When modeling a structural component in COMSOL, the type of support you choose—fixed, pinned, roller, or simple contact—can dramatically change your results. Even if you’re only analyzing one part of a larger assembly, its boundary conditions define how it behaves.
Fixed supports allow no rotation or movement.
Pinned supports permit rotation but restrict translation.
Roller supports prevent vertical movement but allow sliding.
Simple (contact) supports mimic a frictionless resting surface—but are the most computationally expensive to solve.
Simulation results show that these supports produce very different displacements and stresses, with fixed being the stiffest and simple contact allowing the most movement. Roller supports often look similar to contact, but they artificially shorten the span as the boundary slides—leading to under-predicted displacement and stress.
Rollers can be useful shortcuts when displacements are small, but for realistic sliding behavior, contact modeling is best—even if it costs more compute time.
In this Tips and Tricks video, you’ll see why it’s important to choose supports that reflect the real world by letting physics-based simulation guide your intuition—not the other way around.
There’s no reason to learn software programs through trial and error, so in addition to our tips and tricks, our upcoming training classes will support you, the COMSOL® user, in utilizing best practices with many of the COMSOL® modules. Please visit our training calendar and find the course that is best for you.
[VIDEO TRANSCRIPTION]
Intelligent selection of structural supports is important
Hi, welcome to this COMSOL Tips and Tricks video on how intelligent selection of structural supports is important.
Four (4) standard support definitions: So what do we mean by structural supports? Well, let’s think of a scenario where you have a structural member represented by this gray beam here that needs to be supported. You don’t want to model the whole system, but you just want to solve for the displacement stresses and structural behavior of just that particular component in the assembly, and so you need to support that. And the black arrows represent the various supports that’s provided by the adjacent structure. And so here on the top left, we see the roller condition that would allow the beam to slide horizontally but not move up and down, so it would restrain the motion vertically and then provide a reaction force coming from the rest of the structure in the vertical direction. The fixed support constraint would provide resistance to horizontal and vertical motion, as well as a moment resistance, so restrict the structural component from rotating as well, and so the reaction force would restrain that rotational motion. The pinned condition would restrain vertical and horizontal motion, but allow rotation of the end of the structural member. And then finally, the simple condition would be a frictionless surface condition where the beam would just be resting – it’s not welded or glued to the support structure, instead it just rests on top of it and it’s free to slide. And in that scenario, you would be giving a vertical reaction force here. Now the roller and the simple condition are, at first glance, appear to be identical. But we’ll see in a few slides how, from a modeling perspective using COMSOL Multiphysics or another analysis package, you could see there is a distinction between the roller and the simple or the contact condition that we’ll examine here in a little bit.
Examples in the wild: Alright, let’s look at some examples in the wild of how something might be needing to have intelligent support constraints. Let’s think of the example of a bridge. Here’s a wooden bridge that has a restraint here where the end of the bridge, if you’re going to do a model of this, it would not be allowed to move vertically or horizontally, neither would it be allowed to rotate because of the strong nature of this fixed condition. And so you would want to model this with a fixed boundary condition.
Another example would be the same bridge resting on the ground. Now, in this scenario you might expect that if you put a load in the center of this, and you stood in the center of the bridge, that the bridge might try to expand horizontally. And depending on the how much friction there was between the asphalt and the corner of the wood here, you might expect maybe a roller condition to be a good approximation of what’s happening here, or the simple condition where it’s allowed to just rest on the surface here where it can move horizontally, but it can’t move vertically.
Here’s an example of a designed bridge, a bridge that was designed to have a roller constraint in it where movement can happen horizontally, but not vertically. And here’s an example of fixed constraint where these beams are locked in here and in the surrounding structure would not allow for motion horizontally, vertically or any rotation.
Old Jenks Arkansas River Bridge in Tulsa: Here’s an example: This Old Jenks Arkansas River Bridge in Tulsa, Oklahoma. You can see the bridge here on the left, and if you zoom in here on the edge of the bridge where it’s mounted into the ground here, has a designed pinned constraint. So this hinge would be able to rotate so there’s no moment reaction, but it would resist vertical and horizontal displacement, provide a reaction force there.
Well, here’s another example where we have a marine bridge and the idea here is that this would be, you know, a simple condition – most likely.
4 standard support definitions: But one of the things that you’ll know from a model perspective is that, ironically, implementing the simple condition here is actually the most complicated numerically and introduces additional computational expense, having to solve for the contact pressures and for very complicated analyses, you, it can be tempting to use a roller support as a surrogate for a contact condition so that you don’t have to solve the nonlinearities associated with, and the difficult, and try to have that overcome convergence difficulties associated with difficult to solve contact problems. So, it can be very attractive and tempting to use a roller condition as a surrogate for the simple condition, and this might be an example where you might want to do that, not having to solve for the contact between the bridge and the soil here, but instead just assume a roller condition which can be a good approximation.
But let’s look at this numerically. I think a lot of the videos that I’ve seen on YouTube, and if you go to your structural mechanics 101 classes, a lot of this, maybe you thought of already, but let’s go ahead and do some simulation work here and use COMSOL to develop some model results and help inform our intuition on this.
Bridge case study schematic: So here’s a span that we’re going to try to cover this chasm distance – from this block to this block, and we’re going to place a bridge across it, alright, and this simple 2D analysis, but we’re going to test out the various structural supports. Here’s a 200-pound load that we’re going to place in the center of the bridge and see how the displacements and stresses are affected by the choice of the structural support type that we pick on the end. Now again, best practice would be to pick the condition that’s most like the real world, and let’s examine what this bridge would look like if we constrained it in these four different ways.
Standard support schemetics: Okay, so here’s the standard support schematics, roller, fixed, pinned, simple.
COMSOL/FEA Implementation: And then here would be the COMSOL/FEA finite element analysis implementation. We would want to implement our roller condition, we could take this point and constrain the displacement just in the Y direction, that’s the vertical direction if you look at the coordinate system here. But we allow movement of the end of the bridge horizontally. Okay, so that would be a roller condition.
A fixed would be taking this boundary where the bridge would be touching the ground and assume that it was welded in here or bolted in all along this surface and that would be, you could constrain the displacement in both directions all along this edge here, and to represent a fixed condition that would restrain the motion horizontally, vertically, but also the rotational effect.
Pinned would look like this where you take this point and restrain the displacement both vertically and the Y and horizontally in the X.
Simple contact condition is more complicated to implement, it requires structural mechanics module and applying source and destination boundaries and then solving for the contact pressures. And in that scenario, you’d want to include the soil region and, or whatever material was mounted, this was mounted to, and then provide your structural constraint further away and then soften that displacement and stress all throughout this region up into the bridge. And again, as I mentioned before, ironically this simple condition is the most complicated to implement and solve numerically so it can often be tempting to use the roller condition as a surrogate, it’s similar.
Rank the support type from 4 to 1: At this time, if you’re up for it, you could take this short quiz to test your intuition before we look at the simulation results and you can pause the video if you’d like a little more time. But the idea here is to rank the support type from four to one, with four being the least amount of movement allowed, one being the most amount of movement allowed. So, if your intuition says that the pinned condition will provide the least amount of movement then you would put that as number four and then go all the way down to number one for the type of support restraint that would allow the most amount of movement. Alright, so I challenge you if you want pause the video now, fill this out if you need a little bit more time and we’ll check the simulation results in a second.
How’s Your Intuition: So let’s go ahead and dive into the results, and we’ll be able to test how was your intuition.
Displacement results: So the first, or the number four placed constraint here that allows the least amount of movement is going to be the fixed condition. So here’s the bridge results for the fixed condition. It’s, the color scale here represents displacement in inches and so we have .124 inches displacement in the center under the fixed condition. The next amount of displacement allowed, so the pinned condition allows more total displacement in the center, right, which that makes sense because rotation is now allowed. So if you look here close at the edge, this point is fixed in the pinned condition, but the edge of the bridge is allowed to rotate up like this, and allowing for more displacement. So if you had number four fixed, number three 3 pinned, you are right so far. This is where it got a little bit hazy in my mind, and the simulations were particularly helpful because the number two amount of displacement allowed is going to be the roller condition. Okay, so this point here is allowed to translate horizontally. The corresponding point on the other side is allowed to move in, to the right, or is this point is allowed to move in, to the left, and you can see we get a peak displaced; quite a bit more displacement when you apply a roller condition. And then finally, the most amount of displacement of this particular bridge, one-inch total, this little demo model would come with a simple condition. This is where we’re solving for the structural contact forces between the edge of the chasm and the undersurface of the bridge. And, so if you got that right, good job – your intuition has been well-tuned.
Quote – Anthony J. D’Angelo: This person Anthony J. D’Angelo said, “Listen to your intuition. It will tell you everything you need to know.” But one thing we like to do at AltaSim here is not do that. We like to rely upon our physics-based modeling expertise and COMSOL Multiphysics to be able to inform our intuitions where they might be wrong with physics-based reality-based insights.
Roller support: And so let’s dive a little bit more into these results about how this bridge would perform under a roller support. And so if you look at the animation here where the load is increasing in the center with time, you can see the bridge sliding here. And if you focus in on one edge of the bridge, there’s a support reaction force that slides to the left. Okay, so the assumed point where the vertical restraint is applied under the roller condition is allowed to slide to the left. So, the red arrow represents the COMSOL Multiphysics results for the reaction, vertical reaction force increasing there, but you see that yellow – the red arrow sliding to the left, away from the corner of the chasm there. And so that’s going to artificially shrink the chasm span that the bridge, numerically, is modeling.
Simple supports: Alright, so this is the exact reason why the simple support is better if that’s the scenario that we have here. So, here, when you solve with the simple contact analysis, it allows that point on the underside of the bridge to slide up. So this is more realistic, so you don’t have that artificial gap reduction that happens.
Four (4 ) support types behave differently: And so when you look at the, comparing the four types of restraints: fixed (at the top), pinned, roller, and simple, you’ll see this gap that’s formed numerically. And so some of the implications here, I know here at AltaSim we’re trying to develop not only accurate results, but fast results. You know, if you’re doing hundreds of different design iterations and the geometry was very complicated, you know, runtimes for structural analysis may be on the order of overnight, half a day, a day; and it can be worth using a roller condition as a surrogate for a simple or contact condition. But if you do that, a word of caution as we’ve just shown it, you’re not being conservative in your displacement stress prediction, right? So, this, the simulation is going to output lower peak displacement than the actual scenario if it’s not an actual roller constraint, but instead, a frictionless surface or surface with friction, where the bridge is allowed to slide along the edge here. Alright, so this effect is going to be amplified with larger and larger displacement magnitudes where the shortening of the chasm is going to affect your results more with the larger displacement magnitude. So maybe you could get away with using a roller condition assuming small displacements, and that’s often a good approach.
Stress comparison: We look at this from a stress perspective, and maybe this is new to your intuition as well, but with the fixed and the pinned conditions ,you end up putting a good bit of stress at the bridge near where the edges of the chasm that you’re trying to span. In contrast with the roller and simple conditions where you’re allowed to slide horizontally, most of the stress, and in fact the stress is almost doubled, or more than doubled here, peak stress prediction. You’re gonna have a lot of stress and strain in the center of the bridge in this scenario – over 50 megapascals. And whereas if you are constraining with a fixed or pinned condition, you have them, a couple regions of concern, potentially both in the center of the bridge where you going to have the most of the tensile stresses on the underside of the bridge, but also where you put these constraints can cause high regions of high stress that you want to be aware of near the edges.
COMSOL Bridge Support Study: So maybe just a couple of words on how to implement this in COMSOL. So here’s a half symmetric bridge that we just showed where the symmetry condition here is in blue. And if we focus in on how we would restrain this to implement it in COMSOL, we would just right-click on “Solid Mechanics,” select “Prescribed Displacement.” Whoops, that’s the boundary condition version; we want the point condition, so I’ll delete that. We’ll select under “Points,” “Prescribe Displacement,” and then we can grab this one point right here.
If we wanted to implement a roller condition, we would select “Prescribed Displacement,” the X direction “Free,” that’s the default. So we want it to be able to move freely horizontally in the X, but in the Y we would select “Prescribed” equal to zero, that’s going to be your roller condition.
If you wanted to do pinned, you could make both the X and the Y zero, like that. If you wanted to do something like a fixed condition, you could restrain the whole boundary here with zero in the X and Y, or you could restrain both of these points with “Free” – very similar to a fixed condition.
To solve the contact analysis, you would need to add in additional geometry for the additional structure; and then I’ll go to form assembly mode and activate the contact condition between these and solve for the displacements and stresses in both the bridge, as well as the support structure, and then solve for the contact pressure that would develop. That’s going to be the most accurate way if you were gonna be modeling this type of bridge structure that was allowed to slide horizontally here, but you could also get away with a pinned condition assuming you don’t have a lot of displacement here in the center.
Summary: So hopefully this has been a helpful video to you. And in summary, trustworthy structural modeling requires wise selection of support constraints. It’s important to understand where you’ve turned your model off and are making boundary condition assumptions, that it’s representative of the real world, and maybe intuition alone is unreliable, contrary to what the Internet gurus might tell us. We recommend here at AltaSim, letting your intuition be trained by physics-based simulation where you’re exercising good practices and picking the structural support type that most matches the real-world scenario, but also taking into account the computational expense. If we’re trying to optimize on a design it can be valuable to be strategic and using a roller condition instead of a full contact or simple condition it can be very computationally expensive. We wouldn’t be able to get as many simulations done in the, within the design timetables. And, if you’re gonna do that, though, be careful, particularly if you have high displacement of your members, you could expect a lot of sliding that would be neglected if you used a roller condition as a surrogate. So if you do that, be aware you’re not being conservative there, you might want to put like a five-percent, ten-percent safety factor on your results just to make sure you don’t run into issues.
Connect with AltaSim Technologies: So if you found this video helpful and want to learn more, please reach out to us if you need additional COMSOL training or consulting support for any of your analysis needs. Thank you.