Up and Running with Autodesk Nastran In-CAD

PRESENTER: [INAUDIBLE] something. Well it didn't use [INAUDIBLE] simulation all the time. So I've got Ilogic driving the simulation.

So to prove a point, I've got these guys here who will claim they haven't used Nastran a lot. So I hope we'll see what happens. So the code is work in progress. That's like a hint. If something happens, it's work in progress. But we'll see what happens.

And just to get the formalities out, guys, this is for people who had started off in life. I mean, simulation. And I see some experts. They might hopefully pick one or two tips.

So it is for beginners. It's for people I'm looking for at least less than a year who have dabbled with Nastran. So if you've got any questions, please fire away. And if it's something simple, I'll answer them. If they're difficult, I've got [INAUDIBLE] guys here who will be happy to answer them.

OK. So why are we all here? So the ultimate goal is to gain confidence. This is a $6 million question. I get many, many customers who sort of ask me lots of questions. Where do you start from? But this is the reason we're all here, how to gain confidence. So I will try my best to see how you guys feel about it.

And the reason for it is that we can start creating great products. But the real reason is we don't tend to over engineer. And maybe you may not claim that you do that, but many companies do over engineer, just to save their own back. And that's normal. So this is what we're trying to avoid.

Now, many, many years ago, maybe some of you guys, when you start thinking about your simulation, one of the frustrations which come across your mind is why the results are always different every time I do something? It's the same model. I change the mesh. Something changes. And the result is not what I expected. So these are lots of questions which come across your mind. And that is something which we will try to unfold.

And if not, you probably end up not bothering with simulation. That's not what we're here for. And you might go back to your old methods, start using hand calculations, or motion [INAUDIBLE] intuition.

So let's just take this point for a moment. I'll come back to all the other questions. Let's take a very simple example. You may have recognized that one from another popular book somewhere.

So if we take hand calculations, guys, the first thing we have to do as engineers, we have to simplify the shape to something like this, like a box section, so that we can start doing hand calculations. OK? So that's the sort of thing we have to assume is [INAUDIBLE]. So that's the first thing we have to assume. And as a comparison, just to prove a point, if we do a simple sanity check with FEA, we'll get similar results. So it does actually give us the correct results, from that point of view.

But the real reason for using simulation, if you still doubted, is we can add a lot more detail into the real stuff. Now, it's a lot more difficult.

Can I use this? Yes. It's a lot more difficult to determine the stresses at hotspots. And that's the real reason a lot of customers, people like yourselves, will be using the simulation to determine these sort of stresses.

And take you one step further, guys. Am I going the wrong way? Yes. This is the real benefit, reducing weight. A bit like the generative design stuff.

And it becomes impossible sometimes to mimic these stress results without trying to break it down into simple shapes. So this is the reason why intuition or hand calculation is not sufficient. And this is the way it's going. And as an example, generative design will have this simulation built into the code so you don't have to think about it.

Now, going back to this, lots of questions on your mind. I don't know. I am assuming everybody has come across this. If they don't expect what they expected, maybe somebody's wrong. Or these stress results keep changing every time I change the mesh.

And the reality is, guys, you already know what to expect. You know the answers. You know where it's going to be. And you might disagree with me, but we are going to do a little quiz here, guys.

So the rules of engagement is I don't want you to shout the answers. I want you to raise your hand. And I'll pick someone. And if you give the correct answer, I'll throw a book at you.

[LAUGHTER]

OK. Is everybody OK with that? Yeah? Right. So please, do not shout the answers out. And if the experts, if you can keep your hands down. If you don't, I'll pick on someone, just like school.

All right. Let's go. OK. Anyone? Hands where the answer is.

So I'm just going to tell you, everybody has an idea where the stress is going to be. A, B, or C. Hands.

You can't have twice the books. So OK. Well, you're going to get one anyway. OK. You then. What's the answer?

AUDIENCE: B.

PRESENTER: Are you sure?

AUDIENCE: Yes.

PRESENTER: Well done. That was easy. It's going to get harder and harder, guys.

OK. How about this one? Come on. Come on. OK. The lady at the back. Where do you think the highest stress is going to [INAUDIBLE]. Here?

A. There you go. I was just testing then. Well done.

Now, this is an interesting one. No hands? Come on. No one? Come on.

AUDIENCE: [INAUDIBLE]

PRESENTER: All right. I've got this one. Right. This is going to get easy. Isn't it? OK. You said A? OK. Let's take another guess.

[LAUGHTER]

Didn't I say not to shout out?

AUDIENCE: [INAUDIBLE]

PRESENTER: Go on then. I'll let you out. You could disagree, but that's right.

OK. Now, then, don't know. Hands up. No shouting. [INAUDIBLE]. Go on.

AUDIENCE: [INAUDIBLE]

PRESENTER: Aw. Let's try it with you first.

AUDIENCE: Oh. I'll try it.

PRESENTER: No. No. [INAUDIBLE]. Just give me an educated guess.

AUDIENCE: A. [INAUDIBLE]

[LAUGHTER]

PRESENTER: Well, the answer was going to be B and C. And what I was hoping for was another answer. But who got it? Are you sure it was you?

AUDIENCE: I was going to say, it's got to be A because B and C are the same.

PRESENTER: They are the same. Yes.

OK. See, more or less.

AUDIENCE: [INAUDIBLE]

PRESENTER: I could have done it as a half model. But the idea was just to give you an idea. But this is a myth, that you guys, when you're designing your own products, you know where the answer is going to be. This software, or any software will give you an idea what the actual stress value is going to be.

OK. Well, well done, guys. I was hoping to give two books away for this one. But since you shouted the answer out, I have to [INAUDIBLE].

OK. Right. So this is the first tip, guys. When you pick any model, simple, complicated, you have-- well, not have to. I would suggest you run three different analyses with three different mesh sizes.

What you're looking for here is that the stress value between this second and third one is within 10%. Why 10%? It's just a number which is being used by a lot of experts. And it has to be left below 10%. It could be 5%. It could be 2%. But someone like myself would ideally want to be able to compare it between the first one and the last one, 10%. That's the ideal scenario.

But between second and third is a good enough. So that's the first tip.

OK. Now, when we've got complicated shapes, so when we've got something very large, it becomes a little bit difficult to do a mesh size across the whole model. So what I'd suggest we do is, well, we know where the stress is going to be. We've already proven that. Well, at least 90% of them have.

We then go and split a face using the split command. I mean, just as an idea, how many people are aware of this split command?

Yeah. It's more like a plastic tool, isn't it? Whenever I've done FEA training, it's very rare that anybody's used a split command to split faces, or even split parts.

So basically use that command. And you can see, this is the area which we're going to split. And we can refine the mesh around that area. OK? Using the local mesh control you can see here. And as a comparison, you can see the results are identical, but with a smaller model.

Because we'll be looking at high stress areas. We know where it's going to be. So you might as well do a mesh control around the area. OK?

This methodology can be applied to any software. It's nothing to do with Nastran, or Fusion. Or it could be any software. It's a very generic process.

OK. Now, this can get confusing. In Nastran we have the ability to look at results based on elements. And elements in Nastran calculate centroidal results and corner results. There's two types of results.

Now, which one's right, which one's wrong, it's personal preferences. But I will use both. And I'll show you what I mean by that.

So let's assume we've got four elements here. Now, what you can see there, that when it calculates these are the four values in the middle of the elements. They are centroidal values.

And in Nastran you have an option to display them. So, for example, in here there is data type, centroidal, and central. I don't know if you've used it.

Very rarely. I don't know. Has anybody changed it from cone to centroidal? By default corner is the default value. What you're looking for here, guys, is the file displays the maximum. You're looking at the maximum values in all the adjoining elements. OK?

And then you flick through the three different options, maximum, minimum, and average. And in an ideal world, in addition to what I showed you before when you run three different analyses is that value, and that value, the maximum, minimum should also be within 10%.

These are sanity checks, if that's the right word. You're looking at different ways to be more comfortable. OK? I mean, three analyses is good enough. But you can also use this as well.

So this is centroidal results. And the next one is corner. And this is your default. So every time you've used Nastran already, corner results are the ones you're already, by default.

And the same thing here. There you go. If I look at this node here, these are the four values. So you'll pick the maximum, minimum, and average.

Again, if you're looking at the corner results, you're looking at values between maximum and minimum also within 10%. So these are all checks you can actually go. And ideally, the difference between average centroidal and average corner should also be within 10%. So these are all the things you can do.

And what if they're not different? The idea is, if you think of a square shape, square shape is central. As they get smaller, and smaller, and smaller, they'll eventually become more or less the same. A square can almost become a dot. So they will become within 10%. OK?

Now, if you thought that was all confusing, a lot of gibberish, this is a visual check to check whether your results are converging.

So what you see at the top there is a stress plot. And what you see at the bottom there is a stress plot again, with this little button switched on. It's not averaging.

And what you're looking for, guys, and ladies, is visually there's no distinct difference. And that also is a good way to check your results are converging. If you thought centroid on corner was too technical, this is another sanity check, guys, you can actually look at.

OK? So there's no right and wrong. They are just simple tools you can actually use to be able to-- go on.

AUDIENCE: [INAUDIBLE]

PRESENTER: You're looking at that value there?

AUDIENCE: [INAUDIBLE]

PRESENTER: Centroidal and corner gone.

AUDIENCE: [INAUDIBLE]

PRESENTER: Yes. Yes. Absolutely. OK.

Right. This is the big killer. Does anybody know what stress singularities are?

AUDIENCE: Errors.

PRESENTER: What was that? Errors? Well, they're not errors. They're a pain in the bubble.

They're a nightmare. And a lot of experts basically will actually learn to ignore them. So this is the reason why it's important that when you're designing something you know exactly what to expect.

And this is basically one of the reasons why a lot of designers, not analysts, designers probably don't want to use that simulation. Because they do everything I've told them. And they still get hotspots everywhere. So how do we get around it? Or we might end up doing that on the computer.

OK. It is frustrating. So what you can do, as a suggestion, is that inside Nastran we have the ability to look at section stress plots. Now, when we get these hotspots or singularities [INAUDIBLE] having a column singular-- there's lots of different words for it-- basically what normally the stress is typically will be on the surfaces of your metal.

And what you're looking at here is how much of that high stress is actually going through the thickness. Now, the question is, where do you stop? It's just basically experience behind it. So this is one way of doing it.

And if you guys use Inventor, there's no way we can actually-- what's the word I'm looking for? When you do a section, you're not meant to, but I've seen CAD guys sectioning an Inventor model, and it becomes hollow, and they get scared. They say, wow, this is wrong. My solid model is now hollow. That's what they've done.

But here, in Nastran, this is one way to check it. And the other one. So I'm going to go through all these little examples live. OK? And the other one I can use is something called fringe stress plots. What this one does is that if you get a hotspot, a stress singularity, it'll be a little dot somewhere in a corner somewhere. And if you want to find out the stresses in the area of interest, when you look at this fringe plot, you can basically change the high values, the maximum value. And it grays out everything. And that's something cool.

In Inventor FEA, you can't grade it out. It always stays there, and you can't get rid of it. Unless you tip X it and you don't want to show it to any customers.

So this one is the one. So I will show you what I mean by that. OK. OK. Now, this one, guys.

I used to get answers, questions. How good are the Nastran results compared to ANSYS? Because all of my customers, our customers, your customers are using ANSYS, because that's been used there for years.

And I used to go back and say, well, why are you comparing it with ANSYS? What if ANSYS is wrong? And they just look at my face, and I've been using it for 15 years.

But what you should be comparing it with is your hand calculations, your intuition. And now my biggest joke is that if you fly, then you should trust them, because a lot of planes are analyzed using Nastran. That's where they came from. And then they go quiet. So if you still don't believe it, I won't fly back to wherever you came from. OK?

On a serious note, guys, with Nastran in CAD, it's installed. There is a massive document. I've seen it, but I've not read through it. It's basically standards on white papers, hand calculations. And it's all there, guys.

Don't compare it with your own model. These are test case examples. Pick them. And then compare the results.

And it came from these guys who tend to know what they're talking about. Well, we think. Rocket scientists.

And as I said, airplanes, most of them, if not all of them, being used in Nastran. And, again, like ANSYS, Abacus, they're all respected solvers out there, unlike our Inventor. I think Inventor uses something else. But now you've got a choice, guys. You've got a respected solver to do it.

OK. Does anybody still not believe that Nastran gives correct answers? I won't take it personal. OK.

All right, guys. So this is the live stuff. Now, basically Inventor FEA is pretty good for what it does. Now, these are the common applications, when customers now have access to Inventor and Nastran. And these are the common applications where they use Nastran for.

So what I'm trying to say is, people still continue using Inventor. Now, the example I'm going to go through is basically this one here. This is a real life example from a customer. And they want to analyze that.

Now, the thing is, I don't want to analyze any [INAUDIBLE] that coming with it. It's just that little yolk there. And this is what I want to go through live. And I will show you the example of all the things I've been talking about.

And then my volunteers, the brave guys, in two examples. One example is where they're going to create a model from scratch. And they're going to go through an Ilogic. it's like a simple form. Click, click, click, click. Three button click. And you have a result, rather than doing it inside Nastran.

And another example would be basically this one here. And just apply a lot of constraints using the [INAUDIBLE] and then get some results. The idea is just basically to have a look at what you can do with Ilogic and simulation working together.

OK. Any questions, anybody, yet? It's all self explanatory? OK. Right.

So I am going to do it live. So I'm going to be brave, cause Inventor never crashes. OK. So that was serious then.

OK. Right. So we've got an example here, guys. It's an assembly example.

Now, what I've got here is a dummy part, because I want to apply a moment load, or a torque, transmitted by the prop shaft. I don't want to analyze everything. So I've got a dummy part in there, which is this one here.

Now, what you can do inside Nastran, I can exclude that. And I can do something clever in Nastran, and use that to transport the moment load. OK? Because I don't have anything appropriate here.

So if I just open that part. So let's have a quick look here. OK. You can see this here. So I've got a single part there. Now, I don't have a cylindrical face here, guys, to be able to apply the load here.

So I have to create something in Nastran to transmit the torque or the moment onto this example here. So now, before I do anything, guys, now we want to simplify this. Now, inside Inventor we have something called direct modeling. So I've done something already.

So if I scroll this down, guys, you can see I've got rid of the little holes. They're non structural. They're something I don't need to care about. Why have I done that is because I want the model to be small. I want to do a quick analysis.

Now, just to give you an idea, guys, I think I'm going to sit down. So let's just do something else here. So this is something called direct modeling. I don't know how many guys use that.

So I'm going to go in here. I'm going to pick this little feature here. And let's just try something else, if I was just going to pick that chamfer here. for example. Well, I just picked it up. It chamfered the edges. Oh, bugger. How can I deselect that one?

OK. I did the right thing. I pick it again. Now, if I press Apply, can you see it? It's got rid of the chamfer. It's little features.

How many guys use this direct modeling stuff? I think it's awesome. OK?

Now, watch this here. Now, can you see this detailed feature inside? There are lots of little grooves and teeth. Now, I'm going to try this. I'm going to see. I will pick that little chamfer there or that feature there.

And OK. I'm going to try my luck. And it won't get rid of it. The only way to get rid of that, guys, is I've got to go and pick all them little features.

Now, this was a step file. So I don't have the history to be able to press it. So what I'm going to show you here is that one way to get rid of this is inside Nastran. I'll show you a technique which I rarely use, but it can be very useful in this case.

So I think I've simplified it to how I want it. So if I get rid of this here and just have a look, click on a section view here, there we go. Can you see here, guys? There we go. OK?

So we got rid of that little feature there. And the outside feature. So I'm happy with this. That's the most simplification I can do here.

So let's go back into a Nastran environment. OK. The first thing I'm going to do here, guys, is the material has already been specified inside the Inventor environment. You can see it's come across there.

I'm just going to change the color of the mesh so that we can all see it. OK. So the first thing I'm going to do here, guys, is because it's painful and tedious for me to go and pick every single face while in a demonstration environment, inside Nastran I've already done it. I've done the hard work. And I'll just drag and drop it to the top here.

And this is really cool, because I can create lots of loads and constraints in the model tree. And I can have 14, 10, 15 sub cases. So I can run it as one analysis.

Now, the next thing here, guys, I got rid of my slave model, for example. So what can I do here to be able to transmit the moment?

Inside Nastran we have special connectors. And there's something called here rigid body, which I think was mentioned this morning in the class as well. I'm going to transfer the load through this face and this face. And then I want the software to create the center point.

This is where I'm going to transmit my load. So if I drag it down to this way here, and that basically represents my rigid body to transmit the load.

So if we go to our loads here, and let's go to my moment, moment, moment. And because I'm from the UK everything is in metric. So I do apologize.

So 3,000 Newton. For some reason it likes capitals. Let's try M. And pick that little point there. And if I just move my cursor over the lower, you can see it's converted it to newtons per millimeter.

OK. So I've done the easy bit. Now, if I go to my mesh settings, let's put a nice round number of six, and generate the mesh. OK.

Now, the first thing I've noticed, guys, you might notice that when you're looking at curved objects, it's sort of done something there. I think it's not colored it. So if I'm going to go and hide my CAD body, it's basically converted my radius fillet into a chamfer.

So the first thing you got to do here is you've got to be able to represent your geometry with the best possible mesh size. So I'm going to go in here. And if you have been using Nastran, or [INAUDIBLE], there's something called here, if I click on Settings, this little button here. And this will force the mesh to follow the curved geometry.

So let's try that and see if I can see the mesh. OK. Now, I'm a little bit nervous. I don't want the thing to fall apart.

There we go. That looks amazing. So OK. It's taking the fillet into account.

Now, going back, I don't want all these little grooves there. I don't want to account for that in my analysis. OK? So what I'm going to do here, guys, inside here you can suppress features, which you do not want the mesh to include.

So if I go into here, 20% a number, and if I now generate the mesh, you'll notice when I get rid of my CAD body it will get rid of all the grooves. OK. So the best way to visualize this, guys, is you go and hide-- there we go.

OK. This looks a bit jagged here. So the way to get rid of that one is three. And let's see if it's any smoother, guys. So you see how it actually got rid of all the other features. Otherwise you have to go back in Inventor and start to pick every single little feature in there.

OK. And that's a little bit nicer there. OK. So I think I'm happy with that, guys. So let's click on the magic button. Unless we don't get any warnings, or errors, anything like that. I don't like warnings.

OK. Now, you see all these warnings? How many have I got altogether? 30.

Has anybody been running Nastran have a book at hundreds, thousands of warnings? Yeah. Well, you see, I've got 37 here. That doesn't mean my results are wrong. It just basically makes me nervous. OK. To put it bluntly.

So what I'm going to do here guys here, forget the results for a moment, in [INAUDIBLE] Nastran let's look at two things here. It's only using two cores. I've got eight on my machine. So by default it'll use two cores. So I'll show you the little parameter which you can actually change.

The thing which worries me-- well, I get a bit nervous-- is a lot of these warnings-- can you see here? Look at all of them. We can switch them off.

So let me show you how. These are two famous parameters. And there are thousands of parameters, guys, in there. So I don't know everything. There's a lot there.

So let's check. So I'm going to type in processor. There we go. That's the one. Change that to eight. And then the other one was warnings. Whoops. Warnings. And then switch that to off.

And let's see what happens now. It's a small model. It'll be the same. But it's not going to be slower. Let's put it that way.

So if I now run that, and when I look at my log file, it should say there are eight rather than two being used. Let's have a look.

OK. Oh, crikey. I still got that little warning there. OK. So let's come back to that in a moment. And if I go here again, let's have a look.

So up here, can you see there? It's using the eight cores now. And it's got rid of all them warnings which it popped up. Now, if you still don't want to see that, I'm not saying I recommend it. But I'll show you how you can switch it off as well.

What I'm doing here, basically I'm relaxing my default settings. That's what I'm doing here. So let me show you what it is. It's using tetrahedral elements. So if I type in something with tet, I'm guessing here.

The reason why I think it's that one there, guys-- I'm just being naive here-- it was 10. So I'm going to change that. I'm going to relax it. It needs a real number. And let's do it once more. And let's see if I get any warnings.

OK. Ah, that's perfect. Spot on. No warnings. Zero. Zero again. Typically you'll get zero, zero with simple shapes. The more complex the shapes are, you'll probably get warnings. And they're probably looking at elements in a lower stress environment that you don't care about.

If you put a very fine mesh across the whole model, it will reduce. That's the idea behind it. If you don't want to switch them parameters off, you have to put a very fine mesh on your model to get rid of them. Or simplify your shape. That's the way it goes.

So I'm happy with that. Now, we need to look at the results I was talking about just before this. So let me get rid of that. OK. Right. So let's have a look.

If I go into edit, so this dialog box here-- so first of all, let's have a look at these centroidal results, guys. This is one of them. So I've got an average value of 12.36. And if I now look at the maximum value it's 12.99. I've got a minimum value here of 11.54. So what you're looking for here, guys, is a difference within 10%.

If it's outside it, what I would do here normally-- well, the good thing is, guys, they're all in the same area. That's the good sign. They're all on that fillet. OK. So if I just cancel this.

If they were outside 10%, let's say for the sake of our discussion I'm going to try and do this. Ah. Where's it gone? Mesh control.

I'm going to reduce the size now. Now, I'm not going to go and reduce it all over the model. The high stress area is around the fillet. I'm going to pick-- oh, see. I can't pick it. It's because I've got rid of CAD bodies. You can only pick CAD geometry.

So I'm going to go in here and pick that. And then let's generate the mesh again. And I'm hoping the difference between the maximum and minimum for centroidals will be within 10%. It will get closer and closer.

OK. So let's run this again. OK. It will take a little bit longer than before because we've got more elements in here.

Ah, see. That makes me feel good. Zero and zero. I like that. OK. Now, let's see if the difference is actually going to get any smaller. So if a look at von Mises.

OK. So let's just have a quick look again. Right. Minimum is 13.12. And I'll show you the maximum. There you go. They're getting closer and closer.

So I'm happy with that. Now, let's try this one. This is where things can get interesting. Now, the corner maximum-- oh, crikey. Now, this is where the frustration comes, guys. It's jumped from around here where I thought the high stress is going to be, somewhere else. But let me come back to it in a moment. Let's do the easy one first. Minimum.

Well, that's a good sign, guys. It's on the fillet. OK? Even the average is on there. So let's have a look what's going on here. Now, the question is, even the average is there. I will take that value. But I'm going to go and look at this maximum for a moment.

And this will happen. Now, I can't see inside. So what I'm going to do here, guys, I'm going to click on the section view. And it always goes the wrong way.

Let's see. Oh, this went the right way actually. You can see there. And if I zoom in, guys, it's because I've got these sharp, little corner edges. And I can't do anything about that.

I can simplify it, but I'm going to be naive here. The six results, all it's telling me it's on the fillet. That's what I'll pick.

Now, if you don't believe that, so what you can do here, guys, I'm going to leave the maximum value for a moment. And this fringe business I was talking about earlier. So if I go into fringe here and pick on contour options here--

OK. Click here. And I want to change the values. Now, I'm going to go and type in-- I think everything was around 1,400-- and display that.

OK. You see there. OK. So I'm going to go ahead and click on probe and see if the value is around 1,400 around here. So if I zoom here, now you got 14 to click on that there. You can see so the value is more or less the same. And that is a really good way to check. I always use the fringe. And that's what I recommend to customers.

If you get hotspots-- I guarantee you'll get them if you have sharp edges-- use the fringe, and then do a probe. And this is the reason why you're looking at the area of interest. And there's no way. I mean, a simple example of an L plate with no fillet in there, that will give you high stresses, as a very simple example, guys.

OK. So I hope that gives you a little bit. Any questions before we have the fun bit? Did that make sense? Or are you confused further?

AUDIENCE: [INAUDIBLE] Maybe I missed why you [INAUDIBLE].

PRESENTER: Why I traded from two to eight?

AUDIENCE: Why you [INAUDIBLE].

PRESENTER: Well, if it was a complicated model, for example, and if it takes, I don't know, a certain amount of time to do it, I'm asking for it to use all the cores.

AUDIENCE: [INAUDIBLE] I get that [INAUDIBLE]. I was curious as to why you would do that [INAUDIBLE]?

PRESENTER: I wanted to show you the option of where it is. Yes. So that's the reason why I did it, guys.

AUDIENCE: That's not a [INAUDIBLE].

PRESENTER: I think it's two by default.

AUDIENCE: So if you go in there and you modify those parameters for a certain case study, does it remember your changes the next time you go in there? And then I saw there was a reset button there. And that resets everything back to default.

PRESENTER: I don't know. [INAUDIBLE]

AUDIENCE: [INAUDIBLE]

PRESENTER: You said yes. And David said no.

AUDIENCE: [INAUDIBLE] your settings in the next [INAUDIBLE]. And you can use the reset [INAUDIBLE] for this model. So if you started a new model, everything would be back to [INAUDIBLE].

[INTERPOSING VOICES]

PRESENTER: You could. There is a way. Yes. Yes.

AUDIENCE: [INAUDIBLE]

PRESENTER: Yes.

OK. There is a way. Yeah. Go on.

AUDIENCE: Is there something like H adaptive meshing? We don't have that in Nastran, but it's [INAUDIBLE]? How does that compare when you were kind of troubleshooting [INAUDIBLE] or is it [INAUDIBLE]?

PRESENTER: You have that ability, for example, in Inventor FEA. And I don't know how many guys use it. Personally, I do it this way. I go and manually change the mesh three times. But if you had it there, you can use it. The trouble is, if you have geometries with very sharp edges, it won't do it.

OK. I think we'll have some fun. Any more questions before I get the volunteers to break my machine? All right. So first up we've got, it's Darren, isn't it? Yes?

Right. OK. So I think [INAUDIBLE]. I don't know how close this one down. [INAUDIBLE]

All right. OK. [INAUDIBLE]. So what Darren's going to do here is he's going to open a new part file. So I'll do the talking. Yeah. OK.

Now, what you'll notice straight away when he is setting the template, you've got this little form here. This is Ilogic form. So we are basically going through all this. Now, in this example here, we are going to only do one analysis, one sub case. But the real beauty of this method is that you can set three analyses, like the thing I was talking about, as many sub cases as you want. And then hopefully press on button, and it'll go and run all three, rather than trying to do it one by one inside Nastran.

So what Dan's going to do here, he's going to create a very simple shape. No pressure here. So it'll basically be a simple plate with a hole inside it. And what I want to demonstrate here is that every time you open a new part file, guys, the form comes up. And then we'll get Andrew to open an existing file to be able to use the form.

Now, this one here, we've spent about two, three weeks trying to experiment with Ilogic, but the potential of it is phenomenal. You can get the whole process.

This [INAUDIBLE] thing you're talking about, we can basically build it using the Ilogic stuff here. OK. And it's not a sketch. OK.

How many guys actually use Ilogic for their general CAD work? I mean, can you use Ilogic as well, or get someone else to do it? You Ilogic coders as well? OK. OK.

Fantastic. They're almost done. OK. We got a simple shape there, guys. So we're going to do one analysis, one sub case. If you go to the constraints button there, Darren, on the form. Fantastic.

So we're going to click on Add First Constraint. And we're going to pick one of the flat faces. Fantastic.

And as you can see there, you can have multiple sub cases. So that constraint can be applied to all of them. In this case, if you press No, we've just got one sub case. Press OK. And then go to the loads button on the form again, next to constraints.

OK. And if you put the load on the other side, the opposite side-- so if you spin it around it'll be easier. I don't have the fancy space [INAUDIBLE] if you used it. There and spin it around. There we go.

Doesn't matter if it's wrong. It'll be a different face. OK. So if you click on Add Load and pick that face. OK. And now, you have two options here. You can have as many options as you want. But we can apply it normal to the face. Or we can actually specify the X, Y, and Z orientations.

We're here. We'll put in the default options. So we press OK and then specify a value. It's in newtons. You can have it set in pounds. OK. Anything. 10, 100, 1,000, whatever you feel like.

AUDIENCE: Does it grab the values from the default units [INAUDIBLE]?

PRESENTER: Yeah. Press OK. And we are ready to run.

Let's see. So if you click on the Run button, Run Simulation on the form-- down. There you go. Oh, up. Up. Up. Up. Up. Up. Above the logo. There we go.

First time I've seen it. OK. Now, it'll crash. OK. It's a new file. So if you go and save the file in the default directory where it's going to be, just call it what you like, Darren.

OK. Right. Click on run again. Just being as naive as possible. Oh, crikey. No mesh size. It's already built into the system. But it won't let you run the analysis without specifying. So you press OK.

Now, that was 10 mil thick just put 2 in there. And there you go. So this will fire off now.

What I was saying is you can have three analyses, multiple sub cases. And runs it. And then it goes straight into Nastran. And whether we can control parameters, I don't know.

And guess what. All zero warnings as well. Simple shape. Press OK. And we have the results. And then from here you can start interrogating the results.

Any thoughts, guys? Well, thank you very much, Darren.

[APPLAUSE]

You've been very brave. There you go, mate. You deserve one of these.

AUDIENCE: Thank you.

PRESENTER: Right, Andrew. Your turn now.

Yeah. Let me open it. There's another example [INAUDIBLE] there.

OK. So what we got here, guys, is an existing file already opened. And if you go to the button next to it in the model where it says Ilogic, it's built in. So if you click on that, and if you go and press the bottom button there where it says clear or reset, keep going. That's it.

Now, this example, I'm going to create two sub cases. OK. So go into a sub cases button. No. Where it says next to analysis. That's it. And then click on Add Sub Case.

And let's just check if it's created one and two. There you go. Two sub cases. OK. Now, go to the constraints. And if you add face constraint, and on the back face there.

Yes. This time we're going to have the constraint specify to all sub cases. And then if you go to loads button here, right, let's add a load on to the top face there, that little box there. Let's pull it up.

So that's fine. Normal to face. 100, as an example. I say OK. And then change the sub case to edit to number two.

Oh, no. No. No. Go back to load. That's it. And add load. And then put it on that flat face. And that's OK. And let's say minus 100. So it's going to go in the opposite direction, as an example.

OK. If you go click on Run again, let's see what it does. OK. No mesh specified. That's fine.

Don't make it too small. 1.5. And what I'm looking for here, guys, when he presses run we're going to have two sub cases.

So in Nastran I would have to create a new sub case, and then move and apply the loads. Now, I think this is useful when you got 15 sub cases or 10 sub cases. People have multiple load cases.

And we'll see what warnings we get. OK. OK. We got 43 warnings. Ignore them. OK.

OK. So at the top when it comes up, can you see there where it says under sub case two on the ribbon? Andrew? Where it says Previous? Can you see it?

No. No. Press the Previous button, underneath it. Underneath it. That. Yeah. You can see the two different results. There you go. OK.

So that basically gives you an idea of what you guys can achieve with Ilogic. This is basically two weeks of work. Got my friend to do it in his evenings and weekends.

Everyone OK [INAUDIBLE]. Yeah? Found it all right? So the potential is that-- I'm not an Ilogic guy, guys. So don't ask me too many questions. It was given to me.

[LAUGHTER]

AUDIENCE: So just a few Nastran questions [INAUDIBLE]. Does the sub case meaning I could load it this way, but another case, because they're not combined?

PRESENTER: No. They'll be separate sub cases.

AUDIENCE: OK.

PRESENTER: So if I had the two loads in the same sub case, then they'll be combined. But if they are separate, they'll be separate sub cases.

AUDIENCE: OK. Thank you.

PRESENTER: OK. Thank you very much, Andrew.

[APPLAUSE]

Before you go away. There you go. Thank you very much.

OK. So OK. There you go. So are we more confident? Did we learn something new, even the experts?