PTQ Q2 2023 Issue

1.2E-02

3.5E-03

3.0E-03

1.0E-02

2.5E-03

8.0E-03

2.0E-03

6.0E-03

1.5E-03

4.0E-03

1.0E-03

2.0E-03

5.0E-04

0.0E+00

100

120

100

120

COS

Mercaptan

6.0E-04

4.E-05

3.E-05

5.0E-04

3.E-05

4.0E-04

2.E-05

3.0E-04

2.E-05

2.0E-04

1.E-05

1.0E-04

5.E-06

0.0E+00

100

120

100

120

Time on-line / months

No poisoning

Gas plant

Nominal poisoning

RGG burner

Reformer hydrogen

AGE o-gas

Figure 6 Reactor outlet gas concentrations as a function of catalyst age at various poisoning levels

It is noteworthy that for the idealised case without poi- soning, almost all the exotherm still occurs in the top quar- ter of the bed, even for fully aged catalyst. Slip of non-H 2 S sulphur compounds increases from 25 ppm for fresh to 100 ppm for fully aged catalyst because catalyst activ- ity is reduced to the expected spent half-fresh level, with conversion decreasing for all reactions. The exotherm shift typically observed in operations reflects poisoning in the top of the bed, effectively reducing the amount of catalyst and the conversion, which further increases sulphur slip. Exposure to mild levels of contaminants, as encountered in a gas plant or refinery, and moderate levels (as expected) with a unit operating on an RGG burner causes catalyst deactivation at the inlet, moving the exotherm away from the top of the bed. At shorter operating periods, a less deactivated catalyst offsets the effect of poisoning and moderates sulphur slip. In the latter two cases, the model predicts that the first quarter of the bed is completely deac - tivated after about 100 months and 40 months, respec- tively. These poisoning contributions leave very little room for other plant upsets and shorten the operating life. In the most severe cases, such as reformer hydrogen or AGE off-gas, even meeting typical turnaround objectives of 3-5 years (36-60 months) is precluded. Figure 6 shows that in the case of reformer hydrogen and AGE off-gas, the bed activity declines rapidly compared to the other cases owing to the significant levels of con - taminants in the feed. Catalyst is exhausted by 60 and 20 months, respectively, with an even shorter useful operating

life. Outlet concentration plots also indicate the onset of trace sulphur slippage as COS and mercaptans in addition to carbon monoxide. All these compounds typically escape the TGU amine loop and reach the thermal oxidiser, thus increasing sulphur emissions. With severe poisoning, SO 2 slip also occurs early in the life cycle. Even minute levels of SO 2 slip can lead to its gradual accumulation in the quench system and TGU amine loops downstream. SO 2, being a relatively strong acid (compared to the reduced sulphur compounds), can severely reduce the amine solvent’s ability to remove other acid gases, in turn increasing overall emissions. Furthermore, the build- up of SO 2 can lead to severe fouling and corrosion in the quench system. This rigorous high-fidelity kinetic model can help designers and operators forecast the life expec- tancy of the catalyst bed. Figure 7 shows composition profiles along the reactor bed for various species. These would otherwise not be available from normal operating data. For fresh catalyst, all the SO 2 is converted in just the first 20% of the bed. This increases to 40% at 48 months without poisoning and 60% with an RGG burner for a 48-month-old catalyst. The significance of the water-gas shift reaction as a source of hydrogen can be inferred from the minima in the H 2 con- centration profile. The sudden initial drop can be attributed to the fast reduction of SO 2 to H 2 S, which stoichiometrically consumes 3 moles of hydrogen per mole of SO 2 converted. The hydrogen concentration then gradually starts increas- ing through the conversion of carbon monoxide by the

PTQ Q2 2023