Latest / Mechanical Engineering Made Simple / (#52) Master Structural Design Geometric Stiffness to Composites

Transcript

(#52) Master Structural Design Geometric Stiffness to Composites · Mechanical Engineering Made Simple

• 0:00
  Welcome back, design junkies, to Mechanical Engineering made
• 0:02
  Simle. Our Dee dive today is, well,
• 0:05
  it's basically a master class in structural behavior and material
• 0:08
  science. We're pulling out the critical
• 0:10
  concepts, the formulas, the correction factors, all the
• 0:14
  things engineers use every single day to make sure things
• 0:16
  don't break for the next 15 minutes or so.
• 0:19
  We're essentially giving you a shortcut through an entire
• 0:22
  undergraduate curriculum. We'll cover everything from
• 0:25
  fundamental geometry right up to complex composite failures and
• 0:29
  dynamic loads. Yeah, this is your guide to
• 0:31
  mastering the Y behind stress and strain.
• 0:34
  We'll start with how the basic geometry of a part, it's very
• 0:37
  shape governs its strength. Then we'll move into some
• 0:39
  cutting edge materials and finally wrap up with the really
• 0:42
  tricky business of, you know, dynamic stability.
• 0:44
  OK, let's do it. Let's start at the absolute
• 0:46
  beginning, the foundation of it all.
• 0:49
  When a load hits a beam, it's resistance isn't just about how
• 0:52
  much material is there, it's critically about where that
• 0:55
  material is. We're talking about the physical
• 0:58
  properties of a shape that determine how it resists bending
• 1:01
  and torque. Precisely.
• 1:03
  And that brings us to the moments of inertia.
• 1:07
  These are the absolute cornerstone measures of
• 1:09
  stiffness. The math behind This is why an
• 1:13
  I-beam looks the way it does. Right.
• 1:15
  So instead of just reciting the integral formulas, which you
• 1:18
  know, our listeners can look up, what's the core concept?
• 1:21
  Why does putting material farther away from the center
• 1:23
  suddenly make something so much stiffer?
• 1:26
  It's because the moment of inertia for bending, let's say
• 1:29
  about the X axis, is calculated by integrating the area
• 1:32
  multiplied by the square of its distance from that axis, the
• 1:36
  square squared term. That's the magic.
• 1:38
  It means if you double the distance of the material from
• 1:41
  the neutral center, you actually quadruple its contribution to
• 1:44
  stiffness. Wow, so that's why a flat sheet
• 1:46
  is so flimsy when you bend it one way.
• 1:48
  To turn it on its side. And it becomes incredibly stiff.
• 1:50
  That makes perfect sense. Now if we move from simple
• 1:53
  bending to, say, twisting, we get to the polar moment of
• 1:57
  inertia. And this is one of those really
• 2:00
  satisfying connections in statics J which is absolutely
• 2:04
  essential for analyzing. Portion in shaft is simply the
• 2:07
  sum of those two rectangular moments.
• 2:10
  So J equals IX plus IY. Just that simple.
• 2:13
  It connects the resistance to bending in two directions to the
• 2:17
  resistance to twisting. Beautifully simple.
• 2:20
  And tying into all of this, both stiffness and buckling is the
• 2:23
  radius of gyration. Why is this specific
• 2:26
  calculation, the square root of the ratio of inertia to area, so
• 2:31
  important for an engineer? Because it distills a complex
• 2:34
  shape down into a single representative dimension.
• 2:37
  It's like an effective radius for where the stiffness is
• 2:39
  located. When we analyze column buckling,
• 2:42
  the key metric is the slenderness ratio, right?
• 2:45
  And that ratio depends directly on the radius of gyration.
• 2:48
  It tells you how prone a column is to buckling purely based on
• 2:51
  its geometry, before you even consider what it's made of.
• 2:54
  And all of this hinges on finding the neutral axis, the NA
• 2:57
  correctly. For a simple rectangle, it's
• 2:59
  easy. It's right in the middle.
• 3:00
  Yeah. But for something complex like
• 3:02
  an asymmetrical T section, you have to do the math.
• 3:04
  You absolutely have to. The neutral axis is so critical
• 3:06
  because that's the plane where there's zero strange ring
• 3:09
  bending. For asymmetrical shapes, you
• 3:10
  have to calculate the centroid of the area precisely if you're
• 3:14
  off by even a little bit. Your whole stress calculation is
• 3:17
  wrong. Completely wrong.
• 3:18
  All the subsequent bending stresses depend on the distance
• 3:21
  from that axis. A small error there leads to a
• 3:24
  big and potentially unsafe error in your final design.
• 3:28
  OK. So that's a perfect transition.
• 3:30
  Geometry is king for uniform materials, but what happens when
• 3:35
  the material itself isn't uniform?
• 3:37
  Let's move from simple metal beams to advanced composites.
• 3:40
  We're talking about laminates made of individual plies, all
• 3:43
  stacked at different angles. How on earth do engineers even
• 3:47
  begin to analyze that? With a lot of math, the tool we
• 3:50
  use is called Classical Laminated Plate Theory or CLPT.
• 3:54
  Since each ply is strongest along its fiber direction, it's
• 3:57
  anisotropic. You first have to transform the
• 3:59
  properties of each individual ply into the overall coordinate
• 4:02
  system of the whole laminate. So it's like a coordinate
• 4:04
  transformation for stiffness. Exactly.
• 4:06
  It's conceptually a bit like using Moore's circle for stress
• 4:09
  and strain, but you're doing it for the stiffness properties of
• 4:12
  each layer. And CLPT gives you this
• 4:14
  relationship, the ABC matrices. They connect the forces and
• 4:18
  moments to the strains and curvatures.
• 4:20
  Which ones are the? Let's say the well behaved ones.
• 4:23
  That would be matrix A and matrix D Matrix A is the
• 4:27
  membrane stiffness. It relates in plane loads like
• 4:30
  simple stretching to the in plane strains.
• 4:32
  Straightforward. Matrix D is the bending
• 4:35
  stiffness. It connects the bending moments
• 4:38
  to the curvatures of the plate. Again, exactly what you'd
• 4:40
  expect. But then there's the mysterious
• 4:42
  matrix B. This is the part that I mean.
• 4:44
  It sounds like a designer's absolute nightmare.
• 4:46
  Tell us about this membrane bending coupling.
• 4:49
  B is where things get really, really weird.
• 4:51
  If that matrix B is not zero, your structural behavior is
• 4:54
  coupled, meaning it means a purely in plain load.
• 4:58
  Imagine just pulling straight on a flat sheet can actually cause
• 5:01
  the whole thing to warp and twist out of plain whoa.
• 5:04
  And the reverse is true too. Pure bending can cause it to
• 5:06
  stretch or shrink. That sounds catastrophic.
• 5:10
  I mean, for something like an airplane wing skin, they don't
• 5:13
  want it twisting when it's under tension, right?
• 5:15
  Is that coupling ever useful? Almost never.
• 5:18
  In large structures, it's almost always undesirable because it
• 5:22
  introduces these unpredictable deflections.
• 5:25
  You want clean, predictable behavior.
• 5:27
  So how? Do you get rid of it?
• 5:29
  You design for symmetry the coupling the B matrix.
• 5:32
  It mathematically vanishes if your laminate is symmetric about
• 5:36
  its mid plane. So the layers above the center
• 5:38
  have to mirror the layers below. A simple design rule to avoid a
• 5:41
  huge headache. It is, and for uniform strength
• 5:44
  we often use what are called quasi isotropic layups.
• 5:46
  These are specific stacking sequences, like a balance
• 5:49
  symmetric arrangement of 0 + -, 60 sub S that are designed to
• 5:53
  mimic the uniform properties of metal.
• 5:56
  The stiffness is the same no matter which way you pull on it.
• 5:58
  From total complexity back to engineered simplicity, I love
• 6:01
  it. OK, Next up, pressure vessels
• 6:04
  and pipes. These figures of revolution are
• 6:06
  everywhere. The analysis starts by
• 6:08
  classifying them right thin versus thick.
• 6:10
  That's the first step. The rule of thumb is if the
• 6:13
  radius to thickness ratio r / t is more than about 10, you can
• 6:17
  treat it as a thin shell. And the key assumption there is
• 6:20
  what? The great assumption is that you
• 6:22
  can ignore the stress in the radial direction through the
• 6:25
  thickness and assume the main stresses, the hoop and the
• 6:29
  longitudinal, are uniform through the wall.
• 6:32
  And if you misclassify if you call a shell with an R / T of
• 6:36
  say 5A thin shell. Your stress calculations could
• 6:39
  be dangerously wrong. The failure mode changes for
• 6:42
  thick shells. That radial stress is
• 6:44
  significant, and you need more complex equations to figure out
• 6:48
  the stress gradient across the thickness.
• 6:50
  They fail differently. I saw an empirical formula in
• 6:53
  the source material for bursting pressure that apparently works
• 6:56
  for both thin and thick vessels. How does that work?
• 6:59
  It does. It's a great practical check
• 7:01
  because it essentially averages the effect across the thickness.
• 7:04
  It gives you a reliable upper limit for failure, a bursting
• 7:07
  pressure that sidesteps the complexity of the precise stress
• 7:10
  gradient. But the real headache aren't in
• 7:13
  the main body of the vessel, are they?
• 7:14
  It's where the geometry changes the junctions.
• 7:17
  Oh. Absolutely.
• 7:18
  Joining a cone to a cylinder, for example, is a classic
• 7:22
  computationally brutal problem. You have to use a technique
• 7:25
  called deformation matching. Meaning you force them to line
• 7:29
  up. Exactly.
• 7:30
  You have to ensure that both the slopes and the radial
• 7:33
  deflections are identical where the two parts meet.
• 7:36
  If there's a mismatch, you get huge localized stresses right at
• 7:39
  the joint. That's where it's going to fail.
• 7:41
  And you mentioned there's a specific detail that's easy to
• 7:44
  miss, something about a step in the thickness.
• 7:48
  Yes, say at a Weld or a reinforcement.
• 7:50
  If you have an internal step, the pressure doesn't just push
• 7:53
  out on the walls, it also pushes directly on the face of that
• 7:56
  step. It's an extra axial load.
• 7:59
  I see. You have to add correction terms
• 8:01
  to account for that pressure acting on that extra little ring
• 8:03
  of area. It's a classic subtle mistake
• 8:06
  that leads to failure right at the joint.
• 8:08
  OK, let's zoom in even further. We know that failures almost
• 8:11
  always started these little localized points where stress
• 8:14
  peaks stress concentration. And we quantify this with the
• 8:18
  stress concentration factor KT. It's just the ratio of the
• 8:22
  maximum stress right at the feature, the corner of a hole,
• 8:25
  say to the nominal stress you'd calculate far away from it.
• 8:29
  The fluid flow analogy here is so useful.
• 8:31
  Imagining stresses streamlines of water crowding around a rock.
• 8:36
  What does that analogy teach us about design?
• 8:38
  It leads to a wonderfully counterintuitive design
• 8:41
  solution. Sometimes you can make a part
• 8:44
  stronger by removing material. This sounds completely wrong.
• 8:48
  How can removing material reduce stress?
• 8:50
  Because you're not trying to add bulk, you're trying to smooth
• 8:53
  the path for the stress to flow. The classic example is drilling
• 8:56
  small relief holes next to a large primary hole.
• 9:00
  You're not adding material. You're just giving the stress a
• 9:02
  smoother path to follow. Precisely.
• 9:04
  You're allowing those flow lines to divert more gently spreading
• 9:08
  the code and lowering that PKT factor.
• 9:10
  It's an incredibly elegant design principle.
• 9:13
  So once the part is built, how do engineers actually go out and
• 9:16
  measure these strains in the real world to see if their
• 9:19
  models were right? The workhorse tool for that is
• 9:21
  the strain gauge. It's basically a tiny piece of
• 9:24
  metal foil you bond to the structure.
• 9:27
  When the structure stretches or compresses the foils, electrical
• 9:30
  resistance changes by a tiny amount, which we can measure
• 9:33
  with incredible precision. But that's just in One
• 9:35
  Direction. What if the maximum stress is at
• 9:38
  some weird angle you didn't predict?
• 9:40
  For that you use a strain rosette.
• 9:43
  It's a set of three strain gauges, all arranged on one
• 9:46
  backing, usually at zero, 45 and 90°.
• 9:49
  By taking those three measurements.
• 9:51
  You get enough data to solve for the whole picture.
• 9:53
  You get the full state of strain.
• 9:54
  At that point you can find the principal strains, their
• 9:57
  magnitude and their direction, even if they aren't aligned with
• 10:00
  the parts main axis. And what if you want a full map,
• 10:04
  not just a point measurement? That's where photoelasticity
• 10:06
  comes. It's a beautiful experimental
• 10:09
  method. You build a model out of a
• 10:11
  special plastic, load it up, and view it under polarized light.
• 10:15
  The stress patterns appear as these colorful fringes.
• 10:18
  It gives you a complete visual map of where the stress is high
• 10:21
  and where it's low. We've talked about yielding, but
• 10:23
  there's another failure mode that can happen long before the
• 10:26
  material itself gives up, especially in slender parts.
• 10:30
  Elastic stability. Buckling.
• 10:33
  Buckling is a failure of the structure, not the material.
• 10:36
  The theoretical critical load, the Euler load for a column.
• 10:40
  It assumes a perfectly straight ideal member.
• 10:43
  Which never exists in the real world.
• 10:45
  Never. In reality, that theoretical
• 10:47
  value is almost always an upper limit, because every real part
• 10:51
  has tiny imperfections. And where do we see the most,
• 10:54
  well, the most terrifying gap between that theory and the real
• 10:57
  world? Thin shells.
• 10:58
  It's a classic case where the initial theories just didn't
• 11:01
  match experiments at all. For thin walled tubes under
• 11:04
  axial compression, tests consistently show that the
• 11:07
  actual buckling strength is often only 40 to 60% of what the
• 11:10
  theory predicts. 40 to 60%. That is a massive difference.
• 11:15
  Why are thin shells so much more sensitive to these little
• 11:17
  imperfections? Because they have very little
• 11:19
  geometric stiffness to resist them.
• 11:22
  Any tiny dimple or deviation from a perfect circle creates A
• 11:25
  leverage point for the load to cause a collapse.
• 11:28
  It's a geometric nonlinearity that the simple theory doesn't
• 11:31
  capture. So you have to use a big safety
• 11:33
  factor a. Huge one.
• 11:35
  Conservative design approximations are mandatory.
• 11:37
  You have to significantly de rate the theoretical strength to
• 11:41
  be safe. All right, let's just gears to
• 11:43
  dynamic loading a sudden impact. The rule of thumb everyone hears
• 11:48
  is that a sudden load causes twice the stress of a static
• 11:51
  load. Is that actually true?
• 11:53
  It's a good conservative and therefore safe assumption for
• 11:57
  design, but no, it's not precisely true.
• 11:59
  OK, The analysis shows the actual factor is usually less
• 12:01
  than two. For instance, if the impacting
• 12:04
  mass is 4 times the mass of the bar it hits, the stress factor
• 12:07
  is more like 1.84. Using 2 just guarantees you're
• 12:10
  on the safe side. And one last thing, thermal
• 12:13
  stress. A uniform bar.
• 12:15
  He'll tighten both ends and you heat it up by delta T What
• 12:18
  happens? If the ends can't move, that bar
• 12:21
  is immediately put into compression.
• 12:23
  The stress is simply the change in temperature delta T times the
• 12:27
  materials coefficient of thermal expansion gamma, all multiplied
• 12:31
  by its modulus of elasticity E. It's a powerful reminder of how
• 12:34
  materials fight back against constraints.
• 12:38
  So today we've really covered the whole life cycle of
• 12:40
  mechanical design. We started with the geometric
• 12:42
  constants like moments of inertia, manage material
• 12:44
  complexity with CLPT, we analyzed stress at critical
• 12:47
  junctions, and we finished by ensuring stability under both
• 12:51
  dynamic and static loads. That was a deep and I think
• 12:54
  essential dive. We talked a lot about how
• 12:57
  materials failed, but let me leave you with the final
• 12:58
  provocative thought that comes straight from the data on
• 13:01
  brittle materials. For a beam made of something
• 13:04
  brittle like cast iron, the modulus of rupture, which is the
• 13:07
  calculated bending stress right at failure, is actually higher
• 13:11
  when the maximum moment is concentrated at a single point
• 13:14
  compared to when that same maximum moment is spread out
• 13:16
  under uniform load. So think about that.
• 13:18
  Spreading the zone of maximum stress out even a little bit
• 13:22
  causes the brittle material to appear weaker at the moment it
• 13:25
  breaks. Why would making the peak stress
• 13:27
  zone larger reduce the materials apparent strength?
• 13:31
  Go forth and use these insights to design stronger, smarter
• 13:33
  structures. We'll catch you next time for
• 13:35
  another deep dive.