Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Heat transfer inside engine combustion chamber

  1. Jan 25, 2010 #1
    Hi, I'm not sure if I've got this in the right section, but I'm hoping someone can point me in the right direction.

    I'm trying to create a model of what's actually happening during the combustion cycle in an engine and where the energy is going.

    As the intake valve closes and the piston begins to compress the air in the combustion chamber, as the temperature of the air begins to rise, what type of heat transfer is happening?

    The best info I can find is off of here http://en.wikipedia.org/wiki/Thermal_conductivity.
    Using the equation where conduction = kA(T/x).. k being the conductivity coefficient, A being surface area, T being temp differential and x being the distance between the two temperatures.

    Is this what I'm looking for? Would the variable x be for thickness of the block/head/piston (obviously needing to do a separate calculation for each of them)? And would all the heat transfered be lost from the compressed air? I realize I would need set up a spreadsheet to calculate it every couple of degrees of rotation to keep it accurate.

    Just hoping someone could let me know if I'm on the right track or not! :smile:
  2. jcsd
  3. Jan 26, 2010 #2


    User Avatar
    Science Advisor
    Gold Member

    Pretty much. Until addition of heat (spark or fuel injection), the temperature increase in the cylinder is nearly all from compression. This is, as you say, lost to the cylinder head, cylinder walls, and piston (and a small amount of gas escapes through 'blowby').

    Determining the heat transfer coefficient for the combustion chamber components after ignition is the tricky bit, because of the rapidly changing in-cylinder conditions (in terms of gas temperature and pressure).

    You'll need more than a spreadsheet to do this (a 3D combustion simulation will be required, correlated to heat release data from the engine in question).
  4. Jan 26, 2010 #3
    You will need more than Fouriers law because the heat transfer from the compressed air is not natural convection, it is forced convection. Most modern engines (if not all) are designed to have turbulent flow inside the combustion chamber upon air intake to promote air/fuel mixing.

    I would start by first modeling the heat transfer in an isothermal compressor, then an isentropic compressor and move on from there.

    As brewnog pointed out, a simple spreadsheet wont cut it if you want your error less than a couple hundred percent.
  5. Jan 30, 2010 #4
    So I got the spreadsheet put together with some interesting results.
    I have ignored blowby for now, it would have made it significantly more complicated and I wanted to first see if I could pull any feasible numbers. It's sampling every 6 degrees (30 samples), or 250us at 4000rpm.
    This is also just the compression cycle; from BDC to TDC with the valves closed the whole time, gasoline mixed in with the air though.... however much of a difference that makes.

    First, with parameters set to estimated full throttle, 2L engine, 4000rpm, 8:1 compression, hot engine.
    Gave me a max of 251PSI and 827deg Celsius. Seems a little high given that gasoline ignites at 280deg C... obviously including blow-by would bring it in the right direction.
    Increasing to 10:1 compression (where engines typically begin to knock (pre ignite the gasoline) brought those numbers up to 487PSI and 1500deg C!
    Considering I hadn't added in any cylinder leakage, I was fairly happy with how that ended up.

    Now, I set it to simulate running a compression test. 1500rpm from the starter, low volumetric eff (closed throttle plate), cold engine, colder intake air.
    Gave me 29PSI at 56deg C...
    Also, dropping volumetric eff (air sucked in each intake cycle) or RPM too low gave me erratic numbers jumping between positive and negative... (I've got fuel/air mass for the temperature/pressure changes given thermal loss calculated from volumetric efficiency). RPM too low obviously makes the sample rate too long for the thermal transfer.

    Something that surprised me is that if I increase intake temp, it increases cylinder temperature, but actually decreases the pressure. I might have to look into that.

    And I found http://en.wikipedia.org/wiki/Isentropic_process" [Broken] on wikipedia about isentropic heat transfer and from a quick read, it looks like that's what I've actually done. I calculated the energy transfer between the combustion chamber and engine bay, multiplied that by specific thermal capacity of the air/fuel mixture to get the change in temperature, then changed the pressure relative to that temp change as well. I'll properly read through it later on and see if I've got it right.

    Given that I seem to be in the general ballpark (in the first simulation at least), I'll try to find some way of adding in blow-by and see what that does. Adding in combustion later on might make things interesting! :devil:

    But... it's midnight and I need to get up early for work tomorrow. The thumbnails below show my input parameters and output data on the first (4000rpm/hot engine/8:1comp) simulation.

    http://xs.to/thumb-8B8A_4B63E428.jpg http://xs.to/thumb-FE24_4B63E428.jpg
    Last edited by a moderator: May 4, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook