design practice: designing a tube-chassis in fusion [funsies]
Sooo as we said before, I want to try doing some actual design here in that I want to learn more about FEA, constraints, etc. Starting with hand calc-through computer models. So yesterday I drew out a very basic design for a "lander" that could "land chip on another planet."
Yes silly design exercise but I want to see what's the lightest frame we can make such that the frame can survive the 100m/s - 0m/s deceleration in just 2km. What kind of forces we need to do that and etc... That starts with modeling the frame! And we'll start with 1/2" tubes and the drawing we did yesterday.
We'll also look through researching and choosing a material to make this thing out of. The first thing we need to do is actually enable 3D sketch in Fusion 360. But anyways. We'll start with a 3D sketch, and do it all in inches to real-size, then we can put chip in it as a final to actually see if it fits. We're going to measure chip in folded configuration first. We need 38"x29" of frame. So we create a rectangle of that size centered at the origin of the sketch. It's always good to center the sketch at the origin as I was taught by some of my SEVT team members!
Now we need to build the rest of the box in height. The height is 12" in the drawing, we're going to check that. Turns out it needs to be at least 15" if we fold the antenna down that will work. And now we have a 3D box.
Now let's figure out (just for kicks) if we assume the rover is 90lbs, the frame is an extra "w" pounds, and the thing needs to survive the accelerations above what the minimum thrust from four rocket engines to slow it would need to be. We're going to calculate that, find those rocket engines, and then add them to the CAD to make sure they work. We'll also determine the minimum safe distance from their exhaust. We're going to assume we're landing on EARTH, to increase the forces the frame needs to take.
Here's some very quick math. Now if we wanted to use rocket motors, we'd be looking at in the range of K-Motors: http://www.buyrocketmotors.com/aerotech-k1100t-14a-blue-thunder/
That's NOT an accurate calculation I just wanted to see what kind of size we were talking about so the cad can be slightly accurate at least. Before we do anything else, I do need to figure out how to figure out minimum safe distance from a rocket motor because this isn't helping. I'm looking at this as well: https://www.nar.org/safety-information/model-rocket-safety-code/
Since we're looking at K-motors these safe distances are the launching safe distances, they say nothing about the actual distances equipment can be stored.
But then again, if we look at how these rockets get assembled:
It shouldn't be an issue as long as its a few inches away because the fins are right there and the body is made sometimes of flammable materials so since the flame is so directed it might not be an issue. The reason I'm going through this is because the 18" we gave it in the design might be way too big. But we can also start with the 18 inches. http://www.buyrocketmotors.com/aerotech-k1100t-14a-blue-thunder/
These are 54mm motors and let's say we were to use something like two of them on each pad of our lander. Let's give us a 10" plate to mount them.
Here's the 3D sketch. Let's turn it into a chassis. We can use sweep to make this happen. We're going to start with a 1" OD and 0.75" ID tube profile, and sweep it as far as we can. That gives the wall thickness of 1/8."
I started with the squarish "sub-frames" because they were the easiest to do. It made the most sense to do the basic frames and then like attach them with the weird tubes.
Then we get to the tubes the finish up the frame unit itself. That central box is now completed and the four "rocket mounts" are completed. Now we need to connect all the tubes.
Now as you can see, there are only a few tubes left to connect to make this first version of the frame complete. I want to see in the sketch if we can change the dimensions of certain things and if the tubes will change.
So this is the frame now that I changed the "d" (safe distance from the thrusters) to 10in and not 18in. And it changed properly. I'll also have the CAD read from a parameter file at some point.
Those are the final support bars, and now for one last touch. The actual plates that we mount the rocket thrusters to. We're going to make those 1/8" thick as well.
Ta-da, that's the CAD. Now we need to just make all the dimensions we care about reference a text file or something so that we don't need to edit 400 sketches each time: https://forums.autodesk.com/t5/fusion-360-design-validate/global-variables-equations-and-dimensions/td-p/5306729 here we go. Dimensioning from variables.
We can also label the cad now with some variables that show our major design parameters. "d" is the minimum safe distance from the thrusters, "w" is the width of the storage box, "h" is the height of the box, "L" is the length of the box, "ID" and "OD" are the inner and outer diameters of the tubes, and we're going to add "t" which is the thickness of the plate. All the parameters we mentioned, especially the tube diameters, we're going to change in the model for easy FEA. We're going to start with the following material: http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=MTP641
Titanium 6AI-4V which is a common alloy in aerospace. bc it has high corrosion resistance and a high strength-weight ratio. Remember we want to keep this frame as light as possible while increasing the strength. Let's use CAD to determine the weight of this frame if it's all titanium. And according to fusion 360, the frame is ~28.353545632 KGS which is a number that at the very least, makes sense.
So now that we know the weight of chip ~50kgs (which is what we assumed) and the weight of the frame in this case is ~28kgs. What we can do is calculate the forces that rocket engines would need to provide to slow the craft from 100m/s to 0m/s in that range where the rockets are firing.
So let's say each of the rocket engines individually can produce that kind of force, can the frame handle it? Let's FEA and see! Let's make it 1000N per rocket engine and 1000N for the gravity force on the center of the frame. Remember this is titanium. In this very first, primary simulation we're going to apply a 1000N shear force to each of the plate faces, turn gravity off and apply the 1000N down force to the top/center of the frame. And to apply forces to the top we're going to draw a little circle in the frame like this:
Now the forces are applied as shown:
Now let's mesh, run the FEA, and see what the analysis says. Then what we will do is basically change the ID and OD of the tubes to be as small as possible within a reasonable safety factor. We'll also do some hand-calc to verify the results of the FEA are expected. Maybe we'll do the hand-calc while the FEA is running? We also want to figure out exactly what kind of hand calc/model we want to do.
NOTE: WE GOT A MESHING ERROR "SELF INTERSECTION" WHEN CREATING THE MESH. I think that has to do with how this body was created? I'm not entirely sure. But the other thing we need to fix is we have no structural constraints here. And I think if we want to simulate the whole structure's load bearing capabilities, we will fix the disk we generated to hold the gravity force as a structural constraint for now. What that means is we'll delete the 1000N downwards force as it doesn't apply.
We're going to try running with the meshing issues. And we've set the solver in fusion to adaptively refine the mesh to get better results. I also want to see the fusion generative design results at some point for something like this but anyways let's see what the results of the titanium frame with 1/8" tubes is. We'll fix the meshing issue later. I also want to see if we can inspect the forces and the directions at the end of the simulation.
So I think we were trying too fine of a mesh - 6 hours later we've meshed with a standard 20mm mesh. Yay! Now let's actually solve this thing and hope my computer doesn't explode. And now I think we're going to solve this on the cloud for time sake.
So this simulation shows a minimum safety factor of 0.78, but I'm not seeing where that is in the design. We are going to look at that further. And at least we know where the biggest problem area is. That circular region in the middle. So what we need to do in designing this is since the safety factor is 8+ every where else, what if we made this out of 1/2" diameter tubes instead. We don't need safety factors of 8. And then we should also add diagonal supports from the center to the four corners of the frame. We're going to pull the wall thickness down to 0.039 inches. The OD = 0.5, the ID = 0.461! Let's make those quick edits.
Here's what the new frame looks like - a lot skinnier! Let's see what this FEA says. We're going to run it under the same exact conditions to see if it would survive. Then we'll make the CAD more complete and do even more accurate FEA. We're goin to change the thrust forces to 800N each. And now our wall thickness is super thin, so we're going to drop the mesh size to 1mm (the wall thickness). This is going to be interesting. We've deeply dropped the tubes size. I think we overdid it. I think we'll need a thicker than 1mm tube. It's 800N of force per "wing." I guess the FEA will show us. In the meantime I want to look up titanium tubes: https://store.tmstitanium.com/products/titanium-tubing/
Tubes we can try: 1"OD with 0.05" wall thickness. But let's look at McMaster: https://www.mcmaster.com/metals
Yea the 1", 0.051" seems legit. The 0.75", 0.039" also seems pretty valid. We want to use tub sizes we know are actually available of course. We would also want to simulate with this version of titanium because that of course would change the FEA. We will change that after this solve.
____ I think that tube is too small to mesh ____
Let's try with the 1" OD and the 0.898" ID and see what happens... we'll make the mesh size 5mm this time and keep increasing it until it doesn't fail. It is warning us while meshing the tubes are thing. Yea... don't know how accurate FEA will be that's why we'll back it up with hand-calc. Make that 10mm mesh... jeez.
So because of how thin the parts here are, it might be worth looking into a better simulation technique: https://forums.autodesk.com/t5/tag/SIMSOLID/tg-p/board-id/124 SIMSOLID is a new type of FEA that lets you simulate things better. Worst case, we'll move over to solidworks as fusion seems to not be handling this well at all.
We re-started fusion and set the mesh size to 1mm because the wall thickness is slightly larger than a few mm. yay. now we get geometry error. I think we're going to re-create the simulation study and just start-over.
Okay I think the tubes are just WAY TOO THIN. We have to find something better. Let's make the OD 1in and the thickness 3mm. At least we'll get a solve. We're going to use 10mm mesh size. Let's see if this thing will at least fix that safety factor.
So here's the location of the minimum safety factor. It's so small I can't even see it so I'm going to have to think about how to optimize this design Everything except this one area is completely fine. So what we'll probably do is change the geometry of the circular connecting unit.
Anything else -- it seems like we get a meshing error. So instead of fixing that first -- we just are going to fix circle area. I'm just going to increase the diameter to 5in (not 3in).
So now we see the safety factor increased by the tiniest amount. So we need to re-think this design overall. The safety factor everywhere is more than 8 because of how strong it is, and then there's this one tiny location with a 1.76 safety factor.
Now here's the thing, the forces the frame is experiencing are not as high as what would be experienced (I can't believe I'm saying that). We want to re-think this design a little bit. The reason we want to use TUBE is because otherwise we can't weld something like this frame. We need to think about how we would make something while we're designing something of course.
If this thing keeps doing this, I'm going to make multiple files with different diameters, export them, and import them into solidworks to run simulations. That might work better (and be more accurate).
In your face fusion. I keep getting this weird geometry error that makes no sense. I think we're going to try some hand calc to figure out what the forces are where and what diameters we should go to.
We're going to work on this more later. You can't design a whole chassis in one day.