Design and Simulation of thermal cracking reactor because of the chemical reactions in the processes and heat transfer in the furnace are very complex. Therefore, this process from the chemical reactions engineering perspective and from the perspective of transfer phenomena is highly regarded. Computational Fluid Dynamics (CFD) tools help us to investigate the development of detailed models of these reactors and by understanding the basic principles, the optimized design of reactor and by saving the cost and time performance of this equipment in a virtual laboratory can be done. In this research, investigate the proposed model of EDC cracking process in an industrial furnace and by operational data from a petrochemical unit to qualify the model and remove the reductions. To begin, by previous researches that worked on this specific occasion were collected. Then they are studied and investigated. Moreover, the reactor mathematical models were provided and solved by numerical software. Thereby, the operation of the reactor at different operating conditions investigated and sensitivity analysis of effective parameters will be done. As e result, this work could design new reactors or increasing the optimization capacity of existing reactors.

Molecular cracking

Molecular cracking was a process that used in the petrochemical industry and to reduce their hydrocarbons molecular weight by breaking their bonds. This process is the main principal of converting crude oil to useful fuels such as gasoline, diesel, jet fuel and kerosene. Thermal cracking, catalyst cracking, hydrocracking and cracking by water vapour were the most useful cracking processes in the industry. This process was done in the high temperature and pressure and without presence of catalyzer or in the low temperature and pressure with presence of catalyzer.

Geometry of thermal cracking reactor

In EDC thermal cracking process, coils containing process fluid, inside a box called furnace, are positioned horizontally. Coils placement inside the furnace shown in Figure 1 [1]. That positioned on the on both sides of the wall furnace, by burning gases of these furnaces the heat produced, that the heat transferred to the coils by convection or radiation model (Table 1). The number of intended coils to simulate 11 rows of pipes (Inconel 600) with overall length of 95 meters and length of each row tubing is 13.5 meters. Inner diameter of coils is 0.134 meter the distance between centers of the two tubes is 0.32 meter (Figures 2 and 3).

Figure 1: Placement of coil in the furnace.

Figure 2: Schematic of one row of the coil in the software.

Figure 3: Prospective a row of coil.

Table 1: Energy of EDC thermal cracking reaction [2].

Mass balance equation in the software proposed in form of equation 1

  (1)

ρ=Density

t=time

u_i=Potential energy

J_i=diffusion term

S_i=heat at point i

R_i=input heat to the reactor

Phrase J_i reflects is the term infiltration, which is not used in this simulation [2,3]. The energy equation in the Fluent software as follows:

   (2)

ρ=Density

p=pressure

t=time

T=Temperature

u_i=input Potential energy

u_j=output Potential energy

T_ij=torque

J_j=Output diffusion term

s_i=input thermal of chemical reactions

S_h=thermal of chemical reactions

K_eff=the effective conductivity

E=total energy

K_t+K=K_eff is the effective conductivity, K_t turbulent thermal conductivity, according to the model used is the perturbation. The first three terms on the right side of the equation as an expression of directed energy transfer, distribution of components and dispersion of viscosity. S_h is thermal of chemical reactions [4-6]. E is expressed as follows:

   (3)

u_i = input Potential energy

= pressure

ρ = Density

E = total energy

u_i = input Potential energy

h = input heat

Mass continuity equation in FLUENT software is in form of eqn 4:

   (4)

Above equation was the general form of the mass balance equation for compressible and non-compressible flows.

   (5)

In the above equation, P static pressure, τ_ij tensor stress, ρg_i and F_i physical forces of gravity and physical external factors are in i direction. Tensor stress is as follows:

   (6)

μ the molecular viscosity and the second semester is the influence of volumetric expansion [3].

Standard k-ε model

This was a comprehensive model that account as the simplest turbulence model. Two educational turbulent models enables solving the two transfer equations that separately determined turbulent length and time scale. Standard k-ε model is useful for powerful engineering calculations. Powerful of this equation, economical and acceptable accuracy for a wide range of turbulent flow as a result of its popularity in industrial flow and heat transfer simulation. This model is a semiempirical model. Standard k-ε model, was a model based on transfer equations for kinetic turbulent energy and dissipation rate (ε). In achieving the standard k-ε model assumes that the fully turbulent flow and molecular viscosity effects have been ignored. Because of the strengths and weaknesses of the equation to improve its performance, k-ε realizable and k-ε RNG model was introduced. K and ε is obtained by solving simultaneously the following two equations:

   (7)

G_k=turbulent kinetic energy generated due to the gradient of speed.

G_b=turbulent kinetic energy generated due to the buoyancy forces.

Y_M=Share Swinging delay in the compressibility turbulence for a total loss of (j)

C_1ɛ=constant parameter of model=1.44

C_2ɛ=1.92

σ_k=turbulent Prantel numbers for k=1

σ _ɛ=turbulent Prantel number for ε=1.3

S_K=term defined source for k

S_ɛ=term defining source for ε

μ_t=turbulence viscosity that can be obtained from the following equation:

   (9)

RNG k-ε model

This model is based on statistical methods, this model is the same as the standard k-ε model but with the following modifications: 1- An additional term (ε) has been added that significantly increases accuracy for flow to layer velocity. 2- The effect of the rotational movement (vortex motion) included in turbulences and increases accuracy for rotary flow. 3- An analytical formula for turbulent Prantel numbers are provided, as standard k-ε model considers a fixed amount. 4- While the standard k-ε model is the model for high Reynolds numbers, the theory of RNG for analysis of a differential equation for the turbulent viscosity can be provided that it considered the effects of low Reynolds. Effective use of this feature depends on proper behavior of the region near the wall of furnace. K-ε realizable model.

This model have limitations on Reynolds stresses that match with the physics of turbulent flows. This is possibly not in the standard k-ε model and not in the k-ε RNG model. In places where a large curvature in flow lines, vortex and rotation of the screw. K-ε RNG and k-ε realizable models have significant progress towards standard k-ε model. Preliminary studies show that the k-ε realizable model between k-ε models have best performance for confirmation numbers of complex flows and segregated flows. For all of this cases the performance of this model is better than the standard k-ε model. As it was observed that in the standard k-ε model, turbulent viscosity achieved the following equation:

   (10)

That C_μ is constant, but in the k-ε realizable C_μ achieved by the following equation:

   (11)

   (12)

   (13)

   (14)

Average rate of tensor with rotational speed is ω_k constants are defined below.

   (15)

   (16)

As it’s be seen C_μ placed a function of the average strain, rotational speeds, the angle speed of rotation systems, and turbulent fields are (k and ε) (Table 2).

Table 2: Thermo-physical properties of coils [4].

Wall of furnace

Profile of heat was applied in the walls. A total of eleven wall was located and each row of coil has a heat profile of its own. To enter the thermal flux resulting from furnaces for applying to the coils level use from proposed profiles were provided by Kegard in 2007. The length of each row of coils can be categorized Profile provided by the amount of heat to be entered in the software (Figure 4).

Figure 4: Thermal profiles calculated by Kegard.

Boundary conditions of cracking reactor

Inlet flow: Input for fluid flow due to the mass flow rate and the percentage of each of the compounds at the entrance of the inlet is used for the mass flow condition that account as boundary condition (Table 3).

Table 3: Output fluid conditions.

Output flow: The output flow used for pressure boundary condition. The boundary conditions are summarized in Table 4. Investigate the conversion factor and the amount of formed coke due to the different COTs. In this part of the research in different COTs EDC conversion factor and the amount of formed coke has been investigated. The amount of COT according to recent issues in review of EDC cracking processes to the values of 760, 770, 780 and 790 are being selected. How to achieve this profile to be the first calculated and then by using the proportion a variable factor is described. As a result, the new temperature profile in a particular COT due to the new data was achieved. In this method, input temperature to the coil doesn’t change and base COT temperature multiply in the and the other base profile points due to eq.17 and eq.18 multiply in a constant parameter to obtain new Profile points.

Table 4: Applied boundary conditions in the software.

   (17)

   (18)

The results of these calculations can be seen in Figure 5, COT different temperature profiles have been obtained. Figure 6 indicates the EDC conversion factor in different COTs, as the figure shows at higher conversion factor it will be higher in the COTs. But according to the Figure 7 if the amount of COT is higher, the amount of formed coke also will be higher. So by regarding the simulation results it can be said that for optimum economic performance of furnace, the temperature profile of the fluid is very important.

Figure 5: Process fluid temperature profile along the reactor.

Figure 6: Process fluid temperature profile along the reactor.

Figure 7: Process fluid temperature profile along the reactor.

Investigating of adding carbon tetrachloride as propulsion

By adding this compound to the feed in addition to increase conversion factor, it has other effects on the process. Due to increasing number of chlorine radicals in the process, number of unwanted compounds and Acetylene production and as a result coke production has been increased. Creating unwanted compounds in the next stages, in addition to VCM treating problem, also an important loss costly. As shown in Figure 8 conversion factor increases with increasing carbon tetrachloride, on the other hand in the Figure 9 formed coke, of the increases in carbon tetrachloride. So regarding to the simulation results it can be said that for optimum economic performance of furnace, the temperature profile of the fluid is very important. Therefore, optimize the amount of added carbon tetrachloride to the feed for economic performance of thermal cracking unit is very important. In this amount of propulsion conversion factor increases as an allowable limit and reactor performance due to forming coke and unwanted compounds are economical. It can be seen in the Figure 5.

Figure 8: Conversion percentage change Ay.dy.sy in different concentrations of carbon tetrachloride.

Figure 9: Change the concentration of coke in different concentrations of carbon tetrachloride.

K-ε realizable model is the most appropriate model for calculation because of these reasons: [7,8]

1- This version contains a new formula for turbulence viscosity.

2- A modified transferring equation for scattering rate, ε, can be considered. Therefore, this model is an exact equation for fluctuations transferring has been made.

In this research, effect of added carbon tetrachloride as a propulsion has been investigated. According to result of this research and other issues, if the amount of this compound does not reach more than 100 ppm, conversion factor will be increased. If this amount will be more than 100 ppm it’s not economically suggested.