Description
Are you an industrial designer (or a CAD user in general)? Are you not doing much with the simulation tools? Come and learn why and how you should be exploiting finite element analysis (FEA) and simulation in Fusion 360 software. Engineers are designers’ teammates, not opponents. Design and validation tools will help you collaborate with engineers more effectively. Those tools will also help you to carry a design to production with fewer changes. You’ll go to engineering with a sound concept, rather than something that needs a lot of help. We will talk about static stress, shape optimization, and maybe even a little generative design.
Key Learnings
- Test designs in Fusion 360 with static stress and shape optimization, and perhaps do a little generative design as well
- Learn how to apply simulation tools to your design process
- Learn how to export simulation reports for engineering
- Discover Fusion 360 Simulation tools
Speaker
- JSJeffrey SmithJeff Smith (RIT 93' ID) is and Industrial Design working at Autodesk in World Wide Sales focused on Automotive. Currently a TSE (Innovation Agent), Jeff has also been on both the Customer Success Team and Education Team focused on Fusion Adoption. Prior to Autodesk, Jeff spent 20 years as a practicing Industrial Designer with experience in a wide range of markets and manufacturing processes. In addition, Smith has been adjunct faculty at three colleges (the Art Institute of Ft. Lauderdale, RIT and Iowa State). Jeff is home based in South East Florida.
JEFF SMITH: Well, welcome, everyone, to my class. My name is Jeff Smith. And I am happy that you have decided to join my Introduction to Simulation for Industrial Designers. Quickly, before we get started, a little bit about me-- I am an industrial designer. I have been working at Autodesk for almost seven years. I've taught several classes here at AU.
I started on our Education team helping colleges and universities adopt Fusion 360. I spent some time on our Customer Adoption team helping customers adopt and use Fusion 360. And currently, I am on our Automotive team, actually on Sales, helping people leverage our software and listen to customers. So it's quite interesting.
Prior to joining Autodesk, I did about 20-plus years as an industrial designer. My CAD background is SolidWorks. I have built products all over the world, spent time manufacturing products, and design for manufacturing is a thing.
I also taught as adjunct faculty at the Art Institute of Fort Lauderdale for Industrial Design. And I'm currently adjunct faculty at the Rochester Institute of Technology, which I am also alumni from. So quickly, if you'd like to learn a little bit about me, head over and give me a follow on Instagram.
This is me here. I post a lot of what I do, most of which is old-school drawings and marker renderings. But I also do some CAD, some simulation and things of that nature over here. So if you'd like, that's something that you can go follow and learn more about me. So with that in mind, let's go forward and jump right into Fusion. So let me switch over there.
So if you're familiar with Fusion 360, you know this side of the coin. If you're familiar with many other parametric CAD modeling tools like SolidWorks, or Inventor, and so on. I'm not going to dive deep on anything on the modeling side. What I'm going to do is hopefully open the door to why you as a designer should be touching on FEA or simulation on a basic level to empower yourself.
As designers, we work in real-world production. And engineering is our mirror image on this team. We both bring assets to this development process. The more I can reach out to engineering, the more I can learn about that, the better my designs make it to production. The less I have an engineer throwing something back over at the fence at me saying, this won't work, I want to reach out with an olive branch to say, let's work together. If I get my things closer to reality, that helps me.
So with that in mind, my demo is pretty straightforward here. I would like to do a simple L bracket, and that's enough context for what we are going to do. So I'll do a quick sketch on the front plane here. And I'll draw myself a rough L bracket. And again, this is straightforward, nothing major. But I find, when you're learning something new in a technology, keep it simple to start the ball rolling.
So we'll just make a gigantic 25-millimeter thickness L bracket here, OK? And we'll also make this 25. I may want to change this later, so I'm not going to make it equal. And let's just say that it's 200 millimeters long, and our L going down the wall is 125 millimeters. Great. Simple, to the point, nothing major. We'll quickly extrude that as a solid.
I am going to set it as symmetric because I want my spine plain to be right down the middle. And we'll say, 25 millimeters as well. We'll get a nice, good size on this thing. All right? You can see, I've already saved this file as my Autodesk University Sim 1. It's on version 1. We'll save it again. Great. I can put my user-saved version description in if I want or not. And it will go forward from there.
OK, so it's saved. And the reason I saved it is, A, I want it on the cloud, and I want it ready to process, OK? If you're using Fusion, you are comfortable with that, and you're comfortable with my Modeling/Design tab. I am going to go from Design down to Simulation. When I switch over to simulation, the first thing Fusion's going to ask me is, hey, what do you want to do? It's trying to guide you towards framing the context.
So static stress is what we are going to do because I want to know what this thing can handle, and will it pass that test? You can do lots of other things, like Modal Frequency, Electronic Cooling, Thermal, and Thermal Stress. The first grouping can be solved locally or on the cloud. When you step to Structural Buckling and then Nonlinear Static Stress, Event Simulation, and Shape Optimization, those are cloud-only. So you have to leverage cloud computing for those. The first group, you can choose yourself what you want to do.
The last one here is Plastic Injection Molding Preview. That is not quite in Fusion yet. You have to turn it on. That's a test for another story or a project for another story. It's a really powerful tool along the lines of what we're talking about today. Read about that one later. So we're sticking with Static Stress. I will create my study.
Notice that I've got my body in here, just like I would in my design workspace. But my browser has changed. I have things that are similar, like Units. And I've got an assembly as needed, depending on what's happening. My component is here. Named Views, Origin, and then I've got my Model Components. I only have a body right now. So I can access those. But there's nothing else here, and I have no timeline because it's not worried about my construction sequences.
I have Study Materials in my first Study, and my Study 1 is activated-- means I'm working on this one. I've got Mesh information. I've got Results-- obviously, I don't have them yet. But it's all contained in this study. When you work in the Simulation environment or Cam, or Generative, for that matter as well, the idea is to go from left to right.
So first, I'll look at Study. I don't need a new simulation. I could add one if I wanted to. But let's work on the one I picked. Now, next is Simplify. You can simplify your model in the context of the simulation. For example, if I had multiple bodies here, I could delete the bodies I didn't want to test, but it wouldn't delete them from the file.
So here I am. Don't need to simplify. I'm good. Now you have to tell the machine what material you're using. Now if I go to Study Materials, it's going to tell me that my category is Metal. My name is Steel. And I've got one component in there.
And my study material is Same as Model. I've got it set to Yield Strength. You can switch it to Ultimate Tensile Strength. My engineering friends have said, Jeff, as a designer, you should be working in yield strength because tensile strength is when it completely fails and will not have any resemblance to what it ultimately was. So I have been directed by my engineer friends. I'm going to listen to them, and I'm going to stay in my lane as a designer and work on yield strength, OK?
My standard material is steel. I can change from same as model to whatever I want here Or I can just simply go back to my design and right click and say Appearance or Physical Material. Those are ways to change the way it looks. And I want Physical because I want to say this is this material, not just looks like it. So let's go with plastic, and let's just go with a simple ABS.
So that's an ABS plastic. We'll say, Close. And now, watch what happens when I go back to Simulation, and I go to Materials, and I say Study Materials, it's already updated ABS. You could do it either way you wanted to. You could leave it the standard and change it. I'm a believer in changing it, but both work. OK, we know what the material is.
Next, we're going to go to Constraints. I am going to start very basic. There are many ways to constrain things. But I'm going to say, for my type, I'm going to say Fixed. There are more. I'm going to stay Basic. I want this entire face to be locked against the wall. I'm going to assume that my attachment fixture or whatever I'm gluing, bolting, screwing for right now that this face is 100% not going to move.
And again, I am just trying to get a baseline. You can tell it to have freedom in any axis you want. So you could turn off one of the axes, two axes, whatever you like. But I want all three axes-- x, y, and z-- to be locked in place for this face because I want to test the structural integrity of this design. I'll say, OK.
Now, I'm going to come to Loads. And notice, I've got Structural loads, Linear Global, Angular Global, Gravity On, Edit Gravity, Point of Mass-- all good. I don't necessarily need any of these right now. I will turn Gravity on just to make sure that this is oriented in the correct direction. I can edit that as needed. And I'm going to add a simple structural load to this-- very similar to my locking for this face. I'm going to pick the top face in its entirety.
Notice that it shows me a general idea. Based on direction, it's going to go perpendicular to that face because it's planar. But I could choose an angle. I could choose a vertice and so on. I can also pick points, edges as well, and you can load them as needed. But I'm going to say, you know what, let's just put a force on this face.
I'm pretty good in metric and in millimeters, but I can't quite think in Newton meters yet. I know that 40 Newton meters is the torque setting on my cassette for my bike, but I can't quite grasp it yet. I'm working on it. So we're going to change the units, and I'm going to change it from Newton meters to pound force. And I'm going to say, let's put 150 pounds on there, all right? And I'll say, OK.
Contacts are not important here because I am testing a singular piece. If you had connections and things that needed to be joined together, there's a time and a place for that, but we're starting basic. This gets more complicated. Let's have success here. Under Display, you can choose to adjust the mesh. It's going to self-generate anyway, so you don't really need to worry about it. Under Management, you can dive in deep for Mesh Control or Attributes, but, again, not important for what we're doing right now.
When you come to the solve tab, there's one here called Pre-check, which is already green. It's already checked. That's giving me a visual cue that we're ready to go. I'm going to hit Pre-check anyway, and it's going to test for me. It's going to say, Jeff, do we have the information we need? Do I have a constraint? Do I have a load case? Do I have a material? That is bare minimum for it to work. I'm going to say, OK, I agree with you.
And then I'm going to come to Solve. And what it's going to do is give me the option because it's one of the first grouping, and I can do it on the cloud, or I can do it locally. It's going to cost me 5 Cloud Credits right now if I do it on the cloud. I'm going to switch the local, so I can talk about that side of the coin as well.
Now, what it's going to do is it will run this test locally. It will use my hardware to run the math. Now if you haven't done this yet, Fusion will ask you to install the processing software for that. And obviously, if you want to do it, you have to say, yes, please install that. Nastran is a solver, and it will say, hey, I need to install this. It takes 40 seconds. So I've already done it. So I'm going to say, Solve.
It's going to start running and processing this test. It's going to let you know what's going on. It's saying Locally. It would also say in the cloud if it was on the cloud. And we're already done here, all right?
So if I come over here to my results, my safety factor is 2.4 for this piece. And it's giving you a little summary, and it's saying 3 is a common number. Yeah, for consumer products, that's probably in the ballpark. I believe aircraft parts are in the 7 range for a safety factor. It's also giving you recommendations. So how can I adjust this? How can I change this?
And it is saying, Show weakest area of design. Deformation Scale is set to Adjusted. I'm going to say close here. I'm still in my results. And you'll notice that I've got a chart here, and it's set to Load Case with Safety Factor. I'm going to switch it to Stress. And now, it's going to show me where the stress is. Blue is less. As it goes up and climbs toward red, red is bad, OK? That's a general intent. And again, we're at 2.34. We're not that far off.
So let's switch it to Displacement. And you'll see how much it's moving. So red shows you the most movement. Now granted, it said that its deformation scale was sent to Adjusted. If I switch this to Actual, it's going to show you the actual movement.
So you can see that the max movement is 3.4 millimeters here. I can also animate that and show it visually. That's the actual predicted deformation. If I switch it to Adjusted two times, it's going to give me a strong visual of what it's doing. Now it's noting the specific dimension, but this is giving me a feel. So keep that in mind.
All right, so let's cancel that for a second. Let's finish those results, and let's learn from what we've got. I'm going to go back to Design, and I'm going to adjust it. So I'm going to turn my sketch information on. And I'm going to say, you know what, let's change this a little bit.
And let's make this maybe a little bit longer. Let's make this 150, OK. And let's make this 25 on here. Let's make that 20. And let's make this 20. And let's say, you know what, let's add a chamfer on here. So let's put a chamfer on that, and we'll pull that down, and we'll say, a 25-millimeter chamfer as well.
OK, so I adjusted the design. Let's see what my geometric changes do to the simulation because it's all linked. So if I come back-- let's Save it just to be safe. And let's go back to Simulation. It's got an error here. I called yellow-orange errors in Fusion, look at me. Red errors are I'm broken. Yellow is look at me. So what it's telling you here is that my results are out of date. So all I've got to do is right click and say Solve again. It's going to run it locally because it remembered what I chose. It's going to take the new geometry data, and it will run that test.
OK, so our safety factor went down because I made it thinner, even though I added that chamfer in there. So we got worse. So let's close that. Let's finish that result. Let's go back to my design environment. And let's say, all right, let's make that 30 millimeters, and let's make this 30 millimeters.
We made that a little longer, but now, we've got this really chunky design. So we'll hit Save. And then we'll go back to Simulation. I'm out of date. Let's rerun it. So we'll say Solve. Great. So now, we're still taller. We're back to our thickness, and we added the chamfer.
So now, our safety factor-- look, it went up to 4.47. So we're actually slightly over-engineered. And our max deformation is 1.2. If you recall, I believe, it was 3.4 and change. So now, we've learned something. We've made our design stronger by changing the geometry.
Now hopefully, you can see the window now that says, wait a second. If I'm testing what I'm building, there's a chance I might be able to save material. I might be able to predict wall thickness, how much I need, what material it is. And now, you come to engineering with a part that's 80% there, 90% there, depending on how much you go back and forth. That took us almost no time.
So let's go here to my results. And I am going to say report. And I'm going to preview that. And it should open it in a browser. And I can save out a PDF, but for me as a designer, this information here, I get it. I can follow it.
But as we start to get down here, you can see that it's printing out the report of-- I'm fixed here, gravity, load case. I start to get a little bit deep in the weeds here, right? So there's sheer forces, reactionary forces, and so on. But think about the target audience for this report. We are not the target audience. You come to engineering with this report, with your design intent.
Now think about the olive branch. Imagine if an engineer came to you concerned about the way a highlight looked on a surface. You think, wow, you are speaking my language. This is the inverse. You go to engineering with this, they are going to look at you and say, wow, you understand me a little bit. Thank you. So extend that. Reach out towards that. It's a super valid process.
OK, I hope that helps you guys see that this is a potential pathway to increase the efficiency of your work and the efficiency of how you work and collaborate with others, OK? So I'm going to finish this results for a second. And I'm going to start a new file. And I want to talk about how you might want to collaborate with the computer as a designer taking a step forward with simulation.
So I'm going to save this one, and we'll save it as-- in my Test Files, we'll say AU, we'll say, Shape Op1-- because we're going to use simulation in a similar way. I'll do a sketch here. We'll look at that, and I'm going to draw a quick rectangle here.
And I think I was somewhere in the 150 by 200, or maybe it was 250. Good enough for me. And I will say, Extrude that. And we're going to say, Symmetric. And our distance-- I think I did 25, great. So we're in a similar ballpark to what we had general volume. But I'm giving the computer a brick to work with.
Let's change the material. We'll make that Physical Material. And we use ABS. We'll make it ABS again-- terrific. Now we got this saved, terrific. I'm going to switch from Design to Simulation again. So our previous test was a Static Stress. I'm going to go to the opposite end of the corner, which is Shape Optimization. It has to be run on the cloud, and this is you collaborating with the computer or artificial intelligence to optimize a shape to solve a problem.
Now this shape optimization is one result based on your test. It is not generative design. That's another question mark. So this is shape optimization. I will create the study, and my principal is very similar. And this will remind you of what we just did. Again, Simulations-- I happen to be in Study 1. Just make sure I hit Static Stress by accident.
Hold on. Let's get rid of that. So a new one, and let's say, Shape Optimization, Create Study, yes. There we go-- shape optimization. I let it click the first one. So again, left to right. Notice that my Shape Optimization is activated. The first Static Stress I did is not because I'm not working in that one right now.
I don't need to simplify. Its fine. Materials-- my study material, ABS Plastic. I'm good. That's what I did before. Constraints, Structural Constraint. I want to lock this face-- x, y, and z. We'll say, OK. Now I'm going to load it. I'll turn gravity on again. I'm going to do a structural load on the top. And I think I did. it was 150 pounds. That sounds right. It does remember what you used last over here.
So we'll say, 150 pounds, OK. Contacts aren't important. Display isn't important right now. Shape Optimization gives you some other aspects. So you can preserve certain regions to say, OK, don't change this. I need this. You have shape optimization criteria for targeting mass or maximizing and adding stiffness. You've got global constraints as well. But we're not going into that right now. We're going to leave it standard for now.
You've got a Symmetry Plane as well, which allows you to say make it symmetry in this plane. I could obviously add our spline plane for now, but I'm not going to go that far. I just want to preserve a region. And I want to say that this face over here, , coming off here, x distance, whatever I choose, I want to retain this amount in the z-axis 25 millimeters. So whatever you do, that has to preserve that region, OK? It has to start from that world. And let's put that at 0-- nope. Oh, I want it at that. Let's do it one more time.
I want to do Preserve Region here. And I'm just going to drag this one over x distance. Lovely. So now, I have an area that it will not adjust, it will not take away from. I'm going to come to my Solve, I'm going to do my Pre-check. It has everything it needs. It doesn't require a preserve region. I just added it. You have to have the constraint, the load case, and the material.
I'm going to go ahead and say Solve. Obviously, my static stress has nothing on here right now. So it's telling me there's an error, but I only want to do this one. Has to do it on the cloud. I only want to run this one. I'm going to check that one. It's going to run 5 Cloud Credits for me, and I'll say Yes.
So what the artificial intelligence is doing is very similar to the static stress. It's running the math based on the constraint, the load, the material, and the geometry. The variable now is that it's saying, OK, I have this volume to work with. So what is processing is that it's pulling information out of that volume that I have said, this is what you're starting with. So it's going to process and take time to do that. And it's iterating down to achieving those goals. I think I said minimize mass, so it's going to push through that and calculate based on that information. And you can see, it's still sending it to the cloud, and it's processing.
What is wonderful about using shape optimization is that you can get a really good idea of bare-minimum. If I contrast it with the other side of the coin, where I was testing and then me adjusting, you could see the process could take some steps-- back and forth, back and forth. That's wonderful, and it teaches you so much. And I hope that you do that. But when you use shape optimization, you're setting up a question mark to drive out what it does. I'm going to pause my recording for a second until this is finished, so we don't waste class time waiting for number crunching.
OK, my results have processed. Let's go ahead and close it. And look what it's given us. It said, hey, Jeff, I think you should make an A-frame through here and put a triangular shape for that. It's going to be more efficient than what you had done. I have the ability to adjust and scale it.
Like, if I really want to be conservative, I could add more material, and you can see it turning blue. It's saying green is good. Target is where you want to go. I'll just be a little conservative, and I'll push this to that level. Now what's super interesting here is that it generates this as a mesh. And I hope all of you as designers are not going to just take this mesh and drive it to a design intent. I hope that you are going to leverage you as a designer, and you're going to come to the Results tool and go to Promote and say, Add mesh object to Design Workspace.
And when it does that, now, you have a mesh body in here that's a really good underlay for your design intent. Whether I print this out, sketch over it, whether I grab a screen image and draw over it digitally, whether I just model over this right now, I have a foundational mass that I know will achieve that structural test based on the material and the geometry.
Think about that from a designer perspective. If I'm using this as an underlay, I am square in the target zone for where I need to go. And now, I'm adding what I can bring to the question of this development. I'm adding the styling, the element, the what if. And then I can test it again in the simulation. And I can say, hey, will this work? What's going on here?
It is now a tool set in your world that you can arm yourself with. And I and I hope that I've inspired you guys to go that way in trying to explore this side of FEA analysis. And again, this is just a basic starting point. There are many ways to deal with it. But I think this is a really, really good starting point.
OK, so with that in mind, I'm going to put a little breadcrumb out there for you guys. And that breadcrumb is Generative Design. Generative design is taking shape optimization and going a whole other level further. I'm not going to go into running, this test. But I want you to see how your knowledge that we just talked about in both static stress and shape optimization fits in here and how generative is different than shape optimization.
So I'm going to create a study. I happen to have this piece of geometry in here because it's the same file. The principles are the same. It's left to right. I have a Study, great. Should look fairly similar as far as the way the browser shows you information. I can edit the model, just like I could simplify prior. What it gives you is choices for green geometry, which is Preserve Geometry, Obstacle Geometry, which is red. Those are fairly direct in the user interface why they picked green and red. Green is go. Green is starting. Red is stop. Red is no-go.
Starting Shape is yellow, saying, hey, use this as a general guideline. Again, Symmetry Plane, similar to before. Obstacle Geometry Offset means I can grow it larger. But then as I come here to Design Conditions, Structural Constraint is exactly the same. Structural Loading is basically the same. You've got Load Case Attributes and Point Mass as well. But these two are needed.
Then generative says, what are your objectives? What are you trying to do here? Minimize mass, maximize stiffness? And instead of giving you a report of what your safety factor is, it's asking you for the safety factor. And then it goes on to do I want to deal with Modal Frequency, Displacement, Buckling, and so on?
Then it wants to know how will you be manufacturing this? Is it unrestricted? Is it additive? Are you milling this? Are you doing 2.5-axis? Are you doing 3-axis? Are you doing 5-axis? What axes is your tool direction? And/or all of the above, what is some tool diameters that you need here as far as minimum tool diameter, tool shoulder length, head diameter, and so on. Are you doing 2-axis cutting? Are you doing Die Casting? Let me just dock this again. There we go.
Several of these tools didn't exist years ago. When this tool came out, it was only Unrestricted. But you can see that it's adding information as we go and as it learns. Then you want to assign materials. What materials am I adding? Now it defaults sometimes because I used the Additive Material Library last. But if I go to Nonlinear Material or the General Material, it's going to give you everything. And you just drag in the specifics that you need for this test.
And obviously, as we mentioned before, my Pre-check says, whoa, you're not ready. I've got a red error here. And red errors are, we have problems, right? So I've got a Previewer, and then I can Generate. And what this will do, as it similar to shape optimization, it takes the combinations of every material you select and every manufacturing methodology and does tests for each one of them to optimize based on the geometry you're starting with and based on geometry where you can't go, and the load cases, and the structural constraints. And it gives you options upon options that you drive out to reality.
So I'm not going to go in any deeper than that. That's just a quick preview of that side of the coin. Well, thank you all for joining me on my class. I hope you were inspired to at least get the ball moving in Simulation. It's a quite powerful tool and a way to collaborate. Thank you all for joining me and have a good day.