I have been reading about optimization for steel trusses. However, anyone experienced in the actual process and would not mind to discuss more with me? Is it true to advancement of FEM has made design of steel trusses a lot easier (where structural engineers can easily play around with different sections) that causes the older optimization method being forgotten?
Hello Patrick L.Y. Tiong, from what I know about statics, mechanical Engineering, mathematics and FEM calculation, I do not think that FEM is needed in most of the cases (only for some more complicated, non-Standard situations). Combining known analytical equations for Standard steel beam design and shapes in combination with e.g. linear differential equation Systems (i.e. Matrix Algebra) you can solve rather complex geometrical designs / Arrangements in quite a short time using todays PC's, without performing an FEM at all. Even for stress, mass, Dimension reduction / changes you would apply rather direct, straight Forward minimizations without the Need for numerical Approximation calculation using FEM. I would apply FEM e.g. for a new type of beam design with non-simple shape or material property distribution. These data could then be used in steel truss design.
The Problem, which may cause the described development to take place, could be that less sufficiently educated mechanical engineers or non-engineers are doing truss design using Software Tools like FEM and Play around with numbers until they get something they like. The results do not Need to be bad or wrong, but take a lot more time to obtain, despite the uncertainty about the results introduced by the People involved.
Thanks Rüdiger Rentsch. This is inspiring. Can you recommend me any nice reading materials with example please? That will be helpful.
Hello Patrick L.Y. Tiong, as pointed out, that I wrote is based on my knowledge and, of course opinion, about this Topic. I haven't seen any publication dealing with this Topic as yet. One difficulty with the Topic might be that it touches many fields. If you want to pin Point it, you have to think about measures (numbers) to derive and then to carry out field studies. Besides the specific Problem of truss optimization and FEM, the underlying Problem is about (engineering) education, the market pressure and the often discussed lack of skilled workers / engineers.
Hi Patrick, there are many ways for analysis of truss structures. The most common method is to use the Finite Element method, simply because it is accurate enough and formulation and analysis of truss problems with the finite element method are very straight forward. About truss optimization, when the problem is indeterminate any change in elements stiffness results in redistribution of loads in elements. As you mentioned, the basic design method for this case is to change elements, perform the analysis (FEM), obtain the elements loads/stresses and nodal displacements and check the design criteria. If you are thinking of using more sophisticated mathematical optimization methods, you can find a lot of them in the literature. The following paper and MATLAB codes might help:
https://www.researchgate.net/publication/260155936_Design_of_space_trusses_using_modified_teachinglearning_based_optimization
http://www.mathworks.com/matlabcentral/fileexchange/51202-teaching-learning-based-algorithm-for-truss-optimization
http://www.mathworks.com/matlabcentral/fileexchange/51250-truss-optimization-with-matlab-genetic-algorithm--ga--function
Article Design of space trusses using modified teaching–learning bas...
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Inorganic Materials
- Metals
- Alloys
- Steel