Combining sub-matrices into global matrix

  • Thread starter Thread starter Aerstz
  • Start date Start date
  • Tags Tags
    Global Matrix
Click For Summary
SUMMARY

This discussion focuses on the process of combining five elemental matrices into a global stiffness matrix for truss analysis. The method involves summing overlapping values from the sub-matrices to form a single value in the global matrix, as confirmed by participants. An example in Fortran illustrates how to add contributions from a triangular element into a larger global matrix. The final solution involves using determinants to relate the stiffness matrix to the load vector, ensuring all numbers are evenly divisible by a common factor.

PREREQUISITES
  • Understanding of global stiffness matrix assembly in structural analysis
  • Familiarity with Fortran programming for matrix manipulation
  • Knowledge of determinants in linear algebra
  • Basic concepts of truss analysis and finite element methods
NEXT STEPS
  • Study the assembly process of global stiffness matrices in finite element analysis
  • Learn about determinants and their applications in structural engineering
  • Explore advanced matrix operations in Fortran for engineering applications
  • Research common pitfalls in truss analysis and how to avoid them
USEFUL FOR

Structural engineers, civil engineering students, and anyone involved in finite element analysis or truss design will benefit from this discussion.

Aerstz
Messages
35
Reaction score
0

Homework Statement



Combine five element matrices into single matrix forming a global stiffness matrix for simple truss analysis (five elements, four nodes).

Homework Equations



-See attached image-

The Attempt at a Solution



Where the four sub-matrices overlap in the global matrix two or three values/numbers occupy the same position in the global matrix (two separate values for AA, three separate values for BB, and so on). Should I simply add these values to form a single value for the global matrix? For example: if AA1 = -10, and AA2 = +5 should I simply combine them as AA = -10+5 = -5?

Is it really as simple as that?

Thanks.
 

Attachments

  • combined matrix.jpg
    combined matrix.jpg
    12 KB · Views: 487
Physics news on Phys.org
Yes, that is the way you build a global matrix from individual elemental matricies.
 
Thank you very much. I'm assuming your post is based on my AA = AA1 + AA2 and not just the matrix image I posted! I very much appreciate your confirmation.
 
Suppose you had a triangular element with nodes 2,6,8 at the corners. You would add it into the global matrix as follows where A could be a 100X100 matrix. a is the 3X3 contribution from the element. Below would be the coding in Fortran.
The first indice is the row, the second is the column.

A(2,2)=A(2,2) + a(1,1)
A(2,6)=A(2,6) + a(1,2)
A(2,8)=A(2,8) + a(1,3)

A(6,2)=A(6,2) + a(2,1)
A(6,6)=A(6,6) + a(2,2)
A(6,8)=A(6,8) + a(2,3)

A(8,2)=A(8,2) + a(3,1)
A(8,6)=A(8,6) + a(3,2)
A(8,8)=A(8,8) + a(8,3)
 
In my last equation there is a typo.

a(8,3) should be a(3,3).
 
Thank you. It was kind of you to post that, and I appreciate it. Apologies for not replying more promptly but I had a couple of deadlines to meet and my mind was away from the matrix problem for a day or two.

Now my mind is back on this issue I am much more confident but stuck at the final stage. I’ve attached an example someone else produced which shows three stages. I can get from stage one to stage two but not two to three. I don’t know where the 160083 came from, or the three bracketed numbers. Any help with this would be very welcome!
 

Attachments

  • Matrix problem.jpg
    Matrix problem.jpg
    24.4 KB · Views: 482
The problem is solved by using determinants where for a1 you divide the determinant of the stiffness matrix with the load vector substituted for the first column by the determinant of the original stiffness matrix. All numbers are evenly divisible by 3 so it is factored out.
 
Thank you once again. Your reply to me is just the sort of thing I want to read in the textbooks, yet it is nearly always missing and so I often learn little from them. Perhaps I've just been reading the wrong books!
 

Similar threads

Replies
2
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 32 ·
2
Replies
32
Views
4K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 19 ·
Replies
19
Views
5K
  • · Replies 1 ·
Replies
1
Views
6K
  • · Replies 6 ·
Replies
6
Views
2K