Latest / Mechanical Engineering Made Simple / (#60) Complex Beam Analysis and Ultimate Limit Design

Transcript

(#60) Complex Beam Analysis and Ultimate Limit Design · Mechanical Engineering Made Simple

• 0:00
  Welcome back, design junkies, to mechanical engineering made
• 0:02
  simple. Today, we're leveling up.
• 0:05
  We certainly are. If you thought analyzing a
• 0:07
  simply supported beam was complex, wait until we dive into
• 0:10
  beams that are curved, tapered, sitting on squishy foundations,
• 0:15
  and pushed right up to the point of catastrophic failure.
• 0:18
  That's right, Every structural analysis begins with a, you
• 0:21
  know, a beautiful set of ideal assumptions, whether we're
• 0:24
  talking about material uniformity or simple point
• 0:26
  supports. The textbook world.
• 0:27
  Exactly. Our goal in this deep dive is to
• 0:30
  equip you with the specialized knowledge needed when real world
• 0:33
  conditions like axial loads, non uniform sections or extreme
• 0:37
  curvature break those foundational rules.
• 0:40
  We are moving from the safe, idealized world straight into
• 0:44
  high stakes practical engineering.
• 0:46
  OK, let's unpack this. Our mission is to take a
• 0:48
  shortcut through the heavy lifting.
• 0:50
  We're starting with the fundamental mechanical
• 0:52
  assumptions for the common case of straight beams and quickly
• 0:56
  progressing into complex situations like beams on elastic
• 0:58
  foundations and the ultimate limit plastic design.
• 1:02
  Right, we need to know what tools to pull out of the toolbox
• 1:05
  when the basic rules stop applying.
• 1:07
  So before we get fancy, where do we start?
• 1:09
  To appreciate the complexity, we must first respect the
• 1:12
  idealization. The cornerstone of common beam
• 1:16
  analysis is the flexure formula. Fiber stress equals negative M y
• 1:20
  / I. It's elegantly simple, but that
• 1:24
  simplicity is predicated on three crucial assumptions.
• 1:27
  Tell us what those red flags are, the things that, if
• 1:30
  violated, make the analysis immediately suspect.
• 1:33
  OK, the first is material uniformity.
• 1:36
  The beam material must be homogeneous and possess the same
• 1:39
  modulus of elasticity E in both tension and compression.
• 1:43
  So no weird composites or materials that behave
• 1:46
  differently when you pull versus when you push.
• 1:48
  Exactly. Second, geometry matters
• 1:50
  immensely. The beam must be straight or
• 1:52
  nearly straight if there's a slight curve.
• 1:55
  The radius of curvature must be at least 10 times the depth of
• 1:58
  the beam, and 3rd the cross section must remain uniform
• 2:00
  along the entire length. And if that curve to depth
• 2:03
  ratio, say drops to 5:00 to 1:00, we're already in trouble
• 2:06
  potentially seeing calculated results off by 5% or more.
• 2:09
  Precisely now, even under ideal assumptions, we are calculating
• 2:13
  2 primary types of internal stresses, the longitudinal fiber
• 2:17
  stress S which is tensile or compressive, and then the shear
• 2:21
  stress. T and fiber stress is usually
• 2:24
  what we worry about. In slender beans, yes, but
• 2:27
  remember that short stubby beams or those with very thin webs
• 2:31
  often fail due to longitudinal shears.
• 2:34
  It always there. This brings us to the geometric
• 2:37
  foundation of all beam deflection analysis.
• 2:40
  The core idea governing deflection and slope is the
• 2:42
  radius of curvature. Indeed, we established that the
• 2:45
  radius of curvature R is directly proportional to the
• 2:48
  stiffness of the beam EI and inversely proportional to the
• 2:52
  applied bending moment. M which gives us the big one.
• 2:54
  The general differential equation of the elastic curve EI
• 2:58
  multiplied by the 2nd derivative of the vertical deflection Y sub
• 3:01
  C with respect to X = a moment M.
• 3:04
  That equation is absolutely essential because it is the
• 3:07
  mathematical key to solving for deflection in any elastic beam.
• 3:10
  What does that equation tell the engineer?
• 3:12
  It connects the geometry of the curve, the second derivative,
• 3:16
  directly to the internal forces. And of course we build on this.
• 3:19
  We know the shear V is the derivative of the moment with
• 3:22
  respect to X and conversely the moment M is the integral of the
• 3:26
  shear VDX. Right, so you can move
• 3:28
  seamlessly between load, shear, moment, slope and deflection.
• 3:33
  That's the ideal, OK. Standard analysis assumes
• 3:36
  uniform cross sections, which is often great for manufacturing
• 3:39
  simplicity. But what happens when a designer
• 3:41
  needs to vary the depth or width, perhaps for aesthetic
• 3:45
  reasons or simply to save weight?
• 3:47
  We're talking about tapered beams.
• 3:49
  Capered beams are a very common challenge.
• 3:51
  If the moment of inertia I vary significantly along the length,
• 3:55
  say if I varies as the quantity one plus KX divided by L raised
• 4:00
  to the power N, You can't use the simple deflection formulas.
• 4:04
  So you're stuck. Not quite.
• 4:06
  Engineers turn to tabulated coefficients, specialized
• 4:09
  multipliers found in technical tables.
• 4:11
  That sounds like a lot of coefficients to track down and
• 4:13
  manage, especially for complex loading scenarios.
• 4:16
  Is there a more intuitive way to model this variable stiffness?
• 4:20
  There is, and this is a fantastic practical nugget.
• 4:23
  You can often simplify the problem dramatically using an
• 4:26
  equivalent loading technique. Go on.
• 4:28
  Imagine a simply supported beam that is stiff in one segment
• 4:32
  with moment of inertia I2 and less stiff in another I-1.
• 4:36
  Instead of solving the complex geometry you model, the entire
• 4:39
  structure is having the stiffness of the stiffer section
• 4:42
  I2. So you just pretend the entire
• 4:44
  beam is uniformly stiff. How do you account for the
• 4:48
  weaker segment? You adjust the load.
• 4:50
  You increase all loads acting on the portion with the smaller
• 4:53
  moment of inertia I-1 by the ratio of I2 divided by I-1.
• 4:58
  By substituting an equivalent magnified load onto a simpler
• 5:01
  constant section beam, you get accurate results.
• 5:04
  That transforms a geometry problem into a simpler loading
• 5:07
  problem that is genuinely useful.
• 5:09
  It's a great trick. Now let's pivot from geometry to
• 5:11
  support. Most textbook problems feature
• 5:13
  supports at discrete points, but many real world structures like
• 5:17
  pipelines, footings or railway tracks rest on ground or
• 5:20
  material that yields continuously.
• 5:22
  Hello Elastic Foundations. Elastic foundations are defined
• 5:26
  by one key characteristic. They develop continuous
• 5:29
  reactions proportional at each position along the beam to the
• 5:32
  deflection of the beam at that position.
• 5:34
  It's a distributed restorative force.
• 5:37
  And crucially, this model is robust enough that even
• 5:39
  discontinuous foundations like a rail supported by periodic cross
• 5:43
  ties can be analyzed this way. They can, provided the supports
• 5:47
  are close enough. Specifically, if you have at
• 5:50
  least three concentrated reaction forces in every half
• 5:53
  wavelength of the deflected beam, the entire analysis is
• 5:56
  governed by the foundation modulus, which gives us the
• 5:59
  characteristic constant beta. And beta is.
• 6:02
  Beta is equal to the 4th root of the quantity B0 K 0 / 4 EI.
• 6:07
  So beta captures the relative stiffness of the foundation
• 6:10
  versus the stiffness of the beam itself, correct?
• 6:12
  Exactly. But here is a critical
• 6:14
  limitation for you. Well, the complex formulas
• 6:17
  derive for finite length beams on elastic foundations are
• 6:19
  theoretically sound. They come with a computational
• 6:22
  trap. You must not use them when the
• 6:24
  product of beta in the beam length L exceeds 6.
• 6:27
  Why the hard limit at beta L greater than 6?
• 6:30
  Is that a structural failure point?
• 6:32
  No, it's a mathematical failure point.
• 6:34
  When beta L gets large, the calculations require subtracting
• 6:38
  2 very large numbers that are nearly equal.
• 6:41
  A round off error issue. A severe one.
• 6:44
  It can introduce huge errors, especially in simpler software.
• 6:48
  You end up with grossly inaccurate results even if the
• 6:51
  physics are right. It's a classic example of
• 6:53
  computation compromising theory. That raises an important
• 6:56
  question. What if we introduce a force
• 6:59
  running straight down the middle of the beam and axial load P
• 7:03
  while it's bending? We now have combined loading and
• 7:06
  this is where standard superposition breaks down.
• 7:08
  This is where things get non linear.
• 7:11
  Axial tension tends to straighten the beam, which
• 7:13
  actually reduces the overall bending moments.
• 7:16
  Conversely, axial compression amplifies the bending moment and
• 7:20
  deflection. The compressive load is
• 7:21
  effectively working to push the beam further out of line.
• 7:25
  Precisely so if tension reduces the moment, you can't just
• 7:29
  calculate the moment separately and add them.
• 7:30
  Up because the axial load affects the deflected shape of
• 7:33
  the beam itself. It's an interactive effect.
• 7:36
  The maximum normal stress is still calculated using the
• 7:38
  familiar combined stress formula S Max equals P / A plus or minus
• 7:43
  MC divided by I. P is positive for tension,
• 7:46
  negative for compression, but the M here is the moment from
• 7:51
  the combined effect. And this leads to the modified
• 7:54
  superposition rule. What's the practical implication
• 7:57
  of that modification for a designer?
• 7:59
  The implication is clarity. In calculation.
• 8:02
  The total moment or deflection must be found by adding the
• 8:05
  moment for each transverse load combined individually with the
• 8:09
  entire axial load. You have to sum up the
• 8:11
  transverse load effects and then add the axial effect at the.
• 8:14
  End. Got it.
• 8:15
  It has to be factored in for every individual bending
• 8:18
  scenario, right? Speaking of ultimate failure
• 8:20
  under compression, what happens when a key component of an
• 8:23
  I-beam under heavy bending decides to fail like a column
• 8:27
  instead of through simple yielding?
• 8:29
  That's the danger of lateral buckling.
• 8:31
  It typically targets the compression flange of an I-beam.
• 8:34
  If it's insufficiently braced, the flange acting like a slender
• 8:37
  column just buckle sideways. Long before the material might
• 8:40
  have actually yielded from the bending stress.
• 8:42
  Exactly. If you have a long, heavily
• 8:45
  loaded beam spanning a factory floor with no lateral supports,
• 8:49
  this is the mode of failure you fear most.
• 8:51
  And our sources give a specific approximation for the critical
• 8:54
  compressive stress S0 for standard I beams in this
• 8:58
  scenario. Yes, the critical compressive
• 9:00
  stress as zero is approximated as 40,000 -, 60 ML divided by R
• 9:06
  That M coefficient is key as it depends on how the beam is
• 9:09
  supported and loaded. How so?
• 9:12
  Well, if you have a cantilever with a load at the end, M is
• 9:14
  1.000. But if you have a simply
• 9:18
  supported beam with a midpoint load, M drops to 0.500, which
• 9:23
  indicates a better moment distribution resisting buckling.
• 9:26
  OK, time to tackle geometry again, but this time the beam is
• 9:29
  curved in its resting state. Think crane hooks or specialized
• 9:32
  machine linkages. Curvature fundamentally changes
• 9:35
  the math. It creates a non linear stress
• 9:37
  distribution even within the elastic range.
• 9:39
  The critical insight here is that the neutral axis shifts
• 9:43
  inward toward the center of curvature.
• 9:45
  So the formula S = m Y divided by I is no longer strictly
• 9:50
  accurate. It's not.
• 9:51
  The assumption of linear strain distribution is violated.
• 9:54
  The forces are no longer balanced symmetrically around
• 9:56
  the centroid. So what are the main stresses we
• 9:59
  now have to track? The three major stresses are
• 10:01
  circumferential stress, which is the dominant tensile or
• 10:04
  comressive 1, radial stress R, and shear stress Tau.
• 10:10
  Interestingly, in most compact thick sections like crane hooks,
• 10:14
  we often neglect the radial stress because it's so small
• 10:17
  compared to the circumferential OK.
• 10:19
  Let's consider curved beams with wide flanges, like those used in
• 10:23
  large rings for pipes or airplane fuselages.
• 10:26
  If the flange is very wide, the stress won't be uniform across
• 10:29
  that width, right? That's a massive problem caused
• 10:31
  by shear lag. The shear stresses near the web
• 10:34
  prevent the remote edges of the wide flange from participating
• 10:37
  fully and carrying the load. So you have to calculate an
• 10:39
  effective width. You do for an indefinitely wide
• 10:42
  inter flange. The maximum effective width be
• 10:44
  zero is calculated using the 1.56 square root of RT rule.
• 10:48
  Let's make sure we internalize that B0 is 1.56 times the square
• 10:53
  root of the product of the mean radius of curvature R and the
• 10:57
  thickness T. This isn't just theory.
• 11:01
  This calculation dictates how much of that material you paid
• 11:04
  for is actually working. Precisely.
• 11:07
  If you ignore shear lag, you grossly overestimate the
• 11:10
  stiffness and underestimate the stress in the flange near the
• 11:13
  web. It's a dangerous mistake and.
• 11:15
  For thin tubular sections, curvature introduces another
• 11:18
  headache, geometric distortion. Yes, thin rings are highly
• 11:22
  susceptible. Radial stress causes distortion
• 11:25
  that can drastically reduce both strength and stiffness,
• 11:29
  requiring an additional correction factor K in the
• 11:31
  formulas. And there was a memorable detail
• 11:33
  to illustrate. This a truly memorable one.
• 11:35
  Imagine a one inch thin pipe. Fix it both ends subjected to a
• 11:38
  heavy load. The theoretical calculation for
• 11:41
  the deflection resulted in a 20 inch total change in the
• 11:44
  horizontal diameter. 20 inches. That is an extreme massive
• 11:47
  deformation. It instantly highlights the
• 11:49
  limitation of linear elastic theory.
• 11:52
  If your calculation predicts a deflection that large, you know
• 11:55
  the underlying assumption that the loaded ring remains nearly
• 11:58
  circular has been completely shattered.
• 12:00
  Your initial linear analysis is useless.
• 12:03
  Butterly you have to move to a large displacement non linear
• 12:06
  analysis. That sets the stage perfectly
• 12:08
  for our final section. We've established all the ways a
• 12:11
  beam fails within the elastic range.
• 12:14
  Now let's discuss what happens when we intentionally push the
• 12:17
  material past its elastic limit, entering the realm of ultimate
• 12:20
  strength design or plastic design.
• 12:23
  This is the critical transition from worrying about the first
• 12:26
  fiber yielding to calculating the point of catastrophic
• 12:29
  collapse. Once the yield strength SY is
• 12:32
  reached, the failure moment is the plastic moment.
• 12:35
  MP calculated by MP equals SYZ. Z is the key here, the plastic
• 12:40
  section modulus. Explain conceptually what Z
• 12:42
  represents beyond just the math. Conceptually, Z represents the
• 12:46
  moment required to make the entire cross section yield.
• 12:49
  It's idealized as having half the area in tension and half in
• 12:52
  compression, both of the yield strength.
• 12:54
  Mathematically, it's the arithmetical sum of the static
• 12:57
  moments about the neutral axis of the parts of the cross
• 13:00
  section above and below that axis.
• 13:02
  So for a simple rectangle of depth D and width B, the plastic
• 13:06
  section modulus Z is 1/4 BD squared.
• 13:10
  It is purely A geometric property tied to the yield
• 13:13
  point. Correct.
• 13:14
  And the phenomenon that enables plastic design is the formation
• 13:16
  of the plastic hinge Collapse occurs not when the first fiber
• 13:20
  yields, that's the elastic limit, but when enough plastic
• 13:23
  hinges form to transform the structure into an unstable
• 13:27
  mechanism. So it can rotate freely under
• 13:29
  load, and this leads to a fundamental disagreement between
• 13:32
  elastic and plastic analysis near the ultimate limit, the
• 13:35
  concept of moment redistribution.
• 13:37
  This is the definitive distinction.
• 13:39
  Consider a beam fixed at one end and supported at the other with
• 13:42
  a center load W Elastic analysis dictates the maximum moment
• 13:45
  occurs at the fixed wall, but as the load W increases, yielding
• 13:49
  starts at the wall. Critically, as yielding
• 13:52
  progresses, that fixed wall can no longer hold the full elastic
• 13:56
  moment. The moment actually decreases at
• 13:58
  the wall and is redistributed to the center where the load is.
• 14:02
  The structure is finding a new way to carry the load, flexing
• 14:05
  its hidden capacity. Exactly.
• 14:07
  Collapse only occurs when the moment at the wall and the
• 14:10
  moment at the center load both reach MP, forming 2 plastic
• 14:13
  hinges. The resulting collapse load WC
• 14:17
  is much higher than what elastic analysis would predict 6 MP
• 14:21
  divided by L. Which is why ultimate strength
• 14:23
  design is often more economical. It accurately accounts for the
• 14:26
  full capacity of the material. That's a whole.
• 14:28
  Point. We've covered a huge distance
• 14:30
  today, moving from the simple assumptions that govern straight
• 14:33
  beams to the nonlinear math required for combined loads and
• 14:36
  complex geometries, and finally the ultimate capacity determined
• 14:40
  by plastic hinging. The key take away for you is
• 14:43
  that engineering analysis must always is adapt the complexity
• 14:46
  of the structure and the loading.
• 14:48
  Whether you are using the beta factor to define foundation
• 14:51
  stiffness, adjusting loads for tapered sections, or calculating
• 14:55
  Z for ultimate strength, these specialized tools are essential.
• 15:00
  Because basic theory would just fail dramatically.
• 15:02
  It would. So what does this all mean for
• 15:04
  you as you dive deeper into structural challenges?
• 15:07
  Well, we talked briefly about shear lag in curved, wide flange
• 15:11
  beams, but it affects straight beams too, particularly with
• 15:14
  short, stiff spans or thin elements.
• 15:18
  Your next step should be to fully explore the impact of
• 15:20
  shear lag in straight, deep, wide flange beams.
• 15:24
  It's a consequence of high shear stresses near the web that
• 15:27
  severely limits how much of the flange is truly effective in
• 15:30
  carrying the longitudinal load, challenging that fundamental
• 15:33
  assumption of uniform fiber stress even in the standard
• 15:36
  elastic range. Understanding that mechanism is
• 15:39
  crucial for designing efficient modern structures.
• 15:42
  That sounds like a powerful way to end this deep dive.
• 15:44
  Until next time.