Revamps 2025 Issue

2025 revamps ptq

FRACTIONATOR MODIFICATIONS

CO-PROCESSING ROUTES

FCC REVAMPS

UPGRADE STRATEGIES

PTQ SUPPLEMENT

UPDATE YOUR SUBSCRIPTION REGISTER FOR A PRINT OR DIGITAL ISSUE TO MAKE SURE YOU CONTINUE RECEIVING A REGULAR ISSUE OF PTQ ANY QUESTIONS? REGISTER CONTACT US CLICK HERE

revamps ptq ptq PETROLEUM TECHNOLOGY QUARTERLY

3 Fluid catalytic cracking (FCC) revamp opportunities Changes to fluid catalytic cracking unit input parameters may require upgrades for maximum unit value Warren Letzsch Warren Letzsch Consulting 9 Troubleshooting a dehydration train Advanced liquid distributor solutions increase reliability and profitability Norbis Velazquez and Michael Krela Koch-Glitsch Juan Ruiz Gas Processing Expert 15 Revamping distillation processes via dividing wall column technology Consider the operational stability, thermodynamic efficiency, improved product quality, and operational stability from high-performance DWC separation technology

Editor Rene Gonzalez editor@petroleumtechnology.com tel: +1 713 449 5817

Managing Editor Rachel Storry rachel.storry@emap.com Editorial Assistant Lisa Harrison lisa.harrison@emap.com

Graphics Peter Harper

Business Development Director Paul Mason Paul.Mason@petroleumtechnology.com tel: +44 7841 699431

Gaurav Agrawal Sulzer Chemtech

Business Development Luke Massingham Luke.Massingham@ petroleumtechnology.com Managing Director Richard Watts richard.watts@emap.com

21 Cartridge tray installation failures A study of the root causes for failed cartridge tray installations on-site, along with an alternative approach to mitigate the risk and optimise installation time Urmilesh Tiwari Engineers India Limited 27 Avoid utility disruptions in refinery operations Using time-delayed trips and dynamic process safety time analysis to avoid utility disruptions, especially where long shutdown and start-up sequences are required Charlie Gould Fluor Ltd 33 Revamp for co-processing: hydroprocessing challenges Co-processing via hydroprocessing represents a low-barrier entry into the production of renewable fuels. Risks and challenges have been segregated

Circulation Fran Havard circulation@petroleumtechnology.com

EMAP, 10th Floor, Southern House, Wellesley Grove, Croydon CR0 1XG tel +44 208 253 8695 Register to receive your regular copy of PTQ at https://bit.ly/370Tg1e

into five areas Woody Shiflett Blue Ridge Consulting LLC

Cover Repsol’s new polypropylene (PP) and linear low-density polyethylene (LLDPE) process units, both with 300 kta production capacity of 100% recyclable new products for highly specialised applications, represent an investment of more than 657 million euros, the largest industrial one in the last 10 years in Portugal. When fully operational in 2026, it will make the Repsol Sines Industrial Complex one of the most advanced in Europe due to its flexibility. Courtesy: Repsol

©2025 The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, mechanical, photocopying, recording or otherwise – without the prior permission of the copyright owner. The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care has been taken in the preparation of all material included in Petroleum Technology Quarterly the publisher cannot be held responsible for any statements, opinions or views or for any inaccuracies.

Rethinking Old Problems

More with Less

Revamp projects are difficult. Limitations imposed by plot space, congested pipe racks, and outdated equipment, to name a few, present unique challenges. Solutions that rely on excessive margins or comfortable designs lead to overspend. Now more than ever, process designers must find solutions that do more with less. P roven M ethods There is growing awareness that better scope definition earlier in the engineering phase saves time, reduces overall engineering cost, and leads to more successful projects. There is no argument that work completed during Conceptual and Feasibility phases is critical to getting a project on the right path. Engineers at Process Consulting Services, Inc. have developed a proven approach that makes the most of this precious time. At site, PCS engineers coordinate rigorous test runs, much of it through direct field measurements. Data collected is invaluable and often leads to low hanging fruit or hidden gems. Some refinery equipment performs better than design, and for various reasons others perform worse. Good test run data allows seasoned engineers to quickly identify what equipment needs investment and what equipment can be exploited. This way, solutions are developed that direct capital expense in the right areas and overspending is avoided. In one example, pressure drop measurements of a long crude oil transfer pipe showed the line could be reused, saving millions of dollars. Contact us today to learn how PCS’ proven methods can help you do more with less in your next revamp.

Projections for global supply and demand of refined products vary greatly depending on the pace of technological progress and degree of government policy enforcement associated with reducing greenhouse gas emissions. Without major advances in technology, it is hard to imagine a future without conventional fossil fuels over the next decade or two. Based on history, continued rationalization of refining assets is likely. Small, low-complexity refineries will struggle, while large, complex ones will thrive. Capacity creep through gradual improvement of refining units will continue to be a differentiating characteristic for remaining players. Focused revamps will play a critical role. Post-pandemic, inflation and a shortage of skilled construction labor have dramatically increased costs for refinery revamps. It is becoming increasingly difficult for many projects to meet corporate return on investment thresholds.

Field Measurements

3400 Bissonnet St. Suite 130 Houston, TX 77005, USA

+1 (713) 665-7046 info@revamps.com www.revamps.com

Fluid catalytic cracking (FCC) revamp opportunities

Changes to fluid catalytic cracking unit input parameters may require upgrades for maximum unit value

Warren Letzsch Warren Letzsch Consulting

T he fluid catalytic cracking unit (FCCU) represents an opportunity to increase the refinery’s profita - bility because of its large capacity and processing flexibility. It can convert heavy oils into gasoline and die - sel, produce olefins for alkylation and petrochemicals, and process a wide variety of feeds. The FCCU is usually oper - ated within multiple limits, which are based on existing feedstocks and product demands with its current equip - ment. When changes occur in the input parameters, a revamp might be necessary to extract the most value from the unit. FCC equipment revamps are frequently not considered for future turnarounds because capital is limited or the eco - nomics do not seem to justify a change. With FCCU run lengths spread out to four to seven years, the operator is ‘playing a pat hand’ for eight to 14 years if no improve - ments are made to the cracker. A revamp producing $0.20/ bbl will yield $10,000/day for a 50,000 bpd unit. This is $3.65 million/year. Therefore, a revamp with a simple pay - back of one to two years is worth considering. Justifying a project The benefits are usually largest for increased feed capac - ity, followed by improved yields and/or product properties. Reducing operating costs normally provides the smallest payback but can still pay dividends. While the FCCU model may show the required value to justify a project, a model of the entire refinery needs to confirm that the benefits will be realised. Optimisation of the FCC operation must be done such that all units impacted by the change, including the alkylation and reforming units, are not bottlenecks to the overall plan. Every refinery unit should be examined to determine capacity and identify bottlenecks preventing a more prof - itable operation. Good mass, heat, and pressure balances are required around the FCCU and individual sections of the cracker for the analysis. FCC units can be broadly classified as gas oil, resid, or petrochemical units. Some designs are a combination of these classifications. A hydrotreater in front of the cracking unit can be part of the cracking complex in that it provides a more desirable feedstock that can allow more versatility regarding yield structure and operating parameters.

Gas oil crackers downstream from a hydrotreater have a feedstock that does not need to produce much coke to sat - isfy the heat balance. Delta cokes can become low enough that the regenerator temperatures stay very low even when full CO combustion is desired. Maximum feed temperatures should be used to lower the coke make, which requires a fired heater for the feed. Maximum feed temperatures using only heat integration with the main column are about 550ºF, while a fired heater increases the feed temperature to about 720ºF. The feed end point will determine how high the feed tem - perature can be raised without the occurrence of coking in the heater tubes. Moving bed catalytic crackers processed gas oils with an end point of about 950ºF with feed heaters that approached 800ºF. A higher feed temperature lowers the coke required by the heat balance. CO₂ emissions are reduced from the process because the hydrogen content of the heater fuel is higher than that in the coke. Coke makes of 4 to 4.5 wt% are possible, translating to higher liquid yields. Higher reactor temperatures can be accommodated with low delta coke operations because the regenerator temper - ature stays within the normal operating range. In the future, refining complexes may turn into petrochemical plants and sites for recycling plastics. Reactor temperatures of 1,000 to 1,150ºF may be common for optimal performance and minimal catalyst deactivation. A hydrotreater designed for high pressures can provide a low delta coke feed regardless of the crude oil choice. Guard beds for these hydroproces - sors can remove impurities that might interfere with FCC yields and/or product properties. Petrochemicals include light olefins and aromatics, which are the main components from a petrochemical cracker. The gasoline from the catalytic cracker and the reformer produce aromatics. Switching the reformer to a benzene, toluene, and xylenes (BTX) operation (special catalyst and Research Octane Numbers (RONs) >100) can greatly increase the desired aromatics. FCCUs make higher yields of xylenes and less benzene than steam crackers, adding extra value to an integrated refinery/petrochemical com - plex. The unconverted hydrocarbons (paraffins and naph - thenes) from the reformer can be recycled to the catalytic cracker to produce more olefins.

Revamps 2025

www.digitalrefining.com

Reactor FCC feed riser revamps should be considered if bottoms conversion is lower than desired. This manifests itself as higher API gravities on the main column bottoms. When the residence time in the reactor is too low, there is insuf - ficient time for the multiple reactions necessary to provide lower APIs for the decant oil. This is usually caused by pre - vious revamps to increase the cracking capacity. Positive APIs are an indicator, especially those above +2.0. Pilot studies can be made to determine how much further cracking can be expected. If recycle of the bottoms is being employed to convert more bottoms, a riser revamp might be in order. Recycle of bottoms makes more coke and dry gas. Extra residence time in the reactor should reduce the need for recycle. Add another riser Adding a second riser to the process, which operates in par - allel to the main system, might be considered. Overcracking naphtha to produce more light olefins and aromatics in the cracked naphtha might be more easily controlled with an independent riser. Reactor temperature, pressure, residence time, and weight hourly space velocity can be manipulated in the added riser to achieve the best results. Control of both catalytic and thermal cracking is necessary to optimise the yields of dry gas, light olefins, and gaso - line octane and composition. Secondary reactions such as hydrogen transfer also need to be considered. Feed riser terminators Riser terminators are designed to both separate the cat - alyst and vapours, which stops the reaction, and to mini - mise the catalyst carryover to the main fractionator. It also must receive the stripping steam and hydrocarbons that are removed in the stripper. All the licensors of FCCUs offer a close-coupled system designed to minimise dilute phase cracking. Two-stage cyclone separators are marketed by several companies. The main difference between these systems is where the stripper gases are introduced into the exit gases. If it goes into the primary cyclone, then the pressure in the reactor vessel must be higher than the cyclone pressure, and the cyclone operates as a negative-pressure cyclone. If the introduction is downstream of the primary cyclone, then the primary cyclone is a positive-pressure cyclone (see Figure 1 ). Each configuration will work, but the cyclones need to be sealed. An unsealed cyclone will have more vapour underflow, which will increase delta coke and make more dry gas. Vapour quench Another technology that can be used is the quenching of the reactor vapours after the reactor temperature meas - urement. Typically, this is after the primary catalyst/vapour separator in the reactor (see Figure 2 ). This can reduce the post-reactor thermal cracking that gives dry gas and diolefins in the C₄ stream and cat gasoline. Quench added to the outlet of a rough-cut cyclone, which separates the products at the end of the feed riser, can instantly drop the

Pcn

Pcp

Hd p

Hdp

Figure 1 Negative and positive cyclones

Unit modifications: improving profitability Feed injection

The feed injection system is an area that can provide a very profitable revamp option. Benefits seen include increased conversion, lower delta cokes, reduced regenerator tem - peratures, and less dry gas. Changes to the feed system can include adding radial or additional nozzles to improve catalyst/oil contacting. Choosing the location and configu - ration of the nozzles is important, as is the instrumentation used. Reliability has to be 100%, otherwise the yields or run length will suffer. Not all feed injectors give the same performance. Earlier injectors were low-pressure drop systems and could be axi - ally or radially oriented. Every series of modern technology for previously installed injection systems showed real ben - efits over earlier designs. In one instance, a refinery looking at new feed nozzles had two offers that promised 19 and 33 cents/barrel, respectively. The higher value was chosen, and a test run indicated a 36-cent/barrel improvement. In retrospect, when asked what if it had chosen the lower bid and achieved the promised 19 cents, would they have been happy? The answer was ‘yes’. Many refiners have stopped looking for further gains and have settled for what they installed. The entire feed system needs to be reviewed if resid is to be added to the feed diet. The problem with resid is the coking that can occur at various locations of the hydrocar - bon circuit, which limits run length and adversely affects yields. This coking is often traced back directly to the feed nozzles, unit operating parameters, and/or design. Wet steam needs to be avoided.

Revamps 2025

www.digitalrefining.com

temperature as much as 75ºF. Since the catalyst is not being cooled, the amount of quench is minimised, and the coke make is unchanged. This quench acts as a pumparound on the main fractionator. More than 1.0 mil- lion bpd has been used in this service. Spent catalyst stripper Stripper modifications can also be made to reduce the amount of steam being used. If the steam rate is above 3 lb/1,000 lb of catalyst circulated, then better contacting with sufficient stripping stages can reduce this number to 2 lb/1,000 lb of catalyst. The main design variables for a stripper are catalyst resi-

External cyclone

Internal cyclone

Quench

Figure 2 Post-riser quench

dence time, pressure, temperature, and steam rate. If the residence time is too low, the heavier hydrocarbons do not have enough time to desorb from the catalyst. If the time is too long, both catalytic and thermal cracking occur, which can increase the delta coke and dry gas. Usual residence times are about 90 seconds, though some strippers per- form well with 60 seconds. Stripping temperature is a key variable. Raising the tem- perature improves the desorption of hydrocarbons, but the stripper temperature is not independently controlled. The reactor temperature is set, and a corresponding stripper temperature occurs, depending on where the reactor tem- perature is taken and the design of the stripper. Usually, the stripper temperature is a few degrees lower than the reactor. In resid cracking units, poor stripping has been observed when reactor temperatures were low. Many of the heavy crackers were run to produce diesel during the winter sea- son and ran with reactor temperatures of 920ºF to 940ºF. These were typical in gas oil crackers but produced poor results in the resid units. Raising the temperature to at least 970ºF greatly improved the operation by dramat- ically reducing the delta coke. Resid contains nitrogen compounds, which act as temporary poisons due to their bonding with the catalyst’s acid sites. Increasing the reactor temperature is a standard technique for dealing with high nitrogen in the feed for all FCCUs. Aromatic compounds do not desorb as fast as the saturated hydrocarbons. One possible benefit of reduced steam is a lower load on the overhead condensers on the main column. This could allow more feed for the unit. Less steam may provide an economic boon, depending on the cost of steam. In extreme cases, the steam rate was halved without any loss in strip- per performance. A hydrogen reduction in the coke should also increase the liquid yield. Another option is to put the steam into the feed injectors and improve the reactor yields. The hydrodynamics of the

feed system and stripper may provide revamp opportuni- ties. Tracer studies may be helpful with such a study. If the sour water stripper is loaded, then this may provide some relief. Regenerator Afterburn If afterburn is chronic, a revamp may be necessary to cor- rect the problem. The causes could be a low regenerator temperature, a short catalyst bed, or the spent catalyst distribution. Low bed temperatures are caused by low delta coke or short gas contact times. Feed hydrotreating can remove coke precursors, or light feeds may not contain enough large molecules that end up on the catalyst. Short-term solutions include raising the feed temperature, the catalyst activity, or the use of a CO burning promoter. If these do not alleviate the problem, then a revamp should be considered. Raising the bed level may also help. Add additional regenerator length Adding more regenerator length can provide more res- idence time for the air in the regenerator bed. This may prevent the afterburn from occurring in the flue gas line, plenum chamber, and/or the regenerator cyclones. Table 1 shows the data from a commercial unit on two different feeds. This can extend the life of the cyclones and prevent damage to the refractory. Extra bed depth also gives more freeboard to the catalyst, which can reduce catalyst carry- over to the flue gas system. This can reduce flue gas slide valve and orifice chamber erosion, as well as the amount of solids going to the wet gas scrubber or electrostatic precip- itator (ESP). This is especially true for units operating with a superficial velocity >3.2 ft/sec. Whenever the cyclones of the reactor or regenerator are going to be placed, the equipment in that vessel should be examined for possible improvements, as the head removal

Revamps 2025

www.digitalrefining.com

Regenerator temperature profiles

Bed level

Dense

Dilute

Normal (A)

7 8

(B) (A) (B)

11 27 22

Low

33 33

Table 1

provides easy access to the internals. The pressure balance can be revisited, and changes might be made to anticipate future feeds and operating modes. Spent catalyst distribution Even distribution of the spent catalyst and air is essential for lowering afterburn and meeting the regulations for emissions of CO, NO, and SOx. Refiners with a design that transfers the spent catalyst into a single quadrant of the regenerator have temperature variations of up to 100ºF in the bed. Often, uneven cyclone temperatures are observed, and afterburn occurs in the cyclones, plenum chamber, and flue gas line. The high CO concentration above the bed where the catalyst enters and the high O₂ in the other sec - tions above the bed make the afterburn of CO inevitable. Changing the air distribution has not been successful in alleviating afterburn when all the spent catalyst is put into a single quadrant. Dispersing the catalyst over the top of the bed has eased the problem. One such distributor is shown in Figure 3 . In the regenerator, the air and catalyst move vertically, up in the middle and down at the walls, with little radial mixing. Adding multiple distribution arms has worked well, as has a single distributor across the entire regenerator.

Figure 3 Regenerator air and spent catalyst distributors

held a lot of catalyst, because the carbon steel available had a limit of 1,200ºF, which limited the coke burning rate. Bed depths of 10 to 15 ft were adequate since the superficial velocities were around 1.5-2.5 ft/sec. Side-by-side designs (reactor/regenerator) were used, and the spent catalyst line entered the regenerator from the bottom with carrier air, which pushed the catalyst through plates that contained holes. The plates were pie-shaped and sealed at the walls with a catalyst seal. If the air broke through the seal at the wall, a hot spot would form and could progress all the way around the regenerator. These were boxed in and cooled with steam until the unit was shut down for a turn-around. An even flow of the gas through the holes was main - tained by ensuring the pressure drop across the plate grid was at least 30% of the bed delta P. Lower pressure drops would allow the grid to weep nearer the walls, and the resulting catalyst flow would cause erosion of the grid due to higher velocities in the active holes. The holes have to be large enough so that they do not plug, and the plate needs to have the required thickness for mechanical considera- tions. Hole spacing needs to provide good coverage while also not allowing the coalescence of gas bubbles above the plate. Air rings could also be applied in these applications, but more than one would have to be used with larger-diameter regenerators. The nozzles used to distribute the air have two orifices. The inner orifice is sized for even flow to all the nozzles, while the outer nozzle controls the velocity into the bed. Velocities need to be high enough to provide the needed coverage of the cross-sectional area, yet not so high as to cause catalyst attrition. The number of rings depends on the maximum allowable jet penetration from the rings. Pipe grids (Figure 3) are also used in these smaller designs,

Even distribution of the spent catalyst and air is essential for lowering afterburn and meeting the regulations for emissions of CO, NO, and SOx

Counter-current catalyst and air flow have proven to be effective in reducing catalyst deactivation and controlling NOx and CO concentrations. Control of excess O₂ in the flue gas is better, which improves burning efficiency. Air distribution The air distribution system is an area that might improve operations and profitability. Hardware used in this service includes flat plates with holes, pipe grids, and air rings. Each design had to overcome issues that caused operating problems, high maintenance costs, uneven air distribution, and short run lengths. The size of the FCCU dictated the type of distributor. Large-capacity units had large-diameter regenerators that

Revamps 2025

www.digitalrefining.com

which can be in a stacked or a side-by-side configuration. These distributors also require the 30% of the bed pressure drop criteria to avoid catalyst entry into the air distributor. Metallurgy has been changed to stainless steel to allow higher bed temperatures and full CO burn in the regenerator. Erosion can occur faster at the higher temperatures, so circu - lating catalyst through any distributor needs to be avoided. Air blowers When more feed is processed, the required coke make will increase. This usually requires more oxygen to process the extra coke produced. Assuming the air blower is at a limit, an auxiliary blower can be added. This is normally the cheapest way to add more oxygen. However, the air distributor might need revamping due to the increased velocity and higher superficial velocity. The latter will increase cyclone loadings while the former might cause catalyst attrition. A revamp of the existing air blower may be possible depending on its type. Centrifugal machines can be rewheeled, or the driver can be changed if more power is required. If the pressure needs to be raised to accommodate the additional traffic through the fractionator, then the pressure balance needs to be checked to handle the increased pressure drops and changes in unit yields. Oxygen enrichment Adding oxygen to the air is another way to increase capac - ity and is very effective when the superficial velocity is 3.3 ft/sec or higher. This reduces the coke make and CO₂ emis - sions, which helps minimise greenhouse emissions. Any unit operating with a hydrotreated feed is a poten - tial candidate for oxygen (O₂) enrichment. These are low delta coke feeds that produce low bed temperatures. With a lower air nitrogen content, more heat is transferred to the catalyst, and the bed temperature goes up about 10ºF for every 1.0% increase in O₂. More optimal burning may occur, allowing a reduction in burning promoter usage. If a CO₂ tax is passed, the reduced coke would provide an additional payback. Depending on the tax rate, oxygen can be added to the air to further reduce the coke needed to operate the process. A higher regenerator temperature also results in a more favourable regeneration operation, since afterburn is lower at higher bed temperatures. CO promoter use might also be reduced. Coke make of around 4.0 wt% would result, and the combination of coke and dry gas should be 7.0 wt% or less (the ‘Letzsch num - ber’) for modern FCCs running with available advanced technologies. If the dry gas is separated into C1s and C2s, the latter can be sent to an ethylene unit and almost quan - titatively converted into ethylene. This reduces the dry gas by about 50-60%. Methane can be used as fuel gas, feed to the hydrogen plant, or converted to electricity. Coke plus dry gas could be less than 6 wt% based on feed. Gas plant If there is a high concentration of diesel in the bottoms stream from the main fractionator, a loss of revenue occurs if this stream is sold as slurry oil. Sometimes, the main frac - tionator bottoms temperature has to be reduced to prevent

coking in the slurry circuit. Some diesel can end up in the bottoms. Another possibility is that the seal oil (LCO) used for the slurry pumps is also lost in the bottoms. A downstream stripper on the bottoms stream can recover this diesel and be both economical and provide more operating flexibility. There may be some modifications to the gas plant that will provide additional value. More cooling of the main column overhead stream, changing trays in part or all of the main col - umn, or changing to packings that reduce the pressure drop are all ways that gas plants have been successfully modified. Better recovery of the liquefied petroleum gas (LPG) might be obtained by cooling a stream to the primary absorber. There are many tweaks that can be made to the gas plant, and a review by experts can identify potential projects. The FCCU will remain a critical part of refining/petrochemical facilities, and it will pay to provide the catalytic cracker with the latest equipment and make it as nimble as possible Catalyst and additive adders The benefits of quasi-continuous catalyst additions are well known. Catalyst deactivation is minimised, and the yields are better. Catalyst losses are lower. A reliable loading sys - tem should be part of every unit. The FCCU should also have a continuous additive load - ing system that allows the unit to load additives as needed. Systems that can add as many as four products at a time are desirable. CO promoter and DeSOx additives might be added regularly, and the use of ZSM-5 and a bottoms cracking additive might be used on an as-needed basis. If ecat is used, a separate silo and loader would give the best performance. Conclusion FCC processing is going to change in the future. Gasoline and diesel demand will decline with more electric vehicles and the integration of refining with petrochemicals. More downstream chemical processes may become part of the processing complex. FCCs will be run at higher tempera - tures that can make both ethylene and propylene, which are fundamental building blocks for petrochemicals. The FCCU will remain a critical part of refining/petrochemical facilities, and it will pay to provide the catalytic cracker with the latest equipment and make it as nimble as possible. Warren S Letzsch has 56 years of experience in petroleum refining, including petroleum catalysts, refining and engineering, and design. He has authored more than 100 technical papers and publications and holds eight patents in the field of FCC. He holds BS and MS degrees in chemical engineering from the Illinois Institute of Technology. Letzsch is a Fellow of the American Institute of Chemical Engineers. Email: wletzsch@verizon.net

Revamps 2025

www.digitalrefining.com

We Make Precious Metals... Sabin Metal We Don’t Make Catalysts. At Sabin Metal Corporation , our focus begins where your catalysts finish. We specialize in recovering precious metals with unmatched precision. Our global precious metal recovery services help you:

Our detailed reporting on spent catalysts helps your team assess reactor performance and plan for future efficiency. SUPPORT ENGINEERING INSIGHTS

By diversifying recovery channels and strengthening your precious metal accountability process from start to finish. MITIGATE SUPPLY CHAIN RISK

We consistently deliver one of the highest recycled precious metal returns in the industry, maximizing value for our clients. PROTECT YOUR BOTTOM LINE

Sabin sources all feed from secondary streams; our 100% recycled precious metal reduces your carbon footprint. REDUCE MATERIAL WASTE

For over 80 years, Sabin Metal has served clients in more than 60 countries. Partner with Sabin for responsible refining and innovative recovery solutions.

Learn more about our services and why leading companies trust Sabin at www.sabinmetal.com

Troubleshooting a dehydration train

Advanced liquid distributor solutions increase reliability and profitability

Norbis Velazquez and Michael Krela Koch-Glitsch Juan Ruiz Gas Processing Expert

I n the current global energy environment, natural gas is expected to hold a strong position as a reliable fuel in the years to come, and it is considered a bridge fuel towards an era of more sustainable energy sources. 1 Although conditioning of natural gas is a mature field, operators, along with licensors and equipment vendors, keep chasing opportunities to improve processes and opti- mise designs to minimise their environmental footprint, improve energy efficiency, increase reliability, and reduce operational costs. However, in a field where everything seems to have already been said and done, finding such opportunities constitutes a challenge of its own. Natural gas dehydration is one of the ubiquitous pro- cesses that almost any natural gas stream undergoes, as the presence of free water in the gas at pipeline conditions can lead to operational issues, including hydrate formation and corrosion. The degree of dehydration depends on the use and the transit of the gas stream from its source to its destination; the colder the temperature the gas is expected to reach, the more stringent the dehydration specification will be. There are different technologies available to desiccate nat- ural gas streams, and most of them require, to some extent, the use of absorption or distillation processes equipped with mass transfer equipment such as trays, packing, and distributors. Choosing the right process depends on several factors, including the initial water content, process charac- ter, operational nature, economic factors, and the water specification required downstream.2 One such method is water removal by compression and cooling, which aims to decrease water saturation content, first by compressing the

gas and then cooling it to an adequate temperature to gen- erate the desired dehydration level. A plant gas operator in Western Canada, which runs sev- eral midstream assets throughout North America, faces the challenges of conditioning and dehydrating its natural gas production. In its Canadian operation, it operates several proprietary Ifpexol units. This technology allows dehydra- tion and dew point control by using a cold process in the presence of a methanol solution as a single solvent. A methanol-water solution is recovered from the dried gas and recycled back to the two strippers, where a par- tial stream of raw wet gas is used to strip out and recover methanol from the recycled methanol-water mixture. The objective of the stripper is to obtain pure water at the bot- tom and an overhead gas stream loaded with methanol, which serves as an antifreeze agent downstream. The Ifpexol process focuses on the economy generated by a highly integrated dehydration process using a com- pression-cooling approach. In this technology, a methanol closed-loop aqueous solution is recirculated to control hydrate formation in the cold section of the unit. If the methanol losses through the bottom of the strippers are low, the process offers attractive operational costs while being environmentally friendlier than other dehydration technologies, as there are no harmful vent emissions.³ Operating principles The methanol regeneration of the Ifpexol process occurs in the stripping section, where a water-saturated hydrocar- bon gas stream is contacted in counterflow with a methanol solution in a packed stripper tower. The stripping process

Sales gas

Methanol make-up

Cold separator

I fpexol stripper

Methanol solution

C 3 + stabilisation and methanol recovery

Wash water stripper

Gas feed

Wash water + methanol

Compression section

High-purity excess water

Figure 1 Overview of Ifpexol process for dehydration

Revamps 2025

www.digitalrefining.com

relies heavily on a high surface area packing to promote the stripping of the methanol from the methanol-water solu- tion coming from the cold section and stabilisation section. Given the different temperature levels and nature of the available methanol solutions, two strippers are required to regenerate the methanol; these are referred to as the Ifpexol stripper and water wash stripper (see Figure 1 ). The technology licensor specified the use of wire gauze structured packing with a bed height of about 6,000 mm, one inlet feed distributor for the saturated wet gas, and an overhead distributor for the methanol solution liquid feed. Although the fluid properties and operational conditions were not optimal for the use of wire gauze structured pack - ing – due to the high surface tension and relatively high vis- cosity of the liquid phase, as the liquid would not be able to harness the advantage of the capillary effect of the woven material – the high surface area of wire gauze structured packing required (in the range of 500 m2/m³) makes it an appealing option. It allows for enough stripping stages from a rather small amount of liquid compared with a sizeable flow of gas. After the commissioning of the unit, the operator started struggling with stripping efficiency in both towers, leading to a higher fresh methanol consumption and an unsteady operation, especially in the Ifpexol stripper. Similar situa- tions presented after the operator commissioned analo- gous units in different locations. According to the licensor models, the methanol recovery performance of the strip- pers was about 50% lower than expected, and the metha- nol losses were at least twice as high as the design target. While studying the preliminary information received from the field, it was highlighted that the fouling of the distribu - tors (and consequently their cleaning) was more prevalent in the stripper, which operates at cryogenic conditions in the overhead of the tower. This, along with evidence from elsewhere in the plant, supported the idea that the fouling was generated by an accumulation of paraffinic compounds with a tendency to solidify, especially at low temperatures. These compounds were inherent to the gas composition, even if they were not expected to be present at the design stage of the plant. As a result of these findings, it was determined that the cause of the unexpected poor performance was due to the inefficiency of the original internals in the strippers, rather than being inherent to the technology or process design. Consequently, The operator decided to collaborate with Koch-Glitsch to leverage its expertise in addressing the plant’s operational issues. Process evaluation The operational challenges were varied, affecting the process in multiple ways. Operators faced a higher-than- expected pressure drop in the packing, poor methanol stripping, freezing in the cold section due to insufficient methanol circulation, blockages in the liquid feed distrib- utor, and increased consumption of make-up methanol. To cope with this situation, operators had to constantly adjust the process to try to optimise mass transfer in the strippers. Some of the usual actions included increasing the use of

fresh methanol upstream of the cold process to mitigate freezing, manipulating the gas split ratio of the gas feed between the strippers to optimise the methanol recovery, and performing ‘blowbacks’ in the tubular liquid distributor to try to improve the quality of the distribution. Preliminary hydraulic studies were conducted at design and operational conditions, and no indications of evident constraints were found in the packing or the internals. However, the saturated gas was not able to produce enough stripping of the water-methanol feed. Even after a series of operational adjustments, the methanol recovery targets were not met. A grid gamma scan was conducted on both strippers but given the very low liquid flux in both towers, the results of the scan were deemed inconclusive. Knowing the packing was not hydraulically constrained, and the issues could be summarised as a lack of efficiency and operational relia - bility, additional efforts were aimed at studying the liquid distribution. In theory, the original design for the tubular liquid dis- tributors made sense, but they generated significant short - comings during the operation of the strippers. To achieve high drip point density, essential for effective liquid distri- bution in tubular designs, the holes in the laterals were set at 1 mm in diameter, which led to persistent clogging and poor distribution, making the operation unreliable, imprac- tical, and inefficient. Even when the original distributor showed a high drip point density of 200 points/m2, a distribution quality assessment using a widely accepted method in the indus- try5 produced a rating of 64%, whereas the expectation for wire gauze packing is to use high-performance distributors with distribution qualities surpassing the 90% threshold. The distributors frequently became plugged, and meth- anol losses grew unsustainable due to decreased stripping efficiency. Operators would stop the methanol-water flow and use the tower’s high pressure to clear obstructions back into the feed piping system, temporarily restoring flow. This process subjected the system to process and reliabil- ity issues, potentially compromising its mechanical stability and losing stripping capabilities for a few hours at a time, at least once a week, with all the attached consequences this brings to the operation downstream. However, even after cleaning, the efficiency was not on target, and it quickly declined from there, starting the cycle again. Although the Ifpexol and wash water (WW) strippers aim for similar results, their operations differ. The Ifpexol strip- per is designed to contact the cold water-methanol mixture (>70 wt% methanol) from the cold process with saturated gas. In contrast, the WW stripper is focused on recovering any remaining methanol that may have migrated from the cold process to the C 3 + stabilisation section. The liquid feed to the WW stripper is largely dependent on the feed gas composition and is not directly linked to the Ifpexol stripper’s overhead liquid feed. Despite these differ- ences, having the same diameter enabled the original equip- ment manufacturer (OEM) to make a design shortcut, where a single-design distributor was used in both towers instead of using two distinct distributors calibrated for each.

Revamps 2025

www.digitalrefining.com

Improper initial liquid distribution can severely impair the efficiency of packed towers. Severe forms of initial maldistribution, such as those induced by inadequate liquid distrib- utors, lead to inconsistent liquid flow across the packing. This maldistribu- tion manifests as chordal, peripheral, and central blanking profiles, each severely affecting column efficiency by creating over- and under-irrigated zones within the packed bed. When a scenario of initial gross maldistri- bution is presented, a packed bed cannot self-correct this situation⁴, and the performance is destined to bear the consequences. Once the issues with the distributor designs were iden- tified, the task was to produce a new set of distributors that aligned with the internal rates and operational ranges of the towers. After receiving feedback on a set of new operational conditions from the technology licensor, a new simulation model of the strippers was generated using operational data, updated gas compositions, and current saturated gas flows to each of the strippers. This new set of information revealed the nuances between the operation of each tower, informing the design choices in each case. Selecting the right liquid distributor Packed towers operate as a system, with the performance of the packing being directly linked with the type and design of the liquid and vapour distributor. Improper feed distributor design accounts for a considerable number of troubleshooting issues in mass transfer columns, even for relatively simple distillation and absorption systems. The choice of a liquid distributor for a packed tower is informed by several factors, such as the packing type, pro- cess application, liquid load, operational range, and fouling tendency. However, sometimes variables like the available overhead space above the packing, existing tower attach- ments, manway size, and installation time constraints can play a role in defining the final design of a liquid distribu - tor, potentially compounding the complexity of the retrofit. Hence, a one-size-fits-all approach is rare, especially in revamp scenarios. Considering that the original liquid distributors in the strip- pers were pressurised, a compact design was possible, as no static head of liquid was required as a driving force. As a result, the vertical spacing for a new installation was limited. Designs for liquid distribution systems require thorough hydraulic knowledge and a deep understanding of the operational windows and process conditions. Therefore, close collaboration between the parties involved is crucial to the success of the design. There are generic rules for the selection of liquid distrib- utor types, but in general, these are just guidelines and should be treated as such. Therefore, looking at the dis- tributor types just in terms of flow range is simplistic, and this approach should not be used as a replacement for a comprehensive evaluation of the system.

Figure 2 Comparison of liquid distributors Models 166 (left) and 156 (right)

The final selection of a liquid distributor must be the result of a judicious assessment of operational conditions, packing type, fouling potential, mass transfer requirements, available installation spacing, and operational flexibility, among others. To adapt to all these parameters, there are several levers that can be moved. Although having two identical distributors in towers with different operational conditions can save money during the project phase, it can prove costly in the long term due to poor performance dur- ing operation. Once the operational conditions and expectations were revised between the operator and Koch-Glitsch, the deci- sion was made to use trough distributors for both tow- ers, as they offered the most comprehensive solution, addressing most of the issues captured during the evalu- ation. However, if a single design was required (following the original design philosophy), high flow variations were expected for the WW stripper at intermediate and low rates, which would produce severe maldistribution and poor stripping efficiency. Redesign approach Several issues needed to be fixed within the confines of the mechanical constraints of the strippers. For instance, the disruptive cleaning process of the liquid distributor using blowbacks (backflushing the feed distributors into the feed line utilising the high internal pressure of the gas in the tower) was addressed by removing the pressurisation from the distributors and using an open pipe to transport the liq- uid feed into the tower. However, without pressure as a driving force to spread the fluid through the packing, the distribution would rely on any static head the new design could generate in the limited available space. Even if one would aim for a very compact design, a precise distribution was not possible at face value, given the overhead space available. The available vertical distance to install the distributor was about 600 mm from the centreline of the feed nozzle to the top of the packing. The preliminary assessment showed that about twice this distance was required to allow spac- ing for pre-distribution feed piping, the main parting box, and the secondary distribution trough, but the overhead space did not allow for this increase. An additional step to complement the solution was pro- posed to the operator, consisting of replacing the packing

Revamps 2025

www.digitalrefining.com

Comparison of the stripper configurations before and after the revamp

Ifpexol stripper (1524 mm ID)

Water wash stripper (1524 mm ID)

Original

Modified

Original

Modified

Packing type

Conventional

Koch-Glitsch’s high-capacity

Conventional

Koch-Glitsch’s high-capacity

Bed h eight (mm)

6,000

5,500

6,000

5,500

Dis tributor type

Tubular

156 trough distributor

Tubular

166 low flow distributor

Flow mechanism

Pressurised

Gravitational

Pressurised

Gravitational

Metha nol consumption (normalised basis) Freq uency of distributor blowbacks

0.48

Weekly

Eliminated

Monthly

Eliminated

Table 1

with Koch-Glitsch’s proprietary high-capacity wire gauze packing and decreasing the bed height by about 500 mm. Although counterintuitive (as removing packing in a low-ef- ficiency scenario yields fewer chances for the phases to interact), Koch-Glitsch was confident that the combination of packing replacement and the finely tuned liquid distribu - tion would offset the decrease in bed height. When considering the design of a trough distributor, each transition of the liquid flowing downwards from the feed pipe to the first packing layer must be thoroughly assessed. In the case of Koch-Glitsch trough liquid distrib - utors, the liquid is received by a pre-distribution channel, which equalises and disperses liquid from the feed pipe into a parting box located underneath. Adequately sized holes along the parting box proportionally distribute the liquid into each trough according to their respective sizes. Depending on the operational range required and the expected distribution quality throughout the design range, a few different design configurations can be provided for the final liquid delivery mechanism after the troughs under the parting box. In this case, orifices on the side of the

troughs facing towards a splash baffle (Model 156) were chosen for the Ifpexol stripper. For the WW stripper, given the lower liquid flux expected at turndown conditions, the analysis showed that one additional stage of equalisation was required. Therefore, the perforations on the troughs were guided towards sec- ondary troughs with narrower dimensions (Model 166) that would allow for a healthier and more equalised liquid head throughout the operational range. Figure 2 shows a com- parison between the two designs of troughs utilised in lieu of following the original design approach of using identical liquid distributors. Results Table 1 shows a comparison between the original and modified design for the strippers. The results after the mod - ifications were markedly positive, reducing the methanol make-up by more than 50%, finally achieving the targets for methanol usage as per the licensor’s design values, which is one of the most significant advantages of the pro - cess (see Figure 3 ).

Before modications

After modications

Days

Figure 3 Methanol-to-gas feed ratio before and after the modifications

www.digitalrefining.com

Revamps 2025

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32 Page 33 Page 34 Page 35 Page 36 Page 37 Page 38 Page 39 Page 40 Page 41 Page 42 Page 43 Page 44 Page 45 Page 46