Latest / Mechanical Engineering Made Simple / (#91) The Four Pillars of Mechanical Integrity

Transcript

(#91) The Four Pillars of Mechanical Integrity · Mechanical Engineering Made Simple

• 0:00
  Welcome back to mechanical engineering made simple.
• 0:02
  It's great to be back. Today we're doing something a
• 0:04
  little a little different. Usually we're obsessed with the
• 0:08
  flashy stuff. You know, we want to talk about
• 0:10
  rocket nozzles, the aerodynamic curves of a Formula One car.
• 0:14
  The shiny exterior. Exactly the stuff that gets all
• 0:17
  the likes on Instagram. But today we are going
• 0:21
  underneath the hood. We're looking at the invisible
• 0:23
  forces. The forces that actually
• 0:24
  determine whether a machine runs for 20 years without a hiccup,
• 0:27
  or whether it explodes in 20 seconds and takes the factory
• 0:31
  with it, right? We are doing a deep dive into
• 0:34
  what we're calling the four pillars of mechanical Integrity
• 0:37
  now, looking at the stack of sources we have here.
• 0:39
  I mean, we've got engineering handbooks, white papers on
• 0:42
  tribology, vibration analysis guides.
• 0:45
  It's a massive amount of data. It is our mission today is to
• 0:49
  sort of synthesize all of this into a comprehensive look at
• 0:52
  what keeps the world moving. Or maybe more accurately, what
• 0:56
  prevents the world from grinding to a halt?
• 0:58
  Because that's the natural state of things, right?
• 1:00
  Wear, tear and breakdown. Entropy is always trying to win.
• 1:04
  That's a good way to put it. And I want to speak directly to
• 1:06
  you, the listener, for a second. We know who listens to this.
• 1:09
  You might be an engineering student, you know, sweating over
• 1:13
  a board exam, trying to memorize formulas.
• 1:16
  Or maybe you're just the kind of person who sits on a plane,
• 1:19
  looks out at the wing shaking in turbulence, and thinks, How does
• 1:23
  that aluminum handle millions of cycles without snapping?
• 1:27
  Or why doesn't my car engine just melt when I'm driving 70
• 1:31
  mph? Yeah, if you've ever wondered
• 1:33
  about the difference between a machine that feels solid and one
• 1:37
  that feels, you know, cheap, or why some designs last for a
• 1:40
  century and others fail in a week, this deep dive is for you.
• 1:44
  We're peeling back the skin of the machine.
• 1:47
  So give us the road map how we're going to tackle this
• 1:49
  beast, because looking at the source stack, this covers
• 1:52
  everything from like basic physics to quantum level surface
• 1:57
  chemistry. And we're breaking it down into
• 1:58
  four specific areas, the pillars.
• 2:00
  First, we have stress analysis. This is basically the art of not
• 2:04
  breaking. Which, looking at the math, is a
• 2:06
  lot harder than it sounds. Much harder.
• 2:09
  Then we move to pillar 2. Lubrication, the science of
• 2:13
  separating surfaces. OK then Pillar 3 seal technology
• 2:18
  and it's not just rubber rings, it's about keeping extreme
• 2:20
  pressure where it belongs and finally vibration and noise
• 2:25
  managing the the heartbeat of the machine.
• 2:28
  It's a road map for mechanical survival.
• 2:30
  OK, I'm ready. Let's jump right into pillar #1
• 2:34
  stress analysis. The foundation of everything.
• 2:36
  If the structure can't hold the load, nothing else matters.
• 2:40
  You can have the best engine in the world, but if the frame
• 2:42
  snaps you aren't going anywhere. OK, so diving into the sources,
• 2:45
  I hit a snag almost immediately. There's this distinction made
• 2:48
  between total stress and unit stress.
• 2:51
  And after lit, in casual conversation, if I say stress,
• 2:53
  I'm thinking about, you know, a heavyweight sitting on a table.
• 2:56
  Sure, but the source material makes a massive deal about this
• 2:59
  difference. Yeah.
• 3:00
  Why is this distinction so vital?
• 3:02
  It's a really common point of confusion, but it's the
• 3:04
  difference between lay terms and engineering terms.
• 3:07
  Total stress is simply the total resultant force acting on a
• 3:11
  section. So let's say you have a bridge
• 3:13
  column holding up £1,000,000. The total stress is £1,000,000.
• 3:18
  OK, that sounds like a lot of stress.
• 3:20
  It's a lot of force, but that number, 1,000,000 lbs is
• 3:23
  completely useless to a designer until you know what is holding
• 3:27
  it up. That's where unit stress comes
• 3:29
  in. It's the total stress divided by
• 3:31
  the area P divided by. AE over a the classic formula.
• 3:35
  Exactly. Imagine that £1,000,000 is being
• 3:37
  held up by, I don't know, a toothpick.
• 3:40
  The area A is tiny, so when you divide 1,000,000 by that tiny
• 3:44
  tiny number, the unit stress is astronomical.
• 3:47
  The toothpick fails instantly. Right.
• 3:49
  It's vaporized. It's not even there.
• 3:50
  Anymore. But now imagine that same
• 3:52
  £1,000,000 is held up by a solid concrete pillar that's 10 feet
• 3:56
  wide. The total stress is the same,
• 3:58
  £1,000,000, but the air is just huge.
• 4:01
  So the unit stress, the load per square inch is actually quite
• 4:04
  low. Totally manageable.
• 4:06
  So when engineers talk about stress or yield strength, they
• 4:10
  almost always mean unit stress. It's a measure of intensity, not
• 4:14
  just weight. That clicks.
• 4:16
  It's about the density of the struggle, so to speak.
• 4:19
  It's not just about how heavy the burden is, it's about how
• 4:22
  broad your shoulders are. That's a perfect analogy,
• 4:26
  really. OK, so the sources then list 3
• 4:28
  main flavors of this stress. We hear these terms all the
• 4:31
  time, but let's define them properly.
• 4:33
  We have tension, compression and shear.
• 4:36
  The Big Three tension is pulling apart.
• 4:39
  Think of a tug of wardrobe. The molecules are literally
• 4:43
  being stretched away from each other.
• 4:44
  Simple enough. Compression is squeezing with
• 4:47
  the legs of your chair right now, supporting your body
• 4:49
  weight. The molecules are being shoved
• 4:50
  closer together. And sheer, that would always
• 4:53
  feels a bit more abstract to me. Sheer is sliding.
• 4:56
  The best way to picture it is with a deck of cards sitting on
• 4:58
  a table. If you put your hand on the top
• 5:00
  card and push it forward while the table holds the bottom card,
• 5:03
  still all the cards in the middle slide past one another.
• 5:06
  OK, I can see that. That's here in the real world.
• 5:10
  Think about a bolt. You have two metal plates bolted
• 5:13
  together. If you pull the plates in
• 5:15
  opposite directions, the bolt isn't really being pulled or
• 5:19
  squeezed, it's being sliced. The plates are trying to, well,
• 5:23
  decapitate the bolt, right? That's sheer stress.
• 5:26
  It's what happens when you tighten a nut too hard, no, and
• 5:29
  you strip the threads. You've literally sheer the metal
• 5:33
  spirals right off the bolt. Ouch, I've definitely done that.
• 5:36
  It's that feeling where it gets tight, tight, tight and then
• 5:38
  suddenly loose. And your heart sinks.
• 5:40
  Yeah, because you just exceeded the sheer strength of the steel.
• 5:44
  OK. So we have stress, which is the
• 5:46
  four supplied per unit area, Yeah.
• 5:48
  And then the result of that force which is strain.
• 5:50
  Correct. Stress is the cause, strain is
• 5:52
  the effect. You apply stress, you get
• 5:54
  strain. Technically, strain is defined
• 5:57
  as the deformation per unit length, so it's the change in
• 6:00
  length divided by the original length.
• 6:01
  So if I have a rubber band that's 10 inches long and I
• 6:04
  stretch it to 11 inches, the strain is that one inch of
• 6:08
  change divided by the 10 inches of original length, so .1.
• 6:11
  Precisely. It's a dimensionless ratio, and
• 6:15
  this leads to a really fascinating phenomenon mentioned
• 6:18
  in our data called Poisson's ratio.
• 6:20
  I saw that it's the idea that when you stretch something it
• 6:22
  gets thinner, right? Like on the piece of taffy.
• 6:24
  Or if you squash it, it gets fatter.
• 6:26
  It's the ratio of the lateral strain, the thinning to the
• 6:30
  longitudinal strain, the stretching.
• 6:32
  It's a measure of how much it next down.
• 6:35
  So what's a typical number for that?
• 6:36
  Well, most metals, like steel have a poison's ratio of about
• 6:41
  .3. So it thins out about 30% as
• 6:44
  much as it stretches. Is that how I should think about
• 6:46
  it? Roughly speaking, yes, it has to
• 6:48
  do with conserving volume more or less.
• 6:51
  But look at concrete in the data tables.
• 6:53
  It has a ratio of only about point.
• 6:55
  One that's surprisingly low. Why is that?
• 6:57
  It is. Concrete is a rigid aggregate
• 7:00
  structure. It's rocks and sand glued
• 7:02
  together. It doesn't neck down or thin out
• 7:04
  nearly as much when you pull on it, mostly because concrete
• 7:07
  really, really hates being pulled.
• 7:09
  It's brittle. It would rather just crack them
• 7:11
  thin out. Which brings us to the
• 7:13
  relationship between the two. We can't talk about stress and
• 7:16
  strain without talking about the rule that governs them.
• 7:19
  The sources bring up Hook's Law. Ah yes, Hook's law.
• 7:23
  It's the rule of linearity. It states that for most
• 7:26
  materials, up to a certain point, stress is directly
• 7:30
  proportional to strain. So if you double the load.
• 7:33
  You double the stretch. If you triple the load, you
• 7:34
  triple the stretch. It's a straight line on the
• 7:36
  graph A. Perfect straight line and the
• 7:38
  slope of that line. That is a really important
• 7:42
  number. It's defined by the modulus of
• 7:44
  elasticity, also known as Youngs modulus.
• 7:47
  Which is basically a measure of stiffness.
• 7:49
  Right. Yes, exactly.
• 7:50
  Think of it as the materials springiness.
• 7:53
  A high modulus like for steel means the line is very steep.
• 7:58
  It takes a huge amount of stress to create just a tiny bit of
• 8:01
  strain. So it's very stiff.
• 8:02
  Very stiff. A low modulus like for rubber
• 8:05
  means the line is floppy. You pull it a little, it
• 8:07
  stretches a lot. But Hook's law is an infinite
• 8:10
  that straight line it breaks down.
• 8:13
  Oh absolutely. It works up to a point we call
• 8:16
  the proportional limit. Shortly after that, you hit the
• 8:18
  yield point and this this is the most critical concept for a
• 8:22
  designer. Why is that?
• 8:24
  Because before the yield point, the material is elastic.
• 8:27
  If you remove the load, it snaps back to its original shape, just
• 8:31
  like a spring. No harm done.
• 8:33
  But after the yield point. It becomes plastic.
• 8:35
  You have permanently deformed it.
• 8:38
  Think of bending a paper clip. If you bend it just a little, it
• 8:41
  springs back. That's elastic behavior.
• 8:43
  If you bend it too far, it stays bent.
• 8:45
  You've passed the yield point. You have permanently rearranged
• 8:48
  the crystal structure of the metal.
• 8:50
  And if I just keep bending it back and forth?
• 8:52
  Eventually you'll reach its ultimate strength, that's the
• 8:55
  peak of the stress strain curve, and then rupture.
• 8:58
  It breaks. So for an engineer designing say
• 9:01
  a bridge, the yield point is usually the big red do not cross
• 9:05
  line. Usually yes.
• 9:07
  If your bridge bends when a truck drives over it, you have
• 9:10
  failed. As a designer, you want the
• 9:12
  bridge to flex and return. Flex and return always in that
• 9:15
  elastic region. But here is where it gets really
• 9:18
  interesting for me. The sources go deep into modes
• 9:21
  of failure, and it turns out things don't just fail because
• 9:25
  you pulled them too hard one time.
• 9:26
  They fail in these sneaky ways they do, and the first one that
• 9:30
  jumped out at me and it sounds like a horror movie villain is
• 9:33
  fatigue. Fatigue is the silent killer of
• 9:36
  machinery. It's absolutely the most common
• 9:39
  failure mode, and it explains how a material can fail even if
• 9:42
  the stress is well below that yield point we just talked
• 9:45
  about. So I'm not bending the paper
• 9:47
  clip all the way. No, you're just bending it a
• 9:48
  tiny bit, maybe just 10% of its capacity, but you're doing it a
• 9:52
  million times or a billion times.
• 9:54
  This is called cyclic. Loading and what's happening
• 9:56
  inside the metal. The material gets tired.
• 9:59
  What's actually happening on a microscopic level is that tiny,
• 10:02
  tiny cracks form, usually at the surface where there's a little
• 10:05
  scratch or an imperfection, a stress riser.
• 10:08
  And with every cycle, every time, the wing.
• 10:11
  Flaps. With every single cycle, that
• 10:13
  crack opens a nanometer wider. It propagates inward, slowly,
• 10:18
  quietly. The effective area of the part
• 10:21
  gets smaller and smaller and smaller until snap.
• 10:24
  It fails suddenly, without any warning.
• 10:26
  The. Data mentions and endurance
• 10:28
  limit. What's that about?
• 10:29
  Yes, this is particularly important for steel.
• 10:32
  The endurance limit is a theoretical stress level below
• 10:35
  which you can cycle the material indefinitely, literally forever,
• 10:39
  and it will never fail from fatigue.
• 10:41
  Infinite life That sounds too good to be true.
• 10:45
  Well, it is theoretical, and for some materials it's not really
• 10:47
  true for aluminum. Interestingly, it doesn't really
• 10:50
  have a true endurance limit. Aluminum will eventually fail if
• 10:52
  you cycle it enough times, no matter how low the stress.
• 10:54
  So that airplane wing it has a finite life.
• 10:58
  It has a very carefully calculated and inspected finite
• 11:01
  life, yes, but for steel, the sources warn us that endurance
• 11:05
  limit is fragile. Surface condition matters
• 11:08
  immensely. If the surface is rough or
• 11:09
  corroded, or if the part is very large, that endurance limit
• 11:13
  drops significantly. Because those rough spots are
• 11:15
  where the micro cracks start. Exactly.
• 11:17
  They are stress concentrators. A tiny scratch can be the
• 11:21
  genesis of a catastrophic failure millions of cycles
• 11:24
  later. I love the detail in the source
• 11:26
  about shot peening. It sounds like a paradox.
• 11:29
  You blast the surface with little metal balls to make it
• 11:32
  stronger. That sounds like you're damaging
• 11:34
  it. It does seem counterintuitive,
• 11:35
  doesn't it? You'd think you're just putting
• 11:37
  a bunch of dents in the surface, but by bombarding the surface
• 11:40
  was shot you are actually introducing compressive residual
• 11:45
  stresses. You are pre squashing the
• 11:47
  surface layer. OK, so you're squishing the
• 11:49
  metal on the outside. How does that help?
• 11:51
  Well, fatigue cracks need tension to open up.
• 11:54
  They need to be pulled apart to grow.
• 11:56
  If the surface is already under tight compression, if it's
• 11:58
  already squashed together, that compressive layer acts like a
• 12:01
  shield. The crack can't open.
• 12:04
  It has to overcome that built in compression first, before it can
• 12:07
  even start to experience tension.
• 12:09
  It increases the endurance limit significantly.
• 12:12
  That is brilliant. Fighting stress with stress.
• 12:14
  OK, so another failure mode, a much slower one, is creep.
• 12:18
  Creep the slow, permanent deformation of a solid material
• 12:23
  under a constant stress. We tend to think of metals as
• 12:26
  solids that hold their shape, but over a long time, especially
• 12:30
  at high temperatures, they flow. Like a glacier.
• 12:33
  Exactly like a glacier or asphalt on a hot day.
• 12:36
  You know how if you park a car on hot asphalt for a long time,
• 12:38
  the tires sink in a little? That's creep.
• 12:41
  Yeah, the source material points out the temperature dependence,
• 12:44
  which is key. Aluminum starts to creep at
• 12:46
  around 300°F. That's not that hot.
• 12:49
  A car engine gets way hotter than that.
• 12:51
  Which is why engine design is so tricky.
• 12:53
  You can't just use standard aluminum alloys for the hottest
• 12:55
  parts like Pistons. Steel, on the other hand, is
• 12:59
  more robust. It holds out until about 650°F
• 13:03
  before creep becomes a major issue.
• 13:05
  But what about something really hot, like a jet engine?
• 13:08
  Right. If you're designing A turbine
• 13:09
  blade for a jet engine which runs way, way hotter than that,
• 13:12
  creep is your number one nightmare.
• 13:14
  You absolutely do not want your turbine blade slowly stretching
• 13:17
  out over time and hitting the engine casing.
• 13:19
  No, you definitely do not. OK, then there's impact.
• 13:24
  This is the difference between placing a weight gently on a
• 13:27
  table and dropping it. The dynamic effect and the math
• 13:31
  in the source provides a really startling rule of thumb.
• 13:34
  A suddenly applied load, and this doesn't even mean dropped
• 13:37
  from a height, just applied instantly creates double the
• 13:40
  stress of a static load. Double just because it was
• 13:44
  sudden. Yes, it's because the material
• 13:46
  has to absorb the kinetic energy and that's minimum.
• 13:48
  If you drop it from any height, the multiplier becomes huge.
• 13:52
  This is where the property of resilience comes in, the ability
• 13:56
  of a material to absorb energy. And that's related to the stress
• 13:59
  strain curve? It's the area under the elastic
• 14:02
  portion of the stress strain curve.
• 14:03
  A resilient material like a good spring steel can absorb a lot of
• 14:07
  impact energy without permanently deforming.
• 14:10
  It acts like a shock absorber. So we have all these ways things
• 14:13
  can break. How do engineers predict it?
• 14:15
  The sources list these theories of failure.
• 14:18
  Why do we need theories? Can't we just, you know, break
• 14:20
  stuff and see what happens? We can and we do, but that's
• 14:24
  expensive and sometimes incredibly dangerous.
• 14:27
  We need to be able to predict failure before we build, and
• 14:30
  different materials fail in different ways.
• 14:33
  The sources highlight three main theories.
• 14:35
  Let's unpack them. The first one was the maximum
• 14:38
  stress theory, also called ranking.
• 14:40
  This is the simple one. It basically says failure
• 14:43
  happens when the principal stress in any One Direction hits
• 14:46
  the ultimate strength of the material.
• 14:48
  It works pretty well for brittle materials like cast iron or
• 14:51
  concrete. Why brittle materials?
• 14:53
  Because they don't really stretch or to form, they just
• 14:56
  snap when the force gets too high in One Direction.
• 14:58
  But what about for something like steel that does stretch?
• 15:02
  For ductile materials like steel, ranking is actually
• 15:05
  dangerous. It doesn't predict failure
• 15:07
  accurately at all. For that, the gold standard, the
• 15:10
  one every engineer learns, is the distortion energy theory, or
• 15:14
  Von Mises. Von Mises.
• 15:16
  I hear that name all the time in engineering circles.
• 15:18
  Check the von Mises stress. What is it actually checking?
• 15:22
  It's a much more robust theory. It basically says the material
• 15:25
  doesn't fail just because stress is high in One Direction.
• 15:28
  It fails when the energy used to distort the shape of the
• 15:31
  material reaches a certain limit.
• 15:33
  Distort the sheep. Yeah, to stretch it or shear it.
• 15:36
  It accounts for the complex interplay of stresses.
• 15:39
  Tension in One Direction, compression another, shear all
• 15:41
  at the same time. It's a much more complete 3D
• 15:44
  calculation of the stress state. And the third one, maximum shear
• 15:48
  theory. That's the guest theory.
• 15:50
  It's a little more conservative, a little simpler than von Mises.
• 15:54
  It assumes failure is caused by the sliding, the shearing of
• 15:58
  atomic layers within the material.
• 16:00
  It's a safe bet for ductile materials if you want to be
• 16:03
  extra careful, but von Mises is generally more accurate.
• 16:06
  OK, so we have the theories. Now let's look at the shapes,
• 16:09
  the structural elements, beams, columns, cylinders.
• 16:13
  The building blocks of the mechanical world with.
• 16:15
  Beams. The source defines this neutral
• 16:18
  axis. What's going on there?
• 16:20
  Right when you bend a beam, say a simple wooden board across a
• 16:24
  Creek, the top fibers are being squeezed together.
• 16:26
  That's compression. Makes sense.
• 16:28
  And the bottom fibers are being pulled apart.
• 16:30
  That's tension somewhere in the middle.
• 16:33
  There has to be a layer that's doing neither.
• 16:36
  It's not being stretched or compressed.
• 16:38
  It's just going along for the ride.
• 16:39
  It's just going along for the ride.
• 16:40
  That's the neutral axis. It's the eye of the storm.
• 16:43
  And the stress distribution changes along the beam too,
• 16:46
  right? It does.
• 16:47
  Vertical shear is highest at the supports where the beam rests on
• 16:51
  the Bank of the Creek, but the bending moment, the actual force
• 16:56
  that's trying to snap the beam, is highest right in the center
• 16:59
  where you're standing. Which explains why if it breaks,
• 17:02
  it usually snaps in the middle, but if it shears it would slice
• 17:06
  off at the edge. Now columns.
• 17:09
  This seems straightforward. It's just a vertical beam.
• 17:11
  Right, you would think so, but columns are dangerous because of
• 17:13
  completely different failure mode buckling.
• 17:16
  Buckling. Yeah, if you have a short stubby
• 17:19
  block, it fails by crushing simple compression, but a long
• 17:24
  skinny column fails by bowing out sideways, long before the
• 17:28
  material itself actually crushes.
• 17:30
  I've seen that with a ruler, if you press on the ends it pops
• 17:32
  out. Exactly.
• 17:34
  The source mentions the slenderness ratio.
• 17:36
  That's the key metric. It's the length divided by a
• 17:39
  property called the radius of gyration.
• 17:41
  The higher that ratio, the skinnier the column, the more
• 17:44
  likely it is to buckle. Euler's formula is what helps us
• 17:47
  calculate that critical buckling load.
• 17:49
  There's a historical note here that's pretty tragic.
• 17:51
  The Quebec Bridge collapse in 19 O 7.
• 17:54
  A devastating failure. It was a massive cantilever
• 17:58
  bridge and the main compression chords.
• 18:00
  These huge columns failed. They buckled.
• 18:04
  The math they were using at the time just wasn't adequate for
• 18:06
  structures of that scale. So what happened after that?
• 18:09
  That disaster forced the entire engineering community to rethink
• 18:12
  how we test and calculate stresses for massive steel
• 18:16
  columns. They literally had to build full
• 18:18
  size replicas of the bridge sections and crush them to
• 18:21
  understand where the old formulas went wrong.
• 18:24
  It's amazing and terrifying how much of our safety rules are
• 18:26
  written in the debris of past failures.
• 18:28
  It's true. OK, finally, cylinders.
• 18:30
  The sources distinguish between thin walled and thick walled.
• 18:33
  Right. Think of a soda can versus a
• 18:35
  cannon barrel in a thin walled cylinder.
• 18:38
  Under pressure like a boiler or a hydraulic pipe, you have two
• 18:43
  main stresses to worry about. There's longitudinal stress
• 18:46
  trying to stretch the pipe longer, and then there's hoop
• 18:49
  stress trying to expand the circumference.
• 18:51
  Make it fatter. Like the metal hoops on an old
• 18:53
  wooden barrel. That's where the name comes.
• 18:55
  From exactly and the math reveals something absolutely
• 18:58
  critical. The hoop stress is double the
• 19:02
  longitudinal stress. Twice as much.
• 19:04
  Yes, the force trying to split the pipe open lengthwise is
• 19:08
  twice as strong as the force trying to pull it apart at the
• 19:10
  ends. This is why if you overcook a
• 19:13
  hot dog, it always splits lengthwise.
• 19:16
  I never thought of that. The hoop stress in the casing is
• 19:19
  too high. Precisely.
• 19:20
  And the source notes that old steam boilers were always double
• 19:24
  riveted on those longitudinal seams for exactly this reason.
• 19:27
  You have to reinforce the weak point.
• 19:28
  Wow. OK, so we've analyzed the
• 19:31
  stress. We know how to keep the machine
• 19:33
  from breaking apart. Now we need to keep it moving.
• 19:35
  Let's move to Pillar 2, Lubrication.
• 19:38
  The science of separation because if metal touches metal
• 19:41
  under load, you get friction, welding, tearing and very rapid
• 19:45
  destruction. We need a fluid in between.
• 19:48
  The sources start by dividing surfaces into conformal and non
• 19:51
  conformal. It feels like we're back in
• 19:54
  geometry class. It is geometry, but it's
• 19:56
  geometry with serious consequences.
• 19:58
  Imagine a journal bearing a shaft spinning inside a bushing.
• 20:02
  The curve of the shaft matches the curve of the whole.
• 20:05
  They fit snugly. That's conformal.
• 20:07
  The load is spread out over a large area.
• 20:09
  OK. And non conformal.
• 20:10
  Think of a ball bearing or gear teeth.
• 20:13
  You have a sphere touching a flat race or two curved teeth
• 20:16
  meeting at 1:00 tiny spot. They touch at a single point or
• 20:19
  maybe a line. So the contact area is tiny.
• 20:22
  Infinitesimally small. Which means the pressure must be
• 20:25
  insane. Enormous.
• 20:26
  Yeah, absolutely astronomical. And that fundamental difference
• 20:30
  dictates the type of lubrication you need.
• 20:31
  We have 3 regimes. First, the one we always want,
• 20:34
  hydrodynamic lubrication. Is the ideal state right?
• 20:37
  Perfect scenario. It is.
• 20:39
  Think of water skiing. The motion of the skier creates
• 20:43
  a wedge of water that lists them up off the surface in a journal.
• 20:47
  Bearing the spinning shaft does the same thing.
• 20:50
  It drags oil into a converging wedge between the shaft and the
• 20:54
  housing and. It creates pressure.
• 20:55
  It creates a pressure wave that physically licks the shaft
• 20:58
  completely off the metal. So it's floating on oil.
• 21:01
  Completely separated, it's floating.
• 21:03
  This is why a properly lubricated engine bearing can
• 21:06
  last almost for forever. In theory there is 0 where
• 21:10
  because there is no contact. But the source notes that the
• 21:12
  film thickness is significantly affected by load.
• 21:15
  If you overload it, you sink. Correct.
• 21:17
  It's like trying to put a semi truck on water skis.
• 21:19
  You're just going to sink. Now contrast that with the
• 21:22
  second regime, elastohydrodynamic lubrication
• 21:25
  or EHL, right? This is what happens in those
• 21:28
  nonconformal contacts, the ball bearings and gears we just
• 21:31
  talked about. Where the pressure is crazy
• 21:33
  high. High, The source mentions
• 21:35
  pressures over 700 mega Newtons per square meter.
• 21:39
  That is an almost unimaginable crushing force concentrated on a
• 21:43
  tiny spot. 700 Yeah, that sounds like it would just squeeze the
• 21:48
  oil out instantly, like stepping on a ketchup packet.
• 21:51
  You would think so. Any normal fluid would just be
• 21:54
  gone. The two things happen.
• 21:56
  First, the metal itself actually deforms.
• 21:59
  It dents elastically to make the contact area a little bigger.
• 22:02
  That's the elasto part. The steel dense.
• 22:04
  It does, just for a microsecond. But the hydrodynamic part is the
• 22:08
  real magic. Under that extreme pressure, the
• 22:11
  viscosity of the oil changes dramatically, gets thicker, it
• 22:14
  gets exponentially thicker. For the split second it's in
• 22:17
  that high pressure contact zone, it essentially turns into a
• 22:20
  solid, becomes glassy. It's hard enough to push the two
• 22:24
  metal surfaces apart. That is mind blowing.
• 22:26
  The oil turns into a solid Shim just when you need it, and then
• 22:29
  it turns back to a liquid as soon as it leaves the zone.
• 22:32
  Physics is handy that way, and what's really interesting is
• 22:35
  that unlike hydrodynamic lubrication, the film thickness
• 22:38
  in EHL is barely affected by load.
• 22:41
  You can pile on the weight and that glassy film just holds up.
• 22:44
  But then we have the third regime, the one we don't want,
• 22:48
  boundary lubrication. This is what happens when the
• 22:50
  film breaks down. Maybe the speed is too low to
• 22:52
  build up a wedge, or the load is too high, or the oil is too
• 22:55
  thin. The bulk oil is squeezed out.
• 22:58
  So the peaks are starting to tuck.
• 23:00
  The asperity peaks are starting to make contact.
• 23:02
  Now you're relying on molecular layers on films that are only 5
• 23:08
  to 10 nanometers thick. We're talking surface chemistry
• 23:10
  here, not fluid dynamics. Exactly.
• 23:12
  This is where oil additives are critical.
• 23:14
  You have these long chain molecules that anchor themselves
• 23:18
  to the metal surfaces like seaweed on a rock, trying
• 23:20
  desperately to keep the peaks of the metal from actually welding
• 23:24
  to each other. Let's talk about the lubricant
• 23:26
  itself. Viscosity is the big word.
• 23:28
  Viscosity is simply the internal friction of the fluid.
• 23:31
  It's resistance to flow. Think molasses versus water, and
• 23:36
  the sources distinguish between dynamic viscosity and kinematic
• 23:39
  viscosity. Kinematic is just the dynamic
• 23:42
  viscosity divided by the fluid's density.
• 23:44
  And we have a choice between grease and oil.
• 23:46
  Why pick one over the other? Seems that oil is always better.
• 23:49
  Not always. Grease is basically oil that's
• 23:52
  trapped in a thickener, like oil held in a sponge.
• 23:55
  It's great because it stays put. It seals out dirt.
• 23:58
  You pack a wheel bearing with grease and it stays there for
• 24:01
  years. It's simple.
• 24:02
  So less maintenance. Much less, but the source gives
• 24:05
  a specific warning. The jewel effect limits grease.
• 24:08
  The jewel effect? I thought that was about
• 24:10
  electricity. In this context it refers to
• 24:12
  internal heating from churning. Grease doesn't flow well, so it
• 24:16
  can't carry away heat. If you run a grease lubricated
• 24:19
  bearing too fast, the source gives a speed factor of over
• 24:23
  200,000. The churning generates heat that
• 24:26
  can't escape. The grease literally cooks.
• 24:29
  It degrades and fails. So for a high speed you have to
• 24:31
  use oil. You need oil.
• 24:33
  Oil flows, it flushes out contaminants, and crucially, it
• 24:35
  acts as a coolant. It carries heat away from that
• 24:38
  critical contact zone. OK.
• 24:39
  So we've analyzed the stress and we've lubricated the moving
• 24:43
  parts, but a machine is full of fluids, oil, coolant, compressed
• 24:48
  air, fuel. We need to keep them where they
• 24:50
  belong. That brings us to Pillar 3, Seal
• 24:53
  technology. Keeping the pressure in and, we
• 24:56
  can divide seals into two main camps, static for parts that
• 24:59
  don't move relative to each other and dynamic for parts that
• 25:02
  do. Static seems simple enough.
• 25:05
  Gaskets. It's just a piece of cork or
• 25:07
  rubber you squash between 2 pipe flanges.
• 25:10
  It seems simple, but there's a real science to the squish.
• 25:14
  The source introduces 2 critical factors for gasket design, the M
• 25:18
  factor and the Y factor M&Y. the Y factor is the seating stress.
• 25:23
  It's the minimum pressure you need to apply to smash the
• 25:27
  gasket into the flange initially.
• 25:28
  To create a seal, you have to deform it so it flows into all
• 25:32
  the little microscopic imperfections in the metal
• 25:34
  surfaces. OK, so that's the initial
• 25:36
  tightening, the brute force to make it conform.
• 25:37
  Right, but once you do that, you turn the machine on.
• 25:40
  The pipe pressurizes, the internal fluid pressure now
• 25:43
  tries to blow the flange apart. The M factor is the maintenance
• 25:46
  factor. It's a multiplier that tells you
• 25:48
  how much residual clamping pressure you need to keep on
• 25:51
  that gasket to ensure it holds against the fluid trying to
• 25:53
  escape. So you can't just tighten it and
• 25:55
  forget it, you have to make sure the bolts maintain that clamping
• 25:58
  force even when the the system is actively fighting back.
• 26:02
  Exactly. Now what about O rings?
• 26:04
  These are everywhere. Everywhere.
• 26:06
  And the jewel effect comes back here, but in a totally
• 26:09
  different, really weird way. This is one of my favorite
• 26:12
  counterintuitive facts in all of engineering.
• 26:15
  Rubber, when it's under tension, when it's stretched, behaves
• 26:19
  strangely. If you heat it up, it contracts.
• 26:21
  It shrinks. Most things expand when they get
• 26:24
  hot. But rubber, if it's stretched,
• 26:26
  contracts. It has to do with thermodynamics
• 26:29
  and entropy and polymer chains coiling up.
• 26:31
  Now imagine you take an O-ring and you stretch it over a
• 26:34
  rotating shaft to seal it. OK, the shaft is spinning inside
• 26:38
  the stretched o-ring. The shaft spins.
• 26:40
  Friction generates heat. The O-ring heats up because it's
• 26:44
  stretched. The heat makes it contract.
• 26:46
  It squeezes tighter on the shaft.
• 26:47
  Which creates more friction? Which?
• 26:49
  Creates more heat, which makes it contract even more.
• 26:51
  It's a catastrophic positive feedback loop.
• 26:54
  The thermal runaway. The thermal runaway of
• 26:56
  destruction. The seal fails, and it fails
• 26:58
  rapidly. So how do you avoid that?
• 27:00
  How do you use an O-ring on a shaft?
• 27:02
  You never ever stretch an o-ring over a rotating part.
• 27:06
  You design the groove so the o-ring is in compression.
• 27:09
  You put the ring in a groove in the housing and the shaft runs
• 27:12
  inside it. You squash it into the hole
• 27:15
  rather than stretching it over the shaft.
• 27:17
  That is a life saving tip for a designer.
• 27:19
  Wow. Now what about dynamic seals for
• 27:23
  when the shaft is spinning at say 10,000 RPM?
• 27:27
  You can't just squeeze a piece of rubber against it, it would
• 27:29
  melt. No, it burn up instantly.
• 27:31
  For high speeds we have to use more sophisticated, often non
• 27:35
  contact seals like labyrinth seals.
• 27:38
  Labyrinth like a maze. Exactly.
• 27:40
  It's a non contact seal. You create a torturous path for
• 27:43
  the fluid, a series of sharp teeth on the rotor and grooves
• 27:46
  in the stator. The fluid has to navigate this
• 27:48
  maze. It tries to rush through the 1st
• 27:51
  tiny gap, then it expands to a large chamber.
• 27:53
  It loses energy and pressure. Then it tries to rush through
• 27:56
  the next. Gap it gets tired.
• 27:57
  Effectively, at each step, at each tube, it loses pressure.
• 28:01
  By the time it gets to the other side, the pressure is negligible
• 28:04
  and the leakage is minimal. And the source which mentions
• 28:07
  that sharp corners are better here.
• 28:09
  Why is that? Yes, sharp corners create
• 28:11
  turbulence. That turbulence dissipates the
• 28:14
  kinetic energy of the leaking fluid very effectively.
• 28:17
  You want the flow to struggle as much as possible.
• 28:20
  Then there are brush seals. These sound like a toothbrush.
• 28:23
  A very, very high tech toothbrush.
• 28:25
  They are made of densely packed bristles, thousands of them,
• 28:29
  made of super alloys. They actually touch the shaft,
• 28:32
  but they're flexible so they can bend and comply with its
• 28:34
  movement. So they close the gap much
• 28:37
  tighter than a labyrinth seal. Much tighter.
• 28:39
  The source says they can reduce leakage by 5 to 10 times
• 28:42
  compared to a labyrinth seal. That's a huge improvement in
• 28:45
  efficiency. That's a massive improvement.
• 28:47
  It is their standard and high performance turbo machinery now,
• 28:50
  like jet engines. OK, we are holding together,
• 28:53
  running smooth and sealing tight, but machines move and
• 28:58
  movement creates vibration. This brings us to Pillar 4,
• 29:03
  vibration and shock. The dynamics of motion.
• 29:06
  This is where things get shaky. Literally.
• 29:08
  The sources start with a classic model that every engineering
• 29:10
  student learns the single degree of freedom system, a mass, a
• 29:14
  spring, and a damper. It's the simplest way to
• 29:18
  understand the world of vibration.
• 29:20
  You have mass, which represents inertia, you have stiffness
• 29:23
  represented by the spring, and you have energy dissipation,
• 29:27
  which is the damper or dashpot. And every system like this has a
• 29:30
  natural frequency denoted as F Diane.
• 29:33
  Right, it's the frequency it wants to vibrate at if you just
• 29:36
  pluck it and let it go. Depends only on the stiffness
• 29:39
  and the mass. Specifically, it's proportional
• 29:42
  to the square root of stiffness over mass.
• 29:44
  So if you make it stiffer, the frequency goes up, if you make
• 29:46
  it heavier, the frequency goes down.
• 29:48
  Exactly. Think of a guitar string.
• 29:50
  Tighter string, higher note thicker, heavier string, lower
• 29:52
  note. Same physics.
• 29:54
  And the danger zone that everyone talks about is
• 29:56
  resonance. Resonance is when the forcing
• 29:58
  frequency, say the speed of your unbalanced motor, matches that
• 30:02
  natural frequency. When that happens, the system
• 30:05
  absorbs energy from the source with maximum efficiency.
• 30:08
  The amplitude of the vibration spikes, sometimes
• 30:11
  uncontrollably. This is the opera singer
• 30:13
  shattering the wine glass. She finds the note that matches
• 30:17
  the glasses natural frequency. That's the perfect.
• 30:19
  Example, the source talks about a term called transmissibility.
• 30:23
  At resonance, Transmissibility spikes.
• 30:26
  The vibration coming out of the system is much much larger than
• 30:29
  the vibration you're putting in. It amplifies.
• 30:32
  So how do we stop it? The goal is isolation.
• 30:35
  Isolation is the art of mismatching those frequencies.
• 30:39
  If your motor is running at 1800 RPM, you want the natural
• 30:43
  frequency of its mounting system to be much lower than that.
• 30:46
  If you can get the transmissibility below 1, you
• 30:48
  are actively isolating the vibration.
• 30:50
  That's why we put engines on soft rubber mounts.
• 30:53
  We're lowering the stiffness of the system, which lowers the
• 30:55
  natural frequency way below the engines operating speed.
• 30:59
  The engine can shake, but the soft mounts prevent that shaking
• 31:02
  from getting to the car's frame. The sources also talk about
• 31:05
  modes of vibration with nodes and antinodes.
• 31:09
  Right. If you vibrate something more
• 31:10
  complex, like a beam or a plate, it doesn't just move up and
• 31:14
  down, it forms a shape. There will be points that don't
• 31:16
  move at all. Those are the nodes and points
• 31:18
  that move the most. Those are the antinodes.
• 31:21
  And the fatigue failures we talked about way back in Pillar
• 31:24
  one, where do they happen? They almost always happen at the
• 31:27
  antinodes where the bending stress is at its maximum, or at
• 31:31
  the attachment points where the vibration is being constrained.
• 31:34
  If you're inspecting a machine for cracks, that's where you
• 31:37
  look first. Balancing is another huge topic
• 31:40
  here, the difference between static and dynamic imbalance.
• 31:43
  Static is the simple kind. If you have a heavy spot on a
• 31:46
  tire and you put it on a frictionless axle, gravity will
• 31:50
  pull the heavy spot down to the bottom.
• 31:52
  You can balance it without even spinning it.
• 31:54
  But dynamic imbalance is trickier.
• 31:56
  It is. It creates a couple a twisting
• 31:59
  force. You might have a heavy spot on
• 32:01
  one side of the tire and another heavy spot on the other side,
• 32:04
  but at the opposite end of the diameter.
• 32:07
  It's statically balanced. It won't roll to one side, but
• 32:10
  when you spin it, it will try to wobble violently.
• 32:13
  You can only detect it when the part is spinning.
• 32:15
  The analogy of balancing a car tire is perfect.
• 32:18
  They spin it up and put little weights on the rim to counteract
• 32:22
  that. Mr. Omega squared 4th.
• 32:24
  Precisely, you're trying to make the center of mass aligned
• 32:26
  perfectly with the center of rotation.
• 32:28
  Now, how do we measure this vibration?
• 32:30
  The accelerometer. And I love the analogy in the
• 32:33
  source. It compares the piezoelectric
• 32:34
  crystal inside to a sponge. It's a great visual, isn't it?
• 32:38
  A piezoelectric crystal is a material that generates an
• 32:41
  electric charge when you squeeze it.
• 32:43
  So think of it as a sponge full of electric juice.
• 32:45
  Inside the accelerometer you have this crystal with a small
• 32:48
  mass attached to it. When the machine vibrates up and
• 32:51
  down, the inertia of that mass squeezes and releases the
• 32:55
  crystal sponge. The harder the acceleration, the
• 32:58
  harder the squeeze, and the more juice comes out.
• 33:00
  We measure the juice, the voltage, and that tells us the
• 33:04
  acceleration. Yeah, it makes it so clear.
• 33:06
  It's brilliant. OK, finally, noise control.
• 33:09
  Vibration makes noise. Sound is just a compressional
• 33:12
  wave in the air created by a vibrating surface.
• 33:14
  To stop it, we look at a property called transmission
• 33:17
  loss. And the source brings up the
• 33:18
  mass law. The mass law is blunt force
• 33:21
  physics. Heavier is better.
• 33:23
  A heavy massive wall stops more sound than a light flimsy wall.
• 33:27
  It's simply harder for the sound waves energy to physically move
• 33:30
  the heavy mass. If you want a quiet room, build
• 33:33
  thick concrete walls. But there's a trap, a weird
• 33:36
  phenomenon called the coincidence effect.
• 33:38
  Yes, this is where physics plays a nasty trick on you.
• 33:41
  At a certain frequency, the wavelength of the sound wave in
• 33:44
  the air can match the natural bending wavelength of the panel.
• 33:48
  When that happens, the panel becomes acoustically
• 33:51
  transparent. It vibrates in perfect sympathy
• 33:53
  with the sound wave. So the sound goes right through.
• 33:55
  It goes right through the transmission loss, which should
• 33:58
  be getting better at higher frequencies, suddenly plummets.
• 34:01
  It's a form of resonance between the air and the structure.
• 34:04
  To fix these things, we use damping.
• 34:06
  The source lists three types, viscous, Coulomb and hysteresis.
• 34:11
  Right. Viscous damping is like a
• 34:13
  dashpot. It's dragged through a fluid.
• 34:15
  The force is proportional to velocity.
• 34:17
  The faster you try to move it, the more resistance you get.
• 34:20
  Coulomb is just dry friction, 2 things rubbing together.
• 34:24
  It provides a constant resistance force regardless of
• 34:26
  speed. And hysteresis is internal
• 34:28
  material damping. It's the material itself
• 34:31
  absorbing energy. Think of a rather ball.
• 34:34
  When you bounce it, it doesn't come back to the same height
• 34:36
  because some energy was lost as heat within in the rubber.
• 34:39
  That's hysteriosis. So what does this all mean?
• 34:42
  We've looked at stress, lubrication, seals and
• 34:45
  vibration. How do they all tie together?
• 34:47
  It means that successful mechanical design is a balancing
• 34:50
  act. It's a system.
• 34:52
  You need strength from your stress analysis, but you need
• 34:56
  fluidity from good lubrication. You need containment from your
• 34:59
  seals, but you need stability from vibration control.
• 35:03
  If you optimize one tiller at the expense of the others, the
• 35:06
  whole machine fails. It's an.
• 35:08
  Ecosystem. They're all interconnected.
• 35:09
  It is you can build the strongest engine block in the
• 35:12
  world. Perfect stress analysis.
• 35:14
  But if the head gasket fails because you didn't calculate the
• 35:17
  M&Y factors correctly, or the bearing CS because you didn't
• 35:21
  account for EHL, the engine is just a very, very expensive
• 35:24
  paperweight. It connects back to that idea of
• 35:26
  the system. A machine isn't just a
• 35:28
  collection of parts, it's an interaction of forces.
• 35:31
  Exactly. You can't isolate them.
• 35:32
  Before we go, I want to leave the listener with a thought.
• 35:34
  The factor of safety. We use it in all these
• 35:36
  calculations. We say the stress is X, so let's
• 35:39
  design the part to handle twice X, right?
• 35:41
  N = 2 or n = 4 for an elevator, maybe even.
• 35:45
  But is the factor of safety just a factor of ignorance?
• 35:49
  As we get better at understanding these 4 pillars,
• 35:52
  as our math gets more precise, our simulations more powerful,
• 35:55
  can we start to reduce that safety factor?
• 35:57
  Can we make machines lighter, more efficient, push them closer
• 36:01
  to the theoretical edge? Or does the chaos of the real
• 36:04
  world, the pothole you didn't see, the bad batch of oil of the
• 36:07
  operator air, always demand that margin for error?
• 36:10
  That is the eternal question of engineering, isn't it?
• 36:13
  It's the constant tension between optimization and
• 36:15
  robustness. How much do you trust your math
• 36:18
  versus how much do you fear the unknown?
• 36:21
  Something to Mull over. Yeah.
• 36:22
  Thank you for diving deep with us today.
• 36:24
  This was fantastic. My pleasure, it was a lot of
• 36:26
  fun. And if you haven't already, get
• 36:27
  over to our YouTube channel Mechanical Engineering Made
• 36:30
  Simple, where we post hands on how to videos, visual
• 36:33
  breakdowns, and practical engineering walkthroughs.
• 36:35
  You can actually use a family that solves together, stays
• 36:40
  together.