Dynamic controlled pattern extraction and pattern-based model predictive control

Zheng, Niannian; Luan, Xiaoli; Liu, Fei

doi:10.1016/j.jprocont.2021.11.010

Cited by 4 publications

(6 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…12−14 This article is licensed under CC-BY 4 Mathematically, a pattern is spanned by several latent variables that are extracted from the measured process variables. 11,12 A pattern model shows the relationships between the process variables and the pattern as well as the dynamic pattern equations. From an engineering perspective, a pattern can concentrate the most operationally relevant information from the high-dimensional process variables into a lower dimensional structure.…”

Section: Introductionmentioning

confidence: 99%

“…10,11 Consequently, it can be concluded that (i) individual variables can only describe the operation status in a decentralized, independent, and partial manner; (ii) the information from the univariate observation is incomplete and biased from the holistic-and-statistical effect of the whole set of process variables; and (iii) the overall operational behaviors of the process are in accordance with a hidden pattern. 12 Mathematically, a pattern is spanned by several latent variables that are extracted from the measured process variables. 11,12 A pattern model shows the relationships between the process variables and the pattern as well as the dynamic pattern equations.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, pattern control is proposed to control, using the manipulated variables, the pattern so that it tracks to a desired set point or trajectory. 12 From the inability of univariate monitoring to extract the desired pattern, multivariate SPM has been developed to create a large number of latent-variable modeling approaches. The classical approach is a geometric projection that projects the raw high-dimensional process variables onto a low-dimensional but informative pattern space.…”

Section: Introductionmentioning

confidence: 99%

“…Examining the geometric features of the process variables, it can be seen that the reason for this problem is that these variables are actually mutually coupled and cross-correlated in their spatial structure. , Consequently, it can be concluded that (i) individual variables can only describe the operation status in a decentralized, independent, and partial manner; (ii) the information from the univariate observation is incomplete and biased from the holistic-and-statistical effect of the whole set of process variables; and (iii) the overall operational behaviors of the process are in accordance with a hidden pattern. − …”

Section: Introductionmentioning

confidence: 99%

“…Mathematically, a pattern is spanned by several latent variables that are extracted from the measured process variables. , A pattern model shows the relationships between the process variables and the pattern as well as the dynamic pattern equations. From an engineering perspective, a pattern can concentrate the most operationally relevant information from the high-dimensional process variables into a lower dimensional structure. , From these perspectives, it can be concluded that a pattern captures the process dynamics and reflects the similarities and differences of the overall operational status of the process.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Dynamic-Controlled Bayesian Network for Process Pattern Modeling and Optimization

Zheng,

Luan,

Shardt

et al. 2024

Ind. Eng. Chem. Res.

Self Cite

View full text Add to dashboard Cite

Capturing the current statistical features of a process and its dynamic evolution is important for controlling and monitoring its overall operational status. In terms of capturing the process dynamics, existing probabilistic latent-variable methods mostly consider autoregressive relationships, and thus, the causality from the control inputs to the pattern, or key hidden variable, remains unmodeled or implicit. To bridge this gap, a model structured by a newly designed dynamic-controlled Bayesian network (DCBN) is proposed in this paper for pattern modeling, especially pattern control and optimization. Significantly, the innovation and advantage of the DCBN lie in explicitly quantifying the impulse response of the pattern under control inputs. As well, the expectation−maximization algorithm is specially designed for learning the DCBN model. Finally, a new framework for pattern-based process control and optimization is presented in which online pattern filtering and control can be implemented. A case study on the combustion process from an industrial boiler illustrates the advantages of the proposed method in that it can capture the controlled dynamics of the process and achieve optimization by tracking the pattern set point or trajectory.

show abstract

Section: Introductionmentioning

confidence: 99%