Version 00 scoping mesh for Rainier Mesa Western Extract hex chunks to view EW slices A thru E /scratch/nts/tam/rainier_mesa/HSU_basecase_Western/hex_merge_slices Ed selected slice C along y= 4116000 meters Crop at 1300 meter near the watertable Unstructured 2D mesh between 500k and 1 million nodes Subset set of RM Basecase surfaces for stack and slice extraction Use these surface to stack a hex mesh Extract the slice from the hex mesh stack.lgi -> stack_hex.gmv slice_tri.inp slice_tri_clr.inp ---****--- Oct 21 2009 Terry A. Miller wrote: Here are fehm files (first version) /scratch/nts/tam/rainier_mesa/HSU_basecase_Western/fehm_2D_slice_C_v00 This is meant to get you started with fehm. The distribution is adequate and there are no neg coupling coefs Edge lengths in lower thick materials are 100 meters, 40 meters up to layer 13, and 30 meters for remaining 18605 nodes and 36357 triangles There is plenty of room for increased resolution. This is located at slice C at y= 4116000 meters Bottom is flat at 1300 meters lower left corner is translated from 566700, 4116000 to 0,0 *********dump_material_lists******** Minimum material ID value = 4 Maximum material ID value = 24 Total possible materials = 21 Material 4 has 668 nodes. #nodes/nnodes is 0.359043267939E-01 No nodes found in material = 5 Material 6 has 2641 nodes. #nodes/nnodes is 0.141951088417 Material 7 has 302 nodes. #nodes/nnodes is 0.162321956463E-01 Material 8 has 175 nodes. #nodes/nnodes is 0.940607363612E-02 No nodes found in material = 9 No nodes found in material = 10 Material 11 has 514 nodes. #nodes/nnodes is 0.276269819941E-01 Material 12 has 1745 nodes. #nodes/nnodes is 0.937919914002E-01 Material 13 has 1739 nodes. #nodes/nnodes is 0.934694974469E-01 No nodes found in material = 14 No nodes found in material = 15 No nodes found in material = 16 Material 17 has 2809 nodes. #nodes/nnodes is 0.150980919108 Material 18 has 939 nodes. #nodes/nnodes is 0.504703036818E-01 Material 19 has 428 nodes. #nodes/nnodes is 0.230045686643E-01 Material 20 has 302 nodes. #nodes/nnodes is 0.162321956463E-01 No nodes found in material = 21 Material 22 has 4186 nodes. #nodes/nnodes is 0.224993281376 Material 23 has 2027 nodes. #nodes/nnodes is 0.108949207202 Material 24 has 130 nodes. #nodes/nnodes is 0.698736898683E-02 ***Construct Regular Sparse Matrix:2D*** ------------ Matbldtri Number of Nodes = 18605 Max. Connection to a Node = 13 Number of Area Coefficients = 128527 Minimum Voronoi area = 1.2737312E-01 Maximum Voronoi area = 4.4942137E+03 Minimum Voronoi edge length = 0.0000000E+00 Maximum Voronoi edge length = 1.5590283E+01 Total Voronoi area = 5.1109011E+06 Total Triangle area = 5.1109011E+06 ---****--- Now work on refinement. This is very subjective and will take multiple attempts at refine choices. Version 3 is a nice start. This version still needs nodes within the very thin layers, use the interface offsets to add nodes and fill these spaces. 300 m spacing in large areas 200 m spacing in med areas 64 m spacing in narrow areas try03.lgi -> 2D_mesh_v03.inp (double defined nodes) Extract interfaces as lines Create 15m spacing version for distance routine Version 4 starts with Version 3 then adds high res and extracts interfaces as singly defined nodes try04.lgi -> 2D_mesh_v04.inp (double defined nodes) poly_interface_v03.inp poly_interface_v04.inp Offset interfaces Remove nodes where interface buffers will go - keep outside nodes. Need to play with this as the very thin layers have interference with buffer lines and adding points at all mk_addpoints.lgi -> poly_pts_filter.inp ******* connect_recon.lgi *********** this creates a decent tri mesh for fehm using recon 1 there are no buffers in this version connect_recon.lgi -> 2D_mesh_nobuffers.inp connect noadd is giving me fits I suspect I will need to start with extracted slice from stack to fix tiny material slivers at top edge, and to add flat top ************************************** Prepare the smooth triangulated slice for point addition by removing node too close to the poly points and re-smoothing Add interface nodes then connect single defined point set color nodes for fehm and write files