3 replies to this topic

[ggxx2013]

Hi

 

I am designing a cylindrical heater for say a visbreaking process (involving thermal cracking processes in general) of the coil/furnace form wherein lies two processes -

 

1) the first being to heat up the feed stream to a specific temperature followed by

2) to maintain the process stream at that temperature so that isothermal cracking can occur.

 

However, in the second step, the stream being in a tubular reactor (PFR) is going to experience concentration change in the axial direction (from high at the inlet to low at the outlet) as such for the process to be isothermal, a variable heat flux is required.

 

I have researched upon the heater design website and was encouraged to see the section involving Radiant section with gas temperature gradients. there is even a calculator to calculate the variable heat flux depending on how many sections the furnace needs to be separated. May I know how valid is this with regards to my design above? (tried searching for boundary conditions to which this method can be applied but to no avail). Thank you, your help is greatly appreciated.

 

V

[thorium90]

I wouldnt say it is applicable (assuming the website you are referring to is what I think it is). You would need to take into account heat of reactions, change in heat capacities. The type of boundary conditions are either Neumann, Dirichlet or Cauchy. Should be easy to find in many textbooks and on google what it means. I understand this is for your design project right? You would need numerical solvers like Matlab or GAMS to do the calculations. You can setup your DE's and conditions and post here. I think many people would be able to check and give advice on your formulation.

Edited by thorium90, 27 February 2013 - 12:38 AM.

[ggxx2013]

Hi thorium,

 

Thanks so much for your speedy response.  Greatly appreciate your insights on this.. What you are suggesting is that a differential approach is very inaccurate should we just separate the furnace into three zones, am i right? As such, the furnace should be considered on an infinitesimal basis where the concentration changes and hence variable heat fluxes at different points of the PFR can be accounted for and calculated. Di I get this part correct?

 

May I just enquire if it is possible, having divided the heater into 3 zones, and that the inlet and outlet process stream concentrations are known, find out the heat duty required for each of these 3 zones followed by subsequent sizing of these three zones as a rough approximation... would it be possible? I am sorry to be harping on this approach.

 

V

[thorium90]

I meant, you should calculate the radiant and convection sections separately. In the radiant section, energy is mostly by radiation (Stefan-Boltzmann, absorptivity..), or you could make some assumptions on heat flux, lumped capacitance. In the convection section, forced convection on tubes at crossflow. This part would have been similar to a typical fired heater, however, you have reactions occurring which require heat, therefore, the temperature on the other side of the tube is affected by the heats of reaction and the heat capacities of the composition of the gas that happens to be on the other side at every infinitesimal length. Therefore, you might realise that the DE across the tube wall is a BVP and the concentration throughout the reactor length is a IVP. Do this for the flow into and out of the radiant section. Transfer the variables at the bridgewall to solve for the convection section.

 

Similar inquiry:

http://www.cheresour...matlab/?p=69948

Edited by thorium90, 27 February 2013 - 05:04 AM.