Latest / Mechanical Engineering Made Simple / Governing Laws of Heat Exchanger Design (156)

Transcript

Governing Laws of Heat Exchanger Design (156) · Mechanical Engineering Made Simple

• 0:00
  Welcome back, design junkies, to mechanical engineering made
• 0:02
  simple. And welcome to the Deep dive.
• 0:05
  Today we are really getting into the weeds on a foundational
• 0:09
  piece of hardware. We are absolutely tackling the
• 0:12
  silent workhorse of, well, pretty much everything, the heat
• 0:15
  exchanger. If you want to understand
• 0:17
  process engineering, you have to get your head around these
• 0:20
  devices, how they're designed, analyze the whole 9 yards.
• 0:23
  Absolutely. Our mission today is to dissect
• 0:28
  the governing laws, the math that defines them, and then jump
• 0:31
  straight into the real world compromises that engineers have
• 0:34
  to make when they choose a design.
• 0:35
  From the huge shell and tube units and refineries to you
• 0:38
  know, the Super compact stuff in modern HVAC.
• 0:41
  Exactly. We're giving you the framework
• 0:43
  to understand how these critical components size up and perform.
• 0:46
  OK, so let's unpack this. We should probably start with
• 0:48
  the simplest definition possible.
• 0:50
  Right At its most basic, a heat exchanger is any device that
• 0:54
  transfers heat between two fluids or, you know, a fluid in
• 0:58
  the environment. But the.
• 0:59
  Classification. That's where the engineering
• 1:01
  really starts. Specifically, the two main types
• 1:04
  of indirect contact exchangers. That distinction is just
• 1:07
  fundamental. The vast majority of what you'll
• 1:10
  see out in the field are recuperators.
• 1:12
  OK, like a car radiator? A car radiator, A furnace?
• 1:15
  Exactly. The heat transfer happens
• 1:18
  directly, continuously across a solid boundary, a tube wall, a
• 1:23
  sheet, hot fluid on one side, cold on the other, and the
• 1:27
  process is steady state. And then you have the other
• 1:30
  type, which is I think a bit more interesting conceptually,
• 1:33
  the regenerator. It is.
• 1:34
  Regenerators are different because the whole process is
• 1:37
  transient. It's time dependent.
• 1:39
  Instead of being separated, the fluids flow alternately through
• 1:42
  the same surface. So wait, the hot fluid heats up
• 1:45
  the surface? For a certain amount of time,
• 1:47
  yes. And then you switch and the cold
• 1:49
  fluid flows through and picks up that stored heat in the next
• 1:52
  interval. So.
• 1:53
  It's a pulsed process, heating, then cooling, all on the same
• 1:56
  material. That transient nature must make
• 1:59
  the analysis completely different.
• 2:02
  Wildly different, but regardless of type, every single heat
• 2:06
  exchanger is dictated by two major physical laws.
• 2:10
  We always have to start with the first principle.
• 2:13
  Conservation of energy. Conservation of energy.
• 2:15
  This is the equation that tells us about the duty, the total
• 2:18
  amount of heat that gets transferred.
• 2:20
  Right. So the total heat Q has to equal
• 2:24
  the capacity rate of the hot fluid times its temperature
• 2:27
  drop. And it also has to equal the
• 2:28
  capacity rate of the cold fluid times its temperature rise.
• 2:32
  It's a perfect enthalpy balance. But, and this is the key thing,
• 2:36
  the energy balance tells you how much heat is moving, but it says
• 2:39
  absolutely nothing about how big the device needs to be to make
• 2:43
  that happen. And that's the job of the second
• 2:45
  governing relationship, the rate equation, this is.
• 2:47
  The one that connects the heat transfer Q to the physical size
• 2:50
  of the exchanger. Yes, the rate equation is Q = U
• 2:54
  ETA S Theta M It's beautiful because it pulls everything
• 2:57
  together. The rate of transfer is defined
• 2:59
  by U, the overall heat transfer coefficient.
• 3:02
  ETA which is the fin efficiency if you have fins?
• 3:05
  S the total surface area that's your size and Theta.
• 3:08
  M the mean temperature difference, the driving force.
• 3:11
  Let's zoom in on that star variable U, because it feels
• 3:14
  like that's where all the real world messiness is hiding it.
• 3:17
  Is it really is? When you calculate U, you're
• 3:19
  basically summing up five different thermal roadblocks,
• 3:22
  all in series. OK, so you've got convection on
• 3:24
  the hot side, convection on the cold side, the resistance of the
• 3:27
  wall material itself, and. Then you have the two ugly step
• 3:30
  sisters of the thermal world, hot side fouling and cold side
• 3:33
  fouling. Yes, fouling.
• 3:36
  This is where the clean theoretical model just slams
• 3:39
  into The Dirty reality of a factory floor.
• 3:42
  It's the elephant in every room. Fowling is just the accumulation
• 3:46
  of junk, scale, crud, algae, sediment, whatever on the heat
• 3:50
  transfer surface and it it immediately creates this huge
• 3:53
  additional resistance Rd. that you have to build into your
• 3:56
  calculation of you. If you don't, or if you
• 3:58
  underestimate it. You've designed a failure.
• 4:00
  It's that simple. The financial hit is massive.
• 4:03
  Your exchanger can't meet its duty, so you have to run your
• 4:06
  whole process at a lower capacity or you're spending a
• 4:09
  fortune on pumping power and chemical cleaning.
• 4:11
  It is the single biggest operational headache.
• 4:15
  OK. So once we have those governing
• 4:16
  relationships, we get to the analytical tools and the classic
• 4:20
  one for figuring out that mean temperature driving force Theta
• 4:23
  arm is the logarithmic mean temperature difference, the
• 4:27
  LNTDL. MTD is incredibly useful for the
• 4:30
  simple cases, you know, pure counter flow, pure parallel
• 4:34
  flow, and systems where you've got a phase change like boiling
• 4:37
  or condensing so one temperature is constant gives you.
• 4:40
  A really clean way to average the temperature difference.
• 4:42
  It does, but the real world is messy as soon as you step
• 4:45
  outside those four simple arrangements.
• 4:47
  Say you've got a normal shell and tube unit where the flow is
• 4:50
  all over the place, or a cross flow setup.
• 4:52
  The simple LMTD is just wrong. It's completely wrong, and
• 4:56
  that's where the complexity starts to ramp up.
• 4:58
  We have to introduce a correction factor which we call
• 5:00
  F So the true mean temperature difference becomes the simple
• 5:04
  counter flow LMTD multiplied by this factor F.
• 5:08
  And F is always less than one, right?
• 5:10
  It's basically a penalty for having a less than perfect flow
• 5:13
  path. Precisely.
• 5:14
  You have to go look it up in charts or tables for your
• 5:16
  specific geometry. But the real catch with the LMTD
• 5:20
  method, even with F, is that you need to know all four
• 5:23
  temperatures, both inlets and both outlets to even start.
• 5:27
  Right. And if you're trying to design a
• 5:29
  system you don't know the outlet temperatures yet, that's what
• 5:32
  you're trying to find. So you're stuck.
• 5:34
  You have to guess, calculate, check and repeat.
• 5:36
  It's a painful trial and error loop.
• 5:38
  Which brings us to the more modern and, frankly, superior
• 5:41
  alternative. The effectiveness, number of
• 5:43
  transfer units, method or and two, it was created specifically
• 5:47
  to kill that trial and error process, which makes it perfect
• 5:50
  for computer modeling. And it all relies on three
• 5:53
  really elegant dimensionless groups.
• 5:56
  It does. The first one is the capacity
• 5:58
  rate ratio C star. That's just the ratio of the
• 6:01
  minimum fluid capacity rate to the maximum semen over C Max.
• 6:05
  It's always less than one, and it tells you which fluid is the
• 6:09
  thermal bottleneck. #2 is the exchanger effectiveness epsilon?
• 6:13
  This is your performance metric. It's the ratio of the actual
• 6:17
  heat. You transferred Q to the
• 6:19
  absolute maximum possible heat transfer Q Max.
• 6:22
  And that maximum possible transfer Q Max is always, always
• 6:26
  limited by that fluid with the minimum capacity rate.
• 6:29
  It's your weak link. And finally, the one that
• 6:31
  answers the big question, how big does this thing need to be?
• 6:35
  The number of transfer units N2 None 2 is defined as U * s
• 6:38
  divided by semen. You should literally just think
• 6:41
  of None 2 as the measure of the thermal size of your exchanger.
• 6:45
  More None 2 means a bigger exchanger.
• 6:46
  So the magic here is that effectiveness epsilon is a
• 6:49
  direct function of none, 2 and C star for any given flow
• 6:52
  arrangement. That's the win.
• 6:54
  If you know the size then two and the fluid relationship C
• 6:57
  star, you can instantly calculate the performance
• 6:59
  epsilon without ever needing to know the outlet temperatures.
• 7:02
  It's a huge leap forward for design engineers.
• 7:05
  OK, so moving from the math to the actual hardware, we have to
• 7:09
  talk about the undisputed king of the industry.
• 7:12
  The shell and tube exchanger. It is the workhorse.
• 7:15
  High pressure, high temperature. This is what you use.
• 7:18
  It seems simple tubes in a shell, but the variations are
• 7:21
  nuts. Youtubes floating heads.
• 7:24
  Why so many different types? It mostly comes down to two
• 7:26
  things, thermal stress and maintenance.
• 7:30
  If you run hot fluid through the tubes and cold fluid in the
• 7:33
  shell, things expand at different rates.
• 7:35
  And that can RIP the joints apart.
• 7:36
  It can literally tear the exchanger apart.
• 7:38
  So designs like the YouTube or a floating head, they allow the
• 7:42
  whole tube bundle to expand and contract independently of the
• 7:45
  shell. It manages that stress.
• 7:48
  And we can't talk about Shell and two without mentioning
• 7:50
  Teema. The Tubular Exchanger
• 7:51
  Manufacturers Association TEEMA is the global language for these
• 7:55
  things. It standardizes everything from
• 7:57
  shell diameter to flange types. It's not just about dimensions,
• 8:00
  it's about safety and interchangeability.
• 8:03
  So if a unit fails in a plant, you can order a Teema standard
• 8:06
  replacement and know it's going to fit.
• 8:08
  And perform safely, Yes. But here's the frustrating part
• 8:11
  of engineering Teema standards allow for gaps.
• 8:14
  You can't make everything perfectly sealed because then
• 8:17
  you could never clean it or take it apart for maintenance.
• 8:19
  And those gaps mean the flow is imperfect.
• 8:22
  Very imperfect. You get leakage fluid leaks
• 8:25
  between the shell and the baffles or around the tubes in
• 8:28
  the baffle hole. Some of it just bypasses the
• 8:30
  tube bundle entirely. Your nice ideal cross flow model
• 8:34
  is just. Gone.
• 8:36
  So how do we deal with that? We can't just ignore it.
• 8:38
  We correct for it. We use the Bell Delaware method.
• 8:41
  This is the established way to get an accurate heat transfer
• 8:44
  coefficient in a real non ideal shell and tube unit.
• 8:48
  OK, so how does it work? Instead of trying to model the
• 8:50
  messy flow directly, you start with the theoretical idealized
• 8:54
  heat transfer coefficient, and then you multiply it by a series
• 8:57
  of five correction factors. Each one accounting for a
• 9:00
  different leakage path. Exactly.
• 9:02
  It systematically penalizes the ideal number for every major
• 9:05
  leakage stream and bypass path. It's the essential step that
• 9:08
  turns a theoretical model into a commercially accurate design.
• 9:11
  OK. So shifting gears from the
• 9:13
  massive industrial units to the other extreme compact designs.
• 9:18
  These are all about maximizing surface area density, the ratio
• 9:22
  of surface area to volume, which we call beta.
• 9:24
  You see these when you need a ton of heat transfer in a really
• 9:28
  small space right? Like an aerospace or high end
• 9:31
  HVAC? Precisely.
• 9:32
  This includes things like flattened tubes, but the real
• 9:35
  genus is in the complex plate fin surfaces.
• 9:38
  Things like louvered fins, strip fins, WAVY fins.
• 9:42
  What's the idea there? Why make them so complex?
• 9:44
  It's all about manipulating the boundary layer.
• 9:47
  A plain flat fin just lets the air flow smoothly over it, which
• 9:50
  builds up a thick insulating boundary layer.
• 9:53
  Which is bad for heat transfer. Very bad.
• 9:55
  But the WAVY and louvered fins. They intentionally interrupt
• 9:58
  that flow. They have offset sections that
• 10:00
  constantly trick the boundary layer and force it to restart.
• 10:03
  And that restarting process. Leads to a massive increase in
• 10:06
  the heat transfer coefficient. You pay for it with a higher
• 10:09
  pressure drop, but you get huge thermal performance in a tiny
• 10:12
  volume. Now what about the plate and
• 10:14
  frame exchanger? That's another compact design,
• 10:17
  but you see it more in like food and beverage industries.
• 10:19
  Yeah, brewing pharma. They use thin corrugated metal
• 10:24
  plates, usually in a herring bone pattern, stacked in a frame
• 10:27
  and sealed with gaskets. The advantages seem huge.
• 10:30
  They're pretty cheap, you can take them apart to clean them
• 10:32
  really easily. You can literally hose them down
• 10:34
  and the corrugations give you extremely high thermal
• 10:37
  performance, but they have two major major limitations.
• 10:43
  And I'm guessing it has to do with those gaskets.
• 10:45
  First one is the gaskets. Yes, it means you are
• 10:48
  structurally limited to low pressures.
• 10:51
  Try to run high pressure steam through 1 and it will leak.
• 10:55
  And the second limitation. The flow passages are incredibly
• 10:58
  narrow. That gives you a tiny hydraulic
• 11:00
  diameter, which results in a massive pressure loss.
• 11:02
  You need a lot of pumping power just to shove the fluid through
• 11:05
  it. And this brings us right to the
• 11:07
  fundamental trade off at the heart of, well, all mechanical
• 11:11
  engineering design. The competition between thermal
• 11:14
  performance and pumping power. This is it.
• 11:16
  This is the core tension. It is every single thing you do
• 11:20
  to enhance heat transfer, cladding, fins, corrugating a
• 11:23
  surface, making passages smaller.
• 11:25
  Everyone comes with an unavoidable penalty in pressure
• 11:28
  loss. Which means you need more
• 11:30
  pumping power, which is an operating cost.
• 11:33
  If your design is thermally brilliant but eats up energy
• 11:36
  just moving fluid, the whole system is a failure.
• 11:38
  You see this perfectly when you look at how we use fins or
• 11:41
  extended surfaces. Take a double pipe exchanger.
• 11:44
  We add fins to balance out the thermal resistance, usually when
• 11:49
  one fluid is way worse at heat transfer than the other, like a
• 11:51
  gas versus a liquid. So you give the poor performer
• 11:54
  more surface area to work with. Right, but there's a point of
• 11:57
  diminishing returns. The source material shows that
• 12:00
  if the composite thermal resistance drops below about
• 12:02
  .00002 meters squared Kelvin per Watt.
• 12:06
  The fins basically stop. Their efficiency just plummets.
• 12:09
  That number, liver .002 is the point where adding more metal,
• 12:13
  more complexity gives you almost no thermal benefit.
• 12:17
  The resistance is already so low that the fin does next to
• 12:19
  nothing, but you still pay the price in friction and cost.
• 12:23
  It's the constant battle. What a deep dive.
• 12:26
  So we started with that foundational difference between
• 12:29
  the continuous recuperator and the cycling regenerator.
• 12:31
  We did. Then we detailed the analytical
• 12:34
  tools making that critical shift from the cumbersome LMTD method
• 12:38
  to the much more flexible into method, which gave us that key
• 12:41
  metric of thermal size than known to.
• 12:43
  And then we got into the hardware.
• 12:45
  The standardization of the shell and tube with Tima, the clever
• 12:48
  boundary layer tricks and compact fins, and the very real
• 12:52
  pressure limits of gasketed plates.
• 12:54
  And it all comes back to that central conflict.
• 12:56
  Heat transfer performance and pumping power performance are
• 12:59
  fundamentally at odds with each other, right?
• 13:01
  The real challenge in thermal design isn't just about
• 13:03
  maximizing efficiency. It's about minimizing what we
• 13:07
  call confined thermodynamic irreversibility by perfectly
• 13:11
  choosing the dimensions of your flow passages.
• 13:13
  That is the true art of it. This deep dive has really armed
• 13:16
  you with a blueprint to see that core tension in mechanical
• 13:19
  design, maximizing energy transfer while minimizing the
• 13:22
  cost to move the fluids. This is knowledge you can use
• 13:26
  right away, whether you're looking at a server cooler or a
• 13:28
  massive chemical lant. Go forth and exlore the nonitude
• 13:31
  of the world.