Give Away of the PhD Thesis second try - liveDr.-Ing. Manuel Ramsaier2021-04-17 | Hey guys, just that you're sure I did not cheat on this, its a live stream.
its time again to look at the newest release of Hyperworks as there are some interesting changes. I had a lot of fun trying out the new features and try to explore them even more in the future.
Thanks for your time and your interest. If you have questions, please let me know :)
In this video I will take a closer look at linear buckling analysis, a pretty common type of analysis. Linear buckling is the easieast FE-calculation method for determining the critical buckling load for your structure. With this easy example of a single column, we can check the FE-Results by looking at the analytical solutions with the Euler method.
In this video I will take a closer look on inertia relief analysis, a pretty common type of analysis for example in the aerospace industry. These are helpful links to dive deeper into this topic:
As I said, please take it with a grain of salt as I am just beginning to learn this topic. If you find something to be off please share it with me via mail or commend. Thank you in advance.
In this video I will take a closer look on modal transient response analysis. This is the second approach following the direct transient response analysis. Basically you are interested in the response of you structure to a time dependent excitation funciton. It is solved by using shape modes. For large models it is more efficient. Be sure to check the frequency of the excitation. If its hight, a lot of modes have to be extracted, which is time-consuming.
In this video I will take a closer look on direct transient response analysis. Basically you are interested in the response of you structure to a time dependant excitation funciton. It is solved by using complex algebra rather than extracting shape modes (that's modal transient response analysis). For large frequency excitations it is more efficient.
In this video I will take a closer look on modal frequency response analysis. This is the second approach following the direct frequency response analysis. Basically you are interested in the response of you structure to a sinusodial excitation funciton. It is solved by using shape modes. For large models it is more efficient.
In this video I will take a closer look on direct frequency response analysis. Basically you are interested in the response of you structure to a sinusodial excitation funciton. It is solved by using complex algebra rather than extracting shape modes (that's modal frequency response analysis). For large frequency excitations it is more efficient.
Some things did not work out. The SET I create should NOT have "no card image" as this broke things. Rather than that, use SET1 card image. In the modal frequency analysis video this problem is resolved and the postprocessing can be seen there (it is the same way which you would want to apply in this video). Sorry about the confusion.
today is a special episode again. Lukasz and me decided to have a talk about different analysis types, their use cases, how to approach them and much more.
It was a fun and enlightening conversation like always and I very much look forward to more of them in the future. :) Thanks for talkinkg, Lukasz.
as recently shown in the livestream I wanted to make a proper annoucement that I am working on some CLI tools for hyperworks. It is still very early in development but if you want to check it out - here is the link to the repository.
In this tutorial (OS-HWX-T: 6010) you will learn about how to set up a Fatiue Analysis with two static loadcases and test date (time history for fatigue loading) and the rainflow method. We will talk about the general approach to high cycle fatiuge with Altair Optistruct.
Signed von Mises is used here as we use a ductile material like Aluminum.
Get the missing load files here: community.altair.com/community (under documentation, Optistruct 2021.2 Tutorial models)
did you see me get excited about the record feature for Hyperview Scripts. Let me tell you: The hype is real. It will make the life of many of us so much easier. Watch me walk you through the basic steps how to see the comannds following your point and click action in Hyperwiew.
if you have checked the update on Hyerworks recently you may noticed a few changes especially an the panels, which some us of got to love (and hate). They will be gone soon and in this vider I wanted to give a little insight how the transition is going.
Thanks for your time and your interest. If you have questions, please let me know :)
as you might have seen from the last video: Altair has brought a new release for Hyperworks. Besides updating to this version I made minor adjustments to the code for better usability.
its time again to look at the newest release of Hyperworks as there are some interesting changes which also demand action at least of some of us. Say goodbye to the panels and this time for real. It seems to be inevitable and therefore better now where you still can fallback with a recent version.
Thanks for your time and your interest. If you have questions, please let me know :)
new feature: create a bunch of new materials and assign the elements of the lattice structure regarding their axial stress values. Last time we saw how we could analyze the structure and get the stress values out in form of a text file. We use this text file in combination with the original .fem file for a mapping: Which element should be assigned with which new material.
Materials can now be created in batch. Given a lower and upper value for the youngs module which controls the stiffness and the number of materials to be created, the process is: Creating the materials in linear order, assigning the rod elements in a proportional mapping to the newly created material.
new feature: read results from hyperview via tcl commands: stress and displacement values from rod elements / nodes implemented. This is the first step to be able to update the material properties of the rods for an optimization method.
Camera: amzn.to/3zxMRol Lights: amzn.to/3zAjW2Q Greenscreen: amzn.to/3zAjW2Q Controller: amzn.to/3zBGIrd Wacom Tablet amzn.to/2Hi1h0HProductivity Tip - Organizing Windows with Microsoft FancyZonesDr.-Ing. Manuel Ramsaier2022-07-31 | Hi gusys, this one is a short video about a productivity enhancement with Microsoft PowerToys and Le Dimmer. Avoid shifting your focus away from your current task ans also enhance the lifespan of your monitor.
New feature: write the .fem file for the lattice mesher directly - this saves so much time (especially with big models). Enjoy also the little, stupid mistake I made on the way and learn from it to not do it yourself, too ;)
Warning: again no new features - "just" code updates and better software development strategies in this video
This time its about how to remember your code comments for setup, build, test, lint and much more. I use a tool called Task for that, which you can find here: https://taskfile.dev/
Warning: no new features - "just" code updates and better software development strategies in this video
So I restructured the code in order to have it as a CLI tool. I explored dependency management with poetry and lots of other things like formatting, linting, CICD actions in github.
so I added tetrahedron support to the HypermeshLatticeMesher as well as come bug fixes. I could save some serious time when creating the lists. Thank god for sets(). Still not optimal I guess but its a decent start. One bridge at a time.
Next up is Hyperview Scripting and importing it back into the data structure to be able to run an own lattice density optimizer. Still there is serious need for different lattice structures but this will be surely implemented in the future. As always, if you have any questions or comments, feel free to ask.
Thank you for watching and your interest in my little program.
in this episode I want to share with you a tutorial on the Ball Balancing Table - an example which was used in the 2021 Student Contest hosted by Altair.
Please find the mentioned additional material in the links below.
in order to generate lattice structures in hypermesh I took the approach to read in a .fem file of a solid hexahedron mesh (e.g. voxels) and create a tcl script which creates the structures as rods.
It is a first stage - different element types are surely a thing to expand. Also, Hyperview Scripting in order to get the stress or displacement values seem like a good next step.
in this episode I walk you through a neat process how to go from fully constructed geometry to a boundary mesh to a hex mesh even if the geometry is not mappable ( that means, it can easily be meshed with hex elements)
This was brought to my attention a while ago by Dr.-Ing. Robert Dienemann, who used it in his dissertation for cantilever structures with draw manufacturing constraints. Voxel meshsing is often used when doing topology optimization, so if you know how to apply it, it serves you well in the Future.
All the best for you and if you have questions / comments, feel free to add them down below.
Altair has brought it to my attention - that currently a global contest is taking place and I would like to share this info with you. "Explore the limitless possibilities of system simulation enhanced by Artificial Intelligence"
It is a short reaction video on the contest and I mention future plans to support participating students.
as promised I will present the hyperview script part of the Bridge Builder series in a little more detail. In this video you will learn about the very basic functionality of Hyperview and how to automate it with *.tcl files / commands. This is extremely helpful if you have to postprocess multiple files at once and also for studies.
Thanks for your time and your interest. If you have questions, please let me know :)
This update covers the hyperview scripting part for creating the screenshots as well as the density file. Additionally, the postprocessing is described here. Rods with lower density than a specified threshold are eliminated. Rods which point into the same direction are combined as long as no different rods are attached on the bearings. Multiple runs for different settings of the bridge and optimization variables are now possible, although I encountered a few bugs in this episode. Mainly the issue was about class variables which were not reset. I left the main parts of me debugging it in the video if you would like to follow me thorugh here. For a quick look, the update in the code is the faster solution.
Next, report generation is due. I plan to do this on a mark down file with screenshots and variables attached for each covered variant.
Thanks for your time and your interest. If you have questions, please let me know :)
This will hopefully sort out my connection issues to the license server for whatever reason. Please keep in mind that for HWU (at least for me) there was a 24 hours max limit for borrowing.
Thanks for your time and your interest. If you have questions, please let me know :)
This update consists of the driving lane logic and the implementation for Altair Hypermesh. Basically, the problem is to build a driving lane which is somewhat symmetric and also has as few elements as possible. The bug mentioned at the end has been fixed now and the resulting bridge is shown in the thumbnail. I go over the changes based on the graph of git in more detail. Hope this helps to understand where I made changes and what the path forward may look like. The driving lane can be found with the help of the A* (A-Star) algorithm. It is quite popular in engineering problems as well as navigation and I am sure in many other areas. The goal is to find a cost-optimal path from A to B. In our case, the cost is the distance.
Next up (and also requested seperately by one of you guys) is hyper view scripting and report generation - which I will do in a lazy version at first but for sure is to be expanded in the future.
Thanks for your time and your interest. If you have questions, please let me know :)
today I am happy to present you another talk with Lukasz. We really enjoy doing some talks about FEA together, so we decided to tell a little bit about our ways into FEA and also discuss the role of automation, coding in our daily business.
I hope you enjoy this talk as much as the last one and we plan on releasing more of these videos in the future.
This video coffers three update topics: First, I have made it possible to toggle the optimization on rods, which are not considered to be taken into account for optimization. This is for the driving lane - these rods shall be always there in the model with full density. The second change is the introduction of a first cost calculation algorithm. It calculates the rods based on two parameters: first segment cost and cost per segment / grid length. This is still not perfect as we will find out in this video but its a start. The costs for all rods are summed up in a seperate class called Bridge. The last and biggest change is regarding the displacement constraint when minimizing the mass. In order to set it to a reasonable value we want to evaluate the model before optimization. To make possible, we need to work with a batch execution of the optistruct analysis and also automatically reading the max displacement value for which we need a file to read for which you want to set a control card ... you see, it was quite a process. But it was worth it. With the recent commit, this is all possible and we will use a lot of it in future updates.
Thanks for your time and your interest. If you have questions, please let me know :)
As the journey continues a sleepless night lies behind me and the result is a much much cleaner datastructure but not much new functionality. But as well as functionality is important (obviously) the structure and documentation is important as well. If you don't take care on such things you end up with large portions of spaghetti code which is difficult to work with and hard to expand with new functionality. So, this is the first big issue closed - a neat starter datastructure for hypermesh classes starting with the property, component, Rod classes and also the load, loadcollector and loadstep classes. They will be used heavily in the future and I can't wait. Expect the upload frequency to drop a little as for now vacation time is over. I chose to release everything at once nevertheless because I don't see value in holding content back.
Hope you have a great time and learned a thing or two and I am thankful for your time :)
Code: bit.ly/3JZic8n As the previous video was lacking some in depth discussion about the script builder code, I decided to make this video for anyone who is interested in the details and where the journey is heading.
Hope you have a great time and learned a thing or two and I am thankful for your time :)
Code: bit.ly/3JZic8n As the previous video was lacking some in depth discussion about the script builder code, I decided to make this video for anyone who is interested in the details and where the journey is heading.
Hope you have a great time and learned a thing or two and I am thankful for your time :)
recently I played a few Bridge Construction Games and thought it could be an easy task to just take this simple problem (2D) and show you the python / tcl automation with Hypermesh. Sounded like an easy job to do but boy, was I wrong. Uncovering the different constraints and possiblities how to model this problem, I thought this could be a good learning journey which we could explore together. So first of all, the code can be found on: bit.ly/3JZic8n
If you want to checkout the game, find it here (steam): bit.ly/3JNIREJ
I hope you liked the video and find the journey as interesting as it is for me. Feel free to comment, like, subscribe as always but also feel free to make suggestions (also on github regarding the code) and lets explore this thing together!
I thought it would be a good idea to dig a little deeper into the non linear FEA topic by presenting a introductionary tutorial. It covers nonlinear material modelling as well as geometric non linearity (i.e. large displacement FEM).
I hope you like it and as always, if you have questions or comments, feel free to add them below :)
today is a special episode. A while back I was honored to be the guest of Lukasz Skotny and talk about topology optimization to his audience. Today I have the honor to have him here to talk about non liniearity in FEA, a topic I ignore to a large extent on this channel, but what I consider to be a very important topic.
I hope you enjoy this talk as much as I did and we plan on releasing more of these videos in the future.
Feel free to comment below what you find interesting and want to be discussed here or on his channel :) I am sure he will gladly answer your questions as he answered mine.
I wanted to show you a workflow how to extract a script for a problem. You will see struggeling with the tcl code, planning for future features and general FEA modelling in Hypermesh.
If you have any questions, comments, feel free to add them below and as always thank you very much for watching.
today we look at the node and edge equivalence tools in Hypermesh and what the line colors mean. You will notice, I struggle a bit with the toogle edge tool - this may be a bug on just my side but we solved it at the end :)
If you have any questions / comments, please fell free to add them below :)
this video is about the update 2020.1 of Altair's Inspire Tool. I hope you like the content I can provide here and if you have feedback or suggestions, you can always contact me on ramsaier@rwu.de