I’d appreciate advice on how to apply a load on a 3 node beam at a location that is not a node, if that can be done. Thanks

If you want to apply a concentrated force or moment in a particular location, you should place a node there. Of course, you can also define a distributed load (*DLOAD) acting over a specific length of the beam.

consider using *DISTRIBUTING COUPLING

That would mean the nodes are not evenly spread. How will that affect the calculations?

Does that spread the load across an element? I think that is not suitable for what I need.

Can you give an example of using *DLOAD, applying a load only to a portion of a beam? Is that what you mean you can do with *DLOAD?

In finite element analysis all loads are in the end transformed into nodal loads (forces and moments). The coupling element allows to distribute a load applied at any point to a set of nodes, thus you have an equivalent set of loads in equilibrium and only local discrepancies may take place dependind on your choice of nodes for the coupling but that’s part of the modelization effort. I recommend performing some simple examples to check its behaviour.

You may have to increase the mesh density (not necessarily globally) to place a node in the desired location. Preprocessors do this automatically.

I’m talking about a distributed load that can be full or partial like this (from the book “Wzory, wykresy i tablice wytrzymalosciowe” by Niezgodzinski):

Check the beamd.inp file in CalculiX test examples library,

I don’t think DLOAD has that capability (I think only uniform distributed load along complete length is available), you should calculate equivalent nodal loads like in here Lecture 10: Equivalent Load on nodes in beam element - YouTube. Commercial preprocessors usually have that capability built-in for productivity but I guess that’s not a primary objective for ccx/cgx developers.

Yes, it’s applied to the whole element but you can have multiple elements modeling a single beam structure. That’s what I was thinking about.

Thanks for your help. I tried the DISTRIBUTED COUPLING which seems to do what I need. To give context, I am modelling a power pole 10m high, using 5 beam elements. I compared applying a 10kN load at 8, 9, and 10m directly to nodes at those positions against using the coupling giving appropriate weights to the nodes, to determine deflection of the pole tip.

There is a good correlation between the 2 methods, and using the coupling method for intermediate heights is also consistent.

*However, I only understand the basics of FEA and Calculix, so I’d appreciate comments about whether I have modelled this correctly.*

hi,

if you mean B32 or B32R element, no it’s required the beam element to be split with new node at the point of concentrate loads.

and some notes, mid point of quadratic beam element can not be used as applied nodal forces.

best,

Thanks for all your help. I think I have found a working solution.