I am currently using Cosmos/Works. I must do design and analysis and manufacturing and testing - so I can not devote all my time to analysis. Cosmos/Works has worked well for linear static and some thermal analysis, and the learning curve is fairly quick. However, I can only mesh with solid elements - so thin materials of large surface area or truss structures are not easily modeled. Will Algor give me more flexibility to use a wider variety of elements. Is the learning curve steep? Can I use my SolidWorks models? Can I train myself? This thread seems to have some straight-shooters. What's the real story?
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations JAE on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Reasons to switch to Algor
- Thread starter jdipal
- Start date
Here's the summary first, so you can skip the rest if you want:
Consider what it would take to get the functionality out of Cosmos. If you decide to pursue Algor, ask for training when you purchase the product. The learning curve should be steep no matter where you go from here. I combine a variety of element types in virtually every model, but Cosmos can do this, too. You can definitely use your SolidWorks files. You can train yourself, but you shouldn't have to.
Here's the "wordy" version:
I've used Algor for about 10 years, but let me address the Cosmos comments first: Are you sure you couldn't just learn to use Cosmos a little better? It is a good package with much the same capability and it is integrated with SolidWorks because of their common parent company. The integration will only get better over time. It will use multiple element types, although the way it sounds like you are integrating may make it a little more difficult. I realize Cosmos/Works may be a little limited, but the packages comparable to Algor in cost are similar in capabilty. I do think Algor is the better value, but if you already have some Cosmos background, you may want to pursue it a little harder.
Now, because of my long-time contact with Algor, I may not be the best judge of "easy to learn", but I can say that the interface has improved remarkably over the last few years. FEA in general is getting more user-friendly...not that I'm comfortable with it getting too easy, but Algor does a fine job of linear static and thermal analyses with multiple element types. It integrates well with SolidWorks and is listed as a Solutions Partner on the SolidWorks website.
Training yourself or training via the web, considering your background with Cosmos, should come fairly easily no matter what package you pick up from here. If you purchase Algor, ask them to throw in a training class, or some one-on-one time with one of their techies via their web training capability. At a minimum, though, the learning curve will be pretty steep.
Meshing thin cross-sections is very easy. Basically, when you bring the CAD model in, you can set a thickness for midplane meshing in such a way that anything under a certain thickness with be mid-planed and anything over that thickness can be specified for solid elements. Including beam or truss elements is more of an after-the-fact meshing. You draw the line elements in after the plate modeling. Algor separates the SolidWorks parts into groups and surfaces that can be selected and manipulated.
Algor used to have a pre-processor named SuperDraw III. It had some features that have not been incorporated into FEMPRO, yet, but these features are more for convenience.
Call Algor: 1-800-482-5467 Tell them you are interested and that you spoke to one of Michael Miller's customers by the name of Garland Borowski...many of them know me. See if you can get a demo to test drive and determine the learning curve for yourself. Happy Shopping!
Garland E. Borowski, PE
Consider what it would take to get the functionality out of Cosmos. If you decide to pursue Algor, ask for training when you purchase the product. The learning curve should be steep no matter where you go from here. I combine a variety of element types in virtually every model, but Cosmos can do this, too. You can definitely use your SolidWorks files. You can train yourself, but you shouldn't have to.
Here's the "wordy" version:
I've used Algor for about 10 years, but let me address the Cosmos comments first: Are you sure you couldn't just learn to use Cosmos a little better? It is a good package with much the same capability and it is integrated with SolidWorks because of their common parent company. The integration will only get better over time. It will use multiple element types, although the way it sounds like you are integrating may make it a little more difficult. I realize Cosmos/Works may be a little limited, but the packages comparable to Algor in cost are similar in capabilty. I do think Algor is the better value, but if you already have some Cosmos background, you may want to pursue it a little harder.
Now, because of my long-time contact with Algor, I may not be the best judge of "easy to learn", but I can say that the interface has improved remarkably over the last few years. FEA in general is getting more user-friendly...not that I'm comfortable with it getting too easy, but Algor does a fine job of linear static and thermal analyses with multiple element types. It integrates well with SolidWorks and is listed as a Solutions Partner on the SolidWorks website.
Training yourself or training via the web, considering your background with Cosmos, should come fairly easily no matter what package you pick up from here. If you purchase Algor, ask them to throw in a training class, or some one-on-one time with one of their techies via their web training capability. At a minimum, though, the learning curve will be pretty steep.
Meshing thin cross-sections is very easy. Basically, when you bring the CAD model in, you can set a thickness for midplane meshing in such a way that anything under a certain thickness with be mid-planed and anything over that thickness can be specified for solid elements. Including beam or truss elements is more of an after-the-fact meshing. You draw the line elements in after the plate modeling. Algor separates the SolidWorks parts into groups and surfaces that can be selected and manipulated.
Algor used to have a pre-processor named SuperDraw III. It had some features that have not been incorporated into FEMPRO, yet, but these features are more for convenience.
Call Algor: 1-800-482-5467 Tell them you are interested and that you spoke to one of Michael Miller's customers by the name of Garland Borowski...many of them know me. See if you can get a demo to test drive and determine the learning curve for yourself. Happy Shopping!
Garland E. Borowski, PE
Since Garland wrote you a mini version of his thesis, I will give you a new users view. I just recently purchased Algor's software after trying a few including Cosmos. Algor's web training is really good for getting you started. For deeper problems, their tech support guys are really good at helping you with a specific analysis. I have yet to go to one of their training classes but I have one scheduled in Houston next month.
I have only been working with it for a couple of months now. I still have lots of questions, but the software is easy to use. The only thing that is difficult to me was dealing with Superdraw. For structural elements, this is the only way to set up you model.
Chris Foley
Midland, TX
I have only been working with it for a couple of months now. I still have lots of questions, but the software is easy to use. The only thing that is difficult to me was dealing with Superdraw. For structural elements, this is the only way to set up you model.
Chris Foley
Midland, TX
- Thread starter
- #5
Thank you for the information. I have been using Cosmos/Works which is very easy to use due to its tight integration with SolidWorks - however, it is limited - unless we step up to the Cosmos/M package which is very powerful, but a completely different interface outside of SolidWorks. Cosmos/Works seems limited to me because I am forced to select either solid elements or mid-plane meshing, but not both in one model. I often need to place a point mass at a location away from a part and tie it to a surface - something I would like to use simple beam elements for - but this is not supported in Cosmos/Works. Someday, C/W may support mixed element types, but not yet.
I have seen Algor in use at my Air Force customer and it seems to fit my needs. Ultimately, I need to apply a base input random vibration spectrum to a structure and calculate the response stresses and the response spectrum.
Thanks again.
I have seen Algor in use at my Air Force customer and it seems to fit my needs. Ultimately, I need to apply a base input random vibration spectrum to a structure and calculate the response stresses and the response spectrum.
Thanks again.
jdipal,
I have performed many random vibration analyses, although not necessarily on models that I've imported from SW. Algor works well. Ultimately, my software choice is often driven by the software with which I will need to interface. In your case, it sounds like Algor is the way to go, since your Air Force customer is apparently using it. The SolidWorks interface is pretty simple. I generally, however, do not run both my CAD package and FEA package at the same time...too much overhead on the hardware, so I usually export a STEP file or IGES file from SW and import it into Algor. This works very well.
Garland
I have performed many random vibration analyses, although not necessarily on models that I've imported from SW. Algor works well. Ultimately, my software choice is often driven by the software with which I will need to interface. In your case, it sounds like Algor is the way to go, since your Air Force customer is apparently using it. The SolidWorks interface is pretty simple. I generally, however, do not run both my CAD package and FEA package at the same time...too much overhead on the hardware, so I usually export a STEP file or IGES file from SW and import it into Algor. This works very well.
Garland
Garland,
The InCAD interface is much easier. Use InCAD to create the DMIT file. Then you can go back and close the CAD program. It saves a few steps in the process. I have noticed that if I need to change the dimensions of a model, I simply open up the model, perform the InCAD transfer a second time. The model will change in Algor and most of the time the surface loads and constraints will stay the same. At least this is how it works using Alibre.
Chris
The InCAD interface is much easier. Use InCAD to create the DMIT file. Then you can go back and close the CAD program. It saves a few steps in the process. I have noticed that if I need to change the dimensions of a model, I simply open up the model, perform the InCAD transfer a second time. The model will change in Algor and most of the time the surface loads and constraints will stay the same. At least this is how it works using Alibre.
Chris
Good to know. I have the INCAD product for Mechanical Desktop, but I've never really used it. Part of me is still stuck in the "old world"...yes, I still occassionally use the DOS shell!
jdipal, DMIT stands for "Direct Memory Image Transfer". It is a file format that Algor has created to pull in a variety of CAD models. Algor works with the CAD companies to provide a seamless transfer of CAD models. I've known Chris for a long time (college buddies)...if he says it works and that it works well, I'm sure it does.
jdipal, DMIT stands for "Direct Memory Image Transfer". It is a file format that Algor has created to pull in a variety of CAD models. Algor works with the CAD companies to provide a seamless transfer of CAD models. I've known Chris for a long time (college buddies)...if he says it works and that it works well, I'm sure it does.
Jdipal:
I was also in your same situation, i was using cosmoworks to do linear static analysis but then was limited by the need to combine elements and run dynamic analysis. So we bought algor and have been using it for the past three years. As mentioned before there have been allot of good upgrades to the algor interface since i started to use it which make it more user friendly.
If you are doing alot of dynamic analysis i would check NEi Nastran package. They have alot of features that i wish Algor would also have.
Rudy
I was also in your same situation, i was using cosmoworks to do linear static analysis but then was limited by the need to combine elements and run dynamic analysis. So we bought algor and have been using it for the past three years. As mentioned before there have been allot of good upgrades to the algor interface since i started to use it which make it more user friendly.
If you are doing alot of dynamic analysis i would check NEi Nastran package. They have alot of features that i wish Algor would also have.
Rudy
Gbor:
Actually I dont have NEi Nastran but am trying to get my boss to buy it.
Most of the analysis I do are random vibration analysis. Beleive me algor has helped out allot but ive seen some features in Nastran that i have not been able to recreat with Algor.
In NEi nastran if you run a Random vibration analysis one of its outputis the PSD at desired nodes. You select a node and get the response output PSD. During random vibe testing you would end up with the PSD response at the location where you attached the accelerometer. Having this data would help out to correlate analysi to tests.
As a user of both software packeges do you know of a way to do this in Algor?, maybe by post processing the displacement data output?
any help would be appriciated, dont want to have to change to NEi nastran when the problem could have been solved with algor.
thanks
Actually I dont have NEi Nastran but am trying to get my boss to buy it.
Most of the analysis I do are random vibration analysis. Beleive me algor has helped out allot but ive seen some features in Nastran that i have not been able to recreat with Algor.
In NEi nastran if you run a Random vibration analysis one of its outputis the PSD at desired nodes. You select a node and get the response output PSD. During random vibe testing you would end up with the PSD response at the location where you attached the accelerometer. Having this data would help out to correlate analysi to tests.
As a user of both software packeges do you know of a way to do this in Algor?, maybe by post processing the displacement data output?
any help would be appriciated, dont want to have to change to NEi nastran when the problem could have been solved with algor.
thanks
PSD comes in many forms. Algor will do an FFT and this could be used to get the PSD. I seem to recall that they even have a PSD function for certain applications, but Random Vibe isn't something I do every day. I do a lot of transient time history type work with steady inputs. I use their DDAM function and other Response Spectrum type analyses.
In Algor, you can select a certain node in the post-processing and graph it (right click). You can perform an FFT on the graph to get PSD information. If you want it relative to frequency...hmmmm, I'll have to look at this.
Is your license current? If so, call Algor and see what you can find out. I would be curious.
Garland
Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
In Algor, you can select a certain node in the post-processing and graph it (right click). You can perform an FFT on the graph to get PSD information. If you want it relative to frequency...hmmmm, I'll have to look at this.
Is your license current? If so, call Algor and see what you can find out. I would be curious.
Garland
Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Garland:
Thanks for the response, Im up to date in with my license and have tried to get some help with this from Algor but was told from there tech support that the results i was looking for could not be obtained in Algor or in another packege. But to my surprise Ive seen these results obtained from Nei nastran, the output I would like to see from random vibration analysis is PSD, Velocity, Acceleration. The question is how?.
I see that there is an option to do an FFT but that option is not available in the Random vibration module.
Here is the link to the details Nei Nastran provided for dynamic analysis, please look under Random vibration, outputs:
Running a frequency response analysis you can determine the acceleration and velocity at any point by ploting and taking the derivative. Maybe by inputing a PSD into the frequecy response analysis I can determine accelerations and velocities? can this be done?
thanks
Rudy
Thanks for the response, Im up to date in with my license and have tried to get some help with this from Algor but was told from there tech support that the results i was looking for could not be obtained in Algor or in another packege. But to my surprise Ive seen these results obtained from Nei nastran, the output I would like to see from random vibration analysis is PSD, Velocity, Acceleration. The question is how?.
I see that there is an option to do an FFT but that option is not available in the Random vibration module.
Here is the link to the details Nei Nastran provided for dynamic analysis, please look under Random vibration, outputs:
Running a frequency response analysis you can determine the acceleration and velocity at any point by ploting and taking the derivative. Maybe by inputing a PSD into the frequecy response analysis I can determine accelerations and velocities? can this be done?
thanks
Rudy
- Thread starter
- #15
I started this thread so I'll jump back in with my 2 cents. The conversion of displacement or velocity or acceleration time histories into an acceleration PSD or a velocity or displacement spectral density can be calculated by closed-form equations - outside of Algor, Nastran, or Cosmos after you get the displacement time history for the node in question. I use a set of Fortran programs that perform these calculations. Check out vibrationdata.com for the sofware - and a quite a bit of other interesting utilities for vibe analysis. If you have a displacement time history in a two-column text format (time vs. displacement), then the software will do the rest, as long as the data time steps are at least 4 times (or more) the frequency you are interested in.
Tom Irvine writes the software - he is a dynamacist at Orbital with much practical experience in test and analysis. I think there is a small fee for all the software.
Tom Irvine writes the software - he is a dynamacist at Orbital with much practical experience in test and analysis. I think there is a small fee for all the software.
Tom is a frequent poster on the vibrations forum.
Rudy,
Tom's software is MUCH less expensive than a new license of FEA...not to consider the learning curve from switching packages and understanding the "quirks" of a new package.
I didn't realize that Random Vibe wouldn't allow the FFT.
You may want to repost this under a different topic, but it sounds as though you have Sam Murgie's attention and that he is from Algor, Inc. Take the opportunity to at least put in a request...
Garland
Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Rudy,
Tom's software is MUCH less expensive than a new license of FEA...not to consider the learning curve from switching packages and understanding the "quirks" of a new package.
I didn't realize that Random Vibe wouldn't allow the FFT.
You may want to repost this under a different topic, but it sounds as though you have Sam Murgie's attention and that he is from Algor, Inc. Take the opportunity to at least put in a request...
Garland
Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Update:
To all reading this post, I did get a call from Sam Murgie at Algor regarding adding extra features to the Random Vibration module; they said they may be able to add this for the October release.
One of the big reasons I changed from Cosmo to Algor because of there customer support, once again Algor seams to be offering the support cosmos lacked.
I will update as soon as we get some answers from Algor.
To all reading this post, I did get a call from Sam Murgie at Algor regarding adding extra features to the Random Vibration module; they said they may be able to add this for the October release.
One of the big reasons I changed from Cosmo to Algor because of there customer support, once again Algor seams to be offering the support cosmos lacked.
I will update as soon as we get some answers from Algor.
- Status
Similar threads
- Locked
- Question
- Locked
- Question