Session VI - Petrochemical Processing

A. Propylene Question 1 Are there any pyrolysis gasoline hydrogenation units utilizing inhibitors to control organic fouling in the feed/effluent exchanger or reactor itself? Question 2 Refiners periodically increase olefin production to increase propylene make. This invariably leads to bottlenecking in the fractionation system. What operational changes can be made to get the most from the light ends fractionation of the FCCU? B. Ethylene Question 3 What new techniques are being developed to minimize coking in the new generation ethylene furnace technologies? Question 4 How much advantage in selectivity have some of the newer acetylene removal catalysts shown in commercial operation? Question 5 What is the best way to ensure that pyrolysis gasoline does not carryover into the water used in the dilution steam system in an ethylene plant?

Session VI - Petrochemical Processing

A. Propylene Question 1 Are there any pyrolysis gasoline hydrogenation units utilizing inhibitors to control organic fouling in the feed/effluent exchanger or reactor itself? BINFORD: Temperature has a direct effect on the fouling rate. The pressure and hydrogen feed rate in these units typically remain constant through the run. However, as the reactor fouls, the temperature on the feed side is increased to compensate for the reduction in reactor outlet temperatures. This increases fouling rate on the feed side of the exchangers. A number of European and U.S. ethylene plants processing pyrolysis gasoline through their hydrogenation units have reduced fouling in the feed/effluent exchangers and reactors with the use of antipolymerant blends. We have documented tripling of equipment run lengths in several applications. Fouling in these systems is due to free radical polymerization of diolefins and unsaturated aromatics, such as styrene. Temperature and concentration of the polymerizable materials control the rate of fouling. In most instances, chemical dosages required to control fouling remain fairly constant throughout the run, and average between 7 ppm and 20 ppm, based on unit liquid charge rate. We monitor performance by tracking the heat transfer coefficient on the feed/effluent exchangers and the normalized pressure drop through the reactor. TASKER: I agree with Mr. Binford. We have also experienced reactor bed fouling due to gums formed during storage or transportation of the feedstock. Pyrolysis gasoline is very reactive, and it must be kept under nitrogen and not be exposed to air. The storage temperature also has an important effect. So it is better to add any of the additives before transportation and to the rundown stream directly from the unit. The reactor outlet temperature is also very important and depending on the unit design, you could have high skin temperatures in the exchange train. As a guide, you should control the reactor outlet temperature to less than about 160 C to avoid polymerization reactions. Question 2 Refiners periodically increase olefin production to increase propylene make. This invariably leads to bottlenecking in the fractionation system. What operational changes can be made to get the most from the light ends fractionation of the FCCU? MALEK: With increased olefin production in the catalytic crackers, capacity problems manifest themselves primarily in loss of propylene yield initially. One of the key parameters in increasing propylene recovery, or maintaining it, is to keep the operating pressure of the absorber/de-ethanizer as high as possible. For a new design, this equipment would probably operate in the 260 psi to 275 psi range. Many units in the industry are operating at considerably lower pressure. One piece of advice on debottlenecking would be to try, by evaluating the pressure drop through the system, to see whether there is any way to operate your existing absorber/de-ethanizer at a higher pressure. Perhaps rerating some of the mechanical equipment might be a possibility. Another approach would be chill the lean oil to the absorber/de-ethanizer. The lean oil can be cooled to about 55 F before water hydrates begin to present plugging problems in this equipment. One approach that has been used quite a few times in the industry is the installation of tail gas recovery units on the back end of the catalytic crackers. These are usually cryogenic type units. This tail gas unit allows operation of the absorber section of the plant in a rather sloppy mode. The tail gas unit takes up the slack and maintains high propylene recovery. SAMUELS: I agree with Mr. Maleks comments. I would add that an overloaded gas concentration unit allows for quick payouts of advanced controls in the form of inferred properties or analyzers. Tower flooding constraints can be pushed but not exceeded a far greater percentage of the time when advanced controls are operated closed-loop. We have some excellent experience with nonconventional controls to maximize unit performance, for example, in our Garyville Refinery on a heat pump drive controller. We have increased throughput and allowed for a 99% recovery of

propylene and a 2.5% yield increase through an advanced control, that controls motor current at the desired optimum. FLETCHER: Maximizing propylene yield on the FCCU is carried out using different strategies depending upon whether the wet gas compressor is a limitation or not. If the compressor is not limited, a combination of increased reactor temperature and ZSM-5 will maximize propylene yield. If the compressor is already at maximum rate, a combination of operational and catalytic moves will be required. These moves include reformulating the FCC catalyst with an increased rare earth level, and possibly, a reduced zeolite-to- matrix ratio. These moves will reduce dry gas and LPG volume, and thereby, back off the compressor limitation. ZSM-5 additive with possibly a reactor temperature increase can then be applied to increase propylene production up to the wet gas compressor limitation. Additionally, most units will benefit by increased diligence at optimizing the temperatures, pressures, reflux, and sponge oil rates in the absorber/de-ethanizer and sponge towers for maximum propylene recovery. This particular strategy will be covered in detail, I believe, in the Heavy Oil Section. ANDREW W. SLOLEY (Process Consulting Services Inc.): Increasing olefins production, especially the propylene, impacts the FCC gas handling system from the wet gas compressor through the debutanizer. Increased propylene production via catalyst additives or catalyst formulations typically produces the olefins by cracking the middle boiling range portion of the FCC naphtha. Increasing propylene from the unit requires recovering the propylene in the absorber. Lowering gasoline production via cracking increases the quantity of material that must be absorbed while decreasing the absorption material (FCC naphtha). If the FCC dry gas rate is held constant and the gasoline is cracked, the absorber propylene recovery will decrease. While higher reactor olefins will always increase the total propylene production, the percent recovery will decrease unless some gas plant operating variables are changed. Specific operating changes that will help the gas plant include: 1. Maximize the sponge absorber operating pressure to the wet gas compressor limitation. Increasing wet compressor discharge pressure will have a relatively small impact on the compressor volume handling capacity, however, it will increase the horse-power requirements. Higher pressure improves recovery at a fixed absorber liquid rate. 2. Minimize the primary absorber lean oil temperature. 3. Maximize the primary absorber inter-cooler heat removal. 4. Minimize the stripper (deethanizer) reboiler duty to the alky feed C 2 limitations. Operational changes may not achieve the desired olefins recovery targets and the gas plant may need to be revamped. Depending on the unit limitations, increasing gas plant liquid handling capacity may be relatively cheap. EDWARD L. HARTMAN (Koch-Glitsch, Argentina): Higher light olefins production, mainly propylene, impacts the process equipment from the main fractionator through the gas concentration unit. The magnitude of the impact depends on the yield of C 3's and C4's, and the existing equipment limitations. In summary, these impacts are: Wet Gas Compressor: increase in wet gas actual volume and power requirements, due higher reactor effluent dry gas and C3's/C4's production. Operational changes to counteract this are discussed below. Inter-stage and high-pressure receiver condensers: increase in duty due to higher total mixed C4's produced (affects inter-stage condenser) and higher quantity of mixed C3's (affects high pressure condenser). Gas Concentration Unit Fractionation System: for the deethanizer, debutanizer, and C3/C4 splitter the reboiler and condenser duties increase (no condenser on deethanizer), along with the internal vapor and liquid traffic. For the depentanizer, the effect on duties and internal traffic is the opposite. Operational changes to handle this are discussed below. On the wet gas compressor, the compressor load is affected by dry gas production, inerts, C 3's/C4's production, inlet pressure and temperature. As dry gas and C3's/C4's production increases, optimization of the unit pressure balance and main fractionator heat balance becomes very important. Increasing the compressor inlet pressure results in significant reductions in wet gas rate, which counteracts the higher olefins yield. This is done by reducing the pressure drop from the reactor to the compressor, normally by packing the main fractionator. On the other hand, lowering the overhead receiver temperature reduces the wet gas rate as well. This can be achieved by modifying the main fractionator heat balance or changing its design, such as by installing a top pumparound in the tower to unload the overhead condensing system.

For the gas concentration unit fractionation towers, the higher internal vapor and liquid loads can be handled with high performance trays. This is generally more economical than building new columns or packing existing columns. The heat integration between the pumparounds in the main fractionator and some of the gas concentration unit reboilers also results in the need to adjust the heat balance or redesign some of the sections in the main fractionator. B. Ethylene Question 3 What new techniques are being developed to minimize coking in the new generation ethylene furnace technologies? MALEK: In the new generation ethylene furnaces, the continuing trend in the design has been: raising temperatures, lowering the pressure drop, and shortening residence time. Thanks to metallurgical advances, tube metal temperatures in pyrolysis furnaces now are in the range of 2000 to 2100F. As is well documented in the literature, there are certain catalytic effects that occur inside the tubes at these temperatures. These catalytic effects lead to coke formation and more frequent decoking cycles. There have been alloying techniques, sulfiding techniques, and other approaches to minimize these effects. I would like to comment on one of the design efforts that are being worked on for minimizing coking. Since the coking rate goes up dramatically as temperatures rise, the ability for coke to form can be minimized by shortening the residence time. It is very important that dead zones be completely eliminated in the piping and transition sections. There has been some interesting work in changing the geometry of the various piping tubes, manifolds, transition sections, and entries into the transfer line exchangers to accomplish this. There are sophisticated computer analysis tools now available that can do three dimensional flow analysis inside of these tubes. These analyses allow the geometry to be modified such that the velocities can held to extremely high levels. All the dead zones, for all practical purposes, are eliminated. This will go a long way toward minimizing coke formation and extending the life of the run on the pyrolysis furnace. There was a very good article published on this subject just recently in Hydrocarbon Processing by J.V. Albano and K.M. Sundaram with ABB Lummus. This appeared in the July 1997 Hydrocarbon Processing. Question 4 How much advantage in selectivity have some of the newer acetylene removal catalysts shown in commercial operation? TASKER: IFP has developed promoted LD catalysts that bring significant ethylene selectivity. These catalysts are supplied by Acreon in North America and ProCatalyse outside of North America. As important as the catalysts, process set up and operating conditions allow you to maximize catalyst cycles and yields. IFP has developed a new process named "Acetex", and, in this process the bulk of the acetylene hydrogenation is carried out in the mixed phase reactor. As for the MAPD hydrogenation, the liquid phase washes the catalyst, dissolves oligimers as they are formed, therefore stopping the polymerization catalyst fouling. The liquid phase is obtained by a liquid solvent readily available in a petrochemical complex. The Acetex is particularly attractive for feeds containing high acetylene concentrations of about 1.5 wt%. Ethylene yield increases in the range of about 1%, can be achievable. Existing units could be revamped to the Acetex process with modest investments. MALEK: Just to reinforce something along the same lines, the high severity operations have increased acetylene yields up considerably in pyrolysis furnaces. Proper acetylene hydrogenation is very critical. These new catalysts have the selectivity to hold the formation of the green oils to a minimum level. This will extend the catalyst life. I think this is one of the more important advances that has been made recently. Question 5 What is the best way to ensure that pyrolysis gasoline does not carryover into the water used in the dilution steam system in an ethylene plant?

BINFORD: The quench water/oil water separator or settling drum should remove virtually all of the pyrolysis gasoline from the quench water used in the dilution steam generator. Proper operation of the quench settling drum will minimize breakthrough. Pyrolysis gasoline quench water emulsions that will not resolve on their own in the quench settling drum can be resolved by feeding a small amount of reverse emulsion breaker, typically less than 5 ppm, into the quench water feeding the separator. These chemicals are typically net cationic solution polymers. The use of the wrong breaker dosage or feed point can actually make the emulsion worse. In this process, feeding more of a "good" thing is not necessarily better. Field testing on fresh emulsion samples is usually the quickest way to determine the best breakers for each individual problem. One common problem in these systems is that, if improperly treated or operated, the contaminants and/or the treatments will cycle up in the system, since the only methods for removal are blowdown and undesired carryover. As a result, the systems may operate well for a period of time until they cycle up to a point where a different chemical treatment or dose must be used. When this occurs, the emulsion breaker feed is typically turned off and the system blow down is increased until water clarity returns to normal. The breaker is then restarted at a low dose rate after doing a quick bottle test to confirm chemical performance. The pH of the system also plays a critical role in the overall performance of the unit, not only relative to natural emulsification in the system, but also due to its impact on the effectiveness of the emulsion breaker itself. The impact of pH on these systems is often overlooked or underestimated, and pH control is very important. Some ethylene plants have installed coalescers or DOX units down stream of the quench water settling drum to remove the last traces of emulsified hydrocarbons from the quench water blow down to the dilution steam generator. CHARLES T. ORLANDO (Pall Corporation): Pall Corporation has developed a new technology liquid-liquid coalescer system which is capable of pyrolysis gasoline emulsion separation down to 1.5 dynes per centimeter. Side stream testing at numerous plants has shown a milky emulsion which goes right through the older technology coalsecer systems and fouls the dilution steam system, will be removed by the Pall liquid-liquid coalescer down to 30 ppm and, in some cases, down to single digits. We have two units due to start up within the next month or so in France and in the United States.