logo


Apr-2018

Automating catalyst withdrawal for improved safety

Catalyst withdrawals can be carried out continuously and reliably, minimising operator risk and maintaining the integrity of withdrawal piping.

KATE HOVEY and RICK FISHER
Johnson Matthey

Viewed : 4860


Article Summary

It is 2018 and we have cars that drive themselves, auto-landing planes, and household thermostats that can detect when you are getting close to home so they switch on the heating before you arrive. Innovations to automate our everyday lives surround us, and they continue to be developed. Yet, if you take yourself onto a refinery you will find that there are some activities that, although easily automated, continue to be carried out manually. One particular task is the intermittent withdrawal of 700°C catalyst from FCC/RFCC and DCC regenerators. This task exposes the withdrawal piping and valves to significant temperature fluctuations, a highly abrasive transferring medium and, ultimately, it places operators at risk. Many process operations on an oil refinery are automated in order to minimise risk and improve the safety of the people working in that environment; pressure control and the refinery flare system is a perfect example of this. So why are some hazardous operations left to be carried out manually when they can be automated by implementing some very simple engineering design?

I recall one evening working as a junior FCC engineer and shadowing a process operator to learn about some day-to-day activities. We stood on the FCC platform, it was dark and the rumbling regenerator was pouring out heat. The operator was instructed to unload some of the catalyst from the regenerator so he proceeded to open the valves on the withdrawal piping. The pipe began to glow red as if the metal was melting before my eyes. It was vibrating slightly from the flow of the scorching catalyst within, and the operator showed me the sections of piping that had been repeatedly patched where pinhole leaks had previously developed. I remember feeling quite impressed at being able to see this but at the same time quite astonished that the operator had to do this on a regular basis and that it was normal to simply weld a patch of metal over areas where erosion had overcome the integrity of the piping. I had accepted that this was an old unit and perhaps that some of the operations were a bit ‘outdated’. But I was wrong; this is not an outdated procedure at all, in fact this is the normal procedure for the majority of FCCs in the field today. Of course, some new designs have engineered in methods to try and reduce the erosion of the piping, but ultimately the procedure is still the same and the catalyst is withdrawn as a batch process.

Current batch withdrawal practices: life as we know it
Catalyst is continuously added to the FCC and it has been proven that costs are minimised when the addition is carried out as continuously as possible. This has been commercially demonstrated in a joint paper between BP and Intercat, Inc.1 Although it is undesirable, catalyst is continuously lost from the FCC, from the regenerator side along with the flue gas, and from the reactor side along with the product vapours. In most units these losses are less than the daily additions and hence it is normal for the inventory in the FCC to gradually build. On many FCCs, the reactor level is controlled by the spent catalyst slide valve and kept at a continuous level to ensure sufficient stripping efficiency, and the regenerator level is allowed to change. Because of this, catalyst needs to be withdrawn from the regenerator in order to maintain a constant inventory capacity, and refiners typically do this periodically by batch withdrawal carried out by operators. In most cases, the catalyst withdrawal rate is not well controlled and the velocities are both unmonitored and often unknown due to the excessive carrier or ‘cooling’ air. It is common for holes to form, especially at elbows and areas of higher velocity, and experience shows that erosion is significantly minimised when the withdrawal velocity is kept below 10 m/s. Additionally, cooling of the catalyst through the finned sections of piping is greater when the velocities are reduced.

FCC licensors will often provide their standard design for catalyst withdrawal piping, which will specify the pipe class, a section of finned piping for cooling, as well as temperature indications for visibility. It is also common to see some method of controlling the withdrawal flow without the requirement to partially close the isolation valves. These include sacrificial orifice plates, or venturis, which do unfortunately erode over time and need replacement. Some refiners actually use a number of manual valves in series which are choked back to control the flow. In this case, once one valve has been significantly eroded the refiner will move on to the next one and will continue to do this until the turnaround cycle when all the valves will need replacement. This is not an ideal operating practice.

As was mentioned before, this procedure subjects both operators and withdrawal piping to unnecessary risk. In addition to this risk, continuous addition and batch withdrawals also mean that the level in the regenerator is fluctuating and not kept steady.

Does batch withdrawal affect FCC operation?
Aside from the above mentioned safety concerns associated with batch withdrawals, unit stability is also compromised by periodic changes in the regenerator bed level. When the regenerator bed level is reduced as a step change, it has an impact on the heat balance, which directly impacts catalyst circulation and, ultimately, unit conversion. An example of this phenomenon is shown at a US refinery that periodically withdraws approximately 5% of its unit inventory. The quantity withdrawn amounts to approximately 6.8 tonnes of equilibrium catalyst and withdrawal takes about eight minutes. This withdrawal rate is only 3.5% of the catalyst circulation rate; however, the impact is still significant. The regenerator temperature increases by 5°C. Figure 1 shows the reduction in bed level during withdrawal, and Figure 2 shows the impact on regenerator temperature.

In addition to the increase in regenerator temperature, the regenerator pressure spikes during the batch withdrawal procedure. Figures 3 and 4 show how the regenerator and reactor pressures respond during the batch withdrawal period.

But what impact does this have? The importance of these changes can be realised by looking at how the product yields have been impacted. Although batch withdrawal only results in a temporary period of instability, the economics associated with these periods are highly significant and should not be overlooked. The regenerator temperature and pressure may be able to recover shortly after the episode, but the same is not realised for the product yields and unit conversion. Figure 5 shows how the FCC slurry yield elevates by 1 wt% during batch withdrawal and takes twice as long as the withdrawal period to recuperate. This temporary reduction in performance, and the economic deficit that results, is highly significant when aggregated over an extended period.

The displayed impact on regenerator conditions does not only have an impact on unit conversion, but also on the regenerator’s combustion kinetics. Excess bed levels will result in longer residence times in the dense phase and can result in poor air distribution and air channelling. Regenerator cross sectional mixing can be affected, resulting in coke combustion issues and localised areas of high temperatures. The same can be said for low regenerator levels following a withdrawal where the dense phase residence time is reduced. Figure 6 shows an example of another US refiner that experienced elevated levels of carbon monoxide as the regenerator level increased. The carbon monoxide concentration in the flue gas trends closely with changes in bed level and, although it is not represented in Figure 6, this also has a direct impact on the flue gas temperature. Fluctuations like these are not desirable and should be avoided where possible. It is a clear indication of how the bed level has a direct impact on combustion kinetics.

Continuous catalyst withdrawal: back to the future
Do we need to stick to historical, outdated, and unsafe practices of batchwise catalyst withdrawal? The answer is no. A fully continuous and automated catalyst withdrawal system has been designed, commercially installed, and has been in use since March 2016 at Marathon Petroleum Corporation’s (MPC) Garyville, Louisiana refinery.2 This system was directly tied into the existing withdrawal piping and is comprised of an Everlasting isolation valve, a positive displacement fan, and three finned pipe-in-pipe heat exchangers to cool the catalyst, and a collection vessel to receive the cooled catalyst. An overview of the installation at Garyville is shown in Figure 7.


Add your rating:

Current Rating: 4


Your rate: