SCOT inline heater combustion and mixing

Mike Henneke and Joseph Smith, CD-adapco
John Petersen and John McDonald, Zeeco, USA
Dave Wilson, Marathon Ashland Petroleum, USA

Refineries processing high sulfur crude oils produce significant quantities of by-product hydrogen sulfide (H2S), also called acid gas. This gas is often processed in a Claus Sulfur Recovery Unit (SRU). The Claus process converts acid gas (H2S) into elemental sulfur in an oxygen-deficient combustion process and then liquid sulfur from the condenser runs through a seal leg into a covered pit from which it is pumped to trucks or railcars for shipment to end users. Approximately 65 to 70 percent of the sulfur is recovered. The SCOT Process (Shell Claus Off-gas Treating Process) was developed by Shell, and introduced in the early seventies as an attractive process for improving the efficiency of a Claus sulfur recovery unit (see Figure 1). The process consists of four combustion processes (as well as catalytic reactors which are not discussed here):


Fig. 1: SCOT (Shell Claus Off-gas Treating) process		Fig. 2: Transparent surface view showing location of gas gun and spin vanes Fig. 3: Base case temperature (°F) contours of on centerline of burner and vessel

1. Reaction furnace
2. Inline reheater
3. Reducing gas generator
4. Tail gas incinerator

The CFD analysis discussed in this article considers only the second process, the inline reheater. The inline reheater heats the acid gas by mixing it with hot reducing products of combustion. An important design consideration is that the combustion products being mixed are under a reducing atmosphere. If O2 slip (uncombusted O2) is available to mix with the acid gas, the H2S can be oxidized to undesirable compounds (e.g., SO3, SO4, H2SO4) that can attack refractories and damage the environment.

$Wet O2 mole fraction (contours from 0-2%) shown 12”, 24”, 36”, and 48”downstream of fuel discharge. This figure shows the fuel/air mixing and indicates that O2 carry over does not occur.$
Fig. 4: Wet O2 mole fraction (contours from 0-2%) shown 12”, 24”, 36”, and 48”downstream of fuel discharge. This figure shows the fuel/air mixing and indicates that O2 carry over does not occur.		Fig. 5: Log10 of C2H2 mole fraction in the mixing zone between the products of combustion and the SRU tail gas. Note that only the mixing zone where the Claus gas enters has been analyzed.

CFD analysis

The purpose of this CFD analysis was to determine if the proposed burner design for the inline reheater would perform as required. In particular, the client was concerned regarding the following issues:

1. O2 slip
2. Soot formation in the reactor
3. Flame length
4. Swirl number of the combustion air
5. Uniformity (mixedness) of SRU tailgas and combustion products leaving reheater

These issues were analyzed using CFD at several operating conditions, but only the maximum liberation case is discussed in this article. Figure 2 shows the geometry of the CFD model as well as the flow inlets and outlets considered.

Chemistry

The chemistry has been approximated using the eddy break-up model. This model assumes mixing-limited chemistry, which is appropriate for most hydrocarbon combustion reactions. The chemical reactions considered are:

H2 + 1/2 O2 H2O (1)
CH4 + 3/2 O2 CO + 2 H2O (2)
CO + 1/2 O2 CO2 (3)
H2S + 3/2 O2 SO2 + H2O (4)

Figures 3 and 4 indicate that the thermal mixing between the SRU tailgas and the products of combustion is sufficient and that the exhaust is well-mixed. The figures also show that the near-burner mixing is very thorough so O2 slip into the SRU tailgas does not occur.


Fig. 6: Predicted gas temperature (°F) for Cases 1, 4, and 5		Figure 7: Predicted Acetylene Concentration Profiles (Log10) for Cases 1, 4, and 5

Soot formation potential

The model as formulated does not directly compute the formation of soot particles in the reheater. However, the model does do a good job of computing the major species profiles and temperatures. To estimate sooting potential in the mixing zone between the SRU tail gas and the products of combustion, we used the equilibrium program CET89 to compute the equilibrium gas composition at locations in the centerplane of the reactor. To do this, we used nine species concentrations (H2, CH4, CO, CO2, H2O, H2S, O2, N2) and the gas temperature and pressure at about 2500 cells in a constant temperature and pressure equilibrium calculation. Figure 5 shows the equilibrium-predicted C2H2 mole fractions in these cells. These calculations predicted extremely low equilibrium levels of C2H2. While equilibrium calculations are known to underpredict acetylene concentrations, we believe that these predictions indicate that the mixing between the combustion products and the SRU tailgas will form negligible amounts of soot.

Table 1. Flow rates for five performance cases used to characterize SCOT unit

Table 2. Comparison of predicted/measured pressure drop through reactor for selected cases

Performance testing:
Additional cases and comparison to experimental data

Besides the base case, five additional cases were used to demonstrate the capabilities of the current SCOT Burner design. Process conditions for these cases (see Table 1) included high Hydrogen flow rates at two stoichiometric conditions (Cases 1-3), refinery fuel (Case 4), and 100% fuel gas (Case 5). To evaluate the model’s ability to predict soot formation potential and performance, a test rig was built and operated at ZEECO. Comparisons between predictions and measurements for Cases 1, 4, and 5 are shown in Table 2. Results show (see Figures 6 and 7) essentially zero soot formation at the stack, which agreed with visual observations. Comparison between predicted and measured pressure drop through the reactor also show good agreement. Based on these comparisons, the proposed design was constructed and is being installed at Ashland-Marathon.