GRID DEPENDENCY ON BEVEL GEAR
* GRID DEPENDANCY -
Generally the grid dependancy phenomenon used in comutational fluid dynamics, such as variation in the wall function ,wall friction,cell division/boundary cells (counting increment) cell refinement.After doing this modification result will be obtained for all the variations in cases and may be it is possible that the results for the same model and same boundary contions varry for diffrent cases.and we have to choose or itterate accurate result .
Such as for static structural cas the varrying parameter can be categorised as , frictional coefficent , mesh size, use of various order element (linear , qudratic, and higher order elements), The parametes listed above are need to be modified and result has to carried out for all the variations in same model and same boundary conditions.
The results obtained ,compared among the others and hence reaching the exact / good solutions , by analysis of those results. IN the current challeng the material, boundary conditions ,force values ,contact definations are same but only the mesh size varries , The results are obtained for three mesh sizes (4mm,5mm,6mm)...And based on that the results are listed below comparing which is better case and what type of variations are there...
Set up for bevel gear analysis in Ansys :-
The Basic arrangement is same as the spur gear challenge but only diffrence is here ,Apllication of joint loads and creation of joints with predefined manner, so that the simulations physics will be good and accurate. First the model is imported In the geometry tab.. Then some geometry editing is required there ,as we want to see the load behaviours dependancy only on gear tooth ,so some internal material was removed from the internal profile of gear. From the bigger gear 45mm of circle is extruded and material removed. And fron the small gear 25mm of circle was extruded and that amount of material was removed.
The in the mechanical model , material is assigned to the gears as structural steel, In this assignment we had to perform same setup analysis with same boundary conditions but by different mesh size , And the set up demonstrations is based on 4mm mesh size .After that all cases are reviewed properly for conclusion.
The defult meshing is varrying but the face sizing is contant for all cases as 1.75mm
-JOINT LOADS CREATION:-
Analysis setting used for all the cases is :-
Results and conclusion:-
1) 6mm mesh size results;
TOTAL DEFORMATION (MAX) STRESS (MAX) STRAIN (MAX) CONTACT PRESSURE (MAX)
47.837MM 3.88 mpa 2.6e-5 10.636 mpa
GRAPH FOR DEFORMATION GRAPH FOR STRESS VARIATION WITH THE RESPECT TO TIME STEP
2) 5mm mesh size results;
TOTAL DEFORMATION (MAX) STRESS (MAX) STRAIN (MAX) CONTACT PRESSURE (MAX)
47.837MM 4.37 mpa 2.98 e-5 10.636 mpa
GRAPH FOR DEFORMATION GRAPH FOR STRESS VARIATION WITH THE RESPECT TO TIME STEP
3) 4mm mesh size results;
TOTAL DEFORMATION (MAX) STRESS (MAX) STRAIN (MAX) CONTACT PRESSURE (MAX)
47.837MM 4.10 mpa 2.83 e-5 10.645 mpa
GRAPH FOR DEFORMATION GRAPH FOR STRESS VARIATION WITH THE RESPECT TO TIME STEP
All the desired result are listed above in a systematic manner :
comparision chart for all the results:-
CASES TOTAL DEFORMATION (MAX) STRESS (MAX) STRAIN (MAX) CONTACT PRESSURE (MAX)
1) 6MM MESH SIZE 47.837MM 3.88 mpa 2.6e-5 10.636 mpa
2) 5MM MESH SIZE 47.837MM 4.37 mpa 2.98 e-5 10.636 mpa
3) 4MM MESH SIZE 47.837MM 4.10 mpa 2.83 e-5 10.645 mpa
By comparing all the results , I observed due to the change in the mesh size there is slight change in some result but , there are some result which are constant for all cases ..
stress and strain values are changing but the deformation and contact pressure values are not changing , Why we see this type of changes in the results because according to the RELAY-RITZ METHOD for FEM, after discretization and all the preprocessing done the strain valus are calculated for nodes, depending on the strain values displacements are calculated and then we go for stress.
As the strain valus is changing it is obvious the stress value nedd to be changed because due to diffrent mesh size the stiffness averaging is diffrent for all cases..
Hence ,i concluded that the more finner mesh we create better and accurate will be the results, but This has one dis- advantage ,If we are going for finner mesh then we need more computational power/time to solve any structutal/ matahametical model.