Documents disponibles écrits par cet auteur

Hammerstein modeling with structure identification for multi - input multi - output nonlinear industrial processes / Chenkun Qi in Industrial & engineering chemistry research, Vol. 50 N° 19 (Octobre 2011) 

Public ISBD [article]  in Industrial & engineering chemistry research > Vol. 50 N° 19 (Octobre 2011) . - pp. 11153-11169 Titre : Hammerstein modeling with structure identification for multi - input multi - output nonlinear industrial processes Type de document : texte imprimé Auteurs : Chenkun Qi, Auteur ; Han-Xiong Li, Auteur ; Xianchao Zhao, Auteur Année de publication : 2011 Article en page(s) : pp. 11153-11169 Note générale : Chimie industrielle Langues : Anglais (eng) Mots-clés : Modeling Résumé : Hammerstein modeling with structure identification for multi-input multi-output (MIMO) nonlinear industrial processes is investigated in this study. The structure identification of the Hammerstein model is very challenging because the model terms are vectors, and some model terms are inputs of other model terms (i.e., model term coupling). An efficient model structure selection algorithm for the Hammerstein model is proposed with the multi-output locally regularized orthogonal least-squares (LROLS), A-optimality design, and a vector model term selection. To enhance the well-posedness of the regressors, estimation robustness, and model adequacy, the A-optimality criterion is integrated into the model error reduction criterion in the multi-output LROLS. To handle the vector model term coupling problem, a vector model term selection rule is synthesized into the multi-output LROLS. After the model structure is determined, to improve the robustness of the parameter estimation, the regularized least-squares method with the singular value decomposition (RLS-SVD) is used. The simple or sparse Hammerstein model structure can be determined from the noisy process data. The structure identification algorithm only includes a few user-designed parameters which are easy to select. Therefore, the ability of automatic construction of the Hammerstein model is enhanced. Three application examples are used to illustrate the effectiveness of the proposed modeling approach, including the simple model structure, the satisfactory modeling accuracy, the robustness of the algorithm to the noise, and the easy selection of user-designed parameters. DEWEY : 660 ISSN : 0888-5885 En ligne : http://cat.inist.fr/?aModele=afficheN&cpsidt=24573314 [article] Hammerstein modeling with structure identification for multi - input multi - output nonlinear industrial processes [texte imprimé] / Chenkun Qi, Auteur ; Han-Xiong Li, Auteur ; Xianchao Zhao, Auteur . - 2011 . - pp. 11153-11169. Chimie industrielle Langues : Anglais (eng) in Industrial & engineering chemistry research > Vol. 50 N° 19 (Octobre 2011) . - pp. 11153-11169 Mots-clés : Modeling Résumé : Hammerstein modeling with structure identification for multi-input multi-output (MIMO) nonlinear industrial processes is investigated in this study. The structure identification of the Hammerstein model is very challenging because the model terms are vectors, and some model terms are inputs of other model terms (i.e., model term coupling). An efficient model structure selection algorithm for the Hammerstein model is proposed with the multi-output locally regularized orthogonal least-squares (LROLS), A-optimality design, and a vector model term selection. To enhance the well-posedness of the regressors, estimation robustness, and model adequacy, the A-optimality criterion is integrated into the model error reduction criterion in the multi-output LROLS. To handle the vector model term coupling problem, a vector model term selection rule is synthesized into the multi-output LROLS. After the model structure is determined, to improve the robustness of the parameter estimation, the regularized least-squares method with the singular value decomposition (RLS-SVD) is used. The simple or sparse Hammerstein model structure can be determined from the noisy process data. The structure identification algorithm only includes a few user-designed parameters which are easy to select. Therefore, the ability of automatic construction of the Hammerstein model is enhanced. Three application examples are used to illustrate the effectiveness of the proposed modeling approach, including the simple model structure, the satisfactory modeling accuracy, the robustness of the algorithm to the noise, and the easy selection of user-designed parameters. DEWEY : 660 ISSN : 0888-5885 En ligne : http://cat.inist.fr/?aModele=afficheN&cpsidt=24573314

Incremental modeling of nonlinear distributed parameter processes via spatiotemporal kernel series expansion / Han-Xiong Li in Industrial & engineering chemistry research, Vol. 48 N° 6 (Mars 2009) 

Public ISBD [article]  in Industrial & engineering chemistry research > Vol. 48 N° 6 (Mars 2009) . - pp. 3052–3058 Titre : Incremental modeling of nonlinear distributed parameter processes via spatiotemporal kernel series expansion Type de document : texte imprimé Auteurs : Han-Xiong Li, Auteur ; Chenkun Qi, Auteur Année de publication : 2009 Article en page(s) : pp. 3052–3058 Note générale : Chemical engineering Langues : Anglais (eng) Mots-clés : Incremental  modeling  Nonlinear  distributed  parameter  systems  Spatiotemporal  Volterra  kernels  Time-space  separation Résumé : In this article, an incremental modeling approach is proposed to model nonlinear distributed parameter systems, with the help of the newly constructed spatiotemporal Volterra kernels. The complex spatiotemporal process is first decomposed into a series of spatiotemporal kernels, upon which the time−space separation can be further conducted with the spatial Karhunen−Loève and temporal Laguerre basis function expansions. These two decompositions can gradually separate the nonlinear time/space coupled dynamics. Finally, the kernels in the spatiotemporal model are estimated from the experimental data incrementally, which can easily achieve satisfactory modeling performance. Simulations of two transport−reaction processes demonstrate the effectiveness of the proposed modeling approach. En ligne : http://pubs.acs.org/doi/abs/10.1021/ie801184a [article] Incremental modeling of nonlinear distributed parameter processes via spatiotemporal kernel series expansion [texte imprimé] / Han-Xiong Li, Auteur ; Chenkun Qi, Auteur . - 2009 . - pp. 3052–3058. Chemical engineering Langues : Anglais (eng) in Industrial & engineering chemistry research > Vol. 48 N° 6 (Mars 2009) . - pp. 3052–3058 Mots-clés : Incremental  modeling  Nonlinear  distributed  parameter  systems  Spatiotemporal  Volterra  kernels  Time-space  separation Résumé : In this article, an incremental modeling approach is proposed to model nonlinear distributed parameter systems, with the help of the newly constructed spatiotemporal Volterra kernels. The complex spatiotemporal process is first decomposed into a series of spatiotemporal kernels, upon which the time−space separation can be further conducted with the spatial Karhunen−Loève and temporal Laguerre basis function expansions. These two decompositions can gradually separate the nonlinear time/space coupled dynamics. Finally, the kernels in the spatiotemporal model are estimated from the experimental data incrementally, which can easily achieve satisfactory modeling performance. Simulations of two transport−reaction processes demonstrate the effectiveness of the proposed modeling approach. En ligne : http://pubs.acs.org/doi/abs/10.1021/ie801184a

A Karhunen-Loève decomposition-based Wiener modeling approach for nonlinear distributed parameter processes / Chenkun Qi in Industrial & engineering chemistry research, Vol. 47 n°12 (Juin 2008) 

Public ISBD [article]  in Industrial & engineering chemistry research > Vol. 47 n°12 (Juin 2008) . - p. 4184–4192 Titre : A Karhunen-Loève decomposition-based Wiener modeling approach for nonlinear distributed parameter processes Type de document : texte imprimé Auteurs : Chenkun Qi, Auteur ; Han-Xiong Li, Auteur Année de publication : 2008 Article en page(s) : p. 4184–4192 Note générale : Bibliogr. p. 4191-4192 Langues : Anglais (eng) Mots-clés : Spatio-temporal  modeling  problem;  Wiener  modeling;  Karhunen−Loève  decomposition Résumé : The spatio-temporal modeling problem from the input and output measurements for distributed parameter processes under unknown circumstances is investigated. The traditional Wiener modeling is extended to nonlinear distributed parameter systems with the help of the Karhunen−Loève (KL) decomposition. The input is a finite-dimensional temporal variable, whereas the spatio-temporal output of the system is measured at a finite number of spatial locations. First, the measured output is used to construct a finite dimensional approximation of the system output which is expanded in terms of KL spatial basis functions. Subsequently, the temporal coefficients are used to identify a Wiener model. The identification algorithm is based on the least-squares estimation and the instrumental variables method. The simulations for parabolic and hyperbolic systems are presented to show the effectiveness of this spatio-temporal modeling method. En ligne : http://pubs.acs.org/doi/abs/10.1021/ie0710869 [article] A Karhunen-Loève decomposition-based Wiener modeling approach for nonlinear distributed parameter processes [texte imprimé] / Chenkun Qi, Auteur ; Han-Xiong Li, Auteur . - 2008 . - p. 4184–4192. Bibliogr. p. 4191-4192 Langues : Anglais (eng) in Industrial & engineering chemistry research > Vol. 47 n°12 (Juin 2008) . - p. 4184–4192 Mots-clés : Spatio-temporal  modeling  problem;  Wiener  modeling;  Karhunen−Loève  decomposition Résumé : The spatio-temporal modeling problem from the input and output measurements for distributed parameter processes under unknown circumstances is investigated. The traditional Wiener modeling is extended to nonlinear distributed parameter systems with the help of the Karhunen−Loève (KL) decomposition. The input is a finite-dimensional temporal variable, whereas the spatio-temporal output of the system is measured at a finite number of spatial locations. First, the measured output is used to construct a finite dimensional approximation of the system output which is expanded in terms of KL spatial basis functions. Subsequently, the temporal coefficients are used to identify a Wiener model. The identification algorithm is based on the least-squares estimation and the instrumental variables method. The simulations for parabolic and hyperbolic systems are presented to show the effectiveness of this spatio-temporal modeling method. En ligne : http://pubs.acs.org/doi/abs/10.1021/ie0710869

Kernel-Based Spatiotemporal Multimodeling for Nonlinear Distributed Parameter Industrial Processes / Chenkun Qi in Industrial & engineering chemistry research, Vol. 51 N° 40 (Octobre 2012) 

Public ISBD [article]  in Industrial & engineering chemistry research > Vol. 51 N° 40 (Octobre 2012) . - pp. 13205–13218 Titre : Kernel-Based Spatiotemporal Multimodeling for Nonlinear Distributed Parameter Industrial Processes Type de document : texte imprimé Auteurs : Chenkun Qi, Auteur ; Han-Xiong, Li, Auteur ; Shaoyuan Li, Auteur Année de publication : 2012 Article en page(s) : pp. 13205–13218 Note générale : Industrial chemistry Langues : Anglais (eng) Mots-clés : Industrial  processes Résumé : Many industrial processes are nonlinear distributed parameter systems (DPS) that have significant spatiotemporal dynamics. Due to different production and working conditions, they often need to work at a large operating range with multiple working points. However, direct global modeling and persistently exciting experiment in a large working region are very costly in many cases. The complex spatiotemporal coupling and infinite-dimensional nature make the problem more difficult. In this study, a kernel-based spatiotemporal multimodeling approach is proposed for the nonlinear DPS with multiple working points. To obtain a reasonable operating space division, an iterative approach is proposed where the operating space division and local modeling are performed iteratively. The working range of the current local model will help to determine the next operating point required for modeling. Utilizing the potential of each local model, the number of regions can be reduced. In the local modeling, the Karhunen–Loève method is used for the space/time separation and dimension reduction, and after that unknown parameters of kernels are estimated. Due to consideration of time-scale properties in the dimension reduction, the modeling approach is particularly suitable for dissipative PDEs, particularly of parabolic type. The multimodeling and space/time separation techniques can largely reduce the complexity of global nonlinear spatiotemporal modeling. Finally, to guarantee a smooth transition between local spatiotemporal models, a scheduling integration method is used to provide a global spatiotemporal model. To design scheduling functions, a two-stage training method is proposed to reduce the design complexity. Compared with direct global modeling, the exciting experiment and modeling for each local region become easier. Compared with one local modeling, the multimodel integration will improve modeling accuracy. The effectiveness of the proposed modeling approach is verified by simulations. ISSN : 0888-5885 En ligne : http://pubs.acs.org/doi/abs/10.1021/ie301593u [article] Kernel-Based Spatiotemporal Multimodeling for Nonlinear Distributed Parameter Industrial Processes [texte imprimé] / Chenkun Qi, Auteur ; Han-Xiong, Li, Auteur ; Shaoyuan Li, Auteur . - 2012 . - pp. 13205–13218. Industrial chemistry Langues : Anglais (eng) in Industrial & engineering chemistry research > Vol. 51 N° 40 (Octobre 2012) . - pp. 13205–13218 Mots-clés : Industrial  processes Résumé : Many industrial processes are nonlinear distributed parameter systems (DPS) that have significant spatiotemporal dynamics. Due to different production and working conditions, they often need to work at a large operating range with multiple working points. However, direct global modeling and persistently exciting experiment in a large working region are very costly in many cases. The complex spatiotemporal coupling and infinite-dimensional nature make the problem more difficult. In this study, a kernel-based spatiotemporal multimodeling approach is proposed for the nonlinear DPS with multiple working points. To obtain a reasonable operating space division, an iterative approach is proposed where the operating space division and local modeling are performed iteratively. The working range of the current local model will help to determine the next operating point required for modeling. Utilizing the potential of each local model, the number of regions can be reduced. In the local modeling, the Karhunen–Loève method is used for the space/time separation and dimension reduction, and after that unknown parameters of kernels are estimated. Due to consideration of time-scale properties in the dimension reduction, the modeling approach is particularly suitable for dissipative PDEs, particularly of parabolic type. The multimodeling and space/time separation techniques can largely reduce the complexity of global nonlinear spatiotemporal modeling. Finally, to guarantee a smooth transition between local spatiotemporal models, a scheduling integration method is used to provide a global spatiotemporal model. To design scheduling functions, a two-stage training method is proposed to reduce the design complexity. Compared with direct global modeling, the exciting experiment and modeling for each local region become easier. Compared with one local modeling, the multimodel integration will improve modeling accuracy. The effectiveness of the proposed modeling approach is verified by simulations. ISSN : 0888-5885 En ligne : http://pubs.acs.org/doi/abs/10.1021/ie301593u

Probabilistic PCA - based spatiotemporal multimodeling for nonlinear distributed parameter processes / Chenkun Qi in Industrial & engineering chemistry research, Vol. 51 N° 19 (Mai 2012) 

Public ISBD [article]  in Industrial & engineering chemistry research > Vol. 51 N° 19 (Mai 2012) . - pp. 6811–6822 Titre : Probabilistic PCA - based spatiotemporal multimodeling for nonlinear distributed parameter processes Type de document : texte imprimé Auteurs : Chenkun Qi, Auteur ; Han-Xiong, Li, Auteur Année de publication : 2012 Article en page(s) : pp. 6811–6822 Note générale : Industrial chemistry Langues : Anglais (eng) Mots-clés : Probabilistic Résumé : Many industrial processes are nonlinear distributed parameter systems (DPSs). Data-based spatiotemporal modeling is required for analysis and control when the first-principles model is unknown. Because a DPS is infinite-dimensional and time–space coupled, a low-order model is necessary for prediction and control in practice. For low-order modeling, traditional principal component analysis (PCA) is often used for dimension reduction and time–space separation. However, it is a linear method and leads to only one set of fixed spatial basis functions. Therefore, it might not be always effective for nonlinear systems. In this study, a spatiotemporal multimodeling approach is proposed for unknown nonlinear DPSs. First, multimodel decomposition is performed, where probabilistic PCA (PPCA) is used to obtain multiple sets of spatial basis functions from the experimental data by maximizing a likelihood function. Using these multiple sets of PCA spatial bases for time–space separation, the high-dimensionality spatiotemporal data can be reduced to multiple sets of low-dimensionality temporal series. Then, multiple low-order neural models can be easily established to model these local dynamics. Finally, the original spatiotemporal dynamics can be reconstructed by multimodel synthesis. Because the proposed spatiotemporal modeling approach involves a multimodeling mechanism, it can achieve better performance than the traditional PCA-based single-modeling for nonlinear DPSs, which is demonstrated by numerical simulations. ISSN : 0888-5885 En ligne : http://pubs.acs.org/doi/abs/10.1021/ie202613t [article] Probabilistic PCA - based spatiotemporal multimodeling for nonlinear distributed parameter processes [texte imprimé] / Chenkun Qi, Auteur ; Han-Xiong, Li, Auteur . - 2012 . - pp. 6811–6822. Industrial chemistry Langues : Anglais (eng) in Industrial & engineering chemistry research > Vol. 51 N° 19 (Mai 2012) . - pp. 6811–6822 Mots-clés : Probabilistic Résumé : Many industrial processes are nonlinear distributed parameter systems (DPSs). Data-based spatiotemporal modeling is required for analysis and control when the first-principles model is unknown. Because a DPS is infinite-dimensional and time–space coupled, a low-order model is necessary for prediction and control in practice. For low-order modeling, traditional principal component analysis (PCA) is often used for dimension reduction and time–space separation. However, it is a linear method and leads to only one set of fixed spatial basis functions. Therefore, it might not be always effective for nonlinear systems. In this study, a spatiotemporal multimodeling approach is proposed for unknown nonlinear DPSs. First, multimodel decomposition is performed, where probabilistic PCA (PPCA) is used to obtain multiple sets of spatial basis functions from the experimental data by maximizing a likelihood function. Using these multiple sets of PCA spatial bases for time–space separation, the high-dimensionality spatiotemporal data can be reduced to multiple sets of low-dimensionality temporal series. Then, multiple low-order neural models can be easily established to model these local dynamics. Finally, the original spatiotemporal dynamics can be reconstructed by multimodel synthesis. Because the proposed spatiotemporal modeling approach involves a multimodeling mechanism, it can achieve better performance than the traditional PCA-based single-modeling for nonlinear DPSs, which is demonstrated by numerical simulations. ISSN : 0888-5885 En ligne : http://pubs.acs.org/doi/abs/10.1021/ie202613t