Petroleum ChemistryPetroleum ChemistryНефтехимия0028-24213034-5626The Russian Academy of Sciences65563710.31857/S0028242123010069UHEEPHArticlesСтатьиResearch ArticleAn Industrial Data-Based Model to Reduce Octane Number Loss of Refined Gasoline for S Zorb ProcessПромышленная компьютерная модель снижения потерь октанового числа очищенного бензина в процессе S ZorbBoChenpetrochem@ips.ac.ruJieWangpetrochem@ips.ac.ruSongLiupetrochem@ips.ac.ruFushengOuyangouyfsh@ecust.edu.cnDaXiongpetrochem@ips.ac.ruMingyangZhaopetrochem@ips.ac.ruInternational Joint Research Center of Green Energy Chemical Engineering, East China University of Science and TechnologySINOPEC Shanghai Gaoqiao Petrochemical Co., LtdSINOPEC Shanghai Gaoqiao Petrochemical Co15022023631NO1 (2023)№1 (2023)677911022025Copyright ©; 2023, Russian Academy of SciencesCopyright ©; 2023, Российская академия наук2023Russian Academy of SciencesРоссийская академия наукhttps://journals.eco-vector.com/0028-2421/article/view/655637

S Zorb process is one of the main technologies for deep desulfurization of gasoline from fluid catalytic cracking (FCC) process, which by the process will also cause some research octane number (RON) loss of gasoline. Establishing a data-driven model with data mining technologies to optimize production is one of the development directions in petrochemical field. Based on the industrial data from a 1.20 Mt/a S Zorb unit in China in recent three years, 422 modeling samples and 22 modeling variables were screened out and then three data-driven models were established by back propagation neural network (BPNN), radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) to predict RON of refined gasoline (r-RON). The results show that the BPNN model has the best prediction effect and generalization ability. Genetic algorithm (GA), particle swarm optimization algorithm (PSO) and simulated annealing algorithm (SA) in combination with the BPNN model respectively were used to optimize the operation variables to reduce the r-RON loss. The results indicate that the optimized performance of PSO-BPNN model is best because of its largest reduction in r-RON loss at 48.55%. The validity of the PSO-BPNN model was verified in the S Zorb unit and the research methods to establish a data-driven model for reducing r-RON loss are also worthy of reference for other S Zorb units.

Метод реактивной адсорбционной десульфуризации S Zorb - одна из основных технологий удаления серы из бензина в процессе жидкостного каталитического крекинга (FCC) на установках Китая, сопряженная, однако, с некоторым снижением октанового числа получаемого бензина (ОЧИ, RON). Для оптимизации рабочих переменных и уменьшения потерь прогнозированного октанового числа бензина (r-RON) были созданы три компьютерно-управляемые модели нейронной сети: с обратной передачей ошибки обучения (BPNN); с радиальной базисной функцией (RBFNN); с обобщенной регрессией (GRNN). Показано, что наилучшим является эффект модели с алгоритмом оптимизации роя частиц PSO-BPNN, обеспечивающей наибольшее снижение потерь r-RON на 48.55%. Методы исследования, использованные для создания компьютерно-управляемой модели снижения потерь r-RON, заслуживают применения на других блоках установки S Zorb.

FCC gasolineRON lossdata-driven modelneural networkoptimization algorithmбензин жидкостного каталитического крекинга FCCпотери ОЧИ (RON)компьютерно-управляемая модельнейронная сетьалгоритм оптимизацииQiu L.M., Xiang Y.J., Xin M.D., Zou K., Zheng A.G., Xu G.T. Structural verification of nickel sulfide on spent S Zorb adsorbent as studied by HRTEM and XPS // J. Mol. Struct. 2020. V. 1202. P. 127215-127215. https://doi.org/10.1016/j.molstruc.2019.127215Ribeiro E Sousa L.R., Miranda T., e Sousa R.L., Tinoco J. The use of data mining techniques in rockburst risk assessment // Engineering. 2017. V. 3. P. 552-558. https://doi.org/10.1016/J.ENG.2017.04.002Ouyang F.S., Zhang J.H., Fang W.G. Optimizing product distribution in the heavy oil catalytic cracking (MIP) process // Petrol. Sci. Technol. 2017. V. 35. P. 1315-1320. https://doi.org/10.1080/10916466.2017.1297826Sadighi S., Mohaddecy R.S., Norouzian A. Optimizing an industrial scale naphtha catalytic reforming plant using a hybrid artificial neural network and genetic algorithm technique // Bull. Chem. React. Eng. Catal. 2015. V. 10. P. 210-220. https://doi.org/10.9767/bcrec.10.2.7171.210-220Zhu W.B., Webb Z.T., Mao K., Romagnoli J. A deep learning approach for process data visualization using t-distributed stochastic neighbor embedding // Ind. Eng. Chem. Res. 2019. V. 58. P. 9564-9575. https://doi.org/10.1021/acs.iecr.9b00975Chang P., Li Z.Y., Wang G.M., Wang P. An effective deep recurrent network with high-order statistic information for fault monitoring in wastewater treatment process // Expert Syst. Appl. 2020. V. 167. P. 114141. https://doi.org/10.1016/j.eswa.2020.114141Martínez-Martínez J.M., Escandell-Montero P., Soria-Olivas E., Martín-Guerrero J.D., Serrano-López A.J. A new visualization tool for data mining techniques // Prog. Artif. Intell. 2016. V. 5. P. 137-154. https://doi.org/10.1007/s13748-015-0079-4Chang Z.H., Zhang Y., Chen W.B. Electricity price prediction based on hybrid model of adam optimized LSTM neural network and wavelet transform // Energy. 2019. V. 187. P. 115804. https://doi.org/10.1016/j.energy.2019.07.134Luor D.C. A comparative assessment of data standardization on support vector machine for classification problems // Intell. Data Anal. 2015. V. 19. P. 529-546. https://doi.org/10.3233/IDA-150730Duan H.M., Pang X.Y. A multivariate grey prediction model based on energy logistic equation and its application in energy prediction in China // Energy. 2021. V. 229. P. 120716. https://doi.org/10.1016/J.ENERGY.2021.120716Fang S.E., Chen S. Model-free damage localization of structures using wavelet based grey relational analysis // Smart Mater. Struct. 2020. V. 29. № 8. P. 085046. https://doi.org/10.1088/1361-665X/ab99daCai Y.G., Xi M.C., Xue Q.H. Study on the applications of neural networks for processing deformation monitoring data // Appl. Mech. Mater. 2014. V. 501-504. P. 2149-2153. https://doi.org/10.4028/www.scientific.net/AMM.501-504.2149Zhang E.L., Hou L., Shen C., Shi Y.L., Zhang Y.X. Sound quality prediction of vehicle interior noise and mathematical modeling using a back propagation neural network (BPNN) based on particle swarm optimization (PSO) // Meas. Sci. Technol. 2016. V. 27. P. 015801. https://doi.org/10.1088/0957-0233/27/1/015801Liu Xm., Liu Jc., Xu Yr. Motion control of underwater vehicle based on least disturbance BP algorithm // J. Marine. Sci. Appl. 2002. V. 1. P. 16-20. https://doi.org/10.1007/BF02921411Potts M.A.S., Broomhead D.S. Time series prediction with a radial basis function neural network // Proc. SPIE. 1991. V. 1565. P. 255-266. https://doi.org/10.1117/12.49782Zhao Y.P., Zhou X.L. K-means clustering algorithm and its improvement research // J. Phys.: Conf. Ser. 2021. V. 1873. P. 012074. https://doi.org/10.1088/1742-6596/1873/1/012074Yousef W.A. Estimating the standard error of cross-validation-based estimators of classifier performance // Pattern Recognit. Lett. 2021. V. 146. P. 115-125. https://doi.org/10.1016/J.PATREC.2021.02.022Liang F., Gao J., Xu L. Thermal performance investigation of the miniature revolving heat pipes using artificial neural networks and genetic algorithms // Int. J. Heat Mass Transf. 2020. V. 151. P. 119394. https://doi.org/10.1016/j.ijheatmasstransfer.2020.119394Ying J.L., Xiao J.C. Simulated annealing algorithm improved BP learning algorithm // Appl. Mech. Mater. 2014. V. 513-517. P. 734-737. https://doi.org/10.4028/www.scientific.net/AMM.513-517.734Wang H.L., Hu Z.B., Sun Y.Q., Su Q.H., Xia X.W. Modified backtracking search optimization algorithm inspired by simulated annealing for constrained engineering optimization problems // Comput. Intell. Neurosci. 2018. V. 2018. article ID 9167414. 27 pp. https://doi.org/10.1155/2018/9167414