Refining Community Logo

Coking rate vs. Heat flux

Home Forums Coking Technical Coking rate vs. Heat flux

This topic contains 5 replies, has 2 voices, and was last updated by  Mike Kimbrell 2 weeks, 2 days ago.

  • Author
    Posts
  • #28032

    Tushar
    Participant

    Any correlation between Coking rate in tubes vs. Heat flux for coker heaters? At higher rate if fire heater harder does the coking rate inside tube will increase?

  • #28035

    Mike Kimbrell
    Participant

    Coke formation is very dependent on temperature. The amount of coke produced is directly proportional to the residence time and is exponential with temperature. Higher feed rates will require higher firing rates that increases the heat flux in the radiant section of the heater. The higher feed rates will lower the residence time in the heater, but not enough to off-set the increase in the oil film temperature, so the heater fouling rates will increase.

  • #28046

    Tushar
    Participant

    Thanks for your response Mike, can we make any rough calculations?

  • #28061

    Mike Kimbrell
    Participant

    Coking of heater tubes is a first order reaction that can be approximated using the Arrhenius equation:

    k = A exp(-Ea/RT)

    The activation energy (Ea) has been reported to be between 40,000 kcal/g-mole and 55,000 kcal/g-mole. T is the absolute temperature in degrees Kelvin. R is 1.987 kcal/g-mole-K. To evaluate relative coking rates, it is common to use a base temperature to compare against. Residence time is represented by the inverse velocity and a constant that makes up all the other items particular to this heater. In this case the relative coking equation becomes:

    Rc = C/V exp(-Ea/R(1/T – 1/Tref))

    Where the reference temperature is 425 C or 698 K (you could use 700 K).

    This would be a reasonable way to estimate the change in heater coking rate at higher film temperatures for the same feed. As feeds are changed, the coke precursor concentration has to be taken into consideration as the base coking rate will be higher for higher coke precursor concentrations.

  • #28077

    Tushar
    Participant

    Thanks Mike for guiding through this basic equation, it makes lot of sense for reasonable estimates. You mentioned C/V “is the inverse velocity and a constant for heater”. Can you please provide some more details how to estimate that constant, I believe the velocity has to be get from flowrate and tube area.

  • #28093

    Mike Kimbrell
    Participant

    To estimate the velocity in each tube of the heater requires knowing the flow rate and the tube cross sectional area as a start. It also requires knowledge of the amount of velocity media (steam or condensate), the amount of vaporization of the hydrocarbons and the amount of resid conversion on a tube by tube basis. This requires knowledge of the temperature profile of the hydrocarbons on a tube by tube basis. Once this information is available, the relative coking value can be calculated on a tube by tube basis. Comparing the relative coking value between cases then will show the relative impact of the change in feed rate. The constant I mentioned is any value you want to use to get the results into a range that is easy to compare. The same constant should be used for each case on a particular heater.

    It is common for higher flow rates to cause the pressure profile to shift to higher pressures in each tube. This will reduce vaporization and velocity which will increase the residence time. The higher flow rate requires higher duty to meet the same heater outlet temperatures, so the film temperatures will be higher with longer residence times. The tube with the maximum relative coking value sometimes shifts from the outlet tube to one four or five tubes from the outlet at the higher flow rates and heat fluxes. That is why the relative coking value needs to be calculated for each tube for each case.

    Conventional fired heater software like FRNC5 by PFR or FH by HTRI will provide tube by tube temperature, pressure and velocity; however, they do not take the decomposition reactions that occur in a Coker heater into consideration. I know that by adjusting the property tables to account for the cracking that occurs can be done with FRNC5 and I understand it only takes a few manual iterations to get the values to converge.

You must be logged in to reply to this topic.

Refining Community