Hi, i have modelled two plates with bolt hole, I have provided compression only support for bolt holes but the file is not running if i make it rigid body connection between the holes the file runs.
PLATE BOLT FACE.pmx (2.3 MB)
Compression-only constraint won’t connect the plates, it’s just a uni-directional (no tension) support acting in the normal directions of the selected faces. Rigid body constraints are more appropriate, but you can also find some other ideas on this forum. Usually, solid bolt models are used for such cases with just one or a few bolts.
1 Like
yes i tried with bolt and nut it works but the thing is it will be time consuming if i have 10 bolts so i thought ill try to simulate bolt contacts, If i make rigid body for node group file is running but the every node on the surface of bolt is attracting force, actually only plate outer surface takes the load and if there is any bending in the bolt then the inside hole surface get compression force.
For joints loaded in shear, there’s this approach (fresh tutorial from yesterday): https://youtu.be/pS-OEiUN488?si=uZSFbJ-DsHtnHm6L
Otherwise, it might be best to use simplified solid bolt models (just cylinders). You can find several examples here.
2 Likes
Sure thank you for the reply. I’ll check the videos..
These forum threads (and some others) can also be helpful:
Quick connection of assembled parts (at the end the approach from the video tutorial is discussed)
Why the top plate displacement is zero? (discusses the approach with spring and beam elements)
I guess you mean only compression material if you are willing to obtain the load on the right side of the hole.
Force from right side plate will be transferred to left plate through bolt ( bolt head contact face) only and if the bolt bends and comes in contact with plate then those contact nodes will have compression force. But, in my model all the nodes are having forces compression and tension which is not simulating the actual force transfer.
The highlighted part transfers all the force from plate to bolt.
As I’ve mentioned before, it might be best to model the bolt as a simplified solid: https://youtu.be/ZSDf1bNXo8g?si=nlqHatJod3hQHhY4
@FEAnalyst sorry I thought I replied to the message from @ANYS. I forgot to tag @ANYS
My response was related to the hole inside contact area. Different think sorry.
I have played with several different approaches to bolted connections. My experience has been that making the effort to use simple, solid bolts has generally been worth while. I added one to the geometry u sent. This approach feels like it gives a reasonable feel for the bolt loads and the integration with the surrounding structure. I have a meshed bolt template that I import and scale up to fit the job, then I just have to copy and past and reposition each bolt - a bit of a fiddle but worth the time I feel
Singlebolt.pmx (4.9 MB)
1 Like
Hi, i have made pipe connection but i am not able to assign contact between plates and between plate and bolt. i used search contact within 20mm distance , but the file is not running. could you guide me how to assign contact.
PIPE CONNECTION.pmx (1.3 MB)
if i am applying little higher load (more than 6000n)the file is not running (MAST AND SLV)
PIPE CONNECTION only MAST AND SLV.pmx (4.7 MB)
I have modelled same thing in Idea statica which shows the stress values
Zipped PDF Files.zip (489.9 KB)
hi, your models may have a lot of working in keyword editor since this approach is not supported yet. Some threads proposed related to simplifying bolt connection is available, bolt response separated by axial and shear, bending actions being neglected. Axial behavior represents by two nodes nonlinear spring (tension only), but shear springs in bidirectional axes is limited to linear only. Slave of node sets at perimeter plate holes connected to master nodes previously defined using equation constraint, fortunately it supports node sets in latest version of CalculiX to simplify the task.
1 Like
This is how it works in Abaqus:
Either node sets or individual nodes can be specified as input. If node sets are used, corresponding set entries are matched to each other. If sorted node sets are given as input, you must ensure that the nodes are numbered such that they match up with each other correctly once sorted. The nodes in an unsorted node set will be used in the order that they are given in defining the set.
If the first entry is a single node, subsequent entries must be single nodes. If the first entry is a node set, subsequent entries can be either node sets or single nodes. The latter option is useful if a degree of freedom at each of a set of nodes depends on a degree of freedom of a single node, such as might occur in certain symmetry conditions or in the simulation of a rigid body.
While in CalculiX 2.23:
The node in the dependent term (i.e. the first term) of an equation may be replaced by a node set. In that case, the equation is applied successively for all nodes in the node set as dependent node. Consequently, as many equations are created as nodes in the node set.
So what CalculiX has now is explained by the final sentence of the Abaqus documentation and conceptually similar to rigid body constraint. Using node sets on both sides would require ordering them in the same way.
I applied the load as surface tractions now its working, I just wanted to know weather the forces in bolts are correct because bolt stress is 70.85Mpa which is on the compression side but he bolts in the compression side should not have force, bolts which are only the tension side of the plate should have stress.
PIPE CONNECTION only MAST AND SLV final.pmx (4.3 MB)
simplified model can be good and faster, but prone of error and/or results accuracy as example below is bolt bending ignored. Any approach being used need to validate with experimental test or full solid model at least. The advantages of solid model are in any loads type, including axial/normal, shear and bending/torsion or combination but computationally expensive for large models.
2 Likes
For axial force, model will give good results because all the bolts have same force, but if you apply horizontal load instead of axial load then only 50% of the bolts will transmit force the remaining 50% of bolts will experience Zero force (at least in theory).
@fatmac model is correct, if the plates are in tension then the bolt gets tension force but if the plates are in compression then the bolts should not attract any force but in my model all the bolts are having force.
Hi, @supreeth
¿Is there internal pressure on your pipe (pipping network) or is it an steel pipe structure?
Regards