Determining the amount of ‘green’ coke generated when co‐processing lipids commercially by fluid catalytic cracking

Abstract

Co-processing biogenic feedstocks in oil refineries will reduce the greenhouse gas emissions normally associated with fossil-derived transportation fuels. The fluid catalytic cracker (FCC) within a refinery is a robust processing unit and will probably be a preferred insertion point if biocrudes, produced by the liquefaction of biomass, are co-processed within a refinery. Fluid catalytic cracking results in a wide range of intermediate products which can be upgraded to gasoline, diesel, heavy fuel oil and liquified petroleum gas blendstocks. Coke is also produced and provides heating for feedstocks, the endothermic catalytic cracking reactions and the regeneration of the FCC catalyst. However, coke combustion also generates carbon dioxide and is a significant source of refinery greenhouse gas emissions. As detailed here, the continuous nature of the process makes the physical evaluation of any biogenic coke fraction, via methods such as C14 isotope analysis, quite challenging. However, quantifying the stack gases provides one way of assessing the renewable content of the carbon dioxide derived from coke combustion. The hourly data from 1 year of commercial operation was assessed using linear and Bayesian ridge regression to quantify the burning coefficient of the coke when co-processing lipids at the FCC.

Tracking the green coke production when co-processing lipids at a commercial fluid catalytic cracker (FCC): combining isotope 14C and causal discovery analysis

“…Energy & Fuels PAPER supplement 14 C monitoring with the hope that "so sensors" can be established at various points within a renery so data can be collected and validated. 16,17 As reported previously, 18 commercial renery operation data combined with multiple linear regression and a bootstrap method was successfully used to establish a "so sensor" to estimate the amount of green coke generated during coprocessing. These values were further assessed by comparing the ratio between the biogenic and fossil feedstock ows.…”

“…However, 14 C analysis is the only method certied by ASTM and it is currently the method used by reners who are co-processing biogenic feedstocks to quantify the carbon intensities of their processes and fuels, with the driver that they generate credits under British Columbia's and California's low carbon fuel standard (LCFS). 20 2.1 FCC co-processing and data retrieval using Seeq As oleochemical/lipid feedstocks are being co-processed by Parkland at their renery in Burnaby, British Columbia, Canada, 14,18 hourly data has been continuously generated over the last 17 months (as compared to the 12 months of data reported in previous work). 18 With the accumulation of more data, the work reported here also assessed whether the coefficients that were used previously to calculate the amount of green molecules had changed in any way.…”

“…20 2.1 FCC co-processing and data retrieval using Seeq As oleochemical/lipid feedstocks are being co-processed by Parkland at their renery in Burnaby, British Columbia, Canada, 14,18 hourly data has been continuously generated over the last 17 months (as compared to the 12 months of data reported in previous work). 18 With the accumulation of more data, the work reported here also assessed whether the coefficients that were used previously to calculate the amount of green molecules had changed in any way. The Seeq datalab was used to connect the renery's process data with open access python library (70% for training and 30% for testing).…”

“…As described earlier, 18,22 the selection of parameters that impacted the amount of coke produced and burned were based on process knowledge and the FCC heat balance, which is at the core of the catalytic cracking reactions. For example, just enough coke is burnt to satisfy the heat demand for the reactions such as heating the feedstocks, providing heat for the endothermic catalytic cracking reactions, etc.…”

“…As reported previously, 18 commercial refinery operation data combined with multiple linear regression and a bootstrap method was successfully used to establish a “soft sensor” to estimate the amount of green coke generated during co-processing. These values were further assessed by comparing the ratio between the biogenic and fossil feedstock flows.…”