First Comparing with ANSYS APDL solid works simulation is user friendly and takes less time for FE modelling a and analysis
eg 1: force can be applied on face in solid works simulation before meshing where as in ANSYS APDL force can be applied only after meshing
eg 2 In solid works simulation GUI is user friendly the load and boundary condition is shown with symbol, force with force symbol, constrain with constrain symbol
Eg 3 Report generation is easy in solidworks simulation than ANSYS APDL
Second Comparing with ANSYS work bench
Solid works simulation and Ansys work bench is almost same all features are same and can be equally evaluated
simply solid works simulation = Ansys work bench
In my personal experience solid works simulation is good in capturing deformation results but not up to mark in capturing von mises stress results
I dont have any experience with Abacus therefore I am not making any comment
MP
MP2015-03-18
Easier to use.
SF
SF2015-03-18
1. they are all validated codes, they will all solve your problem accurately
2. if a particular tool has a physics set that another doesn't have and you need it, this is no brainer
3. if a particular tool has scaling functionality (faster solve time) that you want/need and are willing to pay for, this is a no brainer
4. if you value user interface, solidworks simulation is probably your best bet
so to me it comes down to what type of problem you're going to solve and how fast you need it
solidworks simulation does a lot but if you need explicit, you need to go to another tool
solidworks simulation probably has the best interface
solidworks simulation doesn't have HPC
JR
JR2015-03-18
Abaqus is dumb. However, that makes it entirely configurable. One major drawback in work I frequently do is that in Abaqus you can't apply multiple remote loads to a particular surface/feature - haven't figured out a great work around just yet - while in SW this is a walk in the park.
SW only has two lousy choices for elements and doesn't give you much you can do with them if there is a problem.
BM
BM2015-03-19
I disagree with the characterization of Abaqus as "dumb". I would go with "strict" though. As in your "remote load" issue: If you use the kinematic option in a coupling you technically over constrain the nodes if you have two "remote loads" acting on the same surface (nodes) as that option eliminates the DOF's in question. If you use the distributed coupling option it will work since it does not eliminate the DOF's. I just tested it - works fine. Abaqus expects you to know about these sorts of things. They could do a lot for their usability if they gave you sensible over ride options for these sorts of things - like switch to distributed from kinematic when it encounters an over constrain on a set of nodes from this useage. Abaqus also give this error when you assign conflicting restraints at surface edges - where an edge is common to two BC's - you need to deselect the common edge from one of the BC's. An override option would do two things: allow you to get the job done and highlight the fact that maybe you don't fully understand something you just did to motivate you to go figure it out.
JR
JR2015-03-19
Bill, I meant 'dumb' in that you have to tell the software everything. I'm certainly not well-versed in Abaqus but my understanding is that you can't apply two forces to the same nodes/surfaces. If that is the case I can think of a simple example where this is a problem. Say you have a reaction force in one direction parallel to a surface yet gravity is applied at a different angle to the (non-modeled) bolted on part, i.e. an air cylinder. A user-friendly software wouldn't require you to calculate by hand the resultant vector force before applying it. However, I may be mistaken in that regard and I'm sure you'll correct me in such case. TIA
BM
BM2015-03-19
Hi James,
There are ways to skin that cat as I mentioned but I do agree that on the UI side Abaqus makes you do some drudgery that is well suited to computer automation. The guys running the Abaqus/CAE product appear, to me at least, to not care too much about these types of conveniences. Why you would make people (and they tend to be quite smart and expensive people) engage in drudgery when its easy to automate is perplexing. Everyone is a prisoner of their history I suppose......
JR
JR2015-03-19
Maybe it's just a way to run up billable hours....
BM
BM2015-03-19
If you want to do non-linear (say plasticity) with contact or any highly non-linear problem for that matter .....you are better off moving to another code. It is a bit of a crap shoot in SWX Sim - you cannot forecast how much time it will take or if you can get a reasonable answer on almost any problem involving non-linear with contact, at least in my experience. Also if you work is outside of its sweet spot - i.e. you are beyond linear contact analysis with solid elements, you might want to consider a move as well. If you do a lot of beam, shells or mixed meshes another code will have less issues but typically more complexity to manage but a much less frustrating experience. Personally I like Abaqus as it is far more robust in low stiffness contact and in general non-linear than say ANSYS, in my experience at any rate. And for full disclosure I sell Abaqus in Canada. However, I do a lot of my consulting work with SWX Sim because it is faster in terms of effort hours to answer for certain classes of problem.
The other reason you might ant to make the move is post processing. SWX Sim is so slow in post processing for models with any complexity/intensity in general it is mind boggling. Considering how expensive some people who do FEA are it is often a no brainer on the cost side to justify a move just on time saved due to post processing. If you really want to drag it (SWX Sim) to its knees just turn on the discrete color option for contour plots. However, Abaqus costs, at a minimum about ~$15k/year/user to rent it - it can be more or less depending on what you are doing and how much you want to do. SWX Sim recurring cost in NA run between $1.5k - $3.5k/year/user.
However, for not too intense usage on problems where it fits SWX Sim can be a very effective tool and its ease of use is unsurpassed for problems it is well suited for.
BM
BM2015-03-19
why drive a car when you can ride a horse......
SF
SF2015-03-20
what are the particular use cases that you can't "get away" with the elements in solidworks simulation and what tools do you want when they aren't behaving?
SF
SF2015-07-08
I totally agree with Bill. SW sim nonlinear is an amateur level and ABAQUS is a pro level. SW sim would not be able provide accuracy in solution above 50 % when nonlinear involves contact, geometry and material. Even if I simplify the model extremely leaving only a few components, it usually underestimates permanent deformation. In contrast, ABAQUS provided relatively accurate solution while it takes skills and efforts to set up models imported from other CAD software. I was pretty disapointed with SW 2015 in NA pacakge. After upgrade, it was a disaster and I want my 2014 version back if possible. It could not even get the solution converged for the same model I have completed in 2014 version. Analysis speed got incredibly slower too. Had they offered low cost training, I would have moved to ABSQUS long time ago. Only thing I be interested in testing products from Dassault is flow sim package.
Jong
SC
SC2015-11-02
Jong Lee said: "SW sim would not be able provide accuracy in solution above 50 % when nonlinear involves contact, geometry and material."
If Jong Lee's claim is a fact, this is very concerning and disappointing, since my company has purchased a Simulation Premium license for this purpose.
I have also told clients we have the software capability to analyze these types of nonlinear problems
(where all 3 non-linearity's: plasticity, large displacement/strains, and contact are modeled).
Based on this information, I am going to need to see a benchmark where all 3 non-linearity's are modeled: (plasticity, large displacement/strains, and contact). I sincerely hope this is an oversight by the OP, but now I need to see the proof. Does anyone know of a benchmark for this type of problem?
Steve
JS
JS2015-11-02
To find benchmarks about SOLIDWORKS Simulation go in SW to:
Help
SOLIDWORKS Simulation
Validation
NAFEMS Benchmarks
(Simulation Add-in has to be turned on ofcourse)
I believe accuracy mostly depends on your own settings and FEA knowledge.
First Comparing with ANSYS APDL solid works simulation is user friendly and takes less time for FE modelling a and analysis
eg 1: force can be applied on face in solid works simulation before meshing where as in ANSYS APDL force can be applied only after meshing
eg 2 In solid works simulation GUI is user friendly the load and boundary condition is shown with symbol, force with force symbol, constrain with constrain symbol
Eg 3 Report generation is easy in solidworks simulation than ANSYS APDL
Second Comparing with ANSYS work bench
Solid works simulation and Ansys work bench is almost same all features are same and can be equally evaluated
simply solid works simulation = Ansys work bench
In my personal experience solid works simulation is good in capturing deformation results but not up to mark in capturing von mises stress results
I dont have any experience with Abacus therefore I am not making any comment
Easier to use.
1. they are all validated codes, they will all solve your problem accurately
2. if a particular tool has a physics set that another doesn't have and you need it, this is no brainer
3. if a particular tool has scaling functionality (faster solve time) that you want/need and are willing to pay for, this is a no brainer
4. if you value user interface, solidworks simulation is probably your best bet
so to me it comes down to what type of problem you're going to solve and how fast you need it
solidworks simulation does a lot but if you need explicit, you need to go to another tool
solidworks simulation probably has the best interface
solidworks simulation doesn't have HPC
Abaqus is dumb. However, that makes it entirely configurable. One major drawback in work I frequently do is that in Abaqus you can't apply multiple remote loads to a particular surface/feature - haven't figured out a great work around just yet - while in SW this is a walk in the park.
SW only has two lousy choices for elements and doesn't give you much you can do with them if there is a problem.
I disagree with the characterization of Abaqus as "dumb". I would go with "strict" though. As in your "remote load" issue: If you use the kinematic option in a coupling you technically over constrain the nodes if you have two "remote loads" acting on the same surface (nodes) as that option eliminates the DOF's in question. If you use the distributed coupling option it will work since it does not eliminate the DOF's. I just tested it - works fine. Abaqus expects you to know about these sorts of things. They could do a lot for their usability if they gave you sensible over ride options for these sorts of things - like switch to distributed from kinematic when it encounters an over constrain on a set of nodes from this useage. Abaqus also give this error when you assign conflicting restraints at surface edges - where an edge is common to two BC's - you need to deselect the common edge from one of the BC's. An override option would do two things: allow you to get the job done and highlight the fact that maybe you don't fully understand something you just did to motivate you to go figure it out.
Bill, I meant 'dumb' in that you have to tell the software everything. I'm certainly not well-versed in Abaqus but my understanding is that you can't apply two forces to the same nodes/surfaces. If that is the case I can think of a simple example where this is a problem. Say you have a reaction force in one direction parallel to a surface yet gravity is applied at a different angle to the (non-modeled) bolted on part, i.e. an air cylinder. A user-friendly software wouldn't require you to calculate by hand the resultant vector force before applying it. However, I may be mistaken in that regard and I'm sure you'll correct me in such case. TIA
Hi James,
There are ways to skin that cat as I mentioned but I do agree that on the UI side Abaqus makes you do some drudgery that is well suited to computer automation. The guys running the Abaqus/CAE product appear, to me at least, to not care too much about these types of conveniences. Why you would make people (and they tend to be quite smart and expensive people) engage in drudgery when its easy to automate is perplexing. Everyone is a prisoner of their history I suppose......
Maybe it's just a way to run up billable hours....
If you want to do non-linear (say plasticity) with contact or any highly non-linear problem for that matter .....you are better off moving to another code. It is a bit of a crap shoot in SWX Sim - you cannot forecast how much time it will take or if you can get a reasonable answer on almost any problem involving non-linear with contact, at least in my experience. Also if you work is outside of its sweet spot - i.e. you are beyond linear contact analysis with solid elements, you might want to consider a move as well. If you do a lot of beam, shells or mixed meshes another code will have less issues but typically more complexity to manage but a much less frustrating experience. Personally I like Abaqus as it is far more robust in low stiffness contact and in general non-linear than say ANSYS, in my experience at any rate. And for full disclosure I sell Abaqus in Canada. However, I do a lot of my consulting work with SWX Sim because it is faster in terms of effort hours to answer for certain classes of problem.
The other reason you might ant to make the move is post processing. SWX Sim is so slow in post processing for models with any complexity/intensity in general it is mind boggling. Considering how expensive some people who do FEA are it is often a no brainer on the cost side to justify a move just on time saved due to post processing. If you really want to drag it (SWX Sim) to its knees just turn on the discrete color option for contour plots. However, Abaqus costs, at a minimum about ~$15k/year/user to rent it - it can be more or less depending on what you are doing and how much you want to do. SWX Sim recurring cost in NA run between $1.5k - $3.5k/year/user.
However, for not too intense usage on problems where it fits SWX Sim can be a very effective tool and its ease of use is unsurpassed for problems it is well suited for.
why drive a car when you can ride a horse......
what are the particular use cases that you can't "get away" with the elements in solidworks simulation and what tools do you want when they aren't behaving?
I totally agree with Bill. SW sim nonlinear is an amateur level and ABAQUS is a pro level. SW sim would not be able provide accuracy in solution above 50 % when nonlinear involves contact, geometry and material. Even if I simplify the model extremely leaving only a few components, it usually underestimates permanent deformation. In contrast, ABAQUS provided relatively accurate solution while it takes skills and efforts to set up models imported from other CAD software. I was pretty disapointed with SW 2015 in NA pacakge. After upgrade, it was a disaster and I want my 2014 version back if possible. It could not even get the solution converged for the same model I have completed in 2014 version. Analysis speed got incredibly slower too. Had they offered low cost training, I would have moved to ABSQUS long time ago. Only thing I be interested in testing products from Dassault is flow sim package.
Jong
Jong Lee said: "SW sim would not be able provide accuracy in solution above 50 % when nonlinear involves contact, geometry and material."
If Jong Lee's claim is a fact, this is very concerning and disappointing, since my company has purchased a Simulation Premium license for this purpose.
I have also told clients we have the software capability to analyze these types of nonlinear problems
(where all 3 non-linearity's: plasticity, large displacement/strains, and contact are modeled).
Based on this information, I am going to need to see a benchmark where all 3 non-linearity's are modeled: (plasticity, large displacement/strains, and contact). I sincerely hope this is an oversight by the OP, but now I need to see the proof. Does anyone know of a benchmark for this type of problem?
Steve
To find benchmarks about SOLIDWORKS Simulation go in SW to:
I believe accuracy mostly depends on your own settings and FEA knowledge.
Janko Stellaard
Cadmes BV
The Netherlands