0
$\begingroup$

Friends, I am an EE/SW/Controls eng. Please, pardon me of perhaps incorrect terminologies.
I am working on a R&D project: A chemical lab equipment has a 2" long metal pipe enclosed in insulation material. "One end" of the pipe is attached to a heating element and a temperature sensor. My work is controlling the temperature of the pipe. Including mechanical engineers and a chemist, as a controls engineer I am, our group discusses what material the pipe needs to be, either aluminum or copper.

I saw a few YouTube clips that demonstrate aluminum delivers heat faster than copper. In controls perspective, faster delivery means less delay and better control. Thus, my intuition tells to use an aluminum pipe, while others prefer copper for its higher thermal conductivity.

Could someone explain how one is better than another for this case?


Edited
The YouTube clip is here.
The pipe contains a section of Gas chromatography "Column", between the "Inlet + trap" to the "Oven". Ideally the temperature has to be the same as the Oven, which changes up to 15'C/sec. Otherwise, we try to keep it at a certain temperature(TBD) other than "Cold spot". The pipe can be a pair of aluminum plates with channels, I thought.

$\endgroup$
15
  • 3
    $\begingroup$ What's more important: mass or volume? Aluminum has better thermal conductivity per mass while copper has better thermal conductivity per volume/area. Aluminum is also way cheaper. That might matter most in the end. $\endgroup$
    – DKNguyen
    Aug 6, 2021 at 18:42
  • $\begingroup$ Can you provide the link to the youtube video? The resistivity of aluminum is 1.6 times of copper, and the conductivity is almost 40% lower than that of copper. It causes me to wonder about the reasoning too. $\endgroup$
    – r13
    Aug 6, 2021 at 19:51
  • 1
    $\begingroup$ Assuming it is not a wetted part! For analytical applications, neither Aluminum or Copper would be suitable $\endgroup$
    – Pete W
    Aug 7, 2021 at 14:51
  • 1
    $\begingroup$ @PeteW Glad to hear that! It is not "wetted part". Would you please explain why not to use Aluminum nor Copper? BTW, if I missed to explain, this part does not contact the gas passing, but encloses a glass column of micro-meters ID. $\endgroup$
    – jay
    Aug 7, 2021 at 15:03
  • 1
    $\begingroup$ If heating electric and cooling with water, and fixed geometry, and concerned with round-trip temperature ramping cycle time, then aluminum is probably better than copper. If concerned with stability at steady state, probably copper better. Some silver alloys maybe better still but impractically expensive. Also if temperature swings are extreme, diffrential thermal expansion may be an issue, so a filler material (can be a liquid) is needed or else none of this matters. $\endgroup$
    – Pete W
    Aug 7, 2021 at 15:04

2 Answers 2

2
$\begingroup$

Simplified analysis to compare materials, for different limiting-case applications, given the same geometry:

  • Want max heating ramp rate, with fixed-power electric heater -- minimize Volumetric Heat (i.e. specific heat * density)
  • Want max cooling ramp rate, with fixed-temperature coolant -- minimize (VolumetricHeat / ThermalConductivity)
  • Want best control performance at/near steady state -- maximize Thermal Conductivity (and possibly thermal mass, but depends on location and nature of noise in thermal load)

Aluminum 6061-T6:

  • density (g/cc) = 2.7
  • specific heat (J/gK) = 0.9
  • volumetric heat (J/ccK) = 2.43
  • thermal conductivity (W/mK) = 167
  • CTE (ppm/K) = 24

Copper, pure:

  • density (g/cc) = 8.9
  • specific heat (J/gK) = 0.385
  • volumetric heat (J/ccK) = 3.43
  • thermal conductivity (W/mK) = 385
  • CTE (ppm/K) = 17

Edit #1, from OP @jay

  1. If the volume of the pipe was the same, aluminum would be more effective delivering the heat from one end to another. If the weight of the pipe was the same, copper would be more effective.
  2. Since, the length/shape of the pipe is constrained by the mechanical structure, the volume matters more.
  3. Thus, I would choose aluminum pipe, if I have good trust in my controller.
  4. I would choose copper pipe for less influence from disturbance (ex. from ambient), if the performance of the controller was excluded from the considerations.

Edit #2 from OP jay

However, in practice and mechanical design perspective:

For 2, if you are building from pipe/tube, there are a finite number of specific options wall thickness. So just make a table of the bulk heat capacity in J/K for the options you have given mechanical constraints. That will inform the remaining analysis, but is not necessarily the performance limiting factor. Often not.

Regarding effect of controller, it depends on placement of heater and sensor.

If you are not sure what you are doing, then a big thermal mass with a slow response might actually be more practical than something light and aggressively tuned.

$\endgroup$
9
  • $\begingroup$ It started shaping some base approach in my thoughts. "volumetric_heat = density * specific_heat" gives some insights. I will study more to understand. We are not too concerned about cooling, though it will eventually be something to consider. Cooling uses a fan at the moment. Could you please add a little more explanation in you answer text, why the properties for "max heating ramp" are less effective to "steady state" target? Can it be a different story, if the controller can be tuned well? Wouldn't the controller prefer faster response time anyway? $\endgroup$
    – jay
    Aug 7, 2021 at 16:27
  • $\begingroup$ Electric heat will have some power limit, i.e. max heat flux is constant. (Can locally reach very high temperatures without too much trouble.) With that constraint, peak dT/dt is roughly determined thermal mass of material. Cooling with a coolant has (sloppy approximation) constant T, so max heat flux proportional to temp diff and thermal conductivity of material. In both cases, best ramp time depends (roughly) on max heat flux, and total thermal mass of material. For fixed geometry, thermal mass depends on volumetric heat capacity. $\endgroup$
    – Pete W
    Aug 7, 2021 at 17:05
  • 1
    $\begingroup$ Deep thoughts! I can agree all that. I will put some good efforts to understand more. Though, I cannot say which one to choose definitely yet. Meantime, I now understand why some of my "correct" answers (in other threads) were not accepted, no offense. Problem might be at the questioner's. :) $\endgroup$
    – jay
    Aug 7, 2021 at 17:47
  • 1
    $\begingroup$ yes for 1. For 2, if you are building from pipe/tube, there are a finite number of specific options wall thickness. So just make a table of the bulk heat capacity in J/K for the options you have given mechanical constraints. That will inform the remaining analysis, but is not necessarily the performance limiting factor. Often not. Regarding effect of controller, it depends on placement of heater and sensor. If you are not sure what you are doing, then a big thermal mass with a slow response might actually be more practical than something light and aggressively tuned. $\endgroup$
    – Pete W
    Aug 12, 2021 at 16:28
  • 1
    $\begingroup$ I hear you, Pete! Finally I see we are talking about the same and agreeable idea based on the fundamentals you answered. And, I see how the mechanical engineer concluded to use copper "pipe" from "For 2, if you are building from pipe/tube, there are a finite number of specific options wall thickness". Now, I am understanding the mechanical engineer here better. I will copy both of yours and mine, edit, and accept, will try. You may re-edit it, since that belongs to you. $\endgroup$
    – jay
    Aug 12, 2021 at 16:54
0
$\begingroup$

If you are controlling the temperature of the pipe (to me this means the temperature of the pipe wall), I don't think it matters. A cast iron pipe won't change temperature as fast, but that applies both to your whatever-you're-doing-to-control-pipe-temp and to changes in pipe temp due to environment and the fluids going through it. All it will change is your loop tuning.

$\endgroup$
1
  • $\begingroup$ I agree with you Tiger. I do, eventually, tune the controller. Though, I prefer the process to be easier to control and controlled accurately, if I can choose. $\endgroup$
    – jay
    Aug 7, 2021 at 2:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.