Latest / Mechanical Engineering Made Simple / (#145)Lesson 3: Why Pipes Burst and Pumps Fail

Transcript

(#145)Lesson 3: Why Pipes Burst and Pumps Fail · Mechanical Engineering Made Simple

0:00
Welcome back to mechanical engineering made simple legends.
0:03
I'm your host. Be water, my friends, or drown
0:06
in the math. Today we crack fluid mechanics,
0:09
the subject professors pitched as a lazy river but delivered
0:12
like a tsunami. Bruce Lee got it right.
0:14
Water flows or it crashes. Bottle it.
0:17
It becomes the bottle bad equation.
0:19
Instant nightmare. Be water.
0:21
Whatever the hell I just butchered.
0:22
I think I just improved the quote.
0:23
Rough pipe, drag city, Reynolds number the bouncer at the door.
0:26
Screw it up and cavitation eats your lunch.
0:29
Stick around. Learn to dodge the water
0:31
hammers. Reality hurls like fastballs.
0:33
Master the valve before it leaks.
0:35
Worse than a campaign promise. Be water my friends, or get
0:38
washed away. Rookie class in session.
0:42
Wow. Yeah, that is quite the syllabus
0:44
today. But I mean, it's vital, and we
0:46
are really stripping away the academic fluff here to look at
0:49
the fundamental laws that dictate how, you know, every
0:51
fluid interacts with our built environment.
0:53
Yeah, and that is exactly the entire mission of this deep
0:56
dive. Because look, we have compiled a
0:57
massive stack of textbook data engineering failure reports
1:00
theoretical models, but. But we're not going to sit here
1:03
and read equations at you. Right, exactly.
1:05
We are here to deliver a hard hitting practical lesson
1:09
designed specifically for you. You know, the engineer actually
1:12
trying to survive real world design challenges.
1:15
We're covering the crucial rules of fluid statics, the
1:18
destructive power of unsteady flows like water hammer which is
1:21
terrifying, the sticky reality of viscous pipes, open channel
1:25
nightmares, and the brutal fipses of drag, lift and
1:29
cavitation. But before we dive in, let's
1:31
establish the ground rules here. There are no pretty promises
1:34
today. Nope.
1:35
We are dealing strictly with the math and the physics that govern
1:38
reality. Fluid mechanics.
1:41
It well, it doesn't care about your deadlines.
1:43
It doesn't care about your budget or your architectural
1:45
vision. It.
1:46
Really doesn't. If you ignore these principles,
1:48
your pipes burst, your pumps will literally dissolve
1:51
themselves from the inside out, and your structures fail
1:54
catastrophically. It's unforgiving.
1:56
So let's start the very beginning.
1:58
Because before a fluid even starts moving and causing all
2:01
that chaos and is already laying traps for the unwary designer
2:04
just by sitting there. I'm talking about the invisible
2:07
weight of fluid statics, the trap of pressure.
2:10
Right, and the very first trap you'll encounter in the field is
2:13
just a fundamental misunderstanding of how we
2:15
measure pressure itself. OK, lay it out for us.
2:18
Well, we have to clearly separate absolute pressure from
2:21
gauge pressure. Absolute pressure is measured
2:24
relative to a perfect vacuum, so 0 absolute pressure means
2:28
there's absolutely nothing there like the vacuum of vape space.
2:32
Gauge pressure, however, is simply absolute pressure minus
2:36
the local atmospheric pressure. Because we live at the bottom of
2:39
an ocean of air. Exactly.
2:40
The air is constantly pushing down on us with the pressure of
2:43
roughly 101,325 pascals, or you know, 1 standard atmosphere.
2:50
So if I leave a pressure gauge sitting on my workbench, it
2:52
reads 0. But it's not that there's no
2:55
pressure there, it's just calibrated to ignore the weight
2:58
of the sky. Which leads to one of the most
3:00
common and honestly most dangerous pain points for young
3:03
engineers. The mistake is specifying the
3:07
wrong pressure sensor type for a sealed system.
3:10
Oh, I've seen this happen. Yeah, imagine you design A
3:13
sealed industrial chemical tank. You specify a gauge pressure
3:17
sensor when the math actually required an absolute pressure
3:20
sensor. Big mistake.
3:21
Huge. So you might start pumping fluid
3:24
out of that tank. You watch the gauge pressure
3:26
drop towards 0 and you think OK great, everything is fine.
3:29
But because you are using a gauge sensor, your quote UN
3:32
quote 0 just means you've reached atmospheric pressure.
3:35
Exactly. And if you keep pumping, that
3:37
needle drops below 0, you haven't just emptied the tank,
3:41
you've created a negative gauge pressure, a vacuum inside it.
3:44
And here's where reality hits hard.
3:46
If that tank was designed to hold pressure from the inside
3:49
pushing out, but it was never structurally reinforced to
3:52
withstand external atmospheric pressure crushing inward.
3:56
It implodes. It implodes.
3:57
It folds up like an aluminum soda can.
3:59
This crunches. Yes, that is a terrifying
4:02
reality check for a lot of engineers.
4:04
You have to know what your zero point actually means in the
4:08
physical world. It really emphasizes why we need
4:10
to understand the baseline physics before we draw a single
4:13
line in a CAD program. Which I guess brings us to
4:16
Pascal's law. This is kind of the magic trick
4:19
of fluid statics, isn't it? It is.
4:22
Pascal's law basically states that pressure applied to an
4:25
enclosed fluid is transmitted undiminished to every part of
4:28
the fluid and to the walls of its container.
4:31
And diminish. Right.
4:32
And this is the entire basis for heavy machinery.
4:34
The hydraulic press, the brakes in your car, the massive
4:37
actuators on an excavator. And the underlying logic here is
4:41
just brilliant. Pressure is simply force divided
4:43
by area, and because the fluid is incompressible and it's
4:47
enclosed, that pressure has to remain constant everywhere
4:50
inside the system. So if you have a tiny piston
4:53
connected by a tube to a massive piston, the math says force 2
4:57
equals force 1 multiplied by the ratio of area 2 to area.
5:00
One. Exactly.
5:00
It's a force multiplier. You push down with a small force
5:04
on a small piston and the fluid transmits that pressure to the
5:07
massive piston, creating this massive output force.
5:11
You can lift a two ton truck with the push of a finger.
5:14
But you know, physics always demands a trade off.
5:17
Right, there's no free lunch. Exactly.
5:19
Energy must be conserved, so to get that massive output force,
5:23
you sacrifice distance. You it'll push that small piston
5:26
A vastly longer distance just to move the large piston even a
5:30
fraction of an inch. Makes sense, but that is for an
5:32
enclosed fluid where we are actively applying external
5:36
mechanical pressure. What about a fluid just sitting
5:39
open to the sky, like a reservoir or the ocean?
5:42
Well, there we have to consider how pressure variations scale
5:45
with depth. Because the deeper you go, the
5:48
heavier the water above you get, right?
5:50
The relationship is a simple linear buildup.
5:53
The pressure at any depth relies on a really straightforward
5:56
calculation. You take the pressure at the
5:58
surface, which is usually just atmospheric pressure, and you
6:00
add the weight of the fluid column above you.
6:02
And that weight is dictated by the fluids density, the pull of
6:05
gravity, and the total depth. Yep, and that linear
6:09
relationship is incredibly powerful.
6:12
It tells us that pressure inside a liquid increases strictly in
6:16
proportion to depth. It does not matter if you were
6:19
at the bottom of a pipe that is 1 inch wide or at the bottom of
6:22
a lake that is 5 miles wide. Wait, really?
6:24
The width doesn't matter at all. Not at all.
6:27
If you are 10 meters down in fresh water, the pressure
6:30
squeezing in on you is exactly the same in both cases.
6:33
The total volume of water around you doesn't matter, Only the
6:37
vertical column of water directly above you matters.
6:40
That's wild to think about, which leads to how we actually
6:43
measure these pressure differences out in the field.
6:46
Let's talk about manometers, like the classic YouTube
6:49
Manometer. Oh yeah, it's wonderfully
6:51
simple. You just have a glass tube
6:53
shaped like AU filled with a heavy liquid, usually something
6:55
like mercury, right? You connect one end to your
6:57
pressurized pipe, leave the other end open to the
7:00
atmosphere, and the pressure in the pipe physically pushes the
7:03
mercury down on one side, forcing it up on the other.
7:06
And you just take a ruler and measure the height difference
7:09
between the two columns. It is brilliant in its
7:11
simplicity, but what if the pressure change you need to
7:15
measure is microscopic? Say you are measuring the draft
7:19
in a low velocity HVAC duct. A standard vertical YouTube
7:24
might only show like a millimeter of movement which is
7:28
impossible to read accurately with the human eye.
7:30
So we cheat gravity, the inclined manometer.
7:33
Exactly. We take one leg of that YouTube
7:36
and we tilt it to a shallow angle, say 10 or 20°.
7:39
Now when that tiny pressure pushes the liquid down on the
7:42
vertical side, the liquid on the tilted side has to travel a long
7:45
physical distance up the slope just to gain a tiny bit of
7:48
vertical elevation. Oh wow, so 1mm vertical change
7:52
might translate to a 10cm diagonal movement along the
7:55
glass tube. Right, you have effectively
7:57
created a mechanical amplifier for your eyeballs using basic
8:00
trigonometry. I love that, but let's be real,
8:03
you aren't going to have fragile glass tubes full of toxic
8:06
mercury strapped to the side of a vibrating diesel engine on a
8:08
job set. No, absolutely not.
8:10
For industrial applications, we use elastic deformation to
8:13
measure pressure, specifically the board and tube.
8:16
It's everywhere. Yeah, that classic dial pressure
8:18
gauge on an air compressor or a welder, right?
8:21
And the mechanism inside a board and tube is fascinating.
8:24
It's essentially A flattened metal tube that has been bent
8:27
into AC shape or sometimes a coil.
8:30
OK. When pressurized fluid enters
8:31
the tube, the internal pressure pushes against the flattened
8:35
walls, trying to force the tube back into a perfect circle, and
8:38
as the cross section tries to round out, the entire C shape
8:41
naturally tries to uncoil and straighten.
8:44
It's exactly like those paper party blowouts you get at
8:46
birthdays. You blow air in, it unrolls.
8:49
That is the perfect analogy. And in the gauge, the tip of
8:53
that uncoiling metal tube. The tube is hooked up to a tiny
8:55
gear train that spins the needle on the dial.
8:58
Simple but bulletproof. Yeah, but modern systems take
9:01
this a step further with diaphragm gauges.
9:03
Instead of a curving tube, the pressure pushes against a
9:06
flexible metal membrane. Like a drum head.
9:08
Exactly, and as the membrane bulges under pressure,
9:11
microscopic strain gauges bonded to its surface stretch, which
9:15
changes their electrical resistance.
9:17
This allows us to convert physical pressure directly into
9:20
an electronic signal. Which is essential for measuring
9:23
rapidly fluctuating pressures in like digital control system.
9:28
Absolutely, you need that instantaneous data.
9:30
OK, let's shift from the instruments back to the fluid
9:33
itself and look at the forces a static fluid exerts on massive
9:37
submerged objects. This brings us to the bathtub
9:41
Archimedes and the concept of buoyancy.
9:43
Ah. Yes, the legend of Archimedes
9:46
realizing that a body immersed in a fluid experiences an
9:50
awkward buoyant force equal to the weight of the fluid it
9:53
displaces. It's foundational.
9:55
Right when a ship sits in the water, it carves out a hole in
9:58
the ocean. And the ocean wants to fill that
10:00
hole back in, so it pushes upward.
10:02
And the center of gravity of that invisible displaced volume
10:05
of water is what we call the center of buoyancy.
10:07
But simply floating isn't the whole job, is it?
10:10
Not at all. A floating object must be
10:12
stable. If a ship floats but flips
10:14
upside down the moment a sea lands on it, well, it's a
10:16
terrible ship. Right, which introduces the
10:19
critical concept of the meta center.
10:22
Imagine a ship sitting perfectly level on calm water.
10:25
Gravity is pulling straight down through the ship's center of
10:28
gravity, and buoyancy is pushing straight up through the center
10:31
of buoyancy along the exact same vertical line.
10:35
Everything is in perfect equilibrium.
10:37
But the ocean is never calm. Let's say a massive wave hits
10:41
the side of the hull and the ship rolls 30° to starboard.
10:46
OK, when the ship rolls, the physical shape of the underwater
10:49
hull changes. The right side is pushed deeper
10:51
into the water and the left side list out because the shape of
10:55
the submerged volume has changed.
10:57
The center of buoyancy physically shifts over to the
11:00
right. So the buoyant force is no
11:02
longer pushing straight up through the center of the ship.
11:04
Exactly. Now, if you draw a new vertical
11:06
line straight up from this shifted center of buoyancy, the
11:10
point where that new line intersects the ship's original
11:13
central vertical axis is called the meta center.
11:16
OK, got it. And the absolute golden rule for
11:18
maritime stability is that this meta center must be located
11:22
above the ship center of gravity.
11:24
Wait, why? Because if the meta center is
11:26
above the center of gravity, you have gravity pulling down on the
11:28
center of the ship and buoyancy pushing up on the side that's
11:32
submerged. This creates A restoring couple,
11:36
a rotational torque that naturally tries to twist the
11:39
ship back to a level upright position.
11:41
The ship saves itself. Oh, that makes sense.
11:43
But if the meta center is below the center of gravity, that
11:46
torque acts in the opposite direction.
11:49
It actively pushes the ship further into the role.
11:52
The ship capsizes. The distance between the center
11:55
of gravity and the metacentre is known as the metacentric height.
11:58
OK, wait, let me stop you there because the data on metacentric
12:01
heights threw me for a loop. How so?
12:03
If a high metacenter means a massive restoring force and
12:06
extreme safely from capsizing, then a massive floating city
12:10
like a cruise ship should have the highest possible metacenter,
12:13
right? I would think so, but the
12:14
numbers show warships have metacentric heights around 1.2
12:18
meters while giant passenger ships are only around half a
12:21
meter. That feels totally backward.
12:24
Why make the civilian ship less stable?
12:26
It seems counterintuitive until you look at the human element of
12:29
engineering. You are entirely correct that a
12:31
high metacentric height means the ship is vastly more
12:34
resistant to capsizing, but naval architects refer to a ship
12:39
with a high metacentre as stiff. Yeah, when a stiff ship rolls in
12:45
a wave, that massive, powerful restoring force snaps it back to
12:48
the vertical position incredibly violently.
12:51
The ship aggressively writes itself.
12:53
If. You put thousands of tourists on
12:55
a stiff ship. That fast, jerky, violent
12:58
snapping motion would subject them to immense lateral G
13:01
forces. They would be violently seasick
13:03
before they even reach the buffet.
13:06
So a passenger ship is intentionally designed to be
13:08
what they call tender. Exactly.
13:10
It has a lower meta center so the restoring force is weaker.
13:13
When a wave hits it, it rolls slowly and it writes itself
13:15
slowly and gently. You sacrifice a margin of
13:18
ultimate extreme weather stability to ensure the
13:21
passengers aren't thrown out of their deck chairs.
13:22
You engineer for the use case. Precisely.
13:26
A warship, conversely, does not care about passenger comfort.
13:30
A warship needs to provide a highly stable level platform for
13:33
its radar and weapon systems to calculate firing arcs.
13:37
If a wave rolls a destroyer, it needs to snap back to dead level
13:41
immediately so the guns can stay on target.
13:44
That is fascinating. The physics are dictated by the
13:46
mission OK, keeping with the destructive forces exerted by
13:50
static fluids. There is a massive trap waiting
13:52
for civil engineers designing underwater structures like a dam
13:56
or a steel sluice gate. Oh, definitely.
13:58
We were talking about the difference between the geometric
14:00
centroid and the true center of pressure.
14:03
Right. So if you are an engineer
14:04
looking at a flat rectangular steel gate holding back 20 feet
14:08
of water in a reservoir, you might calculate the total force
14:11
of that water. And the rookie mistake is
14:13
assuming that because the gate is a perfect rectangle, the
14:16
force of the water pushes dead center on the gate, right on the
14:19
geometric centroid. And it is a fatal assumption.
14:22
You have to remember the fundamental rule we discussed
14:24
earlier. Pressure increases linearly with
14:26
depth. Right.
14:27
The deeper you go, the heavier it gets.
14:29
Exactly so. The pressure of the water at the
14:31
top of the gate is relatively low, but the pressure at the
14:34
very bottom of the gate is immense.
14:37
So the pressure distribution isn't a flat square, it's a
14:40
triangle getting wider and heavier at the bottom, yes.
14:44
And because the bottom of the gate is being pushed
14:46
significantly harder than the top, the effective point where
14:49
all that distributed force is mathematically concentrated, the
14:53
center of pressure is always pushed lower than the geometric
14:56
centroid is never in the middle. So if you design the structural
14:59
hinges for that gate, assuming the massive force of the
15:02
reservoir is hitting the middle, but the real force is actually
15:06
concentrated much lower down. Your torque calculations are
15:09
completely wrong. The hinges will experience a
15:12
massive twisting, bending moment that you never accounted for.
15:15
The steel shears, the gate blows out, and you have a catastrophic
15:18
flood on your hands. That is the unforgiving nature
15:21
of fluid statics. Even when water is doing
15:24
absolutely nothing, it's sheer uneven weight is waiting for you
15:27
to make a geometric error. It really is.
15:30
But we've established that fluids at rest follow
15:32
predictable, if dangerous, rules.
15:35
But what happens when you force that fluid to move and then try
15:38
to stop it abruptly? Well then, the physics shift
15:40
from predictable weight to violently destructive shock,
15:44
right? This brings us to unsteady flow
15:46
and the absolute nightmare phenomenon of water hammer.
15:49
We are moving from statics to dynamics now, specifically the
15:53
mechanics of instantaneous valve closure.
15:56
Imagine a long, heavy industrial pipeline moving thousands of
16:00
gallons of water a minute. At the end of that pipeline is a
16:03
massive shut off valve and for whatever reason a power failure,
16:06
a safety trip that valve slam shut instantly.
16:09
The fluid's velocity drops from high speed to absolutely zero in
16:12
a fraction of a second. Exactly.
16:14
And here's where a widespread simplification taught in
16:16
engineering schools proves fatal in the real world.
16:19
What's that? To make the math easier on
16:21
paper, we often treat liquids as perfectly incompressible.
16:25
And for steady everyday flow, assuming water cannot be
16:29
compressed is a perfectly fine approximation.
16:32
But under extreme kinetic shock, that assumption is a lie.
16:36
Oh wow, so it actually compresses?
16:38
Yes, when that valve slam shut, the momentum of thousands of
16:42
pounds of moving water has to go somewhere.
16:45
The water right at the closed valve stops, but the water a
16:48
mile back up the pipe is still moving at full speed, crashing
16:51
into the stopped water. It's like a multi car pile up on
16:53
a highway. The front car hits the wall and
16:56
the car is behind it, smashed into the front car.
16:58
Perfect analogy. The fluid physically compresses
17:01
packing in on itself, and the steel walls of the pipe are
17:04
forced to bulge outward to accommodate it.
17:07
The kinetic energy of the flow is instantly spontaneously
17:10
converted into a massive localized pressure spike.
17:13
But that pressure spike can't just stay at the valve, right?
17:15
No, it rebounds. It propagates upstream away from
17:18
the valve exactly like a shockwave, and the speed at
17:20
which this destructive wave travels is the speed of sound in
17:23
that specific fluid pipe system. And the math here reveals
17:27
something wild. The speed of that shockwave
17:29
relies entirely on a tug of war between how dense the fluid is
17:32
and how elastic both the water and the pipe wall are.
17:35
The stiffer the pipe, the faster and more destructive the wave.
17:38
We calculate this wave speed using a modified bulk modulus.
17:42
OK, break that down for us. The regular bulk modulus tells
17:45
you how compressible the fluid itself is, but because the
17:49
intense pressure wave also expands the pipe wall outward,
17:52
the fluid effectively acts even more compressible than it really
17:55
is. We have to account for the
17:57
elasticity of the steel or the concrete.
18:00
OK, so how fast are we talking? In a typical municipal water
18:03
main, this high pressure shockwave rips up the pipe at
18:07
speeds between 601,200 meters per second.
18:11
That's insane. Let's put that in perspective
18:13
for you. If I'm driving a car at 60 miles
18:16
an hour and hit a brick wall, the metal of the car crumples,
18:19
absorbing the energy. But if a massive column of water
18:22
in a pipe hits a suddenly closed valve, the water cannot crumple.
18:26
The pipe cannot crumple, so the kinetic energy blasts outward
18:29
into the pipe wall itself. It is literally a supersonic
18:32
shockwave firing backwards through your plumbing.
18:35
And the magnitude of that initial pressure spike is
18:38
brutal. You calculate the change in
18:40
pressure by multiplying the fluids density by the initial
18:43
flow velocity and multiplying that by the massive wave speed.
18:47
This shockwave can easily generate internal pressures 1020
18:51
or 30 times higher than the normal operating pressure of the
18:54
pipeline. And the nightmare doesn't end
18:56
after one hit, does it? The wave bounces.
18:58
It bounces. It travel all the way upstream
19:01
until it hits a reservoir or an open tank.
19:03
Once it hits the open tank, the high pressure suddenly relieves
19:06
itself into the open space, but nature overcorrects.
19:10
What happens? A low pressure wave then
19:12
reflects and travels all the way back down the pipe to the closed
19:15
valve. The fluid rushes back in, slams
19:18
in the closed valve a second time, and the cycle repeats.
19:21
The pipe is hit over and over again until friction finally
19:23
dampens the energy. So the critical timing here is
19:26
the waves round trip time, the time it takes the wave to travel
19:30
from the valve to the reservoir and back.
19:32
If your valve closes faster than that round trip time, the fluid
19:35
experiences it as an instantaneous closure and you
19:38
get the maximum possible pressure spike.
19:40
Exactly. To prevent your system from
19:42
blowing apart, you must control the closure time.
19:46
If you close the valve much slower than that round trip
19:48
time, the initial pressure wave has time to travel up, reflect,
19:52
and come back as a low pressure wave.
19:54
Oh, I see. Yeah, that returning low
19:56
pressure wave meets the still closing valve and counteracts
20:00
the pressure that is currently building up, canceling out the
20:02
worst of the spike. And the textbook data sites.
20:05
Allevi's equation for this Lorenzo Allevi, an Italian
20:08
hydraulic scholar, formalized the complex mathematics for slow
20:11
closing valves in the early 20th century.
20:14
He gave engineers the tools to calculate the exact pressure
20:18
curves if they extend the closing time safely.
20:20
He did, and the overarching lesson for the designer is this
20:23
never size your pipes, your welded joints, or your concrete
20:27
pipe supports based solely on the steady state operating
20:30
pressure. Right, because if a pump runs
20:32
AT50PSI and you use 50PSI supports, you're setting a trap.
20:36
You must calculate the transient spikes from water hammer when a
20:40
pump trips offline. In the real world, the water
20:42
hammer will physically tear inadequate pipe supports right
20:45
out of the concrete walls. It is the invisible
20:48
sledgehammer. OK, so we've just talked about
20:51
the catastrophic force water exerts when we try to stop it
20:54
abruptly, but what's happening when we just try to push it
20:57
steadily through a pipe? It turns out the water fights
20:59
back then too. Yeah, it fights the pipe walls
21:02
and it fights itself. This brings us to the nightmare
21:06
that is viscosity and viscous flow.
21:08
We now enter the realm of the Naver Stokes equations.
21:11
Oh boy. Yeah, these are the fundamental
21:13
governing equations of fluid motion.
21:16
They were essentially Newton's second law.
21:18
Force equals mass times acceleration applied to a
21:21
continuous fluid. OK, So what forces are we
21:24
looking at? They account for the pressure
21:26
forces pushing the fluid body, forces like gravity pulling on
21:29
it, and crucially, the internal viscous friction forces dragging
21:33
it down. The sources mentioned convective
21:35
acceleration and vorticity in relation to Napier Stokes.
21:39
Can we translate that out of textbook jargon?
21:41
Of course, convective acceleration represents how the
21:44
velocity of a fluid particle changes as it moves through
21:47
space, even if the overall flow remains steady over time.
21:50
OK, give me an example. Think of a wide, slow river
21:54
suddenly squeezing through a narrow rocky gorge.
21:58
The water has to speed up as it enters the gorge.
22:00
That spatial speeding up is convective acceleration.
22:04
Got it. And vorticity.
22:06
Think of vorticity like a kayaker paddling down a fast
22:09
river. If the water on the right side
22:11
of the kayak is moving faster than the water on the left side,
22:14
the kayak naturally spins in a circle.
22:17
Vorticity is the mathematical measurement of that local
22:20
rotation or spin within the fluid at a microscopic level.
22:23
OK, that makes sense. Because the Navier Stokes
22:26
equations contain complex non linear terms for this convective
22:29
acceleration and vorticity, they are famously difficult to solve.
22:33
In fact, finding a general analytical solution is one of
22:36
the Millennium Prize problems in mathematics.
22:39
But you know, engineers can't wait for a math prize to be one.
22:41
We have to build things now, so we solve them for simplified
22:44
cases, right? Like pure, perfectly smooth
22:47
laminar flow. Exactly.
22:49
In a strictly laminar regime, viscous forces completely
22:52
dominate. The fluid flows in perfectly
22:55
smooth parallel layers, sliding over each other like a deck of
22:58
cards, without any lateral mixing or disruption.
23:01
When we apply Naver Stokes to laminar flow in a circular pipe,
23:04
the complex math collapses beautifully into the Hagen
23:07
Poiseux formula. And the Hagen Poiseux formula is
23:10
a designer's dream. It proves that in perfect
23:13
laminar flow, the pressure drop along a pipe is directly
23:17
linearly proportional to the velocity of the fluid.
23:21
If you want to double the flow rate of thick oil through a
23:23
pipe, you simply double the pressure from the pump.
23:26
It is clean, predictable math. But this brings us to the
23:30
massive pain point we warned about in the opening.
23:32
Here it is. The trap is miss applying the
23:35
Hagen Pozzu formula to a flow that is not actually laminar.
23:39
An engineer might calculate a pressure drop assuming perfectly
23:42
smooth laminar flow size of pump based on those clean linear
23:46
numbers and sign off on the installation.
23:48
And then what happens? They turn the system on and the
23:50
pump stalls out. It cannot move the fluid.
23:53
Because in the real world, flow is almost never perfectly
23:56
laminar. It is chaotic.
23:58
It is turbulent. Exactly.
24:00
And how do we predict if a flow will be smooth or chaotic?
24:02
We use the Reynolds number. The bouncer at the door.
24:04
The ultimate arbiter of flow regimes.
24:07
It is a dimensionless ratio comparing the inertial forces,
24:10
the momentum pushing the fluid forward against the viscous
24:13
forces, the internal friction trying to hold the fluids layers
24:17
together. Osborne Reynolds visualized this
24:19
beautifully. In the 1880s, he injected a thin
24:22
stream of dye into a glass pipe of flowing water.
24:25
At low speeds, low Reynolds numbers, the dye stayed in a
24:28
perfectly straight, crisp line. The viscous forces were strong
24:32
enough to keep the water layers orderly.
24:34
But as he increased the speed, the inertia overcame the
24:37
viscosity. The dye line began to waver,
24:40
then suddenly burst into chaotic swirls, mixing completely with
24:43
the water. And that transition point in the
24:45
circular pipe typically occurs at a Reynolds number of around
24:48
2320. Below that, the flow is kept
24:52
orderly and laminar. Above it, inertial forces tear
24:55
the smooth layers apart and the flow becomes fully turbulent.
24:58
And once that flow goes turbulent, the clean math of the
25:01
Hagen Poiseau formula is totally useless.
25:04
Completely useless in turbulent flow.
25:07
The pressure drop no longer scales linearly with velocity,
25:11
it scales roughly with the square of the velocity.
25:13
Oh wow. So if you double the flow rate
25:15
in a turbulent pipe, your pressure drop does not double,
25:18
it quadruples. The friction skyrockets.
25:21
Your pump has to work four times as hard just to fight the chaos
25:24
inside the pipe. So what exactly is that chaos?
25:27
What is physically happening inside the pipe when it goes
25:30
turbulent? Well, it is not just random
25:32
static, it is highly structured multi scale mixing in turbulent
25:37
flow. We see the introduction of what
25:39
we call the Reynolds stress. In addition to the normal
25:42
viscous friction sliding between fluid layers, you have
25:46
microscopic erratic fluctuations and velocity in all directions.
25:49
So things are just crashing into each other.
25:51
Fluid particles are moving laterally, constantly colliding
25:54
and mixing across the entire diameter of the pipe.
25:56
Ludvig Prant tried to explain this in the 1920s with his
25:59
mixing length theory. He did rant.
26:02
Dawdle envisioned small parcels of fluid or eddies traveling a
26:06
certain transverse distance, which he called the mixing
26:08
length, before violently colliding with other fluid
26:11
parcels and transferring their momentum.
26:14
Turbulent flow is a physical mechanism of momentum transport,
26:18
right? Because these eddies are
26:19
physically crashing into each other, turbulence is vastly more
26:23
effective at transferring momentum, heat, and dissolved
26:26
chemicals than lameter flow ever could be.
26:28
Let's visualize that turbulence is like a chaotic relay race.
26:33
Massive eddies form in the center of the pipe, possessing
26:36
huge kinetic energy. They break apart into smaller
26:39
eddies, handing that kinetic energy down the line, breaking
26:42
into smaller and smaller vertices, right down to the
26:44
microscopic level where the fluid's viscosity finally acts
26:48
like friction breaks, burning that kinetic energy off as raw
26:50
heat. That's a great way to put.
26:52
It but there is a catch right at the physical wall of the pipe
26:55
the party stops. The sources call it the viscous
26:57
sub layer. It is a critical boundary.
27:00
Yeah. Even in the most violently
27:02
turbulent flow, the fluid velocity right at the solid
27:06
steel wall of the pipe must be exactly 0.
27:10
This is known as the no slip condition.
27:12
Zero. It doesn't move at all.
27:14
Right, because the velocity must drop to absolute zero at the
27:17
surface, there is a microscopic ultra thin layer of fluid right
27:21
against the wall where viscous friction completely dominates,
27:25
forcefully suppressing all turbulence.
27:28
This is the viscous sublayer. If I am designing an industrial
27:31
heat exchanger, trying to boil a fluid inside a pipe, that
27:34
sublayer sounds like a nightmare.
27:36
It's like trying to scrape burnt milk off the bottom of a pot
27:39
while stirring the top. The sublayer is stagnant.
27:42
It acts as an incredibly stubborn insulating barrier.
27:45
The turbulence in the center of the pipe is mixing heat
27:47
beautifully, but that microscopic viscous sub layer
27:50
acts like a thermal blanket against the wall, preventing the
27:53
heat from getting out. So how do you fix it?
27:55
To improve heat transfer, thermal engineers spend their
27:57
entire careers designing pipes with internal ridges, fins, or
28:02
artificial roughness to intentionally disrupt and tear
28:05
apart that viscous sub layer, violently dragging the
28:08
turbulence closer to the hot wall.
28:10
OK, so we've mapped out how fluids behave when they are
28:13
trapped and pressurized inside a closed pipe.
28:16
But when you RIP the roof off the pipe and expose the water
28:19
directly to the atmosphere and gravity, the rules change
28:22
entirely. We are moving into open channel
28:24
flow and the raw power of hydraulic jumps.
28:27
Yeah, when you remove the pressure boundary of a closed
28:29
pipe, you lose the ability to push the water.
28:32
With a pump. You enter the world of open
28:34
channel flow, where the fluid has a free surface exposed to
28:37
atmospheric pressure. And it's all gravity from there.
28:40
The entire flow is driven strictly by gravity, pulling the
28:43
mass of the water down a physical slope.
28:45
This dictates the engineering of everything from ancient Roman
28:48
aqueducts to modern sewer mains, irrigation canals, and concrete
28:52
flood control channels. If we look at the historical
28:55
data, the Romans were masters of this.
28:58
They built aqueducts dropping only a few inches per mile over
29:01
vast distances. But mathematically calculating
29:04
how fast that gravity driven water will actually move is
29:07
tricky because you aren't dealing with smooth pressurized
29:11
pipes, you are dealing with rough terrain.
29:13
Right, because the physical scales of rivers and large
29:16
channels are so massive, the Reynolds numbers are incredibly
29:20
high. The flow is almost entirely
29:22
turbulent. Therefore, the friction holding
29:25
the water back depends heavily on the physical roughness of the
29:28
channel wall. The sources point us to the
29:31
cheesy formula and the heavily utilized Manning equation, and
29:35
the Manning equation relies on a specific empirical roughness
29:38
coefficient, denoted as N. The roughness is the defining
29:41
variable here. Absolutely.
29:42
If you build a smoothly troweled perfect concrete drainage
29:45
channel, your in value might be a very low 0.012.
29:50
The water flows fast and efficiently, but if you were
29:53
analyzing a natural riverbed choked with weeds, large
29:56
boulders and winding turns, the in value might jump to 0.05 or
30:00
higher. That's a huge difference.
30:02
It is that coefficient dictates exactly how much kinetic energy
30:06
is bled off as friction, telling you how fast the gravity driven
30:10
water can actually move. And there's an incredible
30:12
optimization trick hidden in the geometry here.
30:15
If you are a civil engineer tasked with moving the absolute
30:19
maximum amount of flood water through a rectangular concrete
30:22
trench, you don't just make the trench as wide and shallow as
30:25
possible, no. Because you have to optimize
30:28
what we call the wetted erimeter.
30:30
OK, what is that? The wetted perimeter is the
30:32
linear length of the channel's cross section that is hysically
30:35
touching the water. That surface contact area is
30:39
where all the viscous friction happens.
30:41
To get the maximum volumetric discharge for a given
30:44
cross-sectional area of water, you must minimize that wetted
30:47
perimeter. You want the most water touching
30:49
the least amount of concrete. And the math gives us a golden
30:52
rule of thumb. For a standard rectangular
30:55
channel, the optimal most hydro dynamically efficient shape
30:58
occurs when the water depth is exactly half of the channel
31:01
width. If it's a 10 foot wide channel,
31:03
you want the water running exactly 5 feet deep for maximum
31:07
flow efficiency. But open channels rarely
31:09
maintain a perfect uniform slope, hills drop off, rivers
31:13
narrow, and spillways plunge. When the geometry changes, we
31:17
must rely on the concept of specific energy in the Froude
31:20
number to predict what the water will do.
31:23
Specific energy, right? Specific energy is the total
31:25
energy of the flow per unit weight, measured locally
31:28
relative to the channel bottom. It is the sum of the potential
31:32
energy the water depth and the kinetic energy the velocity
31:35
head. And because total energy must be
31:37
conserved, if the channel bottom drops suddenly or the walls
31:40
narrow, the water depth and velocity must drastically adjust
31:44
to maintain the energy balance. Exactly, and the furred number
31:46
dictates the personality of that adjusting water.
31:49
William Froude was a naval architect in the 1800s and his
31:53
dimensionless number is the absolute key to open channel
31:56
behavior. So what exactly is this number?
31:58
The Froude number is a ratio. It compares the water's actual
32:01
flow velocity to the speed at which a small surface wave can
32:04
travel through that specific depth of water.
32:07
If the Froude number is less than one, the flow is slower
32:10
than the speed of a surface wave.
32:12
This is called tranquil or subcritical flow.
32:15
OK, think of a deep, slow moving lazy river.
32:20
If you throw a heavy rock into a lazy river, the ripples expand
32:24
outward in perfect circles, right?
32:26
Because the water is moving slower than the wave speed, the
32:29
ripples can actually travel upstream, fighting against the
32:32
gentle current. The flow knows what is happening
32:35
ahead of it because pressure waves can travel upstream to
32:37
warn the incoming water. That's a perfect visual, but if
32:41
the slope steepens and the Froude number exceeds 1, you
32:44
enter rapid or supercritical flow.
32:47
The water is moving so fast, like rushing down a steep
32:49
concrete spillway, that it outruns its own surface waves.
32:53
If you throw a rock into a raging spillway, you don't see
32:56
circular ripples. The ripples are instantly swept
32:58
downstream faster than they can expand.
33:00
The slow is completely blind. It cannot send any pressure
33:03
signals upstream about obstacles or changes ahead.
33:06
Which sets the stage for one of the most spectacular and violent
33:09
collisions in fluid mechanics, the hydraulic jump.
33:13
A hydraulic jump happens when rapid supercritical flow is
33:17
suddenly forced to slow down. Maybe the steep slip of the
33:20
channel suddenly flattens out. Because the rapid flow is blind
33:24
and traveling faster than its own wave speed, it cannot
33:26
smoothly transition back to tranquil deep flow.
33:30
It has to crash. And it happens violently.
33:33
The high speed shallow water slams into the slower water
33:36
ahead of it. The water level abruptly crashes
33:39
upward, creating a chaotic, churning, highly turbulent
33:42
standing wave. Massive amounts of kinetic
33:45
energy are destroyed in this rolling frothing jump.
33:49
There is a massive pain point here for civil engineers
33:52
designing municipal drainage. If you design an underground
33:55
storm drain without accounting for the fraud number, you are
33:58
asking for destruction. You might design A pipe thinking
34:01
the rain water will flow smoothly, but if a steep grade
34:04
causes it to hit supercritical velocity and it suddenly hits a
34:07
flat section, it triggers A hydraulic jump inside a confined
34:10
concrete pipe. And that jump violently expands,
34:13
choking the pipe, compressing the air ahead of it, and it will
34:16
literally blow heavy iron manhole covers right out of the
34:20
street like geysers. It is a destructive force, yes,
34:23
but it is a misconception to think a hydraulic jump is always
34:26
an engineering failure. In the right context it is a
34:30
vital life saving tool. Definitely.
34:33
A hydraulic jump purposely dissipates massive amounts of
34:36
kinetic energy through extreme turbulence.
34:40
Think about the spillway of a massive structure like the
34:42
Hoover Dam or the Oroville Dam. Millions of gallons of water
34:46
rushing down a massive concrete face possessing catastrophic
34:50
world ending kinetic energy. If that rapid supercritical flow
34:53
is allowed to hit the natural riverbed at the bottom of the
34:55
dam at full speed, it would instantly act like a water jet
34:58
cutter. It would erode the bedrock,
35:00
scour the foundation, eventually undermine and collapse the
35:03
entire dam. So how do they stop it?
35:05
Engineers carefully design heavily reinforced concrete
35:08
stilling basins at the bottom of the spillway specifically to
35:12
trigger a hydraulic jump. The intentionally create this
35:15
violent crash to bleed off the kinetic energy harmlessly into
35:19
the water itself, turning it into heat and noise before it
35:22
safely enters the natural river. I want to tie this back to
35:25
something truly mind bending. You look at the math of a
35:28
hydraulic jump in a river, and you look at the math of a
35:31
supersonic shockwave in a compressible gas like a fighter
35:35
jet breaking the sound barrier. It is the exact same
35:38
mathematics. A tranquil river is subsonic
35:41
flow, a rapid spillway is supersonic flow, and the
35:44
hydraulic jump is the literal shockwave created when the fluid
35:47
hits the brakes. The math of a damn spillway is
35:50
the math of an F22 jet nozzle. Fluid mechanics isn't just a set
35:54
of rules, it's a universal language.
35:56
It is a beautiful demonstration of how universal these governing
35:59
equations are. The same physics supply whether
36:01
you are dealing with liquid water or compressible air.
36:03
Up until now we've talked about fluids flowing inside things,
36:06
pipes or along channels. But what happens when you invert
36:09
the problem? What happens when a solid object
36:12
is forced to move through a static heavy fluid?
36:16
Now we are looking at the aerodynamics of a car, the
36:18
hydrodynamics of a submarine or the wind loads trying to tear
36:22
down a skyscraper. Right, this is external flow,
36:25
drag and the deadly rhythm of vortex streets.
36:28
When an object pushes through a fluid, it violently displaces
36:32
it, forcing the fluid to part and flow around its contours.
36:36
This interaction generates 2 primary aerodynamic forces, drag
36:40
which acts parallel to the direction of flow, pushing the
36:43
object backward, and lift which acts perpendicular to the flow.
36:47
Let's focus on drag, the enemy of speed and efficiency.
36:51
The sources break drag down into two distinct mechanisms,
36:54
friction drag and form drag. Friction drag is exactly what it
36:58
sounds like is the direct sheer stress of the fluid physically
37:01
scraping along the surface of the body can find entirely
37:03
within that microscopic boundary layer we discussed earlier.
37:06
If you have a long, incredibly thin, flat metal plate flying
37:10
straight through the air like a blade, it experiences almost
37:13
entirely friction drag. The air is just scraping along
37:17
the metal. But most things aren't flat
37:18
plates. Cars, planes and buildings are
37:21
blunt. That brings in form drag, which
37:23
is pressure drag. It is caused entirely by the
37:26
physical macroscopic shape of the object.
37:29
Right. High pressure violently builds
37:31
up on the front nose of the object as it Rams into the
37:34
stationary air, and behind the object a low pressure wake forms
37:38
where the airflow separates and spirals away.
37:41
So you have high pressure on the front, low pressure on the back.
37:43
The difference between the crushing high pressure on the
37:45
nose and the empty low pressure on the tail effectively sucks
37:49
the object backward. Historically, understanding the
37:52
separation was a massive hurdle, wasn't it?
37:54
The sources highlight D'alembert's paradox.
37:57
In the 1700s, the brilliant mathematician Jean Laurent
38:00
D'alembert used perfect, flawless mathematical equations
38:05
to model an ideal non viscous fluid flowing around a sphere.
38:09
And his math proved that the fluid would part cleanly at the
38:12
front, flow perfectly around the sides, and close perfectly
38:15
behind the sphere, leaving 0 pressure difference.
38:18
Mathematically, in a perfect fluid drag shouldn't exist.
38:22
But if you stick your hand out the window of a moving car, your
38:25
hand gets ripped backward. Drag obviously exists.
38:28
Dale and Bear's perfect math didn't match reality.
38:31
Dale and Bird's paradox highlighted the absolute
38:33
unforgiving necessity of including internal viscosity in
38:37
our calculations. Because real fluids possess
38:40
viscosity, they experience internal friction.
38:43
Right as the boundary layer of air travels over the front of a
38:46
blunt object, it loses momentum friction.
38:49
When it reaches the back half of the object, it encounters an
38:52
adverse pressure gradient. The pressure is trying to push
38:55
the air back toward the front. The boundary layer simply runs
38:59
out of kinetic energy. It stops, peels off, and
39:02
physically separates from the surface of the object.
39:04
That separation point is everything.
39:06
When the flow peels off, it creates a massive, chaotic,
39:09
churning wake of low pressure behind the object.
39:12
That swirling void is the primary source of form drag.
39:15
Let's examine that wake more closely.
39:17
The flow around a simple cylinder is a classic
39:20
engineering benchmark. At extremely low speeds, low
39:23
Reynolds numbers, the fluid sneaks around the cylinder and
39:27
creates 2 perfectly stable, symmetrical little twin vortices
39:31
right behind it. Looks like 2 small whirlpools
39:33
just resting there. But as the speed increases, as
39:36
the Reynolds number pushes past 90, things get wild.
39:40
The wake becomes unstable. What happens?
39:42
Instead of two stable twin vortices, the eddies begin to
39:45
detach or shed alternately from the top and bottom of the
39:48
cylinder. 1 vortex spins off the top creating a pocket of low
39:52
pressure, then one spins off the bottom, then the top again.
39:56
It creates A staggering alternating pattern of swirling
40:00
vortices flowing downstream, yes.
40:03
This phenomenon is known as a Carmen vortex St., named after
40:06
the brilliant aerodynamicist Theodore von Karman.
40:09
And this alternating shedding is where the ultimate nightmare for
40:12
structural engineers lies. Vortex shedding causes
40:15
catastrophic structural resonance.
40:16
When a vortex sheds from the top of the cylinder, it creates A
40:19
localized drop in pressure on the top, physically pulling the
40:23
cylinder up. When the next vortex sheds from
40:26
the bottom, the low pressure pulls the cylinder down.
40:30
The cylinder is subjected to a powerful alternating cyclical
40:34
side to side force. And the frequency of this
40:36
alternating shedding, how fast it beats, is predicted by a
40:40
dimensionless parameter called the straw hole number, which
40:43
relates the shedding frequency to the objects diameter and the
40:46
flow velocity exactly. If that shedding frequency
40:49
happens to match the natural harmonic frequency of the
40:51
structure, say a tall industrial smokestack, a suspension bridge,
40:54
hanger cable, or a high voltage power line, in a steady
40:58
crosswind, the structure begins to sway, and because the wind
41:02
keeps beating at the exact right rhythm, like pushing a child on
41:05
a swing at exactly the right moment, the sway amplifies.
41:08
If it goes unchecked, the structure will violently shake
41:10
itself apart. That is precisely what brought
41:13
down the Tacoma Narrows Bridge on November 7th, 1940.
41:17
The wind across the Puget Sound was blowing at a steady 42 mph.
41:22
That steady wind speed over the blunt shape of the bridge deck
41:25
created a Carmen Vortex St. whose shedding frequency
41:28
perfectly matched the bridges natural torsional frequency.
41:32
The bridge began to twist and gallop wildly, eventually
41:35
tearing itself to pieces in a matter of hours.
41:38
It was a failure not of weight capacity, but a fluid mechanics.
41:41
To stop a structure from singing itself to death, you have to
41:44
break the physics. You either have to stiffen the
41:46
structure to change its natural frequency, or you have to
41:49
disrupt the vortex shedding entirely.
41:51
That is why when you drive past a refinery or a power plant, you
41:55
see long helical strikes, those spiral metal fins welded around
41:59
the top 30 feet of tall smokestacks.
42:01
They purposely break up the uniform wind flow, preventing
42:04
the Carmen vortices from forming synchronously along the whole
42:08
length of the stack. It is an elegant aerodynamic
42:11
solution. Now there is another highly
42:13
counterintuitive phenomenon regarding flow around a blunt
42:17
object. Let us look at what happens at
42:20
the critical Reynolds number. If you take a sphere or a
42:23
cylinder and keep speeding up the flow, eventually around a
42:27
Reynolds number of roughly 3.8 * 10 to the 5th, the drag
42:32
coefficient suddenly drops off a Cliff.
42:34
Right, the drag coefficient plummets from roughly 1.2 down
42:37
to 0.3. It is a massive instantaneous
42:40
reduction in air resistance. But why?
42:43
What happens at this critical high speed is that the ultra
42:47
thin boundary layer itself right on the front curve of the
42:49
cylinder transitions from smooth laminar flow to chaotic
42:53
turbulent flow before it reaches the separation point.
42:56
Wait, if turbulence is chaotic friction, why is a turbulent
42:59
boundary layer suddenly a good thing for reducing drag?
43:01
Because a turbulent boundary layer possesses significantly
43:04
more internal kinetic energy due to all that microscopic lateral
43:08
mixing we discussed earlier, high energy fluid from the outer
43:10
stream is constantly being mixed down into the slow fluid near
43:13
the wall. Right, so it has more energy to
43:15
fight with. Exactly because the turbulent
43:18
boundary layer has more kinetic energy, it can fight against the
43:21
adverse pressure on the back of the cylinder for much much
43:25
longer before it runs out of steam and is forced to separate.
43:29
The separation point is physically pushed significantly
43:32
further around to the back of the cylinder.
43:33
So the wake shrinks. Precisely.
43:36
The flow hugs the cylinder longer, so the chaotic low
43:39
pressure wake behind the object gets vastly narrower.
43:43
A narrower wake means a much smaller area of low pressure
43:47
pulling the object backward, which translates to a massive
43:50
drop in total form drag. OK, let me unpack this.
43:53
You're saying that sometimes making a surface rougher
43:55
intentionally trips the boundary layer into turbulence, which
43:58
actually reduces the total drag. That feels incredibly
44:01
counterintuitive for an engineer who spends all day trying to
44:04
Polish middle surfaces to reduce friction.
44:06
Yeah, I know, I know it sounds completely backward.
44:08
But yes, it is a deliberate trade off.
44:10
Polishing reduces friction drag, but for a blunt object, form
44:14
drag is the vastly dominant force.
44:16
You sacrifice a little extra friction drag to achieve a
44:19
massive reduction in form drag. This is the exact applied
44:23
science of why golf balls have dimples.
44:26
A perfectly smooth golf ball would maintain a laminar
44:28
boundary layer, the flow would separate early, leave a huge
44:32
sucking wake, and the ball would drop out of the sky after 100
44:35
yards. But the dimples change that.
44:37
The dimples intentionally tripped the boundary layer into
44:40
high energy turbulence. This turbulent layer delays flow
44:44
separation, drastically, shrinks the wake, drops the form drag,
44:48
and allows the ball to fly twice as far.
44:51
Fluid mechanics turning poor aerodynamics into a feature.
44:54
It's brilliant. So we know how to cheat drag by
44:56
manipulating the boundary layer. Now we have to look at the other
44:59
major force, lift, because the forces that give us flight hold
45:02
a very dark, incredibly destructive secret when applied
45:05
to liquids. We are talking about cavitation
45:07
and system destruction. Let us first establish the
45:10
fundamental mechanism of lift. The sources point us to the Kuda
45:13
Joukowski equation. It states that lift equals the
45:16
fluid density multiplied by the uniform flow velocity multiplied
45:20
by the circulation. And circulation is a
45:22
mathematical measure of the overall rotation or spin of the
45:25
fluid field around the object. Right.
45:28
It's easy to visualize with a spinning baseball.
45:31
A pitcher throws A curveball. The spinning seams drag the air,
45:34
creating circulation, and the ball experiences lift, curving
45:37
sideways. But an airplane wing or a pump
45:40
blade isn't physically spinning. No.
45:43
An aerofoil relies on its specific geometric shape to
45:46
force circulation to occur naturally.
45:49
It relies on the aquatic condition.
45:51
A wing has a smooth, rounded leading edge and a razor sharp
45:55
trailing edge. When the wing first accelerates
45:57
from a standstill, the air rushing over the top tries to
46:00
wrap around that sharp trailing edge to meet the air underneath.
46:03
But the geometry is too sharp and the flow is too fast.
46:06
The flow cannot make the turn. No, it violently separates from
46:09
the sharp edge spinning off a localized vortex, a starting
46:12
vortex that is left behind spinning in the air on the
46:14
runway. And nature abhors an unbalanced
46:17
equation. The universe demands that
46:19
angular momentum is conserved. Exactly because a vortex was
46:22
shed spinning in One Direction, a mathematical counter
46:25
circulation of equal and opposite strength is instantly
46:29
bound to the wing itself to balance the physical system.
46:33
This bound circulation speeds up the air flowing over the curve
46:36
top of the wing and slows down the air underneath.
46:39
Which brings us to Bernoulli's principle.
46:41
Higher fluid velocity means lower pressure.
46:44
The high speed air on top creates a zone of low pressure
46:47
and the slower air on the bottom creates a zone of high pressure.
46:51
The high pressure pushes up toward the low pressure,
46:54
resulting in a net upward aerodynamic force lift.
46:58
It's pure physics magic. It allows hundreds of tons of
47:01
metal to float in the sky until you get greedy and tilt the wing
47:04
too far back. Correct.
47:05
If the pilot increases the angle of attack too far, trying to
47:09
generate more lift, the air can no longer follow the steep
47:12
adverse contour of the upper surface.
47:14
The boundary layer runs out of kinetic energy.
47:16
The flow separates completely from the wing.
47:18
The low pressure zone collapses and lift plummets instantly.
47:21
This is an aerodynamic stall. It is a catastrophic loss of
47:25
lift, and pilots train extensively to recognize the
47:27
buffeting that precedes it and push the nose down to recover.
47:31
Stalling is terrifying, but the real dark side of Bernoulli's
47:34
principle isn't stalling an airplane.
47:37
It's what happens when you apply these same aerofoil shapes to
47:40
heavy, incompressible liquids. Let's talk about the absolute
47:44
nightmare of cavitation. Let's do it.
47:46
When an impeller blade inside a massive industrial water pump
47:49
spins, it acts exactly like an airplane wing.
47:53
It generates lift to grab and move the water, and just like an
47:57
airplane wing, the water accelerates over the curved
48:00
upper surface of the metal blade, causing the local
48:02
pressure to drop significantly. But water isn't air.
48:05
If you drop the pressure of air, it just gets thinner.
48:07
If you drop the pressure of water far enough, something
48:10
terrifying happens. Fluid thermodynamics dictate
48:13
that if the local pressure drops below the liquid specific
48:16
saturation pressure, it's vapor pressure.
48:18
The liquid is not just flow faster, it instantaneously
48:22
boils. Wait, it boils at room
48:24
temperature? Right there at room temperature,
48:27
it phase changes from a dense liquid into a gas, forming
48:31
thousands of tiny vapor bubbles on the low pressure side of the
48:34
spinning metal blade. This phenomenon is cavitation.
48:38
And boiling water sounds relatively harmless until those
48:41
bubbles move. The problem is that the impeller
48:44
keeps spinning at thousands of RPM.
48:46
The water sweeps those low pressure vapor bubbles past the
48:50
blade and out into the main body of the pump housing, which is an
48:53
area of much higher pressure. When those low pressure bubbles
48:57
hit that high pressure zone, the vapor inside the bubble cannot
49:00
physically sustain itself against the crushing force.
49:03
The bubble collapses. But it doesn't pop like a soap
49:06
bubble, it implodes. It implodes asymmetrically and
49:08
violently. As the bubble collapses, it
49:10
creates A microscopic hypersonic jet of liquid that fires
49:14
straight through the center of the bubble, slamming directly
49:16
into the surface of the steel impeller.
49:18
And the pain point here is catastrophic pump destruction.
49:22
Young engineers will get a panicked call from the plant
49:24
operator saying their brand new, expensive industrial pump sounds
49:28
like it's grinding a load of gravel.
49:29
It sounds like marbles in a blender.
49:31
And the inexperienced engineer might think a bearing is failing
49:35
or that there is physical mechanical rubbing inside the
49:38
housing. That is a fatal misconception.
49:41
It is not mechanical rubbing. It is a phase change
49:44
thermodynamic issue. Those collapsing bubbles are
49:47
creating localized microscopic shock waves generating 100 to
49:51
200 atmospheres of pressure upon impact.
49:54
They are literally blasting microscopic craters into solid
49:57
stainless steel over a period of a few weeks or even days.
50:01
Cavitation can chew a beautifully machined pump
50:03
impeller down to a useless pitted nub of ruined metal.
50:06
It blows my mind every time I think about it.
50:08
Tiny little vapor bubbles popping can eat through a solid
50:11
bronze ship propeller. It's like a billion microscopic
50:14
jackhammers firing at the speed of sound.
50:16
To avoid destroying your infrastructure, you have to
50:19
obsessively monitor the cavitation number.
50:21
You must ensure that your system is designed so that the local
50:24
dynamic pressure, even at the fastest point over the blade,
50:27
never drops below the vapor pressure of the fluid you are
50:29
pumping. You might have to physically
50:31
lower the pump elevation to increase static head, raise the
50:34
supply tank, or decrease the RPM to keep the pressure up.
50:39
If you ignore the thermodynamics of Brunoi's principal and
50:42
liquids, cavitation will rapidly destroy your system.
50:45
Fluid mechanics takes no prisoners.
50:48
All right, let's wrap this up. We have scoured the engineering
50:51
forums, the subreddits, the late night help boards where
50:54
engineers go when the textbook fails them, and we are going to
50:57
rapid fire through the top three most common questions
51:00
professionals are begging for answers to online.
51:03
Let us address the realities of the field.
51:05
Question one. We see this all the time in
51:07
engineer posts. Why did my pipeline blow apart
51:10
when the main pump shut off? I checked the math a dozen
51:13
times. Every valve, pipe and flan was
51:15
rated for the maximum operating pressure.
51:18
Well, this is the textbook definition of water hammer,
51:20
which we covered earlier. The static pressure rating of
51:23
your pipe is entirely irrelevant when dealing with kinetic shock.
51:28
When your pump shut off abruptly, perhaps due to a power
51:30
trip, the massive column of moving water slammed into the
51:34
stationary fluid ahead of it. It converted all its massive
51:37
kinetic energy into a localized pressure spike that traveled at
51:40
the speed of sound up to 1200 meters per second, rebounding
51:44
backward down your pipe. To fix this, you cannot just buy
51:48
thicker pipes, you must alter the dynamics.
51:51
You must install surge tanks to absorb the expanding volume
51:54
safely, or utilize slow closing automated control valves to
51:58
bleed the momentum off gently over a period significantly
52:01
longer than the waves round trip time.
52:03
Moving on to question 2. Why is my incredibly expensive,
52:06
perfectly factory calibrated flow meter reading totally
52:10
wrong? When I install it in my pipe,
52:12
the manufacturer says it's accurate to within 1%, but my
52:15
flay rates are wildly off. This failure comes down to
52:18
ignoring the velocity profiles dictated by viscous flow.
52:22
The flow velocity is not a uniform block across the
52:25
diameter of the pipe. If you have viscous laminar
52:28
flow, the velocity profile is a steep parabola.
52:32
It is fast in the dead setter and dropped to 0 at the walls.
52:36
If the flow is highly turbulent, the profile is much flatter, but
52:39
still drops to 0 at the viscous sub layer.
52:42
If your sensor probe is inserted only halfway into the pipe or
52:46
placed too close to the wall, it is only reading the highly
52:49
specific local velocity at that exact depth.
52:52
It's not reading the average bolt velocity of the entire.
52:55
Pipe, so placement is everything.
52:57
Furthermore, you must ensure the flow is fully developed.
53:00
If you place a flow meter immediately after a sharp elbow
53:03
joint or a half open valve, the turbulence is chaotic and
53:06
asymmetrical. You have placed the meter many
53:08
pipe diameters away from any obstruction to allow the flow to
53:11
straighten out and ensure the meters internal software is
53:14
calibrated for the correct Reynolds flow regime.
53:17
Last one, question 3. How do I stop my cylindrical
53:20
temperature sensor probe from snapping off violently when I
53:23
insert it into a fast flowing river or a high velocity steam
53:26
pipe? This is a classic manifestation
53:29
of the Karman Vortex St. The fast blunt flow is
53:33
separating around your cylindrical probe, shedding
53:35
destructive vortices alternately from side to side.
53:38
And it hits the resonant frequency.
53:39
Exactly the frequency of this alternating shedding dictated by
53:43
the stroll hall number is inadvertently perfectly matching
53:47
the natural physical resonant frequency of your metal probe.
53:51
The fluid is actively vibrating the metal to death through
53:54
fatigue failure. To stop it, you must change the
53:57
math of the system. You either change the physical
53:59
diameter of the probe to forcefully shift the shedding
54:02
frequency away from the resonant frequency, or more commonly, you
54:05
add helical strikes, small spiral ridges, to the outside
54:08
casing of the probe. These ridges break up the
54:11
uniform flow, preventing the carbon vortices from forming
54:13
synchronously, killing the vibration at its source.
54:16
The answers to the hardest engineering problems are always
54:19
hiding in the physics. Next time you pour a cup of
54:21
coffee, watch the way the stream breaks up as it hits the mug, or
54:25
the way the steam forms chaotic eddies in the air.
54:29
The exact same Navier Stokes equations governing the milk
54:32
swirling in that coffee cup are dictating the massive weather
54:36
patterns across the globe, the aerodynamics of an F35 fighter
54:40
jet, and the blood pumping through your own heart.
54:43
Fluid mechanics isn't just a brutal class you pass to get
54:46
your degree, it's the hidden mathematical code of the
54:49
universe. Once you learn to see the math
54:52
moving in the water, the physical world never looks the
54:55
same. If this helped you even a
54:56
little, say something. We're putting this out for free.
54:58
No paywall, no subscription, nothing locked behind a card
55:01
that only works if we know it's actually reaching people and
55:04
doing what it's supposed to do. If you're a young engineer
55:07
trying to figure this stuff out and this gave you clarity you
55:09
weren't getting anywhere else, drop a like, leave a comment.
55:11
It takes 2 seconds on your end, but it tells us this is worth
55:13
continuing instead of just talking into the void.
55:16
We are excited for your feedback.
55:17
Negative or positive.