Simulation in Fusion 360 - Don't Get "Stressed" Out

JAY TEDESCHI: All right, good afternoon. My name's Jay Tedeschi, Senior Technical Marketing Manager. With me today is Mike Smell and Vaclav and, I'm even going to try a hatchet job on your-- How do you pronounce your last name?

VACLAV PRCHLIK: Prchlik.

JAY TEDESCHI: Like that. What we're going to be talking about for the next 90 minutes or so is a simulation in Fusion 360. Just so that I have a good understanding of what you guys, what your level of understanding is, how many of you guys have used simulation in Fusion 360? All right, a couple of you. And so for the rest of you, this is like a new-- you've not used it before, is that safe to say? OK, good.

AUDIENCE: Are you familiar with [INAUDIBLE]?

JAY TEDESCHI: OK cool, that's good to know. All right, so like I just said, we're going to take a look at Fusion 360 simulation capabilities today. We're going to start with static stress which is like the baseline. We'll be discussing some of the common concepts that you'll use regardless of the simulation type that you're doing.

You know, in the real world when I have to simulate something and analyze something, there are physical conditions that exist at a component level or an assembly level, and that is: How this thing constrained? How it is it affixed? How is it loaded? So there are common concepts that we'll go over that are core to simulation regardless of the type that you're doing. So you know we want to make sure you guys have a good understanding of the capabilities of the simulation tools in Fusion 360, and we want to make sure you get the most out of your experience when you're using them.

So the learning objectives for today. Grasp the basics of static stress and learn how to set up a study. Understand when to use modal frequency analysis, so these are the fundamental frequencies of-- Everybody's seen a tuning fork, right? So that's what a tuning fork is. That reaction, that is a that is a fundamental frequency of that design incited by sympathetic vibration and that's modal analysis. These modal frequencies exist in everything we design and it is very helpful to get a good understanding of what stresses, and what reactions they impose on a design when you're working on something. You get comfortable with thermal studies.

So we'll be looking at both thermal and thermal stress because one of the results of a part as it heats, it expands and depending on how that part is affixed and what it's attached to, that expansion causes stress within the design. So these are both very, very important tools to understand how to utilize. And again, I just said. So de-mystify thermal stress and understand why it's important.

So the agenda. Mike is going to go over what's new in simulations of the November release of Fusion 360. Had quite a few new capabilities added into it, and we'll be going over that. I actually, over the last two days, kind of reworked-- We had been working on this presentation for the last couple of months and the stuff came out last week, the November release, and I was like we got 90 minutes, I think we can squeeze this in. So I've kind of reordered things, and we've got a lot of the new capabilities here. We will show you in addition to Mike talking about it. So first up, we're going to talk about linear stress.

I have a structural buckling example. It's not great. We can discuss it. If we need to show it, we will. But I have that on tap if we need to get to it. We'll be doing linear stress first. Then I'll go to motile frequency, and we'll do thermal and thermal stress, then shape optimization. And finally, Vaclav is going to talk about leveraging the power of cloud simulation. So with that, I will turn it over to Mike.

MIKE SMELL: All righty, thanks Jay. So as Jay said, with the November release we've done quite a bit to add to the simulation work space inside of Fusion. Leading up to that, some something you may not have noticed the October 5th release also added a pretty exciting capability into Fusion with how to cloud solve. So we'll start off and talk about a handful of functionality that's really focused on bringing a new level of usability into Fusion.

So I've been working on what was the 11:9 release for about a year, and one of the things that we really wanted to make sure that Fusion users had, as it related to simulation and Fusion, is really a really rich on-boarding and learning experience. We know that some Fusion users aren't super familiar with simulation. We know that in general simulation is often considered as something that is intimidating. So we want to make sure that we give you a very rich environment to learn.

Another productivity enhancement that came along with the 11:9 release is our ability to give you an experience that simplifies the way that you work with setting up contacts and large assemblies. So in the previous version up to 11:9, if you ran the automatic contacts command, you'd get a long list of contacts in the browser.

And it was hard to identify where those were at or if you wanted to change them from bonded to separation, you had to go through that, in a pretty serial process. What we have now is a unified contact manager where we can look at contacts either individually, or we can look at them by the bodies that participate in those pairs. And then we have a cross highlight between working in that contact management dialogue screen and the campus itself. So we can get a really good clear picture of where these contacts are in relation to the model.

The others are around expanding mesh controls, and we had heard about this a lot on the ideas station. Users were asking us for much more control over the mesh that they were creating. So what we've added is the ability to control the mesh at the surface level, at a body level, or at an edge level. So these local mesh controls can be accessed and put on the model in addition to the mesh size and controls that you have for the global part or assembly that you're working in.

The other addition that we've released with the 11:9 release is the ability to do multiple load cases. In a single static stress analysis in the past, if you wanted to look at potentially a tensile load, or a compressive force load, or moment load, all of these things would be different studies. And in that scenario, you would run a study. You would wait for it to complete. You run the next study. You wait for it to complete, and then you have all of your results.

In this environment, we have the ability to create these load cases in the context of a single study. We can then submit that as a simple job locally or to the cloud. Inside of the results environment, we will have a drop down menu right next to the legend bar where we can pick the load case of interest. And then we've also added a compare workspace where we can then look at multiple studies or multiple load cases in a given document and compare them side by side like you see in the image on the screen.

The last thing that we added around productivity and simplification, is bolted connections. So in the past, if we were trying to model a bolted connection in Fusion, you would be limited to modeling that bolt as a physical part. And one, that increases your mesh density. Two, it doesn't allow you to work with pre-loads. So now we have simplified both the connections where we can pick either threaded bolts or screwed-in if want to call it that. Or bolts with nuts where we can define pre-loads, different materials, so that we can actually get a true representation of that bolted connection behavior.

Aside from the productivity improvements, the other side of what we added was really around increasing the capabilities inside of Fusion simulation. So one of the first ones that we're going to talk about is shape optimization. I'm sure everybody here has see that on main stage this morning. So this is a tech preview study type for our Fusion 360 Ultimate users. These studies can be solved on the cloud and what this is going to allow you to do, is start with either an existing design, and go through the process of defining a reduced mass target based on your loads and constraints. Come up with a new shape. Or it's really good for, what we might want to call conceptual engineering, where we're starting with a blank sheet of paper.

So in the image that you see there, my initial shape before coming up with that organic looking bracket was simply an l-shaped extrusion. And based on the load path and the way that the forces are applied the transfer back to the boundary conditions, the shape that you see there is what's predicted.

The other is around structural buckling. And we thought that this is an important study to add as we know that there are a number of Fusion folks working with consumer products that are often thin-walled plastic parts. You heard Steve Hooper talk about the future with sheet metal coming. So we wanted to prepare ourselves to have some analysis types that are going to be very well-suited to the types of things that Fusion's doing.

As Vasha will talk about later in the cloud, we've had asynchronous cloud solving. And this is really important to point out because the nice thing about this, not only can you submit multiple jobs to the cloud at once, this does not consume Fusion. So again, in the case of running jobs locally in Fusion, if you have four or five studies, you run a study. You let it complete. Your run study two. You let it complete, so on and so forth.

In this scenario, once the jobs are submitted into the cloud, we can close Fusion, and disconnect if we want to get on an airplane. Or we could go back in the modeling workspace and go back to making design changes, doing renderings, creating our tool paths, or any of the other associated functionality inside of Fusion. o simulation is not tying up your Fusion experience.

The next around nonlinear static. So again when we start thinking about the types of designs that we see in the Fusion gallery, consumer products, plastic parts, many of those don't behave in a linear way. And we oftentimes have to predict what's going to happen if the material is nonlinear. So if we wanted to simulate a rubber component, naturally we're going to a nonlinear response for that.

And then the last is around event simulation, and this is our first exercise at releasing a non-linear explicit dynamic solver. So what that means in simple terms, is that we can now look at highly nonlinear impacts, complex contact, material failures you see here. We can define at what point the material would break in reality based on maximum principle strain in the model. And we can actually see elements break and separate as you see here.

OK, so this some of the new technology, we'll cover some of it in the presentation that Jay and Vasha will be doing today. Following this class, from 3:00 to 3:30 we'll be doing hands-on exercises in the open lab downstairs. So I would encourage everybody to come down and check that out if you're interested. So Jay, I'll give it back to you.

JAY TEDESCHI: All right, let me just turn that off. All right. I promise I won't burden you guys with too much theory, but there is a lot that we should go over before we dive into the exercises themselves. First and foremost, why do we perform static stress analysis? Typically it's utilized to determine displacement, stresses, and other results resulting from static loads. Does everyone understand what a static load is? It's not dynamic. It's not changing. It's me pushing on this wall with the force of-- I weigh 165 pounds so whatever that resulting force is right there, it's constant, it's static, it doesn't change.

The deformation, the result on the part, does not have significant effect on load direction, load magnitude, or the surface area faces to which the loads are applied. So the part itself is not deforming. And as Mike pointed out, there is a huge difference between linear and non-linear, and I'll get into that with a little single slide. It will be simple for those of you who aren't familiar with simulation, it will be very simple understand. Deformation does not alter the boundary conditions of the part. That is, remember I was explaining before, how something is constrained? What is it attached to? How does it sit? How is it loaded?

So those things don't change and the materials-- again this is getting back to the linear versus non-linear-- the materials behave in a linear manner remain within the elastic region of the material stress strain curve. In other words, stiffness and strength. The material does not change.

So what are we talking about with that? What is linear versus non-linear? How many of you guys when you were kids, did you ever take a paper clip and take the paper clip and then bend it back and forth? Did you notice that it got to a point if you stop before you like actually snapped it, you'd feel it getting really warm in the middle, and that was from friction. But you get to a certain point, if you left it-- Did you ever notice it was it was actually stiffer the next time you tried messing with it? What you actually did was you work hardened that part. OK, as soon as you went past the point where if you let go of the paper clip, it went back to the way it was.

You were in plastic deformation, and that is, on here-- What you're looking at is this is the load applied to a material. OK, this increasing line right here. Now it gets to a point, and that point is 0.2 , and that is the yield strength of the material. And that is where anything past yield, you are now in plus plastic deformation The part is never going to go back to the way it was. It's going to permanently deform. And that region, OK, anything past 0.2 to 0.1, which is the ultimate tensile strength of the part, that's what we refer to as non-linear because the part does not behave in a linear manner any longer.

So for most of you, if you have a part where it's just something that you're not going to destroy. The loads that you're going to see on a component are not going to permanently deform that part, these are the loads that are every day. You know, the suspension of a car takes the weight on each one of the corners, each one of the suspension components takes the weight of the car at that corner, and basically handles it. And the parts are designed using software such as we're about to use right here, to handle those loads on a daily basis. You get in an accident and what happens? All of a sudden, those forces increase, those parts deform, and those are then permanently damaged, and you have to replace them.

But that's when nonlinear analysis becomes important. I need to know, for example, what is going to happen once my load exceeds this point right here. Once these parts start to break, OK, what is the behavior? That's when you need non-linear analysis. And that's what's very important about what Mike was pointing out to you guys about the event simulation and the nonlinear solver is that, you can have a high degree of confidence in what is going to happen to a part that is going to see a load which exceeds the yield strength of that material.

So does everyone understand this chart? Everything we're going to do from here forward in static stress, we're going to be doing everything down in here, in this little square or that little rectangle right there. OK? Any questions? Go ahead.

AUDIENCE: What is the difference between the red line and the blue line?

JAY TEDESCHI: Well, one is stress and the other is strain. So they're two different types of-- You can actually look at these results when you're looking at the actual results when you run an analysis, you can eye-sight [? von Mises, ?] are actually average stresses. You can then take out like, first principle, second principle, and you can actually look at the isolated stresses. OK, I think that chart correct, right?

All right, another thing we're going to talk about, because I am going to use it-- And I highly recommend using this tool, because it's in there. It's easy to use and it gives you a high degree of confidence in the results you're going to get, and that is mesh convergence. OK, now right out of the gate, you are going to have, with a model-- the first step you have to do once you get past putting in the boundary constraints, putting in the loads, is you have to determine the mesh density and the mesh type for a part that you're analyzing.

OK and what you're essentially doing, is you're telling the software these are elements which represent the actual material, the material properties of the type of material you are using, whether it be aluminum, or ABS, or whatever. Those are applied to each one of these little elements and then, essentially equations are used to figure out what the stresses are from node to node and across the node. And those stresses are averaged for each one of the elements.

So if I had essentially-- I've got this chair. OK. And I'm pushing forward on this right here, so there are high stresses right in here right now, right? Does everyone agree? Or if I push it down like that. So if we were to represent this element, this piece of this part right here with big fat maybe five or six elements going up here, we're going to get really quick results. There's not much for the solver to have to deal with from a math perspective, but we're going to get optimistic results, if you will. If I had used a really coarse mesh in here, I would look at this and go, wow, this is going to work great. But how accurate is that going to be? So that's why we have these mesh convergence tools.

So what the mess convergence tool is, it works with adaptive meshing in Fusion 360. And if you enable it, the way it works is, it goes through multiple passes. OK, so the first time through, you can set like a rather coarse-- Like if you look in here, it's kind of hard to see on the slide, but you can see those are rather large elements, OK. And then when you look down into here-- I know it's hard to see. The elements in here down, the fill it, where the higher stresses, are those are significantly smaller.

So what's actually happened here is we went through the first time, I meshed and I said OK, I want like a medium mesh density on this part. But I turned on the adaptive meshing and convergence. So what happens is Fusion goes through the analysis and then looks at the results and says, OK, well, we've got really high stresses in here and in certain areas.

So in those specific areas-- This is what's intelligent about it, and a beautiful part of it, is that it goes in and says in those areas only, I'm going to start making the mesh density greater and greater. So I want smaller mesh elements, which gives me more accurate results. And then over time and you tell it how many how many cycles you want to run it through, over time what it's doing is it's looking for convergence.

Am I getting less and less of a difference in the results between run number one, run number two, run number three? So if you see, for example, if it sees like this line right here, then go up at an angle, and then up even steeper. It says I'm not converging on an answer and it will stop. If it sees convergence, however, if that difference is getting smaller, and smaller, and smaller, then I have a very high degree of confidence that the results that I'm getting are accurate results.

So I recommend using-- Is that a good explanation? I'm always troubled-- This is really important. I mean I've been doing simulation for a long time, having convergence tools it's just such a confidence booster. All right so, let me go over here and stop for a second, and explain what it is we're looking at. Did that pause it? Yes it did.

OK, so what we're doing here, this is a it's a small gas turbine generator. And up at the front, we've got a electric starter motor. So the way this thing works is, this is a gas turbine. There's a generator on the back. There's a second stage turbine that actually drives the generator. This engine needs to be started and has to spin up to about 8,000 RPM to actually reach a point where it's sustainable.

So the way that starter motors typically work they basically engage a shaft at the end of the turbine or the compressor, and they spin that thing up until ignition exists inside the combustion chamber. Once it's running, the starter then has to disengage and release from the shaft. So there's a tiny-- you can see it in there-- there's a tiny little centrifugal clutch. So as soon as that motor engages, that clutch shoots out grabs the end of the turbine shaft or compressor shaft, starts spinning it up. As soon as it gets up to 8,000 RPM and that thing is running, the electric motor stops, and it pulls back in.

So you're seeing a huge torque load on the housing that has to mount this to the rest of the front of the turbine. And that's what we're going to be-- I left that portion in so you can kind of see there's no magic here. It's a standard design using Fusion tools. I used a little airfoil generator that's available free as one of App Store for Fusion. But I wanted you to see before we dove into actually showing the simulation of it, what we're actually doing.

So here I'm exposing the first sketch. And I just used a primitive, excuse me-- Grab the end of the printed wiring board. We'll just create a 26 millimeter diameter by 43 millimeter cylinder. We'll throw a couple of [? fielitz ?] on it. Again I just wanted you to see some background. there has to be some context to what we're doing, and why we're doing it. And the reason we're doing it, so this sketch right there is generated by the free app that I was telling you about-- the airfoil generator.

So once that's done now we're going to create mount point. This is where it's going to interface with the housing at the front of the turbine. We'll throw some rounds on there just to blend it out. The goal of this whole thing is we need a housing which is aerodynamic in nature, but which can resist the torque load that it's going to see when that motor is starting up. That happens very quick and as you can imagine you saw the size of that engine relative to the size of the motor, so there is significant considerable amount of torque, that's going to be applied a twisting load on this housing.

And that's when tools like simulation tools make sense to use. It allows us to explore options as engineers. This is it, a circular pattern. We're going to start off with two of these things. Next step, I'll just extract the outer bodies of the electric motor itself. We'll use those as a sketch. We'll just drive a hole, so we're going to end up with a cavity in this starter housing that is exactly the same size as the starter motor itself.

So we'll do intersect sketch. I picked the body of the electric motor. Boom, that's done. Now we'll just do an extrude symmetric. Punch out the hole, and then we'll begin our analysis. All right, so now let's get into simulations. We'll start our static stress study. Now, first thing we're going to do-- All right, remember I explain this. We're going to check, well actually the first thing I'm going to do is suppress all of the components in the rest of this assembly with the exception of this component.

At the time I did this, I forgot it's much easier to select this component and there's a command in there now that just says suppress everything not selected. If I was shooting this today, that's how I would have done it, but I shot this a while ago, and it just is what it is. So first thing we're going to do is we're going to add some fixed constraints. So these points here where it bolts onto the front of the housing for the turbine. OK so those are going to be my fixed points.

Now, let's go ahead and add a pin. So this thing is actually, this piece of the housing can slide. It's minute, but it can slide in and out, basically as it's being stressed. So to accurately constrain this, we have to allow it to move along that axis. So we'll use a pin constraint to do that.

Finally, we will add a structural load. This is our moment or twisting moment. This is the load that will be imparted when that little tiny electric motor starts. So there is our boundary constraint or boundary environment. Now we're going to go in and set up the mesh. I'm going to use curved parabolic mesh elements because essentially there's a lot of curved faces on this model.

So you note that I haven't turned on the convergence yet. We'll do that later. Right now I'm just trying to get a baseline. What are the stresses that we're seeing on the design as it is without basically complicating anything at this point. So we'll take a look at the stresses first. We're seeing 10 megapascals.

That's not excessive, but what does concern me is when I go and look at the displacement, the displacement is almost half a mil. And like I said, this is a very small-- The engine itself is about the size of a coffee can, so we don't have a lot of leeway. And if we can make a design change that basically didn't impact the design too much, like say changing the array from two to three legs, that probably will stiffen it up considerably.

So I wanted to show you how easy it is. With that physical change, this is-- I mean it's not a considerably big change, but it's a big change or it's a good size change-- And now when we go back into the simulation tool, I literally have nothing to do other than-- m I have to add one single boundary constraint, and that is those fixed constraints that actually mounted to the front. So we'll just go ahead and we'll update the design. We'll add our one last fixed constraint. We'll select the inside face. Nothing else changes. Now all I have to do is rerun the solve, but when I am going to do is turn on convergence.

OK, so we want adaptive mesh refinement. So what it's going to do now is it's going to sit there it's going to iterate. So it's going to go through the local solve or cloud solve. In this case, I was using local. It's going to go through that first one, and then it's going to go through a second one. It's going to compare the two. Do I need to go to a third? And all that's doing is it's reducing the mesh in these high stress areas to get more refined, more accurate results. So our stresses are down, our displacement is down, and we can go in and check the convergence. We're in really good shape. Ah yes let's go to the animation, because it's always nice to--

You work really hard on it, you get to the point, yeah, let me see what it's doing. It's not just a matter of figuring out what the maximum stresses and displacement are going to be. I think I can go to the next, but you guys get the idea right? OK so as Mike pointed out we've thrown a lot of new tools in here. So you heard me a moment ago explaining the boundary conditions, and the boundary conditions are both the constraints. Constraints being the things that we use to tie down a part. You saw me doing it in that last example. These things were fixed. They were actually bolted into location or welded into a location. They can't move whatsoever. The pinned constraint allows for movement.

I can determine whether I want axial movement, tangential movement, so the constraints allow you to describe the environment that the part operates in. As Mike was pointing out, one of the restrictions that we've had was that there was no way of defining pre-load if I had bolted conditions in a set up.

So what bolted connection allows us to do, and we've had bolted connection for a while in the NASTRAM tool, and what it allows you do is it creates a fake-- it's a little placeholder-- But what we're doing is we're describing at those locations, in this case we have an aircraft. This is part of that challenge and grab CAD. So we just grab this existing design, bolted it down, and our alternative before the bolted connection was to actually physically model the bolts, the stack up, that bolts stack up itself, and you still had no way of defining pre-load even if you did that.

And the problem with taking that approach was you had a significantly higher element count in the model, which would take significantly longer to solve. This is a really elegant and beautiful solution because it allows you to virtually describe the bolting condition that exists for holding down a component.

The next tool we're going to take a look at is the compare environment which has been added. And as Mike pointed out this allows you to essentially isolate different load cases and then compare what's going on between, load case one could be static stress. Load case two, I could be looking at strain. Load case three, I could be looking at, I don't know, frequency. All based on the same loads but different behavioral conditions or different results on my model.

So let's take a look at the example. This is an example of both. So first thing we're going to do is look at the bolted connections themselves, and you'll see they're represented here in the browser. Adding a new one, it's up under the constraint dialogue. So we just add-- we go to the connector. We pick the location. You basically pick where the head is going to be. You pick where the other bolt is going to be or the nut, and then you basically describe the environment itself. Am I going to use washers? Am I going to use a pre-load? In this case, we are going to have a torque load on it. Have any of you guys ever working on a motor before? You know how you torque bolts? That's what I'm talking about when I talk about pre-load. That's torque and that basically creates a pre-load condition on that fastener stack up.

OK so once that's done, we'll go to a cloud solve. So this takes place very quickly. Takes place in the background. One of the advantages of the cloud solve, as Mike pointed out, is I'm free to keep working now on something else instead of tying up my own CPU. So magically time has elapsed, and now we're going to look at the results. It does happen really quick. Based on the size of the model. I remember the first time we started doing cloud stuff and a guy wanted me to exit out of a software and then go back in to show that it's still running. By the time I exited out of the software, it was done. I don't know what I'm going to do. Anyway, so here are the results and now that they're in, we will use the compare tool. Did I hit pause or something? There we go.

So now that we have the results on the part we're going to go in and basically check out, well first of all suppress the plate. But now we'll check out the results themselves. We have three different load cases. All right, different loads, different loading conditions, boundary constraints would be the same we have four bolted connections.

But we solved all three of those load cases simultaneously with the cloud. And now we can go into the compare environment and actually arrange all three of these load cases side by side. And one of the advantages of this, is that you can assign sync between all three of the windows. So as I'm rotating and analyzing, looking at all the stresses in one particular area of my component, all three of the windows will rotate and zoom, and anything I do with the navigation wise the component is going to be done to or applied to the other case windows, as well.

OK this is really, really valuable stuff because up until now, you essentially had the one window and you went OK this is load case one. This is lower case two. I frequently would have like a little notepad and I'd rate things down. Just sketch them down so that then I could essentially have a reference that I could look at and say all right, when I'm in the next load case what were the loads in this specific area ? So it's a pretty valuable tool in my opinion.

All right, modal frequency. So we went over the example of the tuning fork. The author of this demo is sitting in the back Jim Byrne joined us a few moments ago. This is a really good example of modal frequency analysis and why you need to do it. So it's from-- have you guys seen the [? sawza ?] demo that we did the data set? It's that green and gray. Everyone knows what I [? sawza, ?] right?

So we made one ourselves, and you know what's going on when you have a motor that is spinning cylindrical and that motion is being converted into a translational force. So there's essentially there are a lot of forces being exerted on the housing that has to hold everything together, and those are going to cause vibrations. And that modal analysis allows us to look at fundamental frequency modes and see how is the part reacting to those vibrations. How is it moving, how is it deforming, at each one of those different fundamental frequencies? And so without further ado, let's go ahead and check that out.

So we're going to simulation. Again this is a different study type. It's going to be modal frequency. It's already selected. We hit OK. And we begin by essentially isolating and this is the technique I didn't use, which was suppress everything except selected. But Jim fortunately uses it.

And essentially for a modal analysis, all you're doing is explaining to the system how the components are analyzing are constrained. There's a lot less work to do. The results are based more on the material properties of the component itself as well as the conditions that exist in the component or assembly that you're analyzing. The degree of freedom tool this is really valuable. I wish I had a better example. This is a good example of it.

If you want, if we have some time at the end, I bet [? generative's ?] thing, that makes a huge difference when it's doing the generative stuff. Anyway once the part is constrained, we use the degree of freedom tool to analyze whether or not my part was fully constrained, parts were partially constrained, and once it's done, I'm able to see the frequency, the results, how this part reacts to the different frequency modes.

So the first mode this thing bounces up and down, and we could we can look at the displacement itself as a result of those modes being applied to the part. OK so now we're refining the design a little bit more. We're going to say we'll add adding frictionless constraints and what those do is they say we can have face to face contact, but we're not going to allow this part to move up and down or side to side anymore.

So you see as a result of that, all the mode one frequency is now up to about 350 Hertz and all that's happening now is it's pivoting about its center as that blade is running in and out. And we could constrain it even further by adding some busses to either side of that housing. So you see how the design is developing? We started off with a loose understanding of what was going to happen. We used the frequency analysis tool to see how the part was going to behave as the saw was operating.

We said, OK, well, we were seeing a lot of motion up the front, so let's add some features into the plastic housing that holds this whole thing. That will hold this in place it. It will restrain it, allow it to move in and out, but you know it basically constrained that. And then we looked at this twisting analysis, on the twisting results on that, and we added these vertical columns which would then fit into features in the housing and the saw itself. Finally, were able to output all of all the results from any analysis and Fusion 360 can always be output to a custom report. Any questions before we move on to the next example?

AUDIENCE: Back when you were talking about the opening, you said you could apply the torque, so what value are you using for clamp loud? Because If I put torque on it, I get 75% of the proof load or the voltage joints?

You said you were applying for pre-load based on your torque?

JAY TEDESCHI: Correct.

AUDIENCE: Well, I've done a lot of studying-- [INAUDIBLE]. You control the statical distribution of [INAUDIBLE]. Most [? books ?] you charge to 75%, 75% is what you're shooting for, but you're going to have some statistical distribution. So, I'm just curious are you using a normal value?

JAY TEDESCHI: We're putting in what you put.

AUDIENCE: So you were just saying so-- based on the torque value. So are you using 75%, 65%.

JAY TEDESCHI: It's the value that you put in. We're not doing [INAUDIBLE].

AUDIENCE: So if you put a torque in it, that doesn't neccesarily give you the exact climate for it. OK, torque gives you the-- in theory, your restrictions factors, all that stuff you're goin to-- right? If I use lubrication, [INAUDIBLE].

JAY TEDESCHI: Yeah, yeah.

AUDIENCE: And so that's what I'm trying to understand, how your methodolody is if I say I have an apppliance of 600 kilograms.

PRESENTER: Yeah, it's right. We don't do any sort of handbook adjustments.

AUDIENCE: So you're assuming it's 75% of the proof load for that type of material.

JAY TEDESCHI: I think it would be 100%.

PRESENTER: We're assuming what you entered in is the torque.

JAY TEDESCHI: Correct.

PRESENTER: It's going to be-- you're calculating the torque, that's what you're saying.

JAY TEDESCHI: That's my understanding, yes.

PRESENTER: What you can't do is [INAUDIBLE].

AUDIENCE: No, no, exactly, exactly. I mean, I would render [INAUDIBLE]. I just-- when I saw that you're using torque, I'm trying to understand what value you're using because if I tell an engineer, hey, just put in what the standard torque is, I need to know what the [INAUDIBLE].

PRESENTER: Yeah, you're right. That's why we [INAUDIBLE] replacement the calculation ourselves.

AUDIENCE: Yes, no, and I [INAUDIBLE] because my struggle is to make sure I achieve a [INAUDIBLE].

PRESENTER: No, that's back there. That may be something that we need to look at, given the user more ability to control how that scale factor is--

AUDIENCE: It's a struggle I have because-- you know, we break [? a lot of holes. ?]

JAY TEDESCHI: What did you-- Jim he's our NASTRAM guy. You've been doing the bottled connection stuff in NASTRAM for a couple of years now--

PRESENTER: And from what I remember, there's more than one option for entering in the free load.

JAY TEDESCHI: Axial and torque is what we've got now in this.

PRESENTER: OK.

JAY TEDESCHI: Was there more than that in the NASTRAM?

PRESENTER: I wish I could help.

JAY TEDESCHI: OK, I was just curious.

PRESENTER: I could try [? to answer ?] You split it. [INAUDIBLE].

JAY TEDESCHI: Like Mike said, that's good feedback. How am I doing on time right? All right, I'm actually on target for once. All right, so the next example we're going take a look at is thermal and thermal stress. So again, I talked about this in the beginning. This is two different things. Thermal analysis is what is, as we have a different thermal environment that a part has to operate in, essentially what temperatures are going to be distributed throughout the design?

OK that first example that I had of the turbine. Part of that is pretty cool up at the front by the engine. You've got fairly close to ambient temperatures, but back aft of that, where that second stage turbine was, you've got 1,300 degree Fahrenheit exhaust gas coming out of there. So there's a fairly wide range of thermal environments that exist in a very small-- thing is only 29 inches from tip to tail.

So thermal analysis software is valuable, point number one, because it allows me to see what the effect is of having those temperatures in that environment are, based on the conductive properties of the materials that they chose for my design. But more importantly, the addition of thermal stress allows me to see as those temperatures are going to affect the components, what is the structural consequences of that?

These parts are going to be expanding. They're going to be changing based on how they are constrained. Those stresses are going to propagate. That expansion is going to propagate through the part based on the constraints scheme or bolting scheme or however that part is bolted in or attached to everything else in the assembly. So this is why this type of simulation is so valuable. It gives you good understanding of A, of what the thermal distribution is going to the be in the design. And B, it let's me see if I'm going to have any negative consequential effects of that thermal expansion.

This is actually an example that Jim put together and it's a really good one. So the first one was just the overall brake assembly itself and right here we're looking at the thermal stresses on the rotor from the effects of the break heating up. And we actually had a really good discussion about this last night. We were talking to Ian Briggs, and one of the options on the [? manone ?] is carbon ceramic brakes.

The problem with the carbon ceramic breaks is that they don't give any feedback, the bite, unless those things are warmed up. Like you drive that thing out of the garage and immediately jumped on it, and expected the brakes to work, you would have a rude awakening because they don't function, period, until they hit a temperature range where the brake pad is actually based on coefficient of friction is actually grabbing something. Now that carbon ceramic is pretty hard when it's cold. I just found it interesting that we're getting these cars now that operate very close to like Formula One racers which are notoriously bad when they're cold.

Anyway so here is the brake assembly. This is actually from one of our customers that was a Formula One team this was years ago. I think I originally did this in Inventor. I think it was an Inventor data set first. Anyway it's a Italian break set we're going do thermal analysis on it first. So let's set up for the thermal run. And we're going to isolate the brake rotor itself. We want to see how the-- Essentially as this brake road or heats up, how is the temperature distributed throughout the rotor?