Latest / Mechanical Engineering Made Simple / (#88) Stress_Deflection_Energy_Finite Element Analysis

Transcript

(#88) Stress_Deflection_Energy_Finite Element Analysis · Mechanical Engineering Made Simple

• 0:00
  Welcome back design junkies to Mechanical Engineering Made
• 0:02
  Simple. Today we are taking a deep dive
• 0:04
  into really the core of structural analysis where all
• 0:08
  the theory starts to meet the messy reality.
• 0:11
  Our sources stack up beautifully around 22 central problems every
• 0:15
  engineer has to solve. First, how do real world forces,
• 0:18
  which are almost never clean, combined to create complex
• 0:21
  stresses? And second, once we know those
• 0:24
  stresses, what are the best methods from classic calculus
• 0:28
  to, you know, modern numerical shortcuts to predict if a part
• 0:31
  is going to bend too much or just fail?
• 0:33
  It really is the difference between the textbook and the job
• 0:35
  site. We're moving beyond just, you
• 0:37
  know, stretching or squishing a simple bar.
• 0:39
  We're getting to the crucial area of combined effects.
• 0:41
  We're talking about structures that feel axial load, bending,
• 0:44
  maybe even torsion all at the same time.
• 0:46
  And solving for that, well, it demands some really rigorous
• 0:48
  deflection analysis. And the good news for you is
• 0:51
  that this necessity LED engineers to invent some truly
• 0:56
  elegant and fast tools, everything from energy
• 0:58
  principles to the matrix algebra that powers all modern
• 1:01
  simulation. OK, let's unpack that, starting
• 1:03
  with that real world complexity. You mentioned combined stress.
• 1:07
  If you look at a structure, let's say a short stocky column
• 1:11
  of rectangular cross section, it's so rare that a load hits
• 1:14
  perfectly down the center line. It's usually off even by a
• 1:17
  little bit, which means we're not dealing with pure
• 1:20
  compression. We've got this mix of bending
• 1:22
  and thrust. That's the key realization right
• 1:24
  there. The resulting stress, which we
• 1:27
  call Sigma, isn't from a single calculation, it's a
• 1:29
  superosition. It's a combination of two
• 1:32
  separate stress systems acting on the material at the same
• 1:35
  time. The first part is the simple one
• 1:37
  we all learned first. It's the average stress from the
• 1:40
  direct axial load P. So that's just a compressive
• 1:43
  force spread over the area. So P divided by.
• 1:46
  AP divided by A easy enough, but the second part is the stress
• 1:50
  that's generated by the bending moment M.
• 1:52
  That stress is M * Y / I sub X. OK.
• 1:56
  And just to refresh, Y is the distance from the neutral axis
• 1:59
  and I sub X is the second moment of area right, the measure of
• 2:02
  its bending resistance. Exactly, it's all about the
• 2:05
  shapes geometry. So the total stress your Sigma
• 2:08
  is simply P / a + m * y / I sub X.
• 2:13
  You just add them up we're. Literally just adding the
• 2:15
  crushing force and the bending force together.
• 2:17
  Precisely. And the most common way this
• 2:19
  happens in the real world is through what we call eccentric
• 2:22
  thrust. If that load P is acting off
• 2:25
  center by some small distance, we'll call it E sub Y.
• 2:28
  That single offset load is statically equivalent to two
• 2:31
  things happening at once. You still have the pure axial
• 2:33
  thrust P right at the center, but you also introduce a bending
• 2:37
  moment M that's equal to P * E sub.
• 2:40
  Y And that's the moment that creates the bending stress.
• 2:43
  So that single trick, converting an offset force into a central
• 2:46
  force in a moment is just fundamental for doing the
• 2:49
  analysis. It is, but it also creates a
• 2:51
  very dangerous situation, especially for materials like
• 2:53
  say concrete or brick. They're amazing in compression,
• 2:56
  but they are absolutely terrible if you try to pull them apart in
• 2:59
  tension. Which brings us to a really
• 3:01
  critical engineering limit. Then no tension rule.
• 3:04
  I mean when you combine bending and thrust, one side is getting
• 3:07
  compressed more and the other side is getting pulled on.
• 3:10
  That's right, and that's the side you worry about.
• 3:12
  The highest tensile stress is always on the side farthest from
• 3:15
  where you applied the load. For structures like short
• 3:18
  concrete columns, avoiding any tensile stress is non
• 3:21
  negotiable, and our sources highlight a very precise limit
• 3:25
  for a rectangular cross section. If you want to guarantee that
• 3:28
  the stress never goes below 0, meaning no tension, the
• 3:32
  eccentricity E sub Y must be less than or equal to 1/6 of H,
• 3:37
  where H is the depth of that column.
• 3:38
  Wait, OK, that's incredibly specific.
• 3:40
  So if a load is pushed just outside that 1/6 H limit even
• 3:44
  slightly, are we talking about an immediate catastrophic
• 3:47
  failure, or is it more of a slow degradation?
• 3:49
  Why is that exact fraction so important?
• 3:51
  It's typically degradation and cracking, but in critical
• 3:54
  applications cracking is failure.
• 3:57
  The reason that 1/6 fraction is so key is purely geometrical.
• 4:02
  If the resultant force falls within that central third of the
• 4:05
  column, often called the kernel, then the entire cross section
• 4:08
  stays in compression. The bending moment is just
• 4:12
  strong enough to reduce the compressive stress to 0 right at
• 4:16
  the edge, but not strong enough to flip it over into tension.
• 4:19
  I see. So once you step outside that
• 4:21
  kernel, you're actively pulling the fibers apart.
• 4:23
  And you get cracks, then water gets in and you have a huge
• 4:26
  structural problem. Over time is an absolutely
• 4:29
  crucial design check. That is a fantastic nugget of
• 4:32
  knowledge. OK, let's pivot a bit now.
• 4:33
  From strength, you know, will it break to stiffness?
• 4:37
  How much will it move? Why is deflection analysis just
• 4:40
  as critical as predicting the failure stress?
• 4:43
  Because an engineer's job is to make sure a structure is both
• 4:46
  safe and functional. A huge steel beam might be more
• 4:49
  than strong enough not to snap in half, but if it sags 3 inches
• 4:52
  under load, it could crack the drywall above it.
• 4:55
  It could jam up machinery mounted on it.
• 4:58
  Functionally, that design has failed, even if the stress
• 5:00
  limits were never even close. So what's the mathematical
• 5:03
  property that defines that resilience against bending?
• 5:05
  What's the term we use? That would be flexural
• 5:08
  stiffness, and it's always denoted as EI.
• 5:11
  It's the product of a material property, Young's modulus E, and
• 5:14
  a shape property, the second moment of area I.
• 5:18
  So we need stiff materials and smart shapes.
• 5:21
  EI is the perfect marriage of those two things.
• 5:23
  You want a high Young's modulus, like steel, and a high second
• 5:27
  moment of area like an i-beam. That puts the material far away
• 5:30
  from the center where it can do the most good.
• 5:32
  Exactly. And this product EI is so
• 5:35
  fundamental because it governs the basic relationship for
• 5:37
  elastic bending. The curvature which is 1 / r is
• 5:41
  directly proportional to the bending moment M and it's
• 5:45
  inversely proportional to the stiffness EI.
• 5:47
  So 1 / R = M divided by EI. That's the one.
• 5:52
  And that relationship leads us directly to the core governing
• 5:54
  differential equation for deflection that engineers have
• 5:57
  been wrestling with for, well, centuries.
• 6:00
  If V is the vertical deflection along the beam's length Z, the
• 6:04
  equation is EI times the second derivative of V with respect to
• 6:09
  Z ^2 = m, the bending moment function.
• 6:13
  OK, now solving that equation sounds like a absolute
• 6:15
  nightmare, especially for a complex load.
• 6:17
  Like what if you have a distributed load that starts 5
• 6:20
  feet in and Subs 10 feet later? You'd have to write different M
• 6:23
  functions for every single segment, integrate each one
• 6:26
  twice, and then use your boundary conditions to sort of
• 6:28
  painfully stitch them all together.
• 6:30
  That calculus is exactly the pinpoint.
• 6:32
  It's messy, it's very time consuming, and it's so prone to
• 6:35
  error. That very problem.
• 6:37
  The need to handle discontinuities and loading
• 6:40
  efficiently is what drove the invention of an incredibly
• 6:43
  clever tool called Macaulay Brackets.
• 6:44
  Macaulay brackets For anyone who hasn't seen them, what are they
• 6:48
  and how do they save us all that time?
• 6:50
  They're essentially special square brackets you use inside
• 6:53
  the integration. The key rule is that the term
• 6:56
  inside the brackets is only counted if its value is
• 6:59
  positive. If it's zero or negative, the
• 7:01
  entire term just vanishes. You ignore it.
• 7:04
  So they're like a mathematical on off switch for your loads.
• 7:08
  That is a perfect way to describe it.
• 7:10
  You write one single continuous bending moment equation for the
• 7:15
  entire beam, using these brackets to turn loads on and
• 7:18
  off. Then you integrate that one big
• 7:20
  equation twice. The brackets automatically make
• 7:23
  sure that the loads only contribute to the solution after
• 7:26
  they physically started acting on the beam.
• 7:28
  So it turns pages of separate calculations into one single
• 7:32
  elegant procedure. It's a lifesaver.
• 7:34
  OK, so that's the classical approach made easier, yeah.
• 7:37
  But even with that, if you have a huge structure or something
• 7:39
  that's statically indeterminate where you can't solve for the
• 7:41
  reactions easily, you need something more powerful than
• 7:44
  doing it all by hand. Is there a way to bypass all
• 7:47
  that integration and find the deflection directly?
• 7:50
  Absolutely. And this is the conceptual leap
• 7:52
  to what we call energy methods. Instead of balancing forces, we
• 7:55
  start balancing work and energy. And these methods are, I mean,
• 7:59
  indispensable for those tricky statically indeterminate
• 8:02
  structures. So what's the foundational idea?
• 8:04
  How does it work? It all starts with the principle
• 8:07
  of virtual work, and it's a surprisingly simple idea.
• 8:10
  It says that for any system that's in equilibrium, if you
• 8:13
  imagine it moves by some tiny arbitrary amount of virtual
• 8:16
  displacement, the total network done by all the forces during
• 8:20
  that imaginary movement has to be 0.
• 8:22
  So when we apply this to a beam, we put a fake unit load, a test
• 8:25
  force of just one, right where we want to find the deflection.
• 8:29
  The external work done by this little unit load moving through
• 8:31
  the real displacement delta has to equal the internal strain
• 8:35
  energy stored up in the beam from the real loads.
• 8:37
  We're just equating the external work from our test load to the
• 8:40
  internal energy from the actual loads.
• 8:42
  Work equals energy, and because that test load is just one, the
• 8:46
  work it does is numerically equal to the the displacement
• 8:48
  delta. It's beautifully concise.
• 8:51
  And here is the really powerful nugget from this the thing that
• 8:54
  turns this concept into the ultimate engineering shortcut.
• 8:57
  It's the idea of complementary energy, which we call C.
• 9:01
  It turns out that if you can define the complementary energy
• 9:04
  function C for your entire structure, the partial
• 9:08
  derivative of that function with respect to anyone external load,
• 9:12
  say W sub one, directly gives you the displacement delta sub
• 9:17
  one at that exact. Point.
• 9:19
  Wait, hold on a second. You're saying I don't have to
• 9:21
  integrate the differential equation twice, solve for
• 9:24
  boundary conditions, and then plug everything in.
• 9:26
  I just define one energy function for the whole system,
• 9:29
  take one partial derivative with respect to the force I care
• 9:32
  about, and the deflection just pops out.
• 9:35
  That's it. That's the shortcut.
• 9:36
  It's an incredibly direct path. You're basically using the
• 9:39
  principle that the rate of change of energy with respect to
• 9:41
  a force gives you the distance. That's a game changer.
• 9:44
  It really is, and it's the gateway to modern computation.
• 9:47
  Methods like the Raleigh Ritz method are all built on these
• 9:49
  energy principles. Which brings us to the really
• 9:52
  massive scale when we're designing an aircraft fuselage
• 9:55
  or an offshore oil rig, structures with thousands of
• 9:58
  parts and load paths. All these hand calculation
• 10:00
  methods, even the elegant energy ones, they just breakdown.
• 10:03
  There are too many variables. Exactly right complexity demands
• 10:06
  automation, and when manual analysis is no longer feasible,
• 10:10
  you have to move to numerical analysis.
• 10:13
  The language of numerical analysis is matrix algebra.
• 10:17
  Right. For anyone who hasn't touched
• 10:18
  matrix math in a while, let's keep it simple.
• 10:20
  What is a matrix in this context?
• 10:22
  Fundamentally, a matrix is just an organizational tool.
• 10:26
  It's a rectangular grid of numbers that lets us organize
• 10:29
  and manipulate huge sets of data all at once.
• 10:31
  If we say a matrix is 3 by 4, it just means it has three rows and
• 10:35
  four columns. It's a container for
• 10:37
  information. And how do these containers help
• 10:39
  us model something this complex as a skyscraper?
• 10:42
  They are the absolute backbone of the matrix displacement
• 10:45
  method and its modern evolution, the finite element method, or
• 10:49
  FEM. Instead of solving one equation
• 10:51
  at a time, matrices let us represent the entire structure
• 10:54
  as one massive interconnected system of simultaneous
• 10:57
  equations. We write it very compactly as
• 11:00
  the force vector P equals the stiffness matrix K multiplied by
• 11:05
  the displacement vector UP equals K * U That single matrix
• 11:10
  equation contains everything about the structure's behavior.
• 11:12
  It links all the applied loads in P, all the structure's
• 11:15
  resistance properties in the stiffness matrix K and all the
• 11:18
  resulting movements in the displacement vector U.
• 11:22
  So if we know the force is P and we can get the computer to build
• 11:25
  that massive stiffness matrix K based on the geometry materials,
• 11:29
  the computer's only job is to solve for the unknown
• 11:31
  displacements. That's it.
• 11:32
  And once you have the displacements you you can work
• 11:34
  backwards to find the strains and then the stresses in every
• 11:38
  single little piece of the structure.
• 11:39
  And FEM, the finite element method, is the tool that brings
• 11:42
  it all to life. It takes that complex structure
• 11:45
  and it subdivides it into thousands of simpler finite
• 11:48
  elements like tiny triangles or bricks.
• 11:50
  Which are easy to analyze individually.
• 11:52
  Then, using that core P = K * U operation, the computer
• 11:57
  assembles the behavior of all those tiny elements into a
• 11:59
  global system, and it gives us the precise displacements and
• 12:02
  stresses everywhere. It completely democratized
• 12:05
  structural analysis. It allows us to safely design
• 12:08
  things that would have been totally impossible to analyze by
• 12:10
  hand just a century ago. What an incredible journey.
• 12:14
  We started with the real world roblem of combining axial and
• 12:17
  bending stress for eccentric loads and that critical 1/6 H
• 12:21
  check for concrete. We saw how hard classical
• 12:24
  deflection analysis can be, which made us appreciate the
• 12:26
  brilliant shortcut of energy and virtual work.
• 12:29
  And finally, we saw how all that conceptual groundwork leads
• 12:32
  directly to the ultimate numerical framework, matrix
• 12:35
  algebra and the finite element method, which lets us tackle
• 12:38
  almost any structural problem you can imagine.
• 12:41
  And the overall utility is clear.
• 12:43
  I mean, whether you're doing a quick hand check on a simple
• 12:46
  beam or running a massive computer simulation,
• 12:49
  understanding the theory is what ensures safety.
• 12:52
  Our sources remind us that understanding the load factor,
• 12:55
  how much you can increase the loads before collapse, is the
• 12:58
  ultimate safety metric. The theory is what anchors the
• 13:02
  design, no matter what tool you use.
• 13:04
  The theory is the bedrock, so let's build on that.
• 13:07
  We touched on composite materials like steel reinforced
• 13:10
  concrete. We know that for compatibility,
• 13:13
  the steel and the concrete have to stretch together, which means
• 13:15
  their internal stresses have to adjust based on their very
• 13:18
  different Youngs by Dooley. So here's something for you to
• 13:21
  Mull over. If you design A steel reinforced
• 13:23
  concrete beam in the winter, and those materials have very
• 13:26
  different thermal expansion properties, what unexpected
• 13:29
  internal stresses arise in that beam before you even apply a
• 13:32
  single external load? It's something to think about
• 13:35
  how those initial stresses can compromise a structure before it
• 13:38
  even goes into service. Thank you for joining us for
• 13:41
  this deep dive into the mechanical world of structural
• 13:43
  analysis. Always a pleasure.