Ensemble learning based latent variable model predictive control for batch trajectory tracking under concept drift

Jeong, Dong Hwi; Lee, Jong Min

doi:10.1016/j.compchemeng.2020.106875

Cited by 10 publications

(11 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to their nature, batch processes are primarily a dynamic type. The latter includes, for instance, modeling of chemical kinetics [12,13], modeling dynamic behavior of a cryogenic expansion unit [42], an ethylene splitter [14], and batch trajectory tracking [43].…”

Section: Industrial Datamentioning

confidence: 99%

The Machine Learning Life Cycle in Chemical Operations – Status and Open Challenges

Gärtler

Khaydarov

Klöpper

et al. 2021

Chemie Ingenieur Technik

View full text Add to dashboard Cite

Artificial intelligence (AI) has received a lot of attention with many publications in recent years. Interestingly related projects in the industry are mostly still in their early stages. We are convinced that progress will only be possible if the entire machine learning (ML) life cycle is considered. Our study focuses on the practical challenges, uses a recent study as foundation and adopts the life-cycle description, highlights the life-cycle practices in other domains and formulates research directions that can help to improve the utilization of AI and machine learning in the process industry.

show abstract

Section: Industrial Datamentioning

confidence: 99%

The Machine Learning Life Cycle in Chemical Operations – Status and Open Challenges

Gärtler

Khaydarov

Klöpper

et al. 2021

Chemie Ingenieur Technik

View full text Add to dashboard Cite

show abstract

“…However, it is difficult to deal with uneven batch problems and multiphase processes because it manages the entire batch data as a single object, and a batch-wise unfolding method is used. By contrast, mean value normalization subtracts the column mean from the variable-wise unfolded data. , Because it maintains a variable trajectory after normalization, it is often applied to process modeling. , Although it is difficult to capture the nonlinearity owing to the large range of the data, it can address uneven batch data because each measurement is considered separately . In addition, the multimodel method can be easily combined with mean value normalization because the data range that each model covers is reduced.…”

Section: Preliminariesmentioning

confidence: 99%

“…Several methods using LV-MPC have been proposed for batch process control over the past two decades. ,, Flores-Cerrillo and MacGregor first combined a dynamic PCA model with the MPC scheme to calculate the control input in the latent variable space . The dynamic PCA model was adaptively constructed using the moving window concept; as such, it can accommodate time-varying dynamics.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Online Synchronization in Latent Variable Model Predictive Control for Trajectory Tracking of an Uneven Batch Process

Lee

2021

Ind. Eng. Chem. Res.

Self Cite

View full text Add to dashboard Cite

The performance of the batch trajectory tracking algorithm is affected by batch-to-batch variation, which includes irregular phase transitions and batch durations. The latent variable model predictive control is modified to obtain the desired specification by applying an online alignment to assign the prediction model and update the future reference. One of the multiple models is selected at every time step to improve the prediction performance using the relationship between the accumulated measurements of the ongoing batch. The future reference is synchronized based on the current measurement of the ongoing batch; therefore, a partially reduced or extended reference trajectory is adaptively applied to the batches. Compared with the conventional latent variable model predictive control algorithm, the proposed method for assigning a prediction model can improve the trajectory tracking performance not only for the entire batch but also for the phase transition region. Compared with the fixed reference case, the proposed method for aligning the reference can reduce the tracking error at the end of the batch. The batch duration is adaptively decided using the updated reference, reducing the batch duration as long as the ongoing batch adheres to the reduced reference. The proposed methods are verified using a case study of the trajectory tracking problem in industrial penicillin production.

show abstract

“…Moreover, the models which are able to retain high prediction accuracy over wide operational ranges would require large computational efforts in receding horizon optimization, which is detrimental to real-time control . Especially in large-scale application, linear or simple nonlinear prediction models usually prove stronger practicality than comprehensive nonlinear models . On these basis, multi-model prediction control (MMPC) is developed.…”

Section: Introductionmentioning

confidence: 99%

Model Predictive Control of a Solar Power System with Microturbine and Thermochemical Energy Storage

Xiao

Yang

2022

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

This paper presents the model predictive control (MPC) application on the solar power system with microturbine and thermochemical energy storage (TCES). To investigate the potential of a solar-powered turbine, a solar receiver and a TCES are introduced to the Brayton cycle as the replacement of the combustor. MPC is applied to offer the constrained multi-variable real-time optimization control. To increase the practicability in the solar industry, a custom-made multi-modeling approach is proposed based on the close relationship between direct normal irradiance and system states. Feedback correction mechanisms are designed to improve the prediction and target tracking accuracies. During the regulatory control under both extreme and realistic conditions, the multi-MPC (MMPC) shows stronger adaptability and reliability than proportional–integration–differentiation (PID) and higher tracking accuracies than the single-linear-model-based MPC. Although the control performance could be further improved by employing nonlinear MPC (NMPC), the much longer optimization time of NMPC was unsuitable for real-time control. MMPC is further adapted to track the grid demand, which is technically unachievable by PID in the current system. While the output power precisely follows the demand, the performance parameters can still stay close to their design values, retaining a high system efficiency. Overall, the proposed MMPC enables power demand tracking operation of solar air turbine systems, and can ensure high stability and system efficiency.

show abstract

Ensemble learning based latent variable model predictive control for batch trajectory tracking under concept drift

Cited by 10 publications

References 22 publications

The Machine Learning Life Cycle in Chemical Operations – Status and Open Challenges

The Machine Learning Life Cycle in Chemical Operations – Status and Open Challenges

Online Synchronization in Latent Variable Model Predictive Control for Trajectory Tracking of an Uneven Batch Process

Model Predictive Control of a Solar Power System with Microturbine and Thermochemical Energy Storage

Contact Info

Product

Resources

About