Artificial neural network based model predictive control: Implementing achievable set‐points

Hassanpour, Hesam; Corbett, Brandon; Mhaskar, Prashant

doi:10.1002/aic.17436

Cited by 10 publications

(13 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These methods also consider applying PCA-based strategies, assuming the process behaves linearly, which can lead to poorer control performance when used for nonlinear systems. Finally, in the previous example, 52 the correlation in the output space was dictated by the process dynamics not by the fact that the region of operation resulted in correlation between the outputs. Motivated by the above considerations, the first objective of this work is to address the problem of RNN-based MPC design, where the collected data from the process contains correlation in both the input and output spaces due to the closed-loop control structure (correlation is due to the closed-loop condition, and the outputs can in principle be controlled independently).…”

Section: Introductionmentioning

confidence: 83%

Section: Introductionmentioning

confidence: 86%

“…The first is the constrained RNN-based MPC, albeit with arbitrarily given set-points, 29 and the next is the standard RNN-based MPC with the achievable set-points. 52 As shown by the solid blue lines in Figures 10 and 11, the given set-points are not achieved by the constrained RNN-based MPC due to the use of the set-points which are selected independently, although the correlation is maintained in the input space. In addition, as shown by the dash-dotted red lines in Figures 5 and 6, the standard RNN-based MPC is not able to drive the outputs to their set-points, although the achievable set-points are implemented.…”

Section: Spmentioning

confidence: 99%

“…In addition, a PCA-based strategy is presented to calculate the achievable set-points for the RNNbased MPC, where data samples related to the process outputs are correlated. 52 The objective is to maintain the correlation existing between the outputs while calculating the set-points. Otherwise, the independently given set-points (prescribed without considering the correlation in the output space) cannot be achieved, either because of the plant−model mismatch (caused when the MPC controller uses the model for extrapolation) or the fundamental process limitations.…”

Section: Introductionmentioning

confidence: 99%

“…An alternative method is also presented by incorporating a new squared prediction error (SPE) constraint into the standard MPC (MPC using the RNN model trained on the original input–output data) to confine the input movements to the PCA model plane developed using the training input data. In addition, a PCA-based strategy is presented to calculate the achievable set-points for the RNN-based MPC, where data samples related to the process outputs are correlated . The objective is to maintain the correlation existing between the outputs while calculating the set-points.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Artificial Neural Network-Based Model Predictive Control Using Correlated Data

Hassanpour

Corbett

Mhaskar

2022

Ind. Eng. Chem. Res.

Self Cite

View full text Add to dashboard Cite

This work addresses the problem of implementing model predictive control (MPC) in situations where the training data available for modeling contains possible correlations, and an artificial neural network (ANN)-based model is being used. In particular, we consider a problem where data sets are collected from a process that operates under the closed-loop condition in which correlations are induced between several input and output variables. In this situation, if the correlation problem is not addressed, manipulated inputs (calculated by MPC without considering the specific correlation in the input space) and independently prescribed set-points may require predictions in regions where the model is not trained, resulting in a poor closed-loop performance. To address this issue, principal component analysis (PCA)-based strategies are applied to both the input and output spaces in a way that maintains model validity. To that end, a new constraint on the squared prediction error (SPE) is incorporated into the ANN-based MPC optimization problem to make control actions follow the PCA model built using the training input data. Next, a PCA model is developed using the training output data, and then an optimization problem subject to the SPE constraint is defined to calculate set-points which are achievable. The effectiveness of the proposed ANN-based MPC to track these set-points is demonstrated using a chemical reactor example. Finally, a new autoencoder-based strategy is proposed to compute the achievable set-points. This is performed by replacing the PCA-based constraint with the autoencoder-based constraint in the optimization problem to calculate the set-points. The results indicate that the ANN-based MPC performance is improved when the autoencoder-based set-points are used.

show abstract

Section: Introductionmentioning

confidence: 83%

Section: Introductionmentioning

confidence: 86%

Section: Spmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Artificial Neural Network-Based Model Predictive Control Using Correlated Data

Hassanpour

Corbett

Mhaskar

2022

Ind. Eng. Chem. Res.

Self Cite

View full text Add to dashboard Cite

show abstract

A practically implementable reinforcement learning‐based process controller design

Hassanpour,

Wang,

Corbett

et al. 2023

AIChE Journal

Self Cite

View full text Add to dashboard Cite

The present article enables reinforcement learning (RL)‐based controllers for process control applications. Existing instances of RL‐based solutions have significant challenges for online implementation since the training process of an RL agent (controller) presently requires practically impossible number of online interactions between the agent and the environment (process). To address this challenge, we propose an implementable model‐free RL method developed by leveraging industrially implemented model predictive control (MPC) calculations (often designed using a simple linear model identified via step tests). In the first step, MPC calculations are used to pretrain an RL agent that can mimic the MPC performance. Specifically, the MPC calculations are used to pretrain the actor, and the objective function is used to pretrain the critic(s). The pretrained RL agent is then employed within a model‐free RL framework to control the process in a way that initially imitates MPC behavior (thus not compromising process performance and safety), but also continuously learns and improve its performance over the nominal linear MPC. The effectiveness of the proposed approach is illustrated through simulations on a chemical reactor example.

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

Artificial neural network based model predictive control: Implementing achievable set‐points

Cited by 10 publications

References 45 publications

Artificial Neural Network-Based Model Predictive Control Using Correlated Data

Artificial Neural Network-Based Model Predictive Control Using Correlated Data

A practically implementable reinforcement learning‐based process controller design

Machine learning meets process control: Unveiling the potential of LSTMc

Contact Info

Product

Resources

About