Polynomial NARX-based nonlinear model predictive control of modular chemical systems

Νικολακοπούλου, Αναστασία; Braatz, Richard D.

doi:10.1016/j.compchemeng.2023.108272

Cited by 4 publications

(1 citation statement)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their work focused on designing a recurrent neural network (RNN) model to approximate the nominal nonlinear system and integrated the model within an MPC framework to stabilize the system. In [7], polynomial nonlinear autoregressive with exogenous inputs (NARX) models were used to build nonlinear MPC, with the relevant polynomial terms to retain selected via sparse regression. The proposed MPC was applied to a multi-input multi-output chemical reactor, and an algebraic modeling language was used to reduce computational time.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

Alnajdi,

Abdullah,

Suryavanshi

et al. 2023

Mathematics

View full text Add to dashboard Cite

In this study, we present a general form of nonlinear two-time-scale systems, where singular perturbation analysis is used to separate the dynamics of the slow and fast subsystems. Machine learning techniques are utilized to approximate the dynamics of both subsystems. Specifically, a recurrent neural network (RNN) and a feedforward neural network (FNN) are used to predict the slow and fast state vectors, respectively. Moreover, we investigate the generalization error bounds for these machine learning models approximating the dynamics of two-time-scale systems. Next, under the assumption that the fast states are asymptotically stable, our focus shifts toward designing a Lyapunov-based model predictive control (LMPC) scheme that exclusively employs the RNN to predict the dynamics of the slow states. Additionally, we derive sufficient conditions to guarantee the closed-loop stability of the system under the sample-and-hold implementation of the controller. A nonlinear chemical process example is used to demonstrate the theory. In particular, two RNN models are constructed: one to model the full two-time-scale system and the other to predict solely the slow state vector. Both models are integrated within the LMPC scheme, and we compare their closed-loop performance while assessing the computational time required to execute the LMPC optimization problem.

show abstract

Section: Introductionmentioning

confidence: 99%

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

Alnajdi,

Abdullah,

Suryavanshi

et al. 2023

Mathematics

View full text Add to dashboard Cite

show abstract

Flexible operation of Post-combustion CO2 capture process enabled by NARX-MPC using neural network

Jung,

Lee

2023

Computers & Chemical Engineering

View full text Add to dashboard Cite

Analysis of Dynamic Characteristics of Reconfigurable Modular Reactor Systems for Renewable Methanol Synthesis

Gao,

Kong,

Kang

et al. 2024

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

A modular production system for renewable methanol synthesis is one viable way for simultaneously realizing scalable storage of renewable energy and carbon dioxide utilization. It is crucial to improve the dynamic performances of the modular reactor systems to rapidly adapt to fluctuations of renewable energy and increase the methanol yield. In this work, four configurations and their dynamic characteristics of reactors in series and parallel were investigated to elucidate the impacts of reactor configurations on the performances of reconfigurable modular reactors, in which a 2-D transient model of fixed-bed reactors for methanol synthesis was established. The dynamic characteristics of the reactor systems during startup, shutdown, disturbance, and switching were explored. The results indicated that modular reactors in series are preferred when the feed rate is less than two times the rated flow rate in a single reactor, whereas modular reactors in parallel are favorable when the feed rate exceeds six times the rated flow rate. The modular reactors in series and parallel are much more favorable for the feed rates in between. The startup and shutdown times of the reactor systems are significantly affected by the stage numbers of the reactors and the feed flow rate in a single reactor. In general, the startup and shutdown times increase with the increase of stage number and the decrease of flow rates of the reactors. The response time to flow rate fluctuations of the reactor system is mainly determined by the stage number of the reactor and the flow rate of a single reactor. Fewer stage numbers and higher flow rates lead to a rapid response. The mode switching time between different configurations is also intimately related to the flow rates. Larger flow rates result in a relatively short residence time and a rapid switching process. These results provide insights into the optimal design and operation of modular reactor systems for renewable methanol production.

show abstract

Polynomial NARX-based nonlinear model predictive control of modular chemical systems

Cited by 4 publications

References 42 publications

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

Flexible operation of Post-combustion CO2 capture process enabled by NARX-MPC using neural network

Analysis of Dynamic Characteristics of Reconfigurable Modular Reactor Systems for Renewable Methanol Synthesis

Contact Info

Product

Resources

About