FILTERFORMAT:filter/ifirst,ilast,istride/[tolerance/min|max/attribute]
filter/element/[search_range]/[nodelete | delete]
Filter is used in either node or element mode. In 'node' mode it is used to mark for deletion nodes that lie within a distance, epsilon, of one another. In 'element' mode it either marks or deletes elements which have the exact same set of nodes. The test is on node numbers, not the geometric position of nodes so for this to work one should filter the nodes first.
Node Mode:
Used to filter (mark for deletion) points that are geometricly close, (mesh object epsilon value), or if the tolerance parameter is given, closer than the tolerance specified by the user. This command changes the node type of the deleted points to type 'dudded' (itp=21) but does not remove them from the point list. Note that at least one point must be specified in the point range (ifirst,ilast,istride) in order for this command to work properly. Dudded points (itp=21) can be removed from the mesh object by calling rmpoint/compress.
When the min|max / attribute option is used, nodes for deletion are detected based on the standard geometric criteria however, the choice about which node is retained is determined based on comparison of the attribute values and the node with either the min or max value is retained.
Element Mode:
Search a mesh object for duplicate elements. A duplicate element is defined as having the exact same set of nodes in the element connectivity list (itet). The order of the nodes in the connectivity does not matter. The element with the larger itetclr value (master) will be kept. The duplicate element will have its material color (itetclr) changed to max(itetclr) + 1. Two new element attributes (iclr1, iclr2) are added to the mesh object to keep track of the correspondence of master(retained)/duplicate(removed) elements and their original material id (itetclr).
For all elements the values of iclr1 are set their original itetclr values.
For any element that is neither master nor duplicate, the value of icr2 is set to its itetclr value.
For an element that is a master, icrl2 is set to the original itetclr value of its duplicate.
For an element that is a duplicate, icrl2 is set to the original itetclr value of its master.
EXAMPLES:For example, consider the 4 element mesh, where element 2 and 3 are duplicates and:
Element 1,2 itetclr = 1
Element 3,4 itetclr = 2
then after:
filter/element/10/nodelete
Maximum material id max(itetclr) = 2
Duplicate Elements will be set to itetclr = 3
search_range = 10
nelements searched = 4
Number of duplicate element found = 1
cmo/printatt/cmohex1/itetclr/1 0 0
Attribute: itetclr
1 1
2 3
3 2
4 2
cmo/printatt/cmohex1/iclr1/1 0 0
Attribute: iclr1
1 1
2 1
3 2
4 2
cmo/printatt/cmohex1/iclr2/1 0 0
Attribute: iclr2
1 1
2 2
3 1
4 2
All elements are tested. The search for each element's duplicate does not occur over the entire element list. The default for search_range is 10 and results in looking at the 10 elements in the element list sequentially above and 10 elements sequentially below the test element. In the example given below the elements are sorted so that elements that are physically close to each other will be close to each other in the element list.
The search_range can be set by the user. Setting search_range to a number larger than the number of elements will cause all elements to be searched.
The algorithm will only detect one duplicate element per element. If there are more than two elements with the same connectivity, they can be found by calling filter/element multiple times.
The default behavior is to not delete the duplicate elements. However the duplicate elements will be deleted from the mesh if the parameter delete is specified.
In general if you are merging together two meshes and then want to delete duplicate elements the commands might be:
* Merge two mesh objects
addmesh / merge / cmohex / cmohex1 / cmohex2
* Create an attribute with the median x,y,z coordinate of each element
createpts / median
* Sort and reorder the elements based on the median points. This will insure that elements that occupy the
* same location will have element numbers near one another.
sort / -def- / index / ascending / ikey / xmed ymed zmed
reorder/ -def- /ikey
* Filter and remove duplicate nodes.
filter / 1 0 0
rmpoint / compress
* Filter and remove duplicate elements.
filter / element / / delete
filter
Filter all nodes and delete duplicates with epsilon tolerance is set automaticly.
filter / 1 0 0 / 1.e-3
Filter all nodes and delete duplicates where epsilon tolerance is set by user to 1.e-3.
filter / pset get point_set
Filter a subset of the nodes and delete duplicates with epsilon tolerance set automaticly.
filter / 1 0 0 / / min / imt
Filter all nodes and delete duplicates with epsilon tolerance set automaticly. When duplicate nodes are detected the imt attribute is examined and the node with minimum imt value is retained.
filter / element
Filter all elements and set itetclr of duplicates to max(itetclr) + 1. Assign values to iclr1 and iclr2 arrays.
filter / element / / nodelete
Filter elements and set itetclr of duplicates to max(itetclr) + 1. Assign values to iclr1 and iclr2 arrays.
filter / element / / delete
Filter elements and delete duplicate elements. Assign values to iclr1 and iclr2 arrays.
filter / element / 1e20 / delete
Filter all elements (assuming there are less than 1e20) with an exhaustive search and delete duplicate elements. Assign values to iclr1 and iclr2 arrays.