A Computer Aided Tool for Heavy Oil Thermal Cracking Process Simulation

10 pages
128 views

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
Computers and Chemical Engineering 25 (2001) 683– 692 www.elsevier.com/locate/compchemeng A computer aided tool for heavy oil thermal cracking process simulation R. Maciel Filho a,*, M.F. Sugaya b a Laboratory of Optimization, Design and Ad6anced Control, Department of Chemical Processes, Faculty of Chemical Engineering, State Uni6ersity of Campinas (UNICAMP), CP 6066, CEP 13081 -970 Campinas, SP, Brazil b Petrobras, Ilha do Fundao, CEP 21949 -900, Rio de Janeiro, RJ, Brazil ´ ˜ Received 10 Ma
Tags
Transcript
  Computers and Chemical Engineering 25 (2001) 683–692 A computer aided tool for heavy oil thermal cracking processsimulation R. Maciel Filho a, *, M.F. Sugaya b a Laboratory of Optimization , Design and Ad  6 anced Control  , Department of Chemical Processes , Faculty of Chemical Engineering  , State Uni  6 ersity of Campinas (  UNICAMP  )  , CP 6066  , CEP 13081 - 970  Campinas , SP , Brazil  b Petrobra´s , Ilha do Funda˜o , CEP 21949  - 900  , Rio de Janeiro , RJ  , Brazil  Received 10 May 2000; accepted 5 January 2001 Abstract This work presents the model for the development of a computer aided tool for heavy oil thermal cracking process simulation.It is proposed a dual plug flow reactor representation for the light pyrolysis of petroleum distillation residues in coil-type reactors.The resulting reactor model consists of two parallel plug flows, one vapor and the other liquid, traveling at different speeds in acoil. Reaction is assumed to be the rate controlling process with equilibrium between the phases. Because of the pyrolysis reactionsand pressure drop, vaporization takes place continuously along the coil so there is a decrease in the liquid hold-up. The modelderived uses a heuristic lumping approach based on pilot plant data and the resulting pseudokinetic scheme presents a certain feedindependence within the range of stocks available for the study. An industrial case study (delayed coking) is explored to provideinsight into the problem of reconciling the kinetics of pyrolysis and carbonization for the upgrade of distillation residues. © 2001Elsevier Science Ltd. All rights reserved. Keywords :  Delayed coking; Plug flow reactor; Pyrolysis; Thermal cracking; Cokingwww.elsevier.com / locate / compchemeng 1. Introduction The presence of metals and asphaltenic molecules inthe distillation residues has been a serious obstacle tothe use of catalytic processes. Progress in this area hasbeen significant but most processes are still limited todeasphalted oils and atmospheric residues of lightpetroleums, together with mixtures of vacuum gasoilswith atmospheric or vacuum residues (Le Page, Chatila& Davidson, 1990). Hence, thermal cracking remainsthe major option and because of the severe carboniza-tion which characterizes the thermal upgrade of heavyhydrocarbon fractions the usual practice is to split theconversion between a coil reactor and a soaking unitdownstream where coke accumulates, as in the delayedcoking process.The delayed coking process is illustrated in Fig. 1.The feed is usually a vacuum distillation residue al-though atmospheric residues and decant oils are some-times used. After mixing with a recycle stream in thebottoms of the fractionator, the combined feed goes toa fired heater where the pyrolysis reactions start. Bycarefully designing the tubes for high velocities and byusing large surface to volume ratios, the reactions canbe delayed until the reactants reach the coking drums,thus restraining the coke deposition in the furnace asmuch as possible. The products are fractionated andsent to other parts in the refinery for further treatment.The process is semi-continuous because condensationreactions cause coke to accumulate in the drum. Whenfull, the output from the furnace diverts to a paralleldrum, which has already been through decoking andthen tested and pre-heated. The drum removed fromthe operation is steam stripped, quenched with water,drained and cleaned.The furnaces are the most important and expensivepieces of equipment in the unit and usually have a boxgeometry with two radiant sections connected to asingle convection zone. The tubes are horizontal andlocated next to the walls in the radiant section. Theyare separated by a wide pitch to improve the angular * Corresponding author. E  - mail address :  maciel@feq.unicamp.br (R.M. Filho).0098-1354 / 01 / $ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.PII: S0098-1354(01)00669-X  R . M  . Filho , M  . F  . Sugaya /  Computers and Chemical Engineering  25 (2001) 683–692  684 distribution of heat and are connected by ‘U’ bends atone side and ‘mule-ear’ heads on the opposite side, soas to facilitate cleaning. Because of this, not all of thefurnace tube length is exposed to heat exchange ineither the radiant or convection sections. The dead endsincrease the pressure drop and soaking volume, whichneeds to be accounted for. The modern approach is tofurther improve the distribution of heat by placing thetubes at the center of the radiant boxes, double fired(Barros, Bernardo, Chan & Bria, 1997). Feed flow isdownwards, counter to the combustion gases, since thisreduces the maximum wall temperatures, which limitdesign.This contribution proposes a reaction-reactorconfiguration framework for the pyrolysis of distillationresidues in coil reactors formed by a 16-lump kineticscheme and a dual plug flow reactor description(DPFR) which is composed of two parallel plug flowsof vapor and liquid traveling axially at different rates.This configuration is important because the pyrolysisreactions take place mainly in the liquid and it isnecessary to take account of the residence times foreach phase.The resulting framework is then assessed with respectto the mild pyrolysis of petroleum distillation residuesin a complex geometry under non-isothermal condi-tions. The application of this model to delayed cokingfurnaces provides a useful case study which not only isnecessary for validation but gives insight to the com-plex interactions in this process technology. It is possi-ble then to explore the significance of design parametersand use the predicted information on the conditions of the feed as input to soaking reactors. 2. Process model 2  . 1 . Feed characterization Reaction modeling for the pyrolysis and carboniza-tion of petroleum distillation residues is characterizedby a considerable uncertainty concerning the represen-tation of the composition of the feed. Sophisticatedanalytical techniques contribute marginally in describ-ing a few structural properties of the charge and theyprovide only average parameters. The characterizationprocedure adopted here uses the methodology devel-oped by Dente, Bozzano and Bussani (1997), Bozzanoet al. (1998) and McGreavy and Sugaya (1998) whichfairly predicts molecular weights and boiling points aswell as some structural characteristics of the feed fromsimple properties, namely, initial boiling point, sulfurcontent, average molecular weight and density. Essen-tial results of the complex Altgelt and Boduszynski(1994) approach are surprisingly well reproduced. Themethod is a valuable tool when complex and lengthyanalyses are unavailable, as is usually the case. 2  . 2  . Characterization of products A material balance from D-86 and D-1160 ASTMdistillations gives the total consolidated effluent indi-cated in Tables 1–3. Small quantities of 2C 4   and1,3C 4   reported in the chromatographic analyses havebeen distributed over the remaining C 4 components.The equivalent TBP distillation of feeds and productsare divided into 50°C intervals which, gives and ade-quate smooth phase equilibrium curve. Other tempera-tures are used depending on the product specification(boiling ranges). 2  . 3  . Kinetics The pyrolysis of distillation residues at low tempera-tures is usually assumed to be a liquid phase phe-nomenon because the heavier and most reactivemolecules concentrate in his phase. The reaction isassumed to follow the first order, 16-lump scheme inFig. 2 and Eqs. (1) and (2) below:d c i  ,T d x = h L A T z L W  T k  i  c R,L (1)d P i  d x = h L A T z L W  L k  i  c R,L (2)where c i  is the mass fraction of species i  , h the hold-up, W  the mass flow rate, and P stands for SPGR or MW.Subscripts L and T refer to the liquid and total streams,respectively.Kinetic parameters are determined from experimentaldata obtained from Bria and Filgueiras (1982) who Fig. 1. The delayed coking process.  R . M  . Filho , M  . F  . Sugaya /  Computers and Chemical Engineering  25 (2001) 683–692  685Table 1Consolidated yields for eight pilot plant runs a VR-1Feed VR-22 3 4 11 2Run nb. 3 4 T  (°C) 480 500 508 485 470 512 520 5301 1 1 11 1 P (atm.) 1 12.07Rate (kg / h) 2.07 2.07 2.07 2.28 1.24 1.24 1.24H 2 2.97 0.00 2.39 0.00 1.13 1.22 1.834.17 4.17 0.004.17 0.00H 2 S 0.00 0.0034.00C 1 31.36 33.31 29.73 31.37 28.00 23.707.77 6.74 15.34C 2   11.167.63 11.75 11.6716.93 16.54 15.0316.63 16.31C 2 15.72 14.8510.41C 3   16.55 10.21 19.91 19.43 17.68 19.87C 3 12.03 12.57 12.16 8.06 10.63 13.85 14.824.91 7.59 4.696.08 4.811C 4 5.82 6.900.77 i  C 4 0.63 0.79 0.72 0.60 0.68 0.685.12 6.11 6.51 n C 4 4.575.32 5.27 5.672.50 2.77 1.27 0.301.72 0.76H 2  –C 4 1.10 1.620.50C 5  –75 0.97 1.40 1.75 0.00 0.11 1.18 1.931.66 2.03 2.37 0.0075–125°C 0.501.40 0.61 0.770.90 1.09 1.33 0.000.82 0.32125–150°C 0.36 0.481.10150–185°C 1.24 1.50 1.90 0.36 0.42 0.54 0.830.71 0.81 1.09 0.30185–204°C 0.300.72 0.33 0.431.89 2.07 2.85 0.631.98 0.93204–250°C 0.94 1.092.21250–300°C 2.47 2.65 3.57 1.22 1.36 1.37 1.722.90300–350°C 2.902.93 4.26 1.44 1.74 1.95 2.283.72 4.08 5.54 2.213.58 2.66350–400°C 2.86 3.585.40400–450°C 5.54 5.51 6.82 2.82 3.28 3.65 5.334.73 5.00 4.97 1.91450–475°C 2.134.63 2.42 3.865.82 6.55 5.76 1.925.61 2.17475–500°C 2.45 3.9317.58500–550°C 18.15 17.24 12.75 3.89 4.43 5.00 8.0546.81 44.39 43.80 83.00 78.88 75.23550°C + 64.1049.81 a Compositions in wt.%, gases in mol %.Table 2Specific gravities for eight pilot plant runsFeed VR-1 VR-22 3 4 11 2Run 3 40.6697C 5  –75 0.69330.5690 0.6560 0.6609 0.7354 0.6594 0.66090.7238 0.7328 0.7432 0.70930.6572 0.750375–125°C 0.7194 0.70930.743125–150°C 0.7678 0.7657 0.7972 0.7313 0.7682 0.7475 0.73130.7963 0.7887150–185°C 0.82030.7905 0.7339 0.7813 0.7707 0.75390.8109 0.8104 0.8443 0.76280.8109 0.7923185–204°C 0.7945 0.77530.8314204–250°C 0.8294 0.8294 0.8676 0.7972 0.8054 0.8086 0.80030.8602 0.8529 0.8990 0.8353250–300°C 0.82410.8660 0.8265 0.82650.8855 0.8751 0.9260 0.85810.8894 0.8398300–350°C 0.8545 0.84280.9141350–400°C 0.9088 0.8939 0.9465 0.8783 0.8576 0.8586 0.86020.9321 0.9170 0.9600 0.8967 0.8800400–450°C 0.87780.9371 0.88330.9478 0.9358 0.9672 0.91000.9516 0.8956450–475°C 0.8944 0.90470.9606475–500°C 0.9593 0.9503 0.9712 0.9182 0.9123 0.9059 0.92180.9752 0.9672 0.9765 0.9285 0.9303500–550°C 0.92420.9752 0.94440.9979 0.9868 0.9868 0.9665 0.98470.9972 0.9861550°C + 0.9972 looked at the thermal cracking of vacuum distillationresidues in a pilot. The overall flowsheet comprised afeed tank, pump, pre-heater, coil reactor, cooler andproduct tank arranged as in Fig. 3. Various crudes havebeen processed and for each of them an average of fourruns at different temperatures have been reported.  R . M  . Filho , M  . F  . Sugaya /  Computers and Chemical Engineering  25 (2001) 683–692  686Table 3Methods for physical properties used in the modelVapor phase Liquid phaseDensity Gunn and YamadaEquation of state(1971)Stiel and ThodosThermal Mallan Michaelian andconductivity Lockhart (1972)(1964)Dean and StielViscosity Twu (1985)(1965) rates and densities of the two phases. This requiressuitable correlations (e.g. Dukler Wicks & Cleveland,1964; Beggs & Brill, 1973; Agrawal Gregory & Govier,1973). The method of Hughmark (1962) is the preferredchoice for handling horizontal and vertical ascendingflows and has been adopted here. It is particularlyeffective when the flow regime is bubble, slug or annu-lar. The flow regime naturally changes as the reactionproceeds, with the increasing vaporization in the coil.In the case of vertical flow, the contribution of thestatic head to the pressure drop is calculated using  d P d x  el = (  L z L +  V z V )sin q (3)and the flow regime is estimated from the map providedby Griffith and Wallis (1961). Pressure drop is com-puted following Dukier et al. (1964).Phase equilibrium predictions are based on theRedlich–Kwong equation of state, as modified bySoave (1972). Flash calculations use a modified Rach-ford and Rice (1952) procedure. Physical properties canbe estimated from the methods listed in Table 3.The contribution of the supercritical components tothe liquid phase density assumes ideal mixture and usesexpansion coefficients from the API Technical DataBook (1988). The effect of the temperature over theliquid viscosities is computed with the ASTM methodmodified by Wright (1969). This allows all physicalproperties for each component and pseudocomponentin the system to be calculated so the total liquid orvapor property can be averaged using appropriate mix-ing rules.Hence, the resulting reactor model consists of twoparallel plug flows, one vapor and the other liquid,traveling at different speeds in a coil (under isothermal Fig. 2. Kinetic scheme. R, residue (550°C + ), G, gases, N, naphtha(C 5  –204°C), L, liglit gasoil (204–350°C), 350–400, yield of gasoilboiling between 350 and 400°C, SPGR 350–400, specific gravity of gasoil boiling between, 350 and 400°C, MW, molecular weight. 2  . 4  . Residence times The conversion process is assumed to be controlledby the reaction in the liquid, with a thermodynamicequilibrium between the phases. However, the vaporphase moves through the coil at higher speeds than theviscous, heavy liquid so the hold-up of the phases isdifferent from the fraction calculated based on flow Fig. 3. Pilot plant CR-0 where most of the experimental work has been performed.
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x