happen when you import/export your model through several softwares due to loss of accuracy when representing the model entities (nodes, edges, faces, etc...). Regards Renato. (*) (unfortunately or fortunately) computers have limited floating point representation.

 July 28, 2006, 12:46 Re: Negative Jacobian (NOTE) #7 Mani Guest   Posts: n/a if you get negative Jacobians in "some" cells of your structured grid, it will hardly be a problem of coordinate orientation. the fact that it happens only with fine grids suggests small inaccuracies in your grid that lead to cross-overs at small cells, maybe related to floating point precision, or simply incomplete convergence of your grid generator. you should inspect the grid.

 July 30, 2006, 05:44 Re: Negative Jacobian (NOTE) #8 R Andom. Guest   Posts: n/a If all your jacobians are negative, then you are fine also.

January 22, 2012, 15:34
#9
New Member

Join Date: Jan 2012
Posts: 2
Rep Power: 0
Quote:
 Originally Posted by Renato. ;44158 You can work out it by changing the nodal connectivity of the negative elements, e.g.: A triangle with nodes i, j, k can be described by: i-j-k or i-k-j one will have positive Jacobian while the other will have negative. You only need to set the nodal mapping accordingly. In finite element method this ordering is usually counter-clockwise for positive Jacobians. ps: It can also be due distorted/invalid elements Cheers Renato.
Hello,
I found this message regarding the naegative Jacobian helpful as I also have a similar problem. I have the following question regarding the triangular elements:
Considering i-j k if I get negative jacobian, I considered i-k-j. In most of the cases it did work. But in few cases, even this did not work. In that case, I just permutated like:
1-2-3
1-3-2
2-1-3
2-3-1
3-1-2
3-2-1
one of the permutation did give a positive jacobian. Can I consider this permutation and proceed with the further calculation?
Any clarification will be highly appreciated.
Vijayendra

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post CFDworker CFX 8 September 27, 2011 18:16 AlmostSurelyRob OpenFOAM 3 June 24, 2011 13:06 balkrishna OpenFOAM Pre-Processing 0 August 17, 2010 02:39 gaottino OpenFOAM Native Meshers: blockMesh 7 July 19, 2010 14:11 Rasmus Gjesing (Gjesing) OpenFOAM Native Meshers: blockMesh 10 April 2, 2007 14:00

All times are GMT -4. The time now is 06:59.