Machine Learning Modeling and Predictive Control of the Batch Crystallization Process

Zheng, Yingzhe; Wang, Xiaonan; Wu, Zhe

doi:10.1021/acs.iecr.2c00026

Cited by 53 publications

(32 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The AERNN-based predictive control scheme is formulated as a realtime optimization problem that calculates the optimal control action for the manipulated input T j in achieving a number of process performance specifications such as maximizing product yield and crystal size while minimizing energy consumption. The formulation of the AERNN-based MPC (AERNN-MPC) is presented as follows 7 :…”

Section: Aernn-mpc Formulationmentioning

confidence: 99%

“…Following the method in Ref. 7, the reduced‐order RNN model is constructed using the simulation data generated from the discretized PBE in the previous sections to describe the dynamics of batch FF crystallization. Specifically, ascribing to the high‐dimensional input (i.e., 43 dimensions consisting of C , T r , T j , and N i , ∀ i = 1, …, 40) and output (i.e., 42 dimensions consisting of C , T r , and N i , ∀ i = 1, …, 40) state variables generated by the crystallization system, an AE is developed as a dimensionality reduction technique to enhance the computational efficiency of the RNN model.…”

Section: Modeling Of Seeded Batch Cooling Crystallizationmentioning

confidence: 99%

“…The AERNN‐based predictive control scheme is formulated as a real‐time optimization problem that calculates the optimal control action for the manipulated input T j in achieving a number of process performance specifications such as maximizing product yield and crystal size while minimizing energy consumption. The formulation of the AERNN‐based MPC (AERNN‐MPC) is presented as follows 7 :

\min_{u \in S (normalΔ)} \int_{t_{k}}^{t_{k + N}} L (\overset{x}{true}, boldu) dt

s . t . \dot{\overset{x}{true}} (t) = f_{d} (F_{italicnn}^{'} ({\overset{x}{true}}_{r} (t), u (t)))

{\overset{x}{true}}_{r} (t_{k}) = f_{e} (x (t_{k}))

u (t) \in U, \forall t \in [t_{k}, t_{k + N})

∣ Δ u ∣ \leq U_{normalΔ}, \forall t \in [t_{k}, t_{k + N}),

where

\overset{false˜}{boldx} = [C, T_{r}, N_{i}] \in {boldR}^{42}

, i = 1, …, 40 and u = T j...…”

Section: Aernn‐based Predictive Controlmentioning

confidence: 99%

“…5 Additionally, the autoencoder (AE) neural network, which is an unsupervised dimensionality reduction technique that minimizes modeling complexity while still maintaining an adequate model fidelity, can be integrated with RNN to further enhance its computational efficiency. Several recent works have utilized RNN and AE-based RNN (AERNN) as surrogate process models in MPCs to control chemical plants in real time and optimize process performance accounting for closed-loop stability, safety, and control actuator constraints, which demonstrated superior computational efficiency compared to using the conventional first-principles model-based MPCs in their implementation to a diffusion-reaction process, 6 a batch crystallization process, 7 and a plasma etch process. 8 However, although pre-trained machine learning models have exhibited desirable computational properties and have shown to be promising substitutions for first-principles models in MPC, uncertainties in crystallization kinetics including intrinsic and exogenous uncertainties could be of major concern to their applications to real chemical processes, which can result in performance deterioration of these MPC schemes over time.…”

Section: Introductionmentioning

confidence: 99%

“…Additionally, the autoencoder (AE) neural network, which is an unsupervised dimensionality reduction technique that minimizes modeling complexity while still maintaining an adequate model fidelity, can be integrated with RNN to further enhance its computational efficiency. Several recent works have utilized RNN and AE‐based RNN (AERNN) as surrogate process models in MPCs to control chemical plants in real time and optimize process performance accounting for closed‐loop stability, safety, and control actuator constraints, which demonstrated superior computational efficiency compared to using the conventional first‐principles model‐based MPCs in their implementation to a diffusion–reaction process, 6 a batch crystallization process, 7 and a plasma etch process 8 …”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Online learning‐based predictive control of crystallization processes under batch‐to‐batch parametric drift

et al. 2022

Self Cite

View full text Add to dashboard Cite

This work considers a seeded fesoterodine fumarate (FF) cooling crystallization and presents the methodology and implementation of a real‐time machine learning modeling‐based predictive controller to handle batch‐to‐batch (B2B) parametric drift. Specifically, an autoencoder recurrent neural network‐based model predictive controller (AERNN‐MPC) is developed to optimize product yield, crystal size, and energy consumption while accounting for the physical constraints on cooling jacket temperature. Deviations in the kinetic parameters are considered in the closed‐loop simulations to account for the B2B parametric drift, and two error‐triggered online update mechanisms are proposed to address issues pertaining to the availability of real‐time crystal property measurements and are incorporated into the AERNN‐MPC to improve the model prediction accuracy. Closed‐loop simulation results demonstrate that the proposed AERNN‐MPC with online update, irrespective of the accessibility to real‐time crystal property data, achieves a desired closed‐loop performance in terms of maximizing product yield and minimizing energy consumption.

show abstract

Section: Aernn-mpc Formulationmentioning

confidence: 99%

Section: Modeling Of Seeded Batch Cooling Crystallizationmentioning

confidence: 99%

\min_{u \in S (normalΔ)} \int_{t_{k}}^{t_{k + N}} L (\overset{x}{true}, boldu) dt

s . t . \dot{\overset{x}{true}} (t) = f_{d} (F_{italicnn}^{'} ({\overset{x}{true}}_{r} (t), u (t)))

{\overset{x}{true}}_{r} (t_{k}) = f_{e} (x (t_{k}))

u (t) \in U, \forall t \in [t_{k}, t_{k + N})

∣ Δ u ∣ \leq U_{normalΔ}, \forall t \in [t_{k}, t_{k + N}),

where

\overset{false˜}{boldx} = [C, T_{r}, N_{i}] \in {boldR}^{42}

, i = 1, …, 40 and u = T j...…”

Section: Aernn‐based Predictive Controlmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Online learning‐based predictive control of crystallization processes under batch‐to‐batch parametric drift

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

Modeling and predictive control of nonlinear processes using transfer learning method

Xiao

2023

AIChE Journal

Self Cite

View full text Add to dashboard Cite

This work develops a transfer learning (TL) framework for modeling and predictive control of nonlinear systems using recurrent neural networks (RNNs) with the knowledge obtained in modeling one process transferred to another. Specifically, transfer learning uses a pretrained model developed based on a source domain as the starting point, and adapts the model to a target process with similar configurations. The generalization error for TL‐based RNN (TL‐RNN) is first derived to demonstrate the generalization capability on the target process. The theoretical error bound that depends on model capacity and the discrepancy between source and target domains is then utilized to guide the development of pretrained models for improved model transferability. Subsequently, the TL‐RNN model is utilized as the prediction model in model predictive controller (MPC) for the target process. Finally, the simulation study of chemical reactors via Aspen Plus Dynamics is used to demonstrate the benefits of transfer learning.

show abstract

Machine learning meets process control: Unveiling the potential of LSTMc

Sitapure,

Kwon

2024

AIChE Journal

View full text Add to dashboard Cite

In the past three decades, proportional‐integral/PI‐differential (PI/PID) controllers and model predictive controller (MPCs) have predominantly governed complex chemical process control. Despite their advancements, these approaches have limitations, with PI/PID controllers requiring scenario‐specific tuning and MPC being computationally demanding. To tackle these issues, we introduce the long‐short‐term‐memory (LSTM)‐controller (LSTMc), a model‐free, data‐driven framework leveraging LSTM networks' robust time‐series prediction capabilities. The LSTMc predicts subsequent manipulated inputs by evaluating state evolution and error dynamics from both the current and previous time‐steps, which proved effective in our dextrose batch crystallization case study. Remarkably, the achieves less than 2% set‐point deviation, three times better than MPCs, and retains robustness even with 10%–15% sensor noise. With these results, LSTMc emerges as a promising alternative for process control, adeptly adjusting to changing process conditions and set‐points, providing efficient computation for an optimal input profile, and effectively filtering out common industrial process noise.

show abstract

Machine Learning Modeling and Predictive Control of the Batch Crystallization Process

Cited by 53 publications

References 53 publications

Online learning‐based predictive control of crystallization processes under batch‐to‐batch parametric drift

Online learning‐based predictive control of crystallization processes under batch‐to‐batch parametric drift

Modeling and predictive control of nonlinear processes using transfer learning method

Machine learning meets process control: Unveiling the potential of LSTMc

Contact Info

Product

Resources

About