Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Mass flow rate vs temperature.
Posted 30 août 2015, 23:47 UTC−4 Fluid & Heat, Heat Transfer & Phase Change, Modeling Tools & Definitions, Parameters, Variables, & Functions, Results & Visualization, Studies & Solvers Version 5.0 1 Reply
Please login with a confirmed email address before reporting spam
Hi, I used Comsol to model a really simple problem. A pipe with 50mm diameter and 25m long with an inlet temperature of 20°C and 0.5kg/s of mass flow rate. Also, the pipe´s wall temperature was 150°C. I used a parametric sweep to increase the mass flow rate with a step of 0.15 until 2kg/s. The fluid was engine oil and in laminar flow.
When I made the plot of temperature vs mass flow rate, it behaves as a parable in which at a point near 1.4 kg/s the temperature begins to increase instead of decrease as it´s expected.
Do you know what is the explanation for this?
Thanks
When I made the plot of temperature vs mass flow rate, it behaves as a parable in which at a point near 1.4 kg/s the temperature begins to increase instead of decrease as it´s expected.
Do you know what is the explanation for this?
Thanks
1 Reply Last Post 31 août 2015, 05:34 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi
there are a few possibilities why you get unexpected results in these types of conjugated heat transfer.
1) what often happens, one miss a BC between the SPF and the HT section, you need to define your SPF inflow and outflow, as well as the HT outflow, and Temperature constant at the inflow area, as well as at the side boundary of your tube
2) most common, the meshing for the very anisotropic shape of your long and narrow tube. You need to be sure your mesh resolves the gradient of both T and the velocity along the wall. If you check the oil heat diffusivity between 20 and 150°C you will see its harly more than some 0.06-0.09 mm^2/s and the flow velocity is betwen 0.3-1.5 m/s. Oil is viscous, hence the no-slip layer is quite thick and the heat will only propagate here by conduction. Resulting all in the requirement of a high density mesh along the warm wall, as your heat and flow velocity changes only in the first 3-5 mm along the wall
I would propose you start with a 2D-axi model, with structured mapped mesh, with a distribution along the axis of some 200 elements, and across (radial) some 15 elements and a ratio of 14 with a geometric sequence
Then you need to check that this mesh is sufficient, by gently densifying/coarsening it and checking the effect on the results, I havent don ethat in the model below.
To help the solver you could use a Poiseuil type pressure drop as initial conditions, I see some 140[Pa/m] along the axial length at 0.3[m/s] or 0.5[kg/s] flow.
Finally you flow needs a meter to stabilise in velocity, so you could perhaps add a meter and not heat this first meter of the pipe wall.
I would also suggest that you use the stationary solver with an auxialliary Sweep and NOT the Parametry Sweep node, as for me in my v5.1.0.180 I notice the Parametric Sweep node did not detect correctly the continuation condition, and restarted the initial conditions from scratch for each parametric step, what takes twice as long to solve, this is not normal ;(
--
Good luck
Ivar
there are a few possibilities why you get unexpected results in these types of conjugated heat transfer.
1) what often happens, one miss a BC between the SPF and the HT section, you need to define your SPF inflow and outflow, as well as the HT outflow, and Temperature constant at the inflow area, as well as at the side boundary of your tube
2) most common, the meshing for the very anisotropic shape of your long and narrow tube. You need to be sure your mesh resolves the gradient of both T and the velocity along the wall. If you check the oil heat diffusivity between 20 and 150°C you will see its harly more than some 0.06-0.09 mm^2/s and the flow velocity is betwen 0.3-1.5 m/s. Oil is viscous, hence the no-slip layer is quite thick and the heat will only propagate here by conduction. Resulting all in the requirement of a high density mesh along the warm wall, as your heat and flow velocity changes only in the first 3-5 mm along the wall
I would propose you start with a 2D-axi model, with structured mapped mesh, with a distribution along the axis of some 200 elements, and across (radial) some 15 elements and a ratio of 14 with a geometric sequence
Then you need to check that this mesh is sufficient, by gently densifying/coarsening it and checking the effect on the results, I havent don ethat in the model below.
To help the solver you could use a Poiseuil type pressure drop as initial conditions, I see some 140[Pa/m] along the axial length at 0.3[m/s] or 0.5[kg/s] flow.
Finally you flow needs a meter to stabilise in velocity, so you could perhaps add a meter and not heat this first meter of the pipe wall.
I would also suggest that you use the stationary solver with an auxialliary Sweep and NOT the Parametry Sweep node, as for me in my v5.1.0.180 I notice the Parametric Sweep node did not detect correctly the continuation condition, and restarted the initial conditions from scratch for each parametric step, what takes twice as long to solve, this is not normal ;(
--
Good luck
Ivar
Attachments: