PTQ Q3 2023 Issue

that can form a cake layer on top of catalyst inside the reac- tor. When that happens, the pressure inside the reactor will rise suddenly, leading to serious problems. However, the proprietary magnetically induced Universal Filter filtration system can remove all types of solid particles down to 7 nm or less with much simpler operation. This filtration system achieves a near-total prevention of solid particles in the liquid stream from entering the reactor, including FeO, FeS, Fe₂O₃, Ni, NiO, Co, and CoO. Furthermore, the need for expendable macropore filtration packings on top of the reactor and filter cartridges at the reactor feed entry is substantially eliminated, saving mate- rial and operational costs. These costs include loading/ unloading and disposal of the spendable materials. More detailed information is available in a previously published article in PTQ Q1 2023 (p87-95), ‘Universal filter for ultra- cleaning of reactor streams’. Q What heat exchanger fouling prediction frameworks do you see benefiting refinery and petrochemical operations? A Debjit Chandra, Manager, Global Technical Services (Refinery), debjit.chandra@dorfketal.com, and Ajay Kumar Gupta, General Manager Global Technical Services, ajay. gupta@dorfketal.com, Dorf Ketal Chemicals Fouling prediction framework for heat exchangers in refin - ery and petrochemical plants can be best achieved with the integration of analytical, statistical, and AI-based predictive algorithm approaches:  Analytical approaches can include the primary foul- ing precursors in the feed stream to the exchanger and deposit sample analysis. Screening the correct antifoulant and adjusting the chemical dosage based on key feed qual- ity variables and fouling precursors can significantly help in fouling management. v For fouling prediction on crude, changes in current feed with respect to baseline saturates, asphaltene, resin, and aromatics content/ratio can help gauge the potential fouling rate. Dorf Ketal uses the proprietary ‘Oil compatibility model (OCM)’ software to predict the stability of the crude blends. w For relative cleaning effectiveness, baseline fouling rate data analysis using good simulation software or statistical tools such as multiple variable regression analysis of previ- ous runs can help to understand the performance deviation. x Analysis of actual heat transfer rate, fouling factor and overall heat transfer coefficient (U d ), viscosity, and fluid velocity within the exchangers can help in understanding the fouling control before and after cleaning. y AI-based models can help identify key operating vari- ables, which are mainly impacting on fouling rate/run length of the exchanger performance. Modulating these variables based on AI model output has helped refiners to maximise the heat transfer rate/yield improvement. A Tina Owodunni, Senior Staff Consultant, tina.owo- dunni@kbc.global, Simon Calverley, Head of Coaching and Capability Development, KBC Heat exchanger fouling can significantly limit operations. It reduces the rate of heat transfer in the preheat train, which

increases furnace duty and sometimes reduces through- put to meet desired product purity. As a result, it increases operating costs and, most likely, increases CO₂ emissions. Additionally, as the fouling layer reduces the tube inlet diameter and the tube pitch, pressure drop increases and can cause bottlenecks in the feed pump. More severe foul- ing can force the plant to shut down. Therefore, any refinery or petrochemical plant needs an effective tool to manage fouling. While fouling is not measured, it can be calculated from kinetics, such as that introduced by Ebert and Panchal in 1997. However, the activation energy and other parameters must be determined for each crude type, which may be dif- ficult to obtain. Moreover, fouling kinetics is mostly applied to chemical reaction fouling, such as coking. It cannot be used to determine deposition fouling, so kinetic modelling is unlikely to give the complete answer. The most practical way of calculating fouling involves comparing experimental data to rigorous exchanger modelling. Rigorous models are available to understand not only the performance of shell and tube exchangers, commonly used in refineries and petrochemical complexes but also other exchangers such as plate heat exchangers. These models calculate the exchanger’s overall heat transfer coefficient when the exchanger is clean (U clean ), and performance is optimised to predict the outlet temperatures. The operat- ing heat transfer coefficient of the exchanger (U actual ) can be calculated from the plant data (such as inlet and outlet temperatures, flow rates, and composition). For a fouling exchanger, U actual will be lower, so its difference, when com- pared to U clean , determines the exchanger’s fouling factor. The complexity of these calculations is compounded because plant data often contain errors. Before calculating fouling factors, data reconciliation covering the whole heat exchanger train is needed to fit the measured data to the geometry model to find and correct errors. As a result, foul - ing calculations become convoluted. Several tools monitor fouling to manage these complexities, such as KBC’s HX Monitor, HTRI’s SmartPM, to name a few. Using a rigorous data reconciliation package, HX Monitor calculates fouling factors, runs cleaning cases, and estimates the benefits of cleaning exchanges. Then, fouling trends calculated for multiple datasets can be used to predict future fouling. A Veerala Hari Krishna, Senior Data Scientist, L & T Technology Services Ltd, veeralahari.krishna@ltts.com In refinery and petrochemical operations, one of the major problems is heat exchanger fouling, causing loss of revenue in terms of equipment replacement cost, maintenance, and cleaning expenses. Moreover, fouling is responsible for productivity losses. Heat exchanger fouling also leads to a reduction of input feed rates to the plant. Heat transfer resistance increases due to increased fouling thickness, which continuously lowers the heat exchanger operating heat duty. This predicates the need for monitoring heat exchanger fouling and forecast cleaning. Heat exchanger operations are monitored only for the rate of heat transfer. Furthermore, the heat transfer rate can be controlled by altering flow rates. An experienced process

PTQ Q3 2023