Data-driven optimal operation of the industrial methanol to olefin process based on relevance vector machine

doi:10.1016/j.cjche.2020.09.040

Abstract

Abstract: Methanol to olefin (MTO) technology provides the opportunity to produce olefins from nonpetroleum sources such as coal, biomass and natural gas. More than 20 commercial MTO plants have been put into operation. Till now, contributions on optimal operation of industrial MTO plants from a process systems engineering perspective are rare. Based on relevance vector machine (RVM), a data-driven framework for optimal operation of the industrial MTO process is established to fully utilize the plentiful industrial data sets. RVM correlates the yield distribution prediction of main products and the operation conditions. These correlations then serve as the constraints for the multi-objective optimization model to pursue the optimal operation of the plant. Nondominated sorting genetic algorithm II is used to solve the optimization problem. Comprehensive tests demonstrate that the ethylene yield is effectively improved based on the proposed framework. Since RVM does provide the distribution prediction instead of point estimation, the established model is expected to provide guidance for actual production operations under uncertainty.

Key words: Methanol to olefins, Relevance vector machine, Genetic algorithm, Operation optimization, Systems engineering, Process systems

摘要： Methanol to olefin (MTO) technology provides the opportunity to produce olefins from nonpetroleum sources such as coal, biomass and natural gas. More than 20 commercial MTO plants have been put into operation. Till now, contributions on optimal operation of industrial MTO plants from a process systems engineering perspective are rare. Based on relevance vector machine (RVM), a data-driven framework for optimal operation of the industrial MTO process is established to fully utilize the plentiful industrial data sets. RVM correlates the yield distribution prediction of main products and the operation conditions. These correlations then serve as the constraints for the multi-objective optimization model to pursue the optimal operation of the plant. Nondominated sorting genetic algorithm II is used to solve the optimization problem. Comprehensive tests demonstrate that the ethylene yield is effectively improved based on the proposed framework. Since RVM does provide the distribution prediction instead of point estimation, the established model is expected to provide guidance for actual production operations under uncertainty.

关键词: Methanol to olefins, Relevance vector machine, Genetic algorithm, Operation optimization, Systems engineering, Process systems

Zhiquan Wang, Liang Wang, Zhihong Yuan, Bingzhen Chen. Data-driven optimal operation of the industrial methanol to olefin process based on relevance vector machine[J]. Chinese Journal of Chemical Engineering, 2021, 34(6): 106-115.

Zhiquan Wang, Liang Wang, Zhihong Yuan, Bingzhen Chen. Data-driven optimal operation of the industrial methanol to olefin process based on relevance vector machine[J]. 中国化学工程学报, 2021, 34(6): 106-115.

References

[1] C.A.J. Silvestri, The conversion of methanol and other O-compounds to hydrocarbons over zeolite catalysts, J. Catal. 47(2) (1977) 249-259.
[2] P. Tian, Y. Wei, M. Ye, Z. Liu, Methanol to Olefins (MTO):From fundamentals to commercialization, ACS Catal. 5(2015) 1922-1938.
[3] A.N.R. Bos, P.J.J. Tromp, H.N. Akse, Conversion of Methanol to lower olefins. Kinetic modeling, reactor simulation, and selection, Ind. Eng. Chem. Res. 34(1995) 134-135.
[4] J.F. Haw, D.M. Marcus, Well-defined (supra)molecular structures in zeolite methanol-to-olefin catalysis, Top. Catal. 34(2005) 41-48.
[5] T.Y. Park, G.F. Froment, Kinetic modeling of the methanol to olefins process. 1. Model formulation, Ind. Eng. Chem. Res. 40(2001) 4172-4186.
[6] S. Wilson, P. Barger, The characteristics of SAPO-34 which influence the conversion of methanol to light olefins, Microporous Mesoporous Mater. 29(1999) 117-126.
[7] D. Chen, K. Moljord, T. Fuglerud, A. Holmen, The effect of crystal size of SAPO-34 on the selectivity and deactivation of the MTO reaction, Microporous Mesoporous Mater. 29(1999) 191-203.
[8] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.P. Lillerud, S. Kolboe, M. Bjorgen, Conversion of methanol into hydrocarbons over zeolite H-ZSM-5:Ethene formation is mechanistically separated from the formation of higher alkenes, J. Am. Chem. Soc. 128(2006) 14770.
[9] C.D. Chang, Methanol conversion to light olefins, Catal. Rev. 26(1984) 323-345.
[10] J. Li, Y. Wei, G. Liu, Y. Qi, P. Tian, B. Li, Y. He, Z. Liu, Comparative study of MTO conversion over SAPO-34, H-ZSM-5 and H-ZSM-22:Correlating catalytic performance and reaction mechanism to zeolite topology, Catal. Today 171(2011) 221-228.
[11] J. Tan, Z. Liu, X. Bao, X. Liu, X. Han, C. He, R. Zhai, Crystallization and Si incorporation mechanisms of SAPO-34, Microporous Mesoporous Mater. 53(2002) 97-108.
[12] G. Yang, Y. Wei, S. Xu, J. Chen, J. Li, Z. Liu, J. Yu, R. Xu, Nanosize-enhanced lifetime of SAPO-34 catalysts in methanol-to-olefin reactions, J. Phys. Chem. C 117(2013) 8214-8222.
[13] B. Lu, H. Luo, H. Li, W. Wang, M. Ye, Z. Liu, J. Li, Speeding up CFD simulation of fluidized bed reactor for MTO by coupling CRE model, Chem. Eng. Sci. 143(2016) 341-350.
[14] L.T. Zhu, M. Ye, Z.H. Luo, Application of filtered model for reacting gas-solid flows and optimization in a large-scale methanol-to-olefin fluidized bed reactor, Ind. Eng. Chem. Res. 55(2016) 11887-11899.
[15] L.T. Biegler, Advanced optimization strategies for integrated dynamic process operations, Comput. Chem. Eng. 114(2017) 3-13.
[16] M.E. Tipping, Sparse Bayesian learning and the relevance vector machine, J. Mach. Learn. Res. 1(2001) 211-244.
[17] M. Christopher, Bishop, Pattern Recognition and Machine learning, Springer, New York, 2006.
[18] C.M. Bishop, M.E. Tipping, Variational relevance vector machines, In:Proceedings of the Sixteenth Conference on Uncertainty in Artificial Intelligence, San Francisco, CA (2000).
[19] Y. Lecun, Y. Bengio, G. Hinton, Deep learning, Nature 521(2015) 436.
[20] X. Zhou, X. He, B. Chen, Method for solving computer-aided product design optimization problem based on back propagation neural network, Chinese. J. Chem. Eng. 12(2004) 510-514.
[21] L.T. Zhu, J.X. Tang, Z.H. Luo, Machine learning to assist filtered two-fluid model development for dense gas-particle flows, AIChE J. 66(2020) 16973.
[22] C. Cortes, V. Vapnik, Support-vector networks, Mach. Learn. 20(1995) 273-297.
[23] E. Shoesmith, V. Vapnik, S. Kotz, Estimation of dependences based on empirical data, Stat. 33(3) (1984) 324.
[24] K. Deb, A. Pratap, S. Agarwal, T. Meyarivan, A fast and elitist multiobjective genetic algorithm:NSGA-II, IEEE. T. Evolut. Comput 6(2002) 182-197.
[25] X. Gao, B. Chen, X. He, T. Qiu, J. Li, C. Wang, L. Zhang, Multi-objective optimization for the periodic operation of the naphtha pyrolysis process using a new parallel hybrid algorithm combining NSGA-II with SQP, Comput. Chem. Eng. 32(2008) 2801-2811.
[26] Y. Tian, R. Cheng, X.Y. Zhang, Y.C. Jin, PlatEMO:a MATLAB platform for evolutionary multi-objective optimization, IEEE Comput. Intell. M. 12(2017) 73-87.
[27] O. James Berges,[Springer Series in Statistics] Statistical Decision Theory and Bayesian Analysis||Bayesian Analysis, Springer, New York (1985).
[28] O. Chapelle, V. Vapnik, O. Bousquet, S. Mukherjee, Choosing multiple parameters for support vector machines, Mach. Learn. 46(2002) 131-159.
[29] Chih-Chung Chang, Chih-Jen Lin, LIBSVM:a library for support vector machines, ACM T. Intel. Syst. Tec. 2(3) (2011) 1-39.
[30] V. Cherkassky, Y. Ma, Practical selection of SVM parameters and noise estimation for SVM regression, Neural Networks 17(2004) 113-126.
[31] E. Zitzler, K. Deb, L. Thiele, Comparison of multiobjective evolutionary algorithms:Empirical results, Evol. Comput. 8(2000) 173-195.
[32] A. Konak, D.W. Coit, A.E. Smith, Multi-objective optimization using genetic algorithms:A tutorial, Reliab. Eng. Syst. Safe 91(2006) 992-1007.
[33] V. Vapnik, V. Vapnik, V.N. Vapnik, Statistical learning theory, Ann. I. Stat. Math. 55(2003) 371-389.
[34] G.E. Dahl, T.N. Sainath, G.E. Hinton, Improving deep neural networks for LVCSR using rectified linear units and dropout, IEEE, Vancouver, BC, Canada (2013) 8609-8613.
[35] J. Bergstra, Y. Bengio, Random search for hyper-parameter optimization, J. Mach. Learn. Res. 13(2012) 281-305.
[36] P.E. Gill, M.M.A. Saunders, SNOPT:an SQP algorithm for large-scale constrained optimization, Siam. Rev. 47(2005) 99-131.
[37] A. Wachter, L.T. Biegler, On the implementation of an interior-point filter linesearch algorithm for large-scale nonlinear programming, Math. Program. 106(2006) 25-57.
[38] J. Holland, Adaptation in Natural And Artificial Systems, University of Michigan Press, Ann Arb, Mich, USA (1975).
[39] D.A. De Jong, An Analysis of the Behavior of a Class of Genetic Adaptive Systems, Doctoral dissertation (1975).
[40] D.E. Goldberg, Genetic algorithms in search, optimization, and machine learning, Choice Rev. Online 27(2) (1989) 27-0936.
[41] A. Lipowski, D. Lipowska, Roulette-wheel selection via stochastic acceptance, Physica A:Stat. Mech. Appl. 391(6) (2012) 2193-2196.
[42] S. Kirkpatrick, C.D. Gelatt, M.P. Vecchi, Optimization by simulated annealing, Science 220(4598) (1983) 671-680.
[43] D.S. Johnson, C.R. Aragon, L.A. Mcgeoch, C. Schevon, Optimization by simulated annealing:an experimental evaluation; Part I. Graph Partitioning, Oper. Res. 37(1989) 865-892.
[44] J. Kennedy, R. Eberhart, Particle Swarm Optimization, in:Icnn95-international Conference on Neural Networks, Australia (1995).