Catalysis 2026 Issue

FCC modelling and catalyst evaluation A compendium of FCC modelling best practices at a Marathon refinery serves as a practical guide for refinery engineers

F luid catalytic cracking (FCC) kinetic modelling is a valuable tool that enables engineers to optimise unit performance and enhance refinery profitability. The following best practices, developed over decades of expe- rience by Marathon and Ketjen FCC modellers, serves as a practical guide for engineers who are new to FCC modelling. Initial sections focus on building confidence and accuracy in model results through consistency checks, proper data han- dling, and calibration validation, concluding with an example demonstrating how adherence to these best practices ena- bles the user to perform complicated and detailed analysis, such as post-auditing a catalyst formulation change. By following this guidance, engineers can leverage FCC kinetic models to make more informed, data-driven decisions with higher confidence and greater accuracy. Heat balance best practices A primary advantage of kinetic models is their ability to rap- idly perform a multitude of interrelated calculations, includ- ing mass, heat, and hydrogen balances. In a dynamic process like FCC, insufficient or low-quality data during the initial unit setup can degrade the accuracy of subsequent calculations. One area that can be particularly challenging is the FCC unit heat balance. All heat required for operation is generated from the combustion of coke in the regenerator(s). Accurately estimating the coke is fundamental to FCC modelling suc- cess. The key inputs for this calculation are the regenerator flue gas composition and the main air blower air rate. The hydrogen content in coke provides insight into the quality and consistency of the flue gas analysis, determined from the oxygen balance around the FCC regenerator. The only independent variables necessary to calculate the hydro- gen in coke are the oxygen content of the air entering the regenerator (including supplemental oxygen injection, if present) and the flue gas carbon monoxide (CO), carbon dioxide (CO 2 ), and oxygen (O 2) concentrations. Most FCC units only have a single regenerator, and the hydrogen content of coke is typically between 6 and 8 wt%. Units with packed strippers in good operational condition tend to fall towards the lower end of the range, while strip- pers having mechanical defects, obsolete designs, or distri- bution problems generally return higher hydrogen in coke. To understand the reason stripper performance affects hydrogen in coke, one must consider the various ways coke is formed in the FCC unit riser. Approximately 50% of the Conradson carbon residue present in typical vacuum gas oil (VGO) or resid feeds converts to coke. These compounds Alan Kramer and Patrick McSorley Ketjen Bridget Cadigan Marathon Garyville Refinery

are the most hydrogen-deficient components of the feed, having hydrogen-to-coke ratios of around 4 to 5 wt%. Dehydrogenation reactions (catalysed primarily by nickel and copper), cracking, hydrogen transfer, and thermal crack- ing reactions are three additional primary coke sources, and these generate coke with similar hydrogen content. Strippable coke is the final major coke source for most FCC units. It consists of liquid or gaseous hydrocarbons that are not removed by the stripper, are entrained with the catalyst, and carried into the regenerator. Since these hydrocarbons are essentially liquid products, the hydrogen content of stripper coke is similar, in the range of 11-12%. As the stripper coke is hydrogen-rich, the overall hydrogen in coke offers direct insight into the proportion of stripper coke relative to all other coke sources. FCC units with two regenerators combust strippable coke preferentially in the first regenerator. Since it is not solidly bound to the catalyst, it is most easily evolved. Hydrogen in coke for the first regenerator is commonly between 8 and 10 wt%, and the second regenerator is around 4-5%. The overall hydrogen content in coke, found by weighting the hydrogen contents from the coke burn in each regen- erator, results in overall hydrogen in coke between 6 and 8 wt%, the same as for single regenerator units. Hydrogen in coke values that are not aligned with perceived stripper performance indicate poor-quality flue gas data. Next, model users should review the reported heat of reaction, as it offers insight into the quality of the other pri - mary heat balance inputs. As mentioned earlier, all the heat required for the FCC process is generated from the combus - tion of coke in the regenerator(s). This is calculated from the flue gas analysis and the air rate. To close the heat balance, models quantify the overall heat from combustion, then sub - tract all heat consumers. On the regenerator side, the largest consumer of energy is the heating of regenerator air up to the flue gas temperature. Just as with the heat of combustion calculation, an error in the main air blower rate affects the calculated energy to heat air. When present, the next largest heat consumer is the cat- alyst cooler duty. Boiler feed water rate measurement is a common source of error. Heat losses to the environment are not measured but are reasonably assumed to be about 2% of the total heat of combustion. The remaining heat generated is transferred by the cat - alyst to the riser. Here, the largest heat consumer is vapor - ising the feed and heating it to the riser outlet temperature. Feed rate and feed temperature are the key variables in this

Catalysis 2026