Latest / Mechanical Engineering Made Simple / (#59) Flat Plate Stress and Failure Rules

Transcript

(#59) Flat Plate Stress and Failure Rules · Mechanical Engineering Made Simple

• 0:00
  Welcome back, design junkies, to mechanical engineering made
• 0:02
  simple. Today we are moving flat.
• 0:05
  We've talked about linear beams and vertical columns, but now
• 0:08
  we're diving into the world of flat plates.
• 0:10
  Right. And this stuff provides the
• 0:13
  fundamental mathematical skeleton you need to predict
• 0:15
  stress and deflection in things that are, you know, wide and
• 0:19
  thin. Exactly.
• 0:21
  Things like floor slabs, pressure vessel caps, or even
• 0:24
  something as small as a circuit board.
• 0:26
  That sounds deceptively simple though.
• 0:28
  I mean, a plate is just a flat piece of material, right?
• 0:31
  But the math gets complicated really fast because the stress
• 0:34
  isn't just running along 1 axis. It's biaxial.
• 0:37
  And our mission today is really to pull out the core insights,
• 0:41
  the formulas, sure, but more importantly, the conditions that
• 0:44
  tell you when you can even use those formulas.
• 0:46
  It's a fine print, the fine print, because these
• 0:48
  foundational formulas for flat plates, they are not universal
• 0:52
  truths. They rely on very specific
• 0:55
  linear assumptions. And if you break those rules,
• 0:58
  the math just it falls apart instantly.
• 1:00
  So why do we always start with these assumptions?
• 1:02
  If you know real world plates are never perfectly flat or
• 1:05
  uniform, how much error are we baking in right from the start?
• 1:09
  It's all about setting a reliable baseline.
• 1:12
  We start with these ideal conditions.
• 1:14
  We can call them the sacred 4 assumptions to make sure the
• 1:17
  math is actually solvable with linear elasticity.
• 1:21
  OK, so what's rule #1? Rule 1 is that baseline the
• 1:24
  plate is perfectly flat, it has a uniform thickness, and it's
• 1:27
  made of a homogeneous isotropic material, meaning it acts the
• 1:31
  same in every direction. Simple enough for a textbook,
• 1:34
  but in the real world materials have flaws.
• 1:36
  And that's why we have safety factors.
• 1:38
  But the next two assumptions are structural constraints you
• 1:40
  really can't ignore #2 the plate has to be relatively slender.
• 1:46
  What does slender mean in this context?
• 1:48
  Its thickness, let's call it T, can't be more than about 1/4 of
• 1:52
  its smallest width or length. If it gets too chunky, sheer
• 1:55
  effects start to take over. And this simple bending theory
• 1:58
  is well, it's just wrong. Slenderness is key and the third
• 2:02
  one that has to do with movement.
• 2:03
  Yes, maximum deflection has to be small, and crucially, that
• 2:09
  deflection should be no more than about half the plates
• 2:12
  thickness. Half the thickness.
• 2:13
  If it starts to sag too much if the deflection is greater than t
• 2:17
  / 2, you get these non linear effects and you have to switch
• 2:21
  to a whole different method of analysis.
• 2:23
  That half thickness rule seems like a critical dividing line we
• 2:26
  need to keep in our heads. It is, and the 4th is the
• 2:29
  classic disclaimer for all this stuff.
• 2:32
  Stresses have to stay within the elastic limit.
• 2:35
  We're talking about deformation that bounces back if the
• 2:38
  material yields. All bets are off for these
• 2:40
  simple formulas. Got it.
• 2:42
  Now let's think about how a plate actually carries a load.
• 2:45
  When it deflects, where does all that stress actually go?
• 2:49
  Well, when a plate bends, you can imagine the material on the
• 2:52
  top surface is being stretched. That's tension, and the bottom
• 2:55
  is compressed. But there's a plane right in the
• 2:57
  middle the middle surface, which remains entirely unstressed.
• 3:00
  The neutral axis of the plate. Exactly.
• 3:03
  It's a fundamental concept. So the stress isn't uniform
• 3:06
  through the thickness at all. It's zero in the very middle and
• 3:09
  then highest on the outside. Correct.
• 3:12
  The stresses are biaxial, acting in two directions on the plane
• 3:15
  of the plate, and their intensity increases linearly as
• 3:19
  you move away from that neutral middle surface.
• 3:22
  This means your maximum stresses are always always at the outer
• 3:27
  surfaces. Which is, of course, where it's
• 3:29
  most likely to fail. Precisely.
• 3:31
  OK, let's move into the messy world of real life loading.
• 3:34
  One huge challenge is dealing with concentrated or point
• 3:37
  loads. I mean, mathematically, if you
• 3:39
  apply a force to an area that's almost 0, you get what?
• 3:43
  An infinite stress. You do the math gives you an
• 3:46
  impossible infinite stress value, which we know can't
• 3:49
  happen. And this is where practical
• 3:51
  engineering steps in with a really clever fix.
• 3:54
  The equivalent radius. Equivalent radius.
• 3:56
  It's A to sort of smooth out that impossible stress
• 3:58
  singularity. OK, so without just reciting the
• 4:00
  formula, what's the concept? How does this solve the problem?
• 4:03
  It basically replaces the actual tiny loaded radius, let's call
• 4:08
  it R sub O, with an artificial slightly larger value.
• 4:11
  Let's call that one R sub zero. This larger value is
• 4:15
  mathematically derived to give you a finite realistic maximum
• 4:20
  stress. So the correction factor depends
• 4:22
  on the plate thickness T heavily.
• 4:24
  So if I have a really concentrated load, the math
• 4:27
  pretends it's spread over a slightly wider circle, and the
• 4:30
  size of that circle depends on how thick the plate is.
• 4:33
  It just keeps the calculation in the real world.
• 4:35
  You've got it if your actual load area.
• 4:37
  Your R sub O is very small, and the rule of thumb is less than
• 4:41
  about half the thickness. Then you use this larger
• 4:44
  equivalent radius. If the load is already spread
• 4:46
  out over a bigger area, you just use the actual value.
• 4:48
  It turns a theoretical impossibility into something you
• 4:51
  can actually calculate. Now, Speaking of theory versus
• 4:54
  reality, we have to talk about edge conditions.
• 4:56
  You know, free, guided, simply supported, or fixed.
• 5:01
  We all love that fixed condition in our models because it gives
• 5:04
  you maximum stiffness. And this is where the source
• 5:06
  material gives a huge, huge word of caution.
• 5:09
  The theory assumes perfect rigidity at a fixed edge, but
• 5:14
  the text is explicit. No exact edge condition is
• 5:17
  likely to be realized in ordinary construction, and true
• 5:21
  edge fixity is especially hard to achieve.
• 5:24
  Just think about welding or bolting something.
• 5:26
  There's always going to be some tiny rotation, some yield.
• 5:29
  And the consequences of that are massive, even a slight amount of
• 5:32
  yielding at what you thought was a fixed edge.
• 5:35
  It greatly relieves the intense bending stresses right there at
• 5:39
  the edge. So that stress doesn't just
• 5:40
  disappear, it has to go somewhere else.
• 5:42
  It gets redistributed inward, correct?
• 5:44
  That relieve edge stress is transferred and you get this
• 5:48
  crucial trade off, an increase in deflection and much higher
• 5:51
  stresses at the center of the plate.
• 5:52
  So if I designed for say 15,000 say I at the fixed edge, but the
• 5:56
  Weld isn't perfect, I might be dangerously under designing the
• 6:02
  center of the plate. That is the critical risk, and
• 6:06
  the pragmatic advice for engineers is clear.
• 6:09
  When you're designing A nominally fixed edge plate, you
• 6:12
  should probably design for somewhat higher center stresses
• 6:16
  than the pure theory suggests. Build in a safety factor for the
• 6:19
  center. Exactly just to accommodate that
• 6:22
  inevitable real world yield at the boundaries all.
• 6:25
  Right, let's pivot to circular plates.
• 6:27
  They're everywhere, but analyzing them requires this.
• 6:31
  This specialized, almost intimidating mathematical
• 6:34
  toolkit is full of constants and functions.
• 6:36
  It looks formidable, absolutely, but the reason for the
• 6:39
  complexity isn't to punish anyone, it's to generalize the
• 6:42
  solution. These equations use a system of
• 6:45
  interchangeable coefficients, like numerical placeholders to
• 6:48
  cover a massive number of combinations.
• 6:50
  Annular plates. Solid plates.
• 6:52
  Right. Different kinds of loading,
• 6:53
  different boundary conditions. So instead of a unique complex
• 6:56
  equation for every single scenario, they made one master
• 7:00
  equation that's modular. You just swap out the
• 7:02
  coefficient values from a table depending on your specific case.
• 7:04
  It's like a lookup architecture. Precisely.
• 7:07
  It allows decades of complex analysis to be boiled down into
• 7:11
  a compact set of tables, and this system lets you get
• 7:14
  creative. If you have a complex load that
• 7:16
  doesn't fit a standard case, you could often approximate it by
• 7:20
  fitting A simpler curve to it, and then Then you break the
• 7:23
  problem into smaller pieces, solve each one using the
• 7:26
  standard cases from the table, and then add the results up
• 7:29
  using superposition. Classic engineering problem
• 7:32
  solving it is. I also find the specialized
• 7:34
  cases fascinating, the ones that are about forcing a lack of
• 7:38
  flatness from the outside. Oh.
• 7:39
  Yeah, those are vital for modeling assembly forces.
• 7:42
  For example, some cases handle a manufactured change in slope,
• 7:46
  which is what you need if you know a specific boundary
• 7:48
  rotation is being applied. So it's not about a load, it's
• 7:51
  about a forced deformation. Exactly.
• 7:53
  Or another case deals with a small manufactured step that's
• 7:56
  forced to fit a boundary condition, and it tells you the
• 7:59
  reactions needed to make that happen.
• 8:01
  Let's try to ground all this theory in a practical example.
• 8:03
  Let's take a solid circular steel plate.
• 8:06
  It's simply supported along its edge and it's got a uniform
• 8:09
  load, maybe from fluid pressure A.
• 8:11
  Classic scenario. So let's say we have a steel
• 8:13
  plate that's 20 inches in diameter, so it's a 10 inch
• 8:17
  radius. It's pretty thin, only 1/5 of an
• 8:19
  inch thick, and it has a uniform load of £3 per square inch.
• 8:24
  OK, we know the properties for steel high modulus of elasticity
• 8:28
  E and A given Poisson's ratio. What do the formulas spit out?
• 8:34
  Well, the first thing we calculate is the plate constant
• 8:36
  D, which is all about the stiffness of the plates geometry
• 8:39
  and material. Once you have D, you can find
• 8:41
  the deflection. For this case, the center
• 8:43
  deflection comes out to be about -0.083 inches.
• 8:48
  So under that pressure, the 20 inch plate sags by less than
• 8:51
  1/10 of an inch. That's a pretty small
• 8:53
  displacement. It is, which is good.
• 8:55
  It confirms we're operating well within that small deflection
• 8:58
  assumption we talked about. OK, what about the stress?
• 9:00
  The formulas give you a maximum moment and Max and from that we
• 9:04
  find the maximum stress at the outer surface.
• 9:06
  Sigma Max is 9240 lbs per square inch. 9240 CSI.
• 9:13
  So for a common steel, the yield strength might be what, 36,000
• 9:17
  SI, SI or even higher? Exactly so.
• 9:19
  The point is that this calculated stress is quite low
• 9:23
  compared to what the material can handle.
• 9:25
  The numbers prove the design has a large safety margin built
• 9:29
  right in. All those complex functions just
• 9:31
  collapse back into a clean, predictable set of numbers.
• 9:35
  And that's why we trust the tables.
• 9:37
  Moving beyond just basic bending, we should probably talk
• 9:40
  about something that's often ignored in thin plates.
• 9:43
  Shear deflection. Right.
• 9:45
  You mentioned earlier that shear is usually ignored because its
• 9:48
  contribution is tiny compared to the deflection from bending.
• 9:51
  It is for thin plates, but for thicker plates, especially
• 9:54
  annular plates, you know, the ones with big holes in the
• 9:57
  middle, shear can become a pretty noteworthy percentage of
• 10:00
  the total displacement. We need a clear threshold for
• 10:03
  when to worry about it. And what is that tipping point?
• 10:06
  If the thickness T is greater than 1/3 of the difference
• 10:09
  between the inner and outer diameters for a simply supported
• 10:11
  plate, you have to include sheer deflection.
• 10:14
  And if the edges are fixed, that threshold gets even smaller,
• 10:17
  doesn't it? Significantly smaller, yeah.
• 10:19
  If the edges are fixed, you have to include sheer deflection if T
• 10:24
  is greater than just one sixth of that difference.
• 10:26
  Which makes sense because the edge fixity is already
• 10:29
  restricting the bending, so the sheer part of the deflection
• 10:32
  becomes a larger piece of the pie.
• 10:34
  That's a great way to put it. It becomes non negligible much
• 10:37
  sooner. OK, let's turn to some
• 10:39
  specialized materials. What about bimetallic plates?
• 10:43
  You laminate two different materials together and you've
• 10:46
  instantly got a non uniform problem.
• 10:48
  A sandwich plate, yeah, it introduces complexity right
• 10:50
  away. If the croissants ratios of the
• 10:53
  two materials are really different, the plate develops
• 10:56
  this internal in plane strain that just complicates
• 11:00
  everything. So how do you handle that?
• 11:01
  How do you simplify it? To manage it, we calculate an
• 11:04
  equivalent plate stiffness and an equivalent Poisson's ratio.
• 11:08
  It's a method that lets you treat the layered plate as if it
• 11:11
  were a single homogeneous plate with slightly tweaked
• 11:14
  properties. It's a mathematical compromise
• 11:16
  to keep the design practical. It is, but here's the really
• 11:20
  crucial insight about bimetallic plates, and it has to do with
• 11:23
  heat. If you take a standard single
• 11:26
  material plate and you heat it up uniformly, it just expands
• 11:31
  stress free. As long as the edges aren't
• 11:32
  restrained, yeah, it just gets bigger.
• 11:34
  But a bimetallic plate, because the two materials have different
• 11:37
  coefficients of thermal expansion, a uniform temperature
• 11:41
  change will cause flexural stresses even with no external
• 11:44
  restraints at all. The material mismatch forces it
• 11:48
  to try and curve. Exactly.
• 11:49
  Temperature change itself becomes a form of self loading.
• 11:52
  You have to design for those thermal curves.
• 11:54
  That's a fundamental shift in how you have to think about the
• 11:57
  design. Completely.
• 11:58
  Let's go back to that boundary condition we started with.
• 12:01
  The linear theory fails when deflection gets bigger than
• 12:04
  about half the thickness. When we break that rule, what is
• 12:07
  actually happening inside the plate?
• 12:09
  We shift into the domain of nonlinear behavior in thin
• 12:13
  plates. When deflection goes past that
• 12:16
  half T threshold, the plates middle surface starts to get
• 12:19
  strained. It's not neutral anymore, it's
• 12:21
  stretching. So it activates a new load
• 12:24
  carrying mechanism. It does.
• 12:26
  It's called diaphragm stress or direct stress.
• 12:29
  So the plate stops acting just like a stiff bending member and
• 12:32
  starts acting more like a like a drum head or a trampoline that's
• 12:36
  stretched in tension. That is the perfect analogy.
• 12:39
  It begins to carry the load partly through membrane tension,
• 12:42
  and the result is the plate becomes significantly stiffer
• 12:46
  than the simple linear theory would have predicted.
• 12:48
  And the important engineering take away is that for a given
• 12:51
  load, the stresses will actually be less than what the simple
• 12:54
  linear theory suggests. Yes, the linear math is
• 12:57
  conservative at large deflections.
• 13:00
  To solve these problems accurately, you have to use
• 13:02
  these complex empirical nonlinear formulas.
• 13:06
  They have a separate coefficients for the bending
• 13:08
  stress and for that new diaphragm stress.
• 13:10
  You have to account for both mechanisms separately.
• 13:12
  You do for any kind of real accuracy.
• 13:14
  OK, let's look at the absolute end of the line.
• 13:17
  Ultimate strength and plate collapse.
• 13:18
  This is past recoverable deformation.
• 13:21
  Right. For brittle materials, failure
• 13:23
  is sudden. It happens when the maximum
• 13:25
  tensile stress hits the materials ultimate strength
• 13:28
  snap. But for ductile materials like
• 13:30
  steel, it's different. For ductile materials, failure
• 13:34
  is defined by excessive plastic deflection.
• 13:37
  It's the point where the structure just collapses.
• 13:41
  And there's a pretty clean formula for predicting that
• 13:43
  ultimate load for a common case, isn't there?
• 13:45
  There is for a circular plate with a uniform load and simply
• 13:49
  supported edges. The ultimate load W sub U
• 13:52
  depends only on the materials, yield, strength, Sigma and the
• 13:56
  plate's thickness squared. It's a clean equation that
• 14:00
  defines the catastrophic limit. It bypasses all the complexities
• 14:03
  of elasticity. Entirely, it's just about the
• 14:06
  materials yield point and how thick it is.
• 14:08
  This has been a really necessary deep dive into the fundamentals
• 14:11
  of flat plates. I mean, we covered the
• 14:13
  foundational constraints, small deflection, uniform material,
• 14:17
  and the practical need for things like the equivalent
• 14:19
  radius for point loads. And we highlighted some huge
• 14:22
  real world challenges, like the fact that true edge fixity is
• 14:25
  almost impossible, which forces engineers to design for higher
• 14:29
  central stresses. And we kept coming back to that
• 14:31
  critical dividing line the moment deflection exceeds half
• 14:34
  the thickness. That's when you have to switch
• 14:36
  from simple linear bending to that more complex nonlinear
• 14:40
  tension membrane analysis, right?
• 14:42
  So the key take away for any design junkie listening is
• 14:45
  really just to understand your boundaries, know the
• 14:48
  assumptions, know when your plate stops being a simple
• 14:51
  bending member and starts being a diaphragm, and know when to
• 14:54
  switch from simple formulas to the big tables.
• 14:57
  And it's that precise understanding of all these
• 15:00
  complex stress distributions, the bending, the shear, the
• 15:03
  tension that allows for true structural optimization for
• 15:07
  sure. The analysis shows, for example,
• 15:10
  that you can strategically place stiffening ribs along a diameter
• 15:13
  of a clamped circular plate, and you can calculate it to achieve
• 15:16
  the minimum possible weight for a given maximum stress.
• 15:20
  If you can calculate every force, every moment, and every
• 15:23
  potential failure point from the elastic limit all the way to the
• 15:26
  ultimate collapse load, you can strip away every single
• 15:29
  unnecessary ounce of material while still guaranteeing
• 15:32
  performance. And that's the ultimate goal,
• 15:34
  isn't it? Maximum performance with minimum
• 15:36
  material. That's what smart design is all
• 15:38
  about. Performance meets efficiency.
• 15:41
  Join us next time for another deep dive into the core concepts
• 15:45
  of mechanical engineering.