Jun-2021

Catalyst switch solves NOx removal issues

Changing to a corrugated honeycomb design from a pellet type for catalytic NOx reduction of a furnace’s flue gas stream overcame a series of operational problems.

KEN WOHLGESCHAFFEN, Chevron Products Company
KLAVS BELDRING, Umicore

Article Summary

Chevron Products Company operated two 1960s vintage hydrogen plants at one site. In 2000, ultra-low NOx burners were installed in one plant and a pellet style selective catalytic reduction unit (SCR) was installed in the other to reduce nitric oxide emissions from the plants.

The SCR consisted of a series of baskets (rectangular shaped modules) which contained 1.6 mm vanadium-titanium based tri-lobe catalyst pellets (see Figure 1), held in place by retaining screens. These baskets are parallel to each other in a grid and installed in the flue gas duct upstream of the furnace ID fan and downstream of the ammonia injection grid (see Figure 2).

Ammonia was supplied from wastewater treatment plants in the refinery (formerly Chevron’s, now owned and licensed by Bechtel Hydrocarbon Technology Solutions, Inc.). These generate anhydrous ammonia from sour water streams. The ammonia is vaporised in an electric heater, mixed with dilution air, and fed to the SCR through the injection grid. The ammonia flow is controlled based on the setpoint for NOx exiting the SCR (see Figure 3). The nitrogen oxides are reduced to elemental nitrogen and water (steam) across the catalyst according to the following reactions:

4 NO + 4 NH3 + O2 = 4 N2 + 6 H2O     [1]

6 NO2 + 8 NH3 = 7 N2 + 12 H2O          [2]

The unit was designed to reduce the ~200 ppm NOx (at 3% O2) in ~700000 lb/h flue gas by more than 90%. The gas flowed downwards and across the catalyst baskets in this lateral flow reactor (LFR, see Figure 4).

The catalyst easily achieved the NOx reduction target. However, pressure drop build was a major problem over the years. It was common to see the pressure drop increase over time due to particulates being collected in the small holes of the catalyst inlet screens and in the catalyst pellets in the baskets. The SCR was acting as a large filter on the furnace flue gas stream, trapping airborne dust, furnace refractory, and rust. Additionally, there was some attrition to the catalyst due to movement of the pellets. There was no way to bypass the SCR, and even if we could we would likely not meet the environmental permit requirement for NOx reduction.

Plant rates were reduced to keep below the maximum pressure drop limit which was set based on the mechanical (collapse) strength of the reactor. Eventually, the plant had to be shut down and the pellet catalyst removed, screened to remove fines, re-installed, and topped up with a fresh catalyst. Additionally, the catalyst baskets were hydroblasted to clear the small screen holes which tended to plug (see Figure 5). This was occurring about every 1-3 years outside of planned turnarounds.

In 2013 (red line in Figure 6), we had an unusually large increase in pressure drop compared with previous years. The pressure drop took a jump upwards following a plant pit stop for unplanned maintenance work on the CO2 removal system of the plant.

Investigation revealed that although we had not done any work in the furnace, sufficient loose refractory, rust, and debris from full thermal cycling of the furnace (hot to cold, then cold to hot again) resulted in a large amount of particulate clogging the SCR. Furthermore, our normal practice following furnace entry and furnace maintenance work is to temporarily install a filter screen upstream of the SCR and ‘cold blow’ air through it by running the ID fan. This allows us to collect and remove particulate generated by maintenance work before starting up the plant. Since no furnace work had been done, the filtration step was not followed. This caused the step up in pressure drop and high rate of pressure drop increase following the 2013 turnaround.

In 2014, we experienced the highest rate of pressure drop build ever seen on the unit (green line in Figure 6). In less than a year, the pressure drop was going to hit the high limit. Our investigation revealed the following causes of the excessive pressure drop:
1) We had Cetek coated the radiant section refractory of the furnace and re-tubed the furnace, resulting in lower temperature of the gas entering the SCR which caused ammonium salt formation
2) Large gaps had developed in the flue gas ducting, drawing cold ambient air into the gas, cooling it, and contributing to salt formation
3) Over-injection of ammonia caused ammonium salt formation. We saw the ammonium salts in the catalyst and inlet screens when we opened the unit up for inspection and confirmed their presence by laboratory analysis of samples.

When we saw this rapid rate of pressure drop increase in 2014, we decided to tackle the problem in a different way. We operate a Umicore corrugated honeycomb style SCR in another of our hydrogen plants. That SCR has never experienced the pressure drop issues we saw with the pellet style catalyst. We decided to convert the pellet style SCR to Umicore’s DNX deNOx catalyst (see Figure 7).

We partnered with Umicore on a fast-tracked project to replace the SCR in October 2014. The mechanical challenge was to fit the corrugated honeycomb catalyst into the existing ducting with little to no modification of the ducting. This required us to work closely with Umicore on the detailed design and installation of its DNX modules.

Another challenge was to define the catalyst design life because the catalyst change was in-between turnarounds, so the first objective was to make it to the next turnaround where the catalyst replacements should follow the turnaround schedule for the plant. It was well documented that ammonia salts were present in the existing catalyst, but it was difficult to quantify the amount and when and where they were formed. Normally, ammonia salts are identified as being a white, sticky substance visible around the catalyst, but in this case a greenish substance was also present. This was identified as ammonium chromium sulphate and was mainly found on metal sheets covering the pelletised catalyst on the inlet side.

The formation of ammonia salts occurs via chemical reactions dictated by thermodynamics, so changing the catalyst to a type with an increased porosity should possibly improve performance because the catalyst can accommodate more ammonia salts and overall provide a higher activity if condensation of ammonia salts occurs while keeping the pressure drop low. Experiences from Umicore on other units operating at or slightly below the ammonia salt dew point have demonstrated that the cumulative accessible surface area in the catalyst is lowered when ammonia salts condense in the micropore structure. (In catalyst types with a homogenous pore structure, the ammonia salts could block the micropore system, resulting in reduced overall catalyst surface area and consequent lower activity. However, if the catalyst has a non-uniform pore size distribution, optimal resistance to ammonia salt inhibition is ensured due to the more porous catalyst structure).