Neural network MPC for heating section of annealing furnace

Cho, Mingi; Ban, Jaepil; Seo, Minseok; Kim, Sang Woo

doi:10.1016/j.eswa.2023.119869

Cited by 8 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The RNN-based MPC successfully maintained the desired concentration of the paste thickener, even in the presence of severe pump failures. Other RNN-based applications include solving a generic nonconvex control problem [22], optimal policy selection [23], fault diagnosis for HVAC systems [24], theory [25] and application [26] of a generic nonlinear system for open-loop simulations, multi-mode process control of a generic system [27], chemical reactor control [28], crystallization processes [29] annealing furnaces [30], N-tank problems [31] and corn production [32]. Achirei et al [33] very recently introduced a modelbased predictive controller that utilized the object map obtained from the convolutional neural network (CNN) detector and light detection and ranging (LIDAR) data to guide an omnidirectional robot to specific positions in a warehouse environment.…”

Section: Introductionmentioning

confidence: 99%

Model Predictive Evolutionary Temperature Control via Neural-Network-Based Digital Twins

Ates,

Bicat,

Yankov

et al. 2023

Algorithms

View full text Add to dashboard Cite

In this study, we propose a population-based, data-driven intelligent controller that leverages neural-network-based digital twins for hypothesis testing. Initially, a diverse set of control laws is generated using genetic programming with the digital twin of the system, facilitating a robust response to unknown disturbances. During inference, the trained digital twin is utilized to virtually test alternative control actions for a multi-objective optimization task associated with each control action. Subsequently, the best policy is applied to the system. To evaluate the proposed model predictive control pipeline, experiments are conducted on a multi-mode heat transfer test rig. The objective is to achieve homogeneous cooling over the surface, minimizing the occurrence of hot spots and energy consumption. The measured variable vector comprises high dimensional infrared camera measurements arranged as a sequence (655,360 inputs), while the control variable includes power settings for fans responsible for convective cooling (3 outputs). Disturbances are induced by randomly altering the local heat loads. The findings reveal that by utilizing an evolutionary algorithm on measured data, a population of control laws can be effectively learned in the virtual space. This empowers the system to deliver robust performance. Significantly, the digital twin-assisted, population-based model predictive control (MPC) pipeline emerges as a superior approach compared to individual control models, especially when facing sudden and random changes in local heat loads. Leveraging the digital twin to virtually test alternative control policies leads to substantial improvements in the controller’s performance, even with limited training data.

show abstract

Section: Introductionmentioning

confidence: 99%

Model Predictive Evolutionary Temperature Control via Neural-Network-Based Digital Twins

Ates,

Bicat,

Yankov

et al. 2023

Algorithms

View full text Add to dashboard Cite

show abstract

“…Although the model differs from measurements, it is suitable for real-time applications of control due to the moderate computational effort. Cho et al [15] proposed a data-driven neural network MPC (model predictive controller) as a fast predictive model for the real-time control of an ACL furnace. He et al [16] developed a first-principle model to determine the strip temperature using the heat balance method.…”

Section: Introductionmentioning

confidence: 99%

Improved Waste Heat Management and Energy Integration in an Aluminum Annealing Continuous Furnace Using a Machine Learning Approach

Andayesh,

Flórez-Orrego,

Germanier

et al. 2023

Entropy

View full text Add to dashboard Cite

Annealing furnaces are critical for achieving the desired material properties in the production of high-quality aluminum products. In addition, energy efficiency has become more and more important in industrial processes due to increasing decarbonization regulations and the price of natural gas. Thus, the current study aims to determine the opportunities to reduce energy consumption in an annealing continuous furnace and the associated emissions. To this end, the heat transfer phenomenon is modeled and solutions for the decreasing fuel consumption are evaluated so that the overall performance of the process is enhanced. A heat transfer model is developed using the finite difference method, and the heat transfer coefficient is calculated using machine learning regression models. The heat transfer model is able to predict the heat transfer coefficient and calculate the aluminum temperature profile along the furnace and the fuel consumption for any given operating condition. Two solutions for boosting the furnace exergy efficiency are evaluated, including the modulation of the furnace temperature profiles and the energy integration by the recycling of exhaust flue gases. The results show that the advanced energy integration approach significantly reduces fuel consumption by up to 20.7%. Sensitivity analysis demonstrates that the proposed strategy can effectively reduce fuel consumption compared with the business-as-usual scenario for a range of sheet thicknesses and sheet velocities.

show abstract

Quantitative Uniformity Evaluation of Steel Slab Heating Temperature by Non-contact Measuring Approach

Yang,

Yu,

Wang

et al. 2024

Metall Mater Trans B

View full text Add to dashboard Cite

Neural network MPC for heating section of annealing furnace

Cited by 8 publications

References 22 publications

Model Predictive Evolutionary Temperature Control via Neural-Network-Based Digital Twins

Model Predictive Evolutionary Temperature Control via Neural-Network-Based Digital Twins

Improved Waste Heat Management and Energy Integration in an Aluminum Annealing Continuous Furnace Using a Machine Learning Approach

Quantitative Uniformity Evaluation of Steel Slab Heating Temperature by Non-contact Measuring Approach

Contact Info

Product

Resources

About