You can subscribe to this list here.
2000 
_{Jan}
(1) 
_{Feb}
(13) 
_{Mar}
(11) 
_{Apr}
(8) 
_{May}
(14) 
_{Jun}
(48) 
_{Jul}
(39) 
_{Aug}
(17) 
_{Sep}
(16) 
_{Oct}
(19) 
_{Nov}
(18) 
_{Dec}
(18) 

2001 
_{Jan}
(16) 
_{Feb}
(11) 
_{Mar}
(5) 
_{Apr}
(19) 
_{May}
(19) 
_{Jun}
(10) 
_{Jul}
(31) 
_{Aug}
(26) 
_{Sep}
(11) 
_{Oct}
(8) 
_{Nov}
(11) 
_{Dec}
(3) 
2002 
_{Jan}
(9) 
_{Feb}
(8) 
_{Mar}
(5) 
_{Apr}
(11) 
_{May}
(4) 
_{Jun}
(12) 
_{Jul}
(30) 
_{Aug}
(18) 
_{Sep}
(13) 
_{Oct}
(8) 
_{Nov}
(5) 
_{Dec}
(2) 
2003 
_{Jan}

_{Feb}
(4) 
_{Mar}
(7) 
_{Apr}
(15) 
_{May}
(15) 
_{Jun}
(11) 
_{Jul}
(8) 
_{Aug}
(1) 
_{Sep}

_{Oct}
(5) 
_{Nov}
(10) 
_{Dec}
(6) 
2004 
_{Jan}
(7) 
_{Feb}
(15) 
_{Mar}
(6) 
_{Apr}
(3) 
_{May}
(1) 
_{Jun}
(7) 
_{Jul}
(4) 
_{Aug}
(2) 
_{Sep}

_{Oct}

_{Nov}
(1) 
_{Dec}
(2) 
2005 
_{Jan}
(6) 
_{Feb}

_{Mar}
(1) 
_{Apr}
(1) 
_{May}
(1) 
_{Jun}
(4) 
_{Jul}
(7) 
_{Aug}
(8) 
_{Sep}
(2) 
_{Oct}
(3) 
_{Nov}
(5) 
_{Dec}

2006 
_{Jan}

_{Feb}

_{Mar}
(3) 
_{Apr}

_{May}
(6) 
_{Jun}
(5) 
_{Jul}

_{Aug}
(2) 
_{Sep}
(3) 
_{Oct}
(9) 
_{Nov}
(4) 
_{Dec}
(1) 
2007 
_{Jan}
(4) 
_{Feb}
(4) 
_{Mar}
(5) 
_{Apr}
(6) 
_{May}
(4) 
_{Jun}
(5) 
_{Jul}
(18) 
_{Aug}
(19) 
_{Sep}
(11) 
_{Oct}
(13) 
_{Nov}
(6) 
_{Dec}
(9) 
2008 
_{Jan}
(18) 
_{Feb}
(5) 
_{Mar}
(12) 
_{Apr}
(9) 
_{May}
(3) 
_{Jun}
(6) 
_{Jul}
(5) 
_{Aug}
(1) 
_{Sep}
(11) 
_{Oct}
(9) 
_{Nov}

_{Dec}
(57) 
2009 
_{Jan}
(52) 
_{Feb}
(29) 
_{Mar}
(41) 
_{Apr}
(41) 
_{May}
(102) 
_{Jun}
(102) 
_{Jul}
(67) 
_{Aug}
(25) 
_{Sep}
(18) 
_{Oct}
(26) 
_{Nov}
(11) 
_{Dec}
(18) 
2010 
_{Jan}
(11) 
_{Feb}
(3) 
_{Mar}
(5) 
_{Apr}
(5) 
_{May}
(1) 
_{Jun}

_{Jul}

_{Aug}
(3) 
_{Sep}
(5) 
_{Oct}
(2) 
_{Nov}

_{Dec}
(2) 
2011 
_{Jan}

_{Feb}

_{Mar}

_{Apr}
(1) 
_{May}

_{Jun}

_{Jul}

_{Aug}

_{Sep}

_{Oct}

_{Nov}
(3) 
_{Dec}
(5) 
2012 
_{Jan}
(1) 
_{Feb}
(5) 
_{Mar}

_{Apr}

_{May}
(1) 
_{Jun}
(5) 
_{Jul}

_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

2013 
_{Jan}

_{Feb}
(2) 
_{Mar}
(2) 
_{Apr}
(3) 
_{May}
(1) 
_{Jun}

_{Jul}

_{Aug}

_{Sep}
(5) 
_{Oct}
(4) 
_{Nov}
(1) 
_{Dec}
(1) 
2014 
_{Jan}
(2) 
_{Feb}

_{Mar}

_{Apr}

_{May}

_{Jun}

_{Jul}

_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

2015 
_{Jan}
(1) 
_{Feb}
(3) 
_{Mar}

_{Apr}

_{May}

_{Jun}

_{Jul}
(1) 
_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

2016 
_{Jan}
(1) 
_{Feb}

_{Mar}

_{Apr}

_{May}

_{Jun}

_{Jul}

_{Aug}
(1) 
_{Sep}

_{Oct}

_{Nov}

_{Dec}

S  M  T  W  T  F  S 





1

2

3

4

5
(1) 
6

7

8
(5) 
9
(1) 
10

11
(1) 
12
(2) 
13
(2) 
14
(2) 
15
(1) 
16

17

18
(2) 
19
(3) 
20
(2) 
21
(5) 
22
(4) 
23

24

25

26
(3) 
27
(2) 
28
(5) 
29
(2) 
30
(5) 

From: UE59264Osman F Buyukisik <absd00t@c1...>  20000630 17:53:03

Tobias, I put in density of 1e3 as suggested by Dennis, and took out relaxation, lowered the penalty term (1e6), and seems to be working. i.e. I get compressive stress at x, y, z corners. control_timestep 100 0.1 1.0 control_timestep_iterations 100 10 Here is what I think: 1). start with a low penalty term. This will result in penetration error, but it will be fixed later. Lower that term, easier to converge. 2). either add density or relaxation, and try to get something that at least makes sense qualitatively. 3). Once you got that working, now increase the penalty term until you get satisfactory numbers. 4). You also may need to refine the mesh to achieve 3. To make sure you get contact with geometry items, make sure you have enough of a tolerance on the geometry object(s). Also check the direction of the normals, and switch using contact_geometry_switch command. And finally, don't forget the animal sacrifice + incantations to the contact solution gods. You all do that, don't you? :) Best, Osman 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000630 16:30:03

Tobias, I think "options_relaxation 0.5 1. 1. " applies relaxation to VX, VY, and VZ primary unknowns respectively, not materi_velocity, materi_strain_total,materi_stress as you said (is this correct Dennis ??). I think tochnog is having some difficulty with multiple targets. May be overcome by just playing with adjustments(?) 1). Final time usually set to 1 for static runs. Usually due to the way bc's are defined. I don't think there is anyting magic (again Dennis ??) 2),3) Don't really know how they are related. I saw that the way you generate the mesh, you'll get some badly distorted elements on the surface. And wrote another simple segment to create a better looking sphere. It is like the 2D one I posted earlier. : 
From: Tobias Hoevekamp <tobias.hoevekamp@il...>  20000630 15:26:58

Hi there, I am still struggling with my contact problem of a sphere confined within a box. By playing around with the option relaxation > options_relaxation > 0.5 (materi_velocity) > 1.0 (materi_strain_total) > 1.0 (materi_stress) I did get a feeling how control_timestep and control_timestep_iterations_automatic do work together. This works perfectly fine when I do consider only one contact plane. (see th5108.dat) As soon, as I activate the other two contact planes things behave strangly (see th5109.dat). (Osman, the extex17.dat you've sent me a week ago runs thru, but it looks as if no deformation due to the contact occurs). Strange. I guess, it is still a matter of me not knowing enough about the parameters I am playing with. Perhaps you can help me in answering some of the follwoing questions. 1) Does the final time_increment of 'control_timestep' has to be 1.0? I've seen at least one example with a very small final time_increment. It was said that this was due to converging aspects. Is there are rule of the thumb for that? What is the big picture behind this parameter? 2) Can > control_relaxation_materi_velocity 10 100 as be expressed in terms of > options_relaxation > 0.5 (materi_velocity) > 1.0 (materi_strain_total) > 1.0 (materi_stress) 3) Can "contact_relaxation 1e2" be somehow inclueded in options_relaxation? Have a nice weekend. Bye Tobias ________________________________________________________________________ _/_/_/_/_/_/_/_/_/ _/ _/ _/ _/ _/ Ecole polytechnique federale de Zurich _/_/_/ _/ _/_/_/_/ Politecnico federale di Zurigo _/ _/ _/ _/ Swiss Federal Institute of Technology Zurich _/_/_/_/ _/ _/ _/ Tobias Hoevekamp Tel: +41 1 632 33 04 Institute of Food Science Fax: +41 1 632 11 55 ETH Zentrum, LFO E21 tobias.hoevekamp@... CH8092 Zuerich http://www.vt.ilw.agrl.ethz.ch/~hoevekam Home: Bucheggstrasse 50 CH8037 Zuerich Tel/Fax: +41 1 350 70 51 ________________________________________________________________________ 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000630 12:06:36

Hello Tochnoggers, The default sourceforge email list mail size is 40KB. If larger than that it would require my approval. Well, I have forgotten my email admin password :( So, even though I get the email messages that says I need to approve some large emails, I cannot do so at this moment, and sourceforge has not helped me yet. So, until I get this fixed large emails won't go through! Sorry bout that Osman 
From: <pachevalier@ne...>  20000630 11:20:44

Hello, It is a second consignment of this message without the file joins because it has been blocked (superior size to 40k). In a first time, have you an answer to the file below. Hello to all, I worked on the file sent by Osman to create a quarter of circle, but cont= act doesn't calculate himself. My file is retailed below. I also join the = file cer4.flavia.res to show you what it spends. I don't understand the mi= stake! Greetings. Patrick Chevalier France (le 29 juin 2000 fichier pour calculer 1/4 de cercle envoyé par Osman F Buyukisik) echo yes number_of_space_dimensions 2 materi_velocity materi_strain_total materi_stress end_initia group_type 0 materi group_materi_elasti_young 0 206.e3 group_materi_elasti_poisson 0 0.3 group_materi_memory 0 updated_without_rotation group_integration_points 0 maximal ( a simple 1/4 circle 2D model ) ( define the xc,yc for the CENTER ) start_arithmetic xc 0.0 = end_arithmetic start_arithmetic yc 254.0 = end_arithmetic ( define radius ) start_arithmetic r 254. end_arithmetic ( define which quadrant to model ) start_arithmetic i_loc 1. end_arithmetic start_arithmetic j_loc 1. end_arithmetic ( end of user defined stuff  ) start_arithmetic sqrt2 1.41421356 end_arithmetic start_arithmetic r_o2 r devide 2. = end_arithmetic start_arithmetic r_osq2 r devide sqrt2 end_arithmetic start_arithmetic toler r multiply 1e5 end_arithmetic start_arithmetic x1 i_loc multiply r_o2 plus xc end_arithmetic start_arithmetic y2 j_loc multiply r_o2 plus yc end_arithmetic start_arithmetic y4 j_loc multiply r plus yc end_arithmetic start_arithmetic x6 i_loc multiply r plus xc end_arithmetic start_arithmetic x5 i_loc multiply r_osq2 plus xc end_arithmetic start_arithmetic y5 j_loc multiply r_osq2 plus yc end_arithmetic node 0 xc yc node 1 x1 yc node 2 xc y2 node 3 x1 y2 node 4 x6 yc node 5 x5 y5 node 6 xc y4 element 0 quad4 0 1 2 3 element 1 quad4 3 1 5 4 element 2 quad4 3 5 2 6 geometry_circle_segment 0 xc yc r i_loc j_loc toler geometry_line 0 xc y4 xc yc toler geometry_line 1 xc yc x6 yc toler geometry_set 0 geometry_circle_segment 0 geometry_line 0 geomet= ry_line 1 (contact line) geometry_line 5 xc y4 x6 y4 toler (symmetry) bounda_unknown 0 geometry_line 0 velx bounda_time 0 0. 0. 100. 0. (prescribed compression) bounda_unknown 1 geometry_line 1 vely bounda_time 1 0. 16. 1. 16. contact_penalty_velocity 1.e8 contact_geometry 1 geometry_line 5 post_calcul materi_stress sizedev post_point 0 0.0 0.0 print_filter 0 post_point_dof 0 sigyy ( define geometry set 0 prior to using the commands below ) control_mesh_refine_globally 9 h_refinement control_mesh_refine_globally_geometry 9 geometry_set 0 control_mesh_refine_globally 19 h_refinement control_mesh_refine_globally_geometry 19 geometry_set 0 control_mesh_refine_globally 29 h_refinement control_mesh_refine_globally_geometry 29 geometry_set 0 (control_mesh_refine_globally 39 p_refinement control_mesh_refine_globally_geometry 39 geometry_set 0) control_timestep 100 0.1 1. control_timestep_iterations 100 5 control_print 100 time_current post_point_= dof end_data  La messagerie itinérante sans abonnement NetCourrier  Web : http://www.netcourrier.com Minitel : 3615 et 3623 NETCOURRIER Tél : 08 36 69 00 21 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000629 13:49:54

Patrick, Quick answer to: 3,4) You may need to add contact_geometry_switch xxx yes. Check the direction of the normal of the geometry item. The normal should point towards the contacter elements. I am also adding a tochnog data file segment to create 1/4 circular models. The exmple17 method ends up with highly distorted elements at (R,45 deg). This one starts with 3 quads. Enjoy! osman ( a simple 1/4 circle 2D model ) ( define the xc,yc for the CENTER ) start_arithmetic xc 0.0 end_arithmetic start_arithmetic yc 254.0 end_arithmetic ( define radius ) start_arithmetic r 254. end_arithmetic ( define which quadrant to model ) start_arithmetic i_loc 1. end_arithmetic start_arithmetic j_loc 1. end_arithmetic ( end of user defined stuff  ) start_arithmetic sqrt2 1.41421356 end_arithmetic start_arithmetic r_o2 r devide 2. end_arithmetic start_arithmetic r_osq2 r devide sqrt2 end_arithmetic start_arithmetic toler r multiply 1e5 end_arithmetic start_arithmetic x1 i_loc multiply r_o2 plus xc end_arithmetic start_arithmetic y2 j_loc multiply r_o2 plus yc end_arithmetic start_arithmetic y4 j_loc multiply r plus yc end_arithmetic start_arithmetic x6 i_loc multiply r plus xc end_arithmetic start_arithmetic x5 i_loc multiply r_osq2 plus xc end_arithmetic start_arithmetic y5 j_loc multiply r_osq2 plus yc end_arithmetic node 0 xc yc node 1 x1 yc node 2 xc y2 node 3 x1 y2 node 4 xc y4 node 5 r_osq2 r_osq2 node 6 x6 yc element 0 quad4 0 1 2 3 element 1 quad4 2 3 4 5 element 2 quad4 1 6 3 5 geometry_circle 0 xc yc r toler ( define geometry set 0 prior to using the commands below ) control_mesh_refine_globally 5 h_refinement control_mesh_refine_globally_geometry 5 geometry_set 0 
From: <pachevalier@ne...>  20000629 11:24:43

Hello Osman and Tobias, Thank you for your answers, the example extex17.dat of Tobias is the answer to my question for the modelling. Besides, I work to the modelling of contacts and it is a good example. On the other hand, I meet problems in calculations. 1  after the calculation of extex17.dat, the part of sphere doesn't distort itself when I visualize extex17.flavia.res in the postprocessor. The superior plan (plane xz, y=254) should move downwards but him the fact not. 2  the same problem puts itself with the quarter of cylinder that I defined in 3D in my previous message. 3  I defined the quarter of cylinder in GiD in 3D. Then, I have add contact in the file *.dat. nodes, group, ... contact_penalty_velocity 1.e18 geometry_cylinder 0 0.0 255. 0.0 0.0 255. 254 254.0 1.e2 contact_geometry 1 geometry_cylinder 0 post_calcul materi_stress sizedev control_print 10 time_current post_point_dof The message: matrix doesn't converges shows off. 4  I calculated the example 17 of the manual with a contact on a circle instead of a plan. I replaced the plan by a quarter of circle as below. geometry_circle_segment 0 0.0 256 256. +1 1 1.e2 contact_geometry 1 geometry_circle_segment 0 Contact doesn't calculate himself, the quarter of circle doesn't move. What it passes? Thank you for your answer. Patrick Chevalier France  La messagerie itinérante sans abonnement NetCourrier  Web : http://www.netcourrier.com Minitel : 3615 et 3623 NETCOURRIER Tél : 08 36 69 00 21 
From: JONAS FORSSELL <VCC2.JOF@MEMO.VOLVO.SE>  20000628 15:51:50

 Inkommet från VCC2.JOF 7659696 000628 17:46 From: absd00t@... To: VCC2.JOF@..., tochnogusers@... Date: Wed, 28 Jun 2000 08:07:55 0400 Subject: [Tochnogusers] Contact Osman, I will be happy to contribute. Dennis sent me a fix. I'll check it out ASAP on my installation and send the results to this fora. /Jonas Jonas, You don't really have to learn c++, most of the code is like c. Just a few I/O is done through c++ constructs. There are no classes/methods. The files are readable once you read the programmers manual first. If dont object I will put both your name in a new task for extending/fixing the contact. I may be able to help too. Why don't you post your latest data file, and we'll see if anything is possible. Osman  000628 17:46  Sänt till  > tochnogusers(a)lists.sourceforge.net 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000628 13:56:31

Dennis, After looking at the dbs file more carefully, I see that even though sigxx develops, the nodes actually moved in the x direction. I tried with density=1, time=0.1, to 1.0 and by hand sigyy should be 20 and it is. For sigxx, min xvel =0 , max xvel = 1, > accel=1, mass=1, > force=1 > sigxx = 1. regards, osman 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000628 12:12:20

Jonas, You don't really have to learn c++, most of the code is like c. Just a few I/O is done through c++ constructs. There are no classes/methods. The files are readable once you read the programmers manual first. If dont object I will put both your name in a new task for extending/fixing the contact. I may be able to help too. Why don't you post your latest data file, and we'll see if anything is possible. Osman 
From: dennis roddeman <dennis.roddeman@fe...>  20000628 11:24:34

Jonas, > 1. Contact with geometry entities AND contact between elements is NOT > POSSIBLE. If both is specified, the node_boundary records are ignored. > I would really like to work on this in the source code, but then I > have to start learning C++ to begin with.... :) Maybe not difficult. I attach a bit changed contact.cc. Could you test it it does both geometry and element targets? > 3. My previous supplied code does not need relaxation (as Dennis > thoughtfully pointed out). In fact, I could even remove the iterations > and just go straight on single time steps. Maybe this is related to > the fact that the time step I use is just 1e6 seconds. Probably small timesteps without iterations is the best thing to do for this type of calculation. > > I would welcome any hints on how to speed up the calculation time. > Right now, a complete run takes about 24 hours on my PII 350 MHz > computer. Do you use 'control_solver index diagonal'? Kind,  Dr. Dennis Roddeman Finite Element Application Technology, FEAT Tel +31 06 226 806 50, Fax +31 045 566 11 41 Email dennis.roddeman@... 
From: JONAS FORSSELL <VCC2.JOF@MEMO.VOLVO.SE>  20000628 08:27:02

 Inkommet från VCC2.JOF 7659696 000628 10:19 Hello Tochies, Some update on the continuing contact mission: 1. Contact with geometry entities AND contact between elements is NOT POSSIBLE. If both is specified, the node_boundary records are ignored. I would really like to work on this in the source code, but then I have to start learning C++ to begin with.... :) Osman, Dennis, How do I "assign" myself for this task and supply you with the updated code when ready? Through the sourceforge webpage? 2. The workaround to 1. is to define all contacts using node boundary records. I redesigned the previous contact wall as a large element instead which I fixed in two of the nodes, thereby creating an impact wall. 3. My previous supplied code does not need relaxation (as Dennis thoughtfully pointed out). In fact, I could even remove the iterations and just go straight on single time steps. Maybe this is related to the fact that the time step I use is just 1e6 seconds. I would welcome any hints on how to speed up the calculation time. Right now, a complete run takes about 24 hours on my PII 350 MHz computer. 4. My dutch friends here at work tells me that Tochnog means something like "In due time". Just in case someone is interested.... Keep well Jonas  000628 10:19  Sänt till  > tochnogusers(a)lists.sourceforge.net 
From: Ted Christiansen <ted@ca...>  20000627 15:15:26

Feng, Links below: http://studbolt.physast.uga.edu/templon/fortran/f2c.html http://www.altavista.com/cgibin/query?pg=q&sc=on&hl=on&act=2006&par=0&q=f2c&kl=XX&stype=stext Interesting.. http://studbolt.physast.uga.edu/templon/fortran/nagf90.html Ted Feng Jili wrote: > Hello, tochnogers, > Who could tell us how to get the f2c.lib? Thanks in advance. > > _______________________________________________ > Tochnogusers mailing list > Tochnogusers@... > http://lists.sourceforge.net/mailman/listinfo/tochnogusers 
From: Feng Jili <fengjl@ma...>  20000627 14:21:17

Hello, tochnogers, Who could tell us how to get the f2c.lib? Thanks in advance. 
From: Tobias Hoevekamp <tobias.hoevekamp@il...>  20000626 12:43:35

Hi Patrick, You should specify at least one initial element. The mesh is usually generated by hrefinment of the initial element(s). I started out looking at a similar problem by changing and thereby getting to know example 17. See: http://tochnog.sourceforge.net/tnu/node51.html and tochnog/test/examp17.dat. With the help of Osman, I was able to extend this 2Dexample to 3D with 3 contact planes. (It's running, albeit I am still evaluating wether the results make sense...) You find the *.dat file attached (extex17.dat) Good luck Tobias pachevalier@... wrote: > > Hello, > > I join a tochnog file which, the mesh doesn't make himself. It is a part of sphere blocked on a face and with a pressure on the superior part. It is a solid in 3D. Are you able to help me? > > thank you > > ( > A 1/8 of sphere with pressure on the upper in 3D > Patrick Chevalier pachevalier@... > ) > echo no > number_of_space_dimensions 3 > materi_velocity > materi_strain_total > materi_stress > end_initia > > ( define geometry ) > geometry_quadrilateral 1 0. 0. 0. 0. 254. 0. 0. 254. 254. 0. 0. 254. 0.02 > geometry_quadrilateral 2 0. 254. 0. 254. 254. 254. 254. 254. 254. 254. 254. 0. 0.02 > geometry_sphere_segment 1 0. 254. 0. 254. +1 1 +1 0.01 > > (Steel material properties, units: N,mm) > group_type 0 materi > group_integration_points 0 minimal > group_materi_density 0 7.8e6 > group_materi_elasti_young 0 2.1e5 > group_materi_elasti_poisson 0 0.29 > group_materi_memory 0 updated_without_rotation > > ( conditions aux bords ) > bounda_unknown 0 geometry_quadrilateral 1 velx vely velz > bounda_time 0 0. 0. 1.e6 0. > bounda_force 1 geometry_quadrilateral 2 vely > bounda_time 1 0. 1000. 1.e6 1000. > > control_timestep 30 1 1. > control_timestep_iterations 30 3 > > end_data > >  La messagerie itinérante sans abonnement NetCourrier  > Web : http://www.netcourrier.com Minitel : 3615 et 3623 NETCOURRIER > Tél : 08 36 69 00 21 > > _______________________________________________ > Tochnogusers mailing list > Tochnogusers@... > http://lists.sourceforge.net/mailman/listinfo/tochnogusers  ________________________________________________________________________ _/_/_/_/_/_/_/_/_/ _/ _/ _/ _/ _/ Ecole polytechnique federale de Zurich _/_/_/ _/ _/_/_/_/ Politecnico federale di Zurigo _/ _/ _/ _/ Swiss Federal Institute of Technology Zurich _/_/_/_/ _/ _/ _/ Tobias Hoevekamp Tel: +41 1 632 33 04 Institute of Food Science Fax: +41 1 632 11 55 ETH Zentrum, LFO E21 tobias.hoevekamp@... CH8092 Zuerich http://www.vt.ilw.agrl.ethz.ch/~hoevekam Home: Bucheggstrasse 50 CH8037 Zuerich Tel/Fax: +41 1 350 70 51 ________________________________________________________________________ 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000626 12:34:08

Hello, I have replied to Patrick's email. But it is either my email system or sourceforge setup, that it only replies to the author not to the list. So, if you want the replies to go the list, you'll have to manually include Tochnogusers@... I have told Patrick to generate the mesh with control_mesh_macro or another preprocessor since he did not have mesh generation commands in the data file. Is anyone interested in modifying the sphere macro to generate 1/8 or hollow models? best regards, osman 
From: <pachevalier@ne...>  20000626 11:25:10

Hello, I join a tochnog file which, the mesh doesn't make himself. It is a part of sphere blocked on a face and with a pressure on the superior part. It is a solid in 3D. Are you able to help me? thank you ( A 1/8 of sphere with pressure on the upper in 3D Patrick Chevalier pachevalier@... ) echo no number_of_space_dimensions 3 materi_velocity materi_strain_total materi_stress end_initia ( define geometry ) geometry_quadrilateral 1 0. 0. 0. 0. 254. 0. 0. 254. 254. 0. 0. 254. 0.02 geometry_quadrilateral 2 0. 254. 0. 254. 254. 254. 254. 254. 254. 254. 254. 0. 0.02 geometry_sphere_segment 1 0. 254. 0. 254. +1 1 +1 0.01 (Steel material properties, units: N,mm) group_type 0 materi group_integration_points 0 minimal group_materi_density 0 7.8e6 group_materi_elasti_young 0 2.1e5 group_materi_elasti_poisson 0 0.29 group_materi_memory 0 updated_without_rotation ( conditions aux bords ) bounda_unknown 0 geometry_quadrilateral 1 velx vely velz bounda_time 0 0. 0. 1.e6 0. bounda_force 1 geometry_quadrilateral 2 vely bounda_time 1 0. 1000. 1.e6 1000. control_timestep 30 1 1. control_timestep_iterations 30 3 end_data  La messagerie itinérante sans abonnement NetCourrier  Web : http://www.netcourrier.com Minitel : 3615 et 3623 NETCOURRIER Tél : 08 36 69 00 21 
From: dennis roddeman <dennis.roddeman@fe...>  20000622 13:10:52

Hi Osman I looked at my linux executable size, that is 2.5 Mb or so. Hmm. Here is my guess: In increased lately two array sizes near input with: MSTRING in input.cc and MRANGE in tochnog.h. (necessary because some people are running large input files.) I bett it is the MSTRING, try a smaller one and see if the problem remains (maybe a stack overflow in your compiler with long static arrays). If not try the MRANGE. (It would be better to make some arrays dynamic, I looked once, but not easy; would give chaos code.) Dennis UE59264Osman F Buyukisik wrote: > > Hello Dennis, Tochnoggers, > I just realized that my tochnog executable is 100MB!! What happened? > input.cc takes long time to compile and really exercises my hard disk, > and creates 100MB input.o file. (june 18 timestamp). This is on > hpux10.20 machine using egcs 2.95. But using HP's own c89 and aCC, > there is no problem. Something in the input.cc is triggering a bug in > gcc now (at least on my platform). Any others having similar problem?? > > Best Regards, > Osman > > _______________________________________________ > Tochnogusers mailing list > Tochnogusers@... > http://lists.sourceforge.net/mailman/listinfo/tochnogusers  Dr. Dennis Roddeman Finite Element Application Technology, FEAT Tel +31 06 226 806 50, Fax +31 045 566 11 41 Email dennis.roddeman@... 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000622 12:47:44

Hello Dennis, Tochnoggers, I just realized that my tochnog executable is 100MB!! What happened? input.cc takes long time to compile and really exercises my hard disk, and creates 100MB input.o file. (june 18 timestamp). This is on hpux10.20 machine using egcs 2.95. But using HP's own c89 and aCC, there is no problem. Something in the input.cc is triggering a bug in gcc now (at least on my platform). Any others having similar problem?? Best Regards, Osman 
From: dennis roddeman <dennis.roddeman@fe...>  20000622 10:17:40

Hi Jonas, Does that index on the velocity relaxation need to be 100 in stead of 0? Kind, > (setup the solutions) > > control_timestep 100 1e3 0.1 (time step 0.1 micro second) > control_timestep_iterations 100 10 > control_relaxation_materi_velocity 0 1e2 > > _______________________________________________ > Tochnogusers mailing list > Tochnogusers@... > http://lists.sourceforge.net/mailman/listinfo/tochnogusers  Dr. Dennis Roddeman Finite Element Application Technology, FEAT Tel +31 06 226 806 50, Fax +31 045 566 11 41 Email dennis.roddeman@... 
From: JONAS FORSSELL <VCC2.JOF@MEMO.VOLVO.SE>  20000622 09:45:22

 Inkommet från VCC2.JOF 7659696 000622 11:40 Hello Tochies, In response to Tobias mail: 1. I recognize the > tochnog th51.dat > Error: the solution seems to diverge for unknown : velx > value: 1.59255e+15 This was the result I got when I turned the contact wall normal the wrong way. The results is that the wall sucks the nodes in the wrong direction instead of repelling them (at least that is my theory for the higs speed). Try reversing the normal by changing the direction of the line or surface. 2. I used velocity relaxation in my simulations. A value of 1e2 seems to work. 3. Here is a suggestion for units and an extract from my current crush problem (by the way, the node boundary doesn't work yet, but the geometry boundary does): ====================================================================== (Simulation of bumper section crush 20000607 Jonas Forssell) (Units are: Node coordinates  mm Velocity  m/s Density  kg/mm3 Youngs modulus  GPa Gravity  9.81e3 Energy  J Answers: Force  kN time  ms Mass  kg Pressure  GPa (Initialize) echo yes number_of_space_dimensions 2 materi_velocity materi_strain_total materi_strain_plasti materi_plasti_kappa materi_stress end_initia (Define a material) group_type 0 materi group_materi_density 0 7.8e6 group_materi_elasti_young 0 210 group_materi_elasti_poisson 0 0.3 group_materi_memory 0 updated group_materi_plasti_vonmises 0 1.100 (Boron steel Yield 1100 MPa) (contact wall) geometry_line 5 52 100 52 100 0.1 contact_geometry 0 geometry_line 5 contact_penalty_velocity 1e12 (Internal contact) node_boundary 9 yes node_boundary 4 yes node_boundary 5 yes node_boundary 13 yes node_boundary 25 yes node_boundary 33 yes node_boundary 41 yes node_boundary 53 yes node_boundary 61 yes node_boundary 75 yes node_boundary 113 yes node_boundary 144 yes node_boundary 168 yes node_boundary 196 yes node_boundary 224 yes node_boundary 252 yes node_boundary 317 yes node_boundary 333 yes node_boundary 353 yes node_boundary 365 yes node_boundary 373 yes node_boundary 385 yes node_boundary 394 yes (Apply velocity to the section) geometry_line 2 50 100 50 100 1 bounda_unknown 0 geometry_line 2 velx bounda_time 0 20 (setup the solutions) control_timestep 100 1e3 0.1 (time step 0.1 micro second) control_timestep_iterations 100 10 control_relaxation_materi_velocity 0 1e2 control_print_gid 120 yes control_print 120 time_current control_repeat 130 250 100 (do 25 ms in all) =================================================================== By the way Tobias, you are more or way asking the same questions as I would like to know. I will follow the answers in this fora with great interest. Hope this was any help to you Tobias. Cheers, Jonas  000622 11:40  Sänt till  > tochnogusers(a)lists.sourceforge.net 
From: Tobias Hoevekamp <tobias.hoevekamp@il...>  20000621 15:30:27

Hi there, I am really amazed and equally pleased by the support that I have experienced so far working with Tochnog. Thanks alot. Before addressing some of my questions about Tochnog I would like to give you an insight about what I intend using Tochnog for. I am more and more convinced that Tochnog is BEST suited for my calculations of deformable (regular) porous media in 3D. Those calculations are supposed to cover round about 50% of my Ph.D. work along with similar experimental investigations. My following brief description of the numerical part within my Ph.D thesis relates to sketchs of the problem (76 MByte) which can be found at: http://www.vt.ilw.agrl.ethz.ch/~hoevekam/Tochnog/PhD_Description.pdf I do look at the smallest 'periodic' region within a regular deformable porous media (Fig 1). In Fig. 1 the porous media is shown without any deformation of itself. The deformation of the porous media is subject to 3 forces depicted in Fig.2 a) pressure imposed by the fluid flow b) the contact problem (and c) body forces). The deformation of the porous media will be dependend on the position within the porous media (neglecting wall effects, assuming periodic bcs). The further downstream we are, the greater the deformations will be (Fig. 3) Working with some of the examples (1 and 17 > extex17.dat) and looking at "Chapt. 8 Interaction analyses..." indicate to me, that Tochnog should be capable of handling my problem. (I am looking forward to extending Tochnogs source code e.g. with periodic boundary conditions, or are they in it already?) In order to get started with the CFD part, I created a geometry given in Fig. 4 by extending example1.dat to the attached th20.dat. Tochnog complains with > > Error: the, default, BICG solver did not converge. > The initial error in the linear equations is: 3.29528e05 > The final error in the linear equations is: 7.25046e+16 > That is not good enough. > > > Probably you have a wrong input file. >  Not enough boundary conditions? >  A diverged calculation? > Hmm. I guess I need your help again :( Let me adress some more general questions about tochnog. 1) The NavierStokes eqt. seems to be treated a little less detailed than the other parts of Tochnog. a) Why? b) Can you recommend any reading that shows the Implementation of the NS (I do like "Chapt 1 Implementation of the convection..." in the programmers guide and would like to see something similar on "Material deformation and flow") c) In "Chapt 1.6 Space discretization..." you say that all fields are assumed to be continious the domain. To my knowledgede, stresses in CFD are normally treated discontiniously with the exception of pressure which can be treated continiously or discontiniously. How is that done? What do you mean by "more diffuse"? 2) Is there a particular FEM book that you could recommend also in terms of the notation used in Tochnog? 3) Units: Sometimes it is not clear to me, which units are used for variables. e.g. tochnog/test/axisym5.dat seems to use [rho: ton/mm^3] tochnog/test/example3.dat seems to use [E: kg/mm^2] (at least that's what is says within the manual) 4) How shall I address little typos etc. within the manuals? 5) How shall I address suggestions, that I do have about the further development of tochnog 6) Finally: What does 'tochnog' stands for. (I've been asked several times) I am feeling likewise happy and sorry about adressing all these questions. Happy, since I know, that you have all the experience to answer them, sorry, that it causes some 'work' for you. Knowing, that I might  eventually  be able to take 'work' away from all of you by answering some of those 'nasty' questions makes me also be kind of happy about my nasty questions ... I am looking forward to your help and response Tobias ________________________________________________________________________ _/_/_/_/_/_/_/_/_/ _/ _/ _/ _/ _/ Ecole polytechnique federale de Zurich _/_/_/ _/ _/_/_/_/ Politecnico federale di Zurigo _/ _/ _/ _/ Swiss Federal Institute of Technology Zurich _/_/_/_/ _/ _/ _/ Tobias Hoevekamp Tel: +41 1 632 33 04 Institute of Food Science Fax: +41 1 632 11 55 ETH Zentrum, LFO E21 tobias.hoevekamp@... CH8092 Zuerich http://www.vt.ilw.agrl.ethz.ch/~hoevekam Home: Bucheggstrasse 50 CH8037 Zuerich Tel/Fax: +41 1 350 70 51 ________________________________________________________________________ 
From: dennis roddeman <dennis.roddeman@fe...>  20000621 15:25:49

Hi, Please notice below that my email address changed. (Till August 1 I will read once in a while the old address, afterwards not anymore). Kind,  Dr. Dennis Roddeman Finite Element Application Technology, FEAT Tel +31 06 226 806 50, Fax +31 045 566 11 41 Email dennis.roddeman@... 
From: roddeman <d.roddeman@wb...>  20000621 14:11:59

Hi, I didn't follow the discussion, busy, just the remark that since some time there now exists the OPTIONS_RELAXATION which is maybe more easy to use than the other ones. (And indeed with relaxation more iterations are needed). Dennis UE59264Osman F Buyukisik wrote: > > Tobias, > > that was control_relaxation_materi_velocity and contact_relaxation. I > have attached the .dat file. At one time I had resuls that were > somewhat OK (needed more mesh refinement I thought), I tried to make > it better but still no good results yet :( I think since it runs and > gives compressive Y stress, but not quite as high as it should, it > needs more iterations due to the relaxation I had to put in. > I also made a tet mesh in gmsh and attaching the input file for that. > Osman > >  > Name: extex17.dat > extex17.dat Type: unspecified type (application/octetstream) > Encoding: base64 > Description: hex mesh > >  > >  > Name: ball1.geo > ball1.geo Type: unspecified type (application/octetstream) > Encoding: base64 > Description: gmsh input file  Dr. Dennis Roddeman, tel. +31 53 4892458, fax +31 53 4893900 University of Twente, Mechanical Engineering, Room WB N111 Postbox 217, 7500 AE Enschede, The Netherlands, d.g.roddeman@..., http://www.tm.wb.utwente.nl/~roddeman 
From: UE59264Osman F Buyukisik <absd00t@c1...>  20000621 13:17:55

Tobias, that was control_relaxation_materi_velocity and contact_relaxation. I have attached the .dat file. At one time I had resuls that were somewhat OK (needed more mesh refinement I thought), I tried to make it better but still no good results yet :( I think since it runs and gives compressive Y stress, but not quite as high as it should, it needs more iterations due to the relaxation I had to put in. I also made a tet mesh in gmsh and attaching the input file for that. Osman 