Students who have used Workbench should be familiar with the ‘weak springs’ feature. This function is very effective for situations where constraints are insufficient. Its principle is that the software automatically applies a spring with a very small stiffness coefficient at the under-constrained boundary, thereby satisfying the solving requirements. Without significantly affecting the solution accuracy, it enables the model to be solved in a short time, allowing users to quickly adjust the model based on preliminary calculation results.
The setting location for enabling weak springs in WB is as follows:
Many students have privately messaged Shuige asking whether APDL has a similar function.
Actually, it does, and the operation method is simple and quick. In terms of interface display, it is more intuitive than WB!
Below, a case study is used to illustrate how to enable it and its effects, for your reference.
As shown below, a metal plate is subjected to equal tension on both sides. In physics, this situation is very common, and the stress in the middle is also easy to calculate – force divided by the loaded area. Of course, due to the presence of holes, there will be stress concentration.
However, if the user directly models and analyzes it this way, it is highly likely that there will be insufficient boundary condition constraints. That is, although the physical phenomenon is reasonable, during model analysis, appropriate boundaries still need to be applied according to the actual model.
To overcome the above situation in the early stage of product design, thereby quickly evaluating work and accelerating the design cycle, reasonable springs can be automatically applied at under-constrained boundaries through the weak spring function. In APDL, the command to apply weak springs is:
WSPRINGS
No special parameters are needed. It can be used in both preprocessing and solving.
The APDL help documentation does not provide much explanation for this command, and it cannot be implemented through GUI. For the above model, the changes after entering this command are as follows:
Constraints are added at the four corners of the model. Note that these constraints are not directly on the plate nodes, but rather a zero-length spring is created. One end of the spring is connected to the plate node, and the other end is fixed.
Opening the element type menu in the main interface, it is found that 3 spring types have been added, which are springs with freedom in three directions, with element type Combin14.
Opening the real constant type, it is found that an additional real constant has been added, with a value of 0.21.
From the above inspection, it can be seen that after using the weak spring command, the software automatically adds springs at under-constrained locations, with element type 14 linear springs, and the spring stiffness is automatically calculated.
The reason why APDL is more intuitive than WB, as mentioned earlier, is that after using this command, you can immediately see where the springs are added, whereas in WB, you can only view them after the solution is complete.
The complete code using weak springs is as follows:
finish
/clear
/prep7
et,1,shell181
mp,ex,1,2.1e5
mp,dens,1,7850e-12
mp,prxy,1,0.3
sectype,1,shell
secdata,10
blc4,,,1500,500
cyl4,750,250,100
aovlap,all
asel,s,,,2
adele,all
allsel,all
lesize,10
amesh,all
!Apply weak springs
wsprings
/solu
allsel,all
nsel,s,loc,x,0
f,all,fx,-10
nsel,s,loc,x,1500
f,all,fx,10
allsel,all
solve
/post1
set,last
plnsol,u,xFor comparative analysis, the correct mechanical model is supplemented here, i.e., applying appropriate symmetric boundary conditions at the axis of symmetry, as shown below:
Displacement comparison:
Weak springs:
Normal constraints:
Stress comparison:
Weak springs:
Normal constraints:
From the results, it can be seen that weak springs have a certain impact, but if the mechanism is complex and debugging is very cumbersome, sacrificing a little accuracy relative to time cost is, to a certain extent, also acceptable!
The above is the content about applying weak springs in APDL. It is worth noting that this method is more effective for static analysis. For modal analysis, weak springs cannot serve as actual boundary conditions. If the user’s boundary conditions are unreasonable, rigid body modes will still appear!
No comments yet