Zhiqiang Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhiqiang Wu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Zhong Zhao
State Key Laboratory of Digital Steel, Northeastern University, Shenyang 110819, China
8 shared publications - Chunyu He
State Key Laboratory of Digital Steel, Northeastern University, Shenyang 110819, China
7 shared publications - Zhijie Jiao
State Key Laboratory of Digital Steel, Northeastern University, Shenyang 110819, China
7 shared publications - Junyi Luo
State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819, China
5 shared publications - Zhiqiang Wang
State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819, China
5 shared publications - Chujie Liu
State Key Laboratory of Digital Steel, Northeastern University, Shenyang 110819, China
3 shared publications
Latest Publications
- Open Access
Observer-Based Finite-Time Adaptive Fuzzy Control for Nonlinear Systems with Full-State Constraints
This paper investigates practical finite-time adaptive fuzzy control for uncertain nonlinear systems subject to full-state constraints and unmeasured state variables. To deal with the inaccessibility of some state variables, an observer is constructed. A state-dependent nonlinear mapping is introduced to ensure that all system states remain within their prescribed bounds. In contrast to typical barrier Lyapunov function (BLF)-based schemes, the presented method manages full-state constraints without imposing any extra prerequisites on virtual control signals. Fuzzy logic systems (FLSs) act as estimators for the unknown nonlinearities emerging in the control law design, while the integration of dynamic surface control techniques helps circumvent the “complexity explosion” characteristic of conventional backstepping approaches. Subsequently, a practical finite-time adaptive fuzzy tracking controller is constructed, which guarantees the semi-global practical finite-time stability of the closed-loop system, with all closed-loop signals remaining bounded for all time and the tracking error converging to a residual set within finite time. Simulation results demonstrate that the tracking error enters a small residual set within finite time and that all system states remain within their prescribed constraints.
Open AccessObserver-Based Finite-Time Adaptive Fuzzy Control for Nonlinear Systems with Full-State Constraints
This paper investigates practical finite-time adaptive fuzzy control for uncertain nonlinear systems subject to full-state constraints and unmeasured state variables. To deal with the inaccessibility of some state variables, an observer is constructed. A state-dependent nonlinear mapping is introduced to ensure that all system states remain within their prescribed bounds. In contrast to typical barrier Lyapunov function (BLF)-based schemes, the presented method manages full-state constraints without imposing any extra prerequisites on virtual control signals. Fuzzy logic systems (FLSs) act as estimators for the unknown nonlinearities emerging in the control law design, while the integration of dynamic surface control techniques helps circumvent the “complexity explosion” characteristic of conventional backstepping approaches. Subsequently, a practical finite-time adaptive fuzzy tracking controller is constructed, which guarantees the semi-global practical finite-time stability of the closed-loop system, with all closed-loop signals remaining bounded for all time and the tracking error converging to a residual set within finite time. Simulation results demonstrate that the tracking error enters a small residual set within finite time and that all system states remain within their prescribed constraints.
- Open Access
Event-Trigger-Based Fuzzy Adaptive Finite-Time Control for Uncertain Nonlinear Systems with Unmeasurable States
This article delves into the fuzzy finite-time adaptive control problem for uncertain nonlinear systems where state measurements are unavailable, nonlinear functions are unknown, and communication is limited. To emulate the unknown nonlinear relationships within the control methodology, we exploit fuzzy logic systems, while also proposing a state observer to address the challenge of unobservable states. To avoid the “complexity explosion” problem intrinsic to conventional backstepping techniques, the controller is developed based on the dynamic surface control methodology, which incorporates first-order filters to successfully alleviate this issue. An event-triggered approach is introduced to alleviate the computational and communication overhead. By leveraging the finite-time control approach, an adaptive finite-time fuzzy control algorithm is constructed using the adaptive backstepping technique. An event-triggered mechanism is designed to reduce communication frequency, while rigorously maintaining closed-loop stability and ensuring a positive minimum inter-event time to avoid Zeno behavior. The proposed finite-time controller achieves finite-time stability of the controlled systems, thereby guaranteeing that all system signals remain bounded within a finite time, despite the presence of unmeasurable states, unknown nonlinear functions, and limited communication constraints. This paper differentiates itself from recent related studies by proposing a co-designed observer–controller framework that rigorously guarantees finite-time stability under an event-triggered communication mechanism, thereby effectively addressing the multiple concurrent challenges of state estimation, rapid convergence, and limited network resources. Simulation examples are conducted to illustrate the effectiveness and feasibility of the derived control algorithm.
Open AccessEvent-Trigger-Based Fuzzy Adaptive Finite-Time Control for Uncertain Nonlinear Systems with Unmeasurable States
This article delves into the fuzzy finite-time adaptive control problem for uncertain nonlinear systems where state measurements are unavailable, nonlinear functions are unknown, and communication is limited. To emulate the unknown nonlinear relationships within the control methodology, we exploit fuzzy logic systems, while also proposing a state observer to address the challenge of unobservable states. To avoid the “complexity explosion” problem intrinsic to conventional backstepping techniques, the controller is developed based on the dynamic surface control methodology, which incorporates first-order filters to successfully alleviate this issue. An event-triggered approach is introduced to alleviate the computational and communication overhead. By leveraging the finite-time control approach, an adaptive finite-time fuzzy control algorithm is constructed using the adaptive backstepping technique. An event-triggered mechanism is designed to reduce communication frequency, while rigorously maintaining closed-loop stability and ensuring a positive minimum inter-event time to avoid Zeno behavior. The proposed finite-time controller achieves finite-time stability of the controlled systems, thereby guaranteeing that all system signals remain bounded within a finite time, despite the presence of unmeasurable states, unknown nonlinear functions, and limited communication constraints. This paper differentiates itself from recent related studies by proposing a co-designed observer–controller framework that rigorously guarantees finite-time stability under an event-triggered communication mechanism, thereby effectively addressing the multiple concurrent challenges of state estimation, rapid convergence, and limited network resources. Simulation examples are conducted to illustrate the effectiveness and feasibility of the derived control algorithm.
- Open Access
Research on Intelligent Control Method of Camber for Medium and Heavy Plate Based on Machine Vision
With the continuous development of intelligent manufacturing in the iron and steel industry, there are increasing requirements for the quality control and precision of steel products. Camber is one of the critical defects affecting product quality in medium and heavy plates. Its occurrence during the rolling process not only reduces the yield of plates but also leads to serious production accidents such as rolling scrap and equipment damage, increasing the operational costs of enterprises. Addressing the difficulties that camber is influenced by complex factors and direct modeling control is challenging, this study proposes a camber detection and control method for medium and heavy plates based on image processing and machine learning algorithms, relying on an actual plate production line. The Optuna-XGBoost model is used to mine and train the production data of plates rolling, extracting the optimal control experience of operators as the pre-control values for camber. The Optuna-XGBoost model achieves an R2 of 0.9999 on the training set and 0.9794 on the test set, demonstrating excellent fitting performance. Meanwhile, a camber detection technology during the plate rolling process is developed based on machine vision. A feedback control model for camber of medium and heavy plates based on distal lateral movement is established. The combined application of pre-control and feedback control reduces the occurrence of camber, ensuring the overall flatness of steel plates during the rolling process. This paper establishes an intelligent control framework for plate camber, synergized by data-driven pre-control and machine vision-based feedback control, offering a novel approach for the online optimal control of complex nonlinear industrial processes.
Open AccessResearch on Intelligent Control Method of Camber for Medium and Heavy Plate Based on Machine Vision
With the continuous development of intelligent manufacturing in the iron and steel industry, there are increasing requirements for the quality control and precision of steel products. Camber is one of the critical defects affecting product quality in medium and heavy plates. Its occurrence during the rolling process not only reduces the yield of plates but also leads to serious production accidents such as rolling scrap and equipment damage, increasing the operational costs of enterprises. Addressing the difficulties that camber is influenced by complex factors and direct modeling control is challenging, this study proposes a camber detection and control method for medium and heavy plates based on image processing and machine learning algorithms, relying on an actual plate production line. The Optuna-XGBoost model is used to mine and train the production data of plates rolling, extracting the optimal control experience of operators as the pre-control values for camber. The Optuna-XGBoost model achieves an R2 of 0.9999 on the training set and 0.9794 on the test set, demonstrating excellent fitting performance. Meanwhile, a camber detection technology during the plate rolling process is developed based on machine vision. A feedback control model for camber of medium and heavy plates based on distal lateral movement is established. The combined application of pre-control and feedback control reduces the occurrence of camber, ensuring the overall flatness of steel plates during the rolling process. This paper establishes an intelligent control framework for plate camber, synergized by data-driven pre-control and machine vision-based feedback control, offering a novel approach for the online optimal control of complex nonlinear industrial processes.
- Open Access
- 9 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (18) , 9856
PoMQ-ViT: Mixed-Precision Quantization Vision Transformer with Pareto Optimization
Vision Transformer (ViT) has established itself as the leading model in image processing, and quantization offers a practical way to deploy ViTs on memory-constrained devices. This technique maps float32 inference to lower bits such as int8. However, existing studies often use uniform bit-width across ViT encoders, leading to sub-optimal allocation and accuracy loss. This study aims to optimize bit-width allocation by leveraging the observed variation in quantization sensitivity among different ViT encoders. We propose PoMQ-ViT, a method that dynamically assigns bit-width to ViT encoders based on their sensitivity. It uses Pareto Optimization to balance accuracy and inference speed. Experiments on the ImageNet Dataset show PoMQ-ViT outperforms uniform bit-width methods across DeiT and Swin models. Within the same computation budget, it also achieves 0.1–1.5 percentage points higher accuracy than other mixed-precision paradigms across different models. This work demonstrates PoMQ-ViT’s effectiveness in ViT quantization, providing a practical solution for resource-constrained deployments.
- 9 September 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (18) , 9856
Open AccessPoMQ-ViT: Mixed-Precision Quantization Vision Transformer with Pareto Optimization
Vision Transformer (ViT) has established itself as the leading model in image processing, and quantization offers a practical way to deploy ViTs on memory-constrained devices. This technique maps float32 inference to lower bits such as int8. However, existing studies often use uniform bit-width across ViT encoders, leading to sub-optimal allocation and accuracy loss. This study aims to optimize bit-width allocation by leveraging the observed variation in quantization sensitivity among different ViT encoders. We propose PoMQ-ViT, a method that dynamically assigns bit-width to ViT encoders based on their sensitivity. It uses Pareto Optimization to balance accuracy and inference speed. Experiments on the ImageNet Dataset show PoMQ-ViT outperforms uniform bit-width methods across DeiT and Swin models. Within the same computation budget, it also achieves 0.1–1.5 percentage points higher accuracy than other mixed-precision paradigms across different models. This work demonstrates PoMQ-ViT’s effectiveness in ViT quantization, providing a practical solution for resource-constrained deployments.
- Open Access
Intelligent Optimization Control of Plate Plan-View Pattern Based on Intermediate Slab Pattern Vision Inspection and BWO-DNN Algorithm
In the production process of plate, the main factors affecting the yield of plate are the crop-cutting and edge losses. It is very important to accurately predict the crop pattern of plates and effectively control the plan-view pattern. In this paper, a detection scheme is proposed to obtain the plan-view pattern of the intermediate slab and finished plate by placing detection devices after the roughing mill and finishing mill, respectively. An image processing algorithm is used to obtain a dataset of the plan-view pattern parameters, and a plan-view pattern prediction and control model for plate is established based on the BWO-DNN (beluga whale optimization–deep neural network) algorithm. The BWO algorithm is used to optimize the hyperparameters in the DNN algorithm to complete the establishment of the intelligent model. In terms of model analysis, goodness of fit (R2) and mean absolute error (MAE) are used as evaluation indicators. The results show that the intelligence model established based on BWO-DNN has good predictive and control performance, realizing intelligent prediction of the crop pattern of plates and parameter optimization of plan-view pattern control. The actual production verification on site shows that the irregular areas of the plate head and tail can be reduced by 17.2% and 22.6%, respectively.
Open AccessIntelligent Optimization Control of Plate Plan-View Pattern Based on Intermediate Slab Pattern Vision Inspection and BWO-DNN Algorithm
In the production process of plate, the main factors affecting the yield of plate are the crop-cutting and edge losses. It is very important to accurately predict the crop pattern of plates and effectively control the plan-view pattern. In this paper, a detection scheme is proposed to obtain the plan-view pattern of the intermediate slab and finished plate by placing detection devices after the roughing mill and finishing mill, respectively. An image processing algorithm is used to obtain a dataset of the plan-view pattern parameters, and a plan-view pattern prediction and control model for plate is established based on the BWO-DNN (beluga whale optimization–deep neural network) algorithm. The BWO algorithm is used to optimize the hyperparameters in the DNN algorithm to complete the establishment of the intelligent model. In terms of model analysis, goodness of fit (R2) and mean absolute error (MAE) are used as evaluation indicators. The results show that the intelligence model established based on BWO-DNN has good predictive and control performance, realizing intelligent prediction of the crop pattern of plates and parameter optimization of plan-view pattern control. The actual production verification on site shows that the irregular areas of the plate head and tail can be reduced by 17.2% and 22.6%, respectively.
- Open Access
- 29 May 2025
- Published by MDPI AG in Preprints.org
Intelligent Optimization Control of Plate Plan View Pattern Based on Intermediate Slab Pattern Vision Inspection and BWO-DNN Algorithm
In the production process of plate, the main factors affecting the yield of plate are the crop cutting and edge losses It is very important to accurately predict the crop pattern of plates and effectively control the plan view pattern. In this paper, a detection scheme is proposed to obtain plan view pattern of intermediate slab and finished plate by placing the detection devices after roughing mill and finishing mill respectively. The image processing algorithm is used to obtain the dataset of plan view pattern parameters, and the plan view pattern prediction and control model of plate is established based on BWO-DNN algorithm. The BWO algorithm is used to optimize the hyperparameters in DNN algorithm to complete the establishment of the intelligent model. In terms of model analysis, goodness of fit (R2) and mean absolute error (MAE) are used as evaluation indicators. The results show that the intelligence model established based on BWO-DNN has good predictive and control performance, realizing intelligent prediction of the crop pattern of plates and parameters optimization of plan view pattern control. The actual production verification on site shows that the irregular area of plate head and tail can be reduced by 17.2 % and 22.6% respectively.
- 29 May 2025
- Published by MDPI AG in Preprints.org
Open AccessIntelligent Optimization Control of Plate Plan View Pattern Based on Intermediate Slab Pattern Vision Inspection and BWO-DNN Algorithm
In the production process of plate, the main factors affecting the yield of plate are the crop cutting and edge losses It is very important to accurately predict the crop pattern of plates and effectively control the plan view pattern. In this paper, a detection scheme is proposed to obtain plan view pattern of intermediate slab and finished plate by placing the detection devices after roughing mill and finishing mill respectively. The image processing algorithm is used to obtain the dataset of plan view pattern parameters, and the plan view pattern prediction and control model of plate is established based on BWO-DNN algorithm. The BWO algorithm is used to optimize the hyperparameters in DNN algorithm to complete the establishment of the intelligent model. In terms of model analysis, goodness of fit (R2) and mean absolute error (MAE) are used as evaluation indicators. The results show that the intelligence model established based on BWO-DNN has good predictive and control performance, realizing intelligent prediction of the crop pattern of plates and parameters optimization of plan view pattern control. The actual production verification on site shows that the irregular area of plate head and tail can be reduced by 17.2 % and 22.6% respectively.
- Open Access
Rectangular Improvement Method for Plan View Pattern of Plates During the Angular Rolling Process
The effect of the angular rolling process on the plan view pattern of a plate was studied, and the rectangular influencing factors and improvement methods for this process were proposed in this paper. DEFORM (v11.0) finite element software was used to simulate the processes of conventional rolling and angular rolling, and the degree of rectangularity of plates under different rolling process conditions was compared. A formula to characterize the degree of rectangularity of plates was established; the closer this value is to one, the better the degree of rectangularity. Considering the actual rolling process conditions, the range of theoretically calculated rectangular rotation angles was extended to obtain the optimum rectangular rotation angle using the finite element simulation method. In the two-pass angular rolling process, the optimal rectangular angle of the second pass was 14.275° when the first pass was 15°. The optimal rectangular angle of the plate was 19.008° when the first pass’ angle was 20°. Two-pass angular rolling is different to four-pass rolling, and the simulation results showed that (1.0012) was less than (1.0015) and (1.0034) was less than (1.0055). The rectangularity degree of the four-pass process was better than the two-pass process. Angular rolling experiments were carried out, and the actual data show that the characteristic rectangular value of the rolled piece was 1.003 during the four-pass process and 1.014 during the two-pass process. This verified that separating the one-group two-pass angular rolling process from the one-group four-pass angular rolling process can improve the rectangular degree of the rolled plate, thereby increasing the yield rate. This provides a theoretical basis for industrial applications.Open AccessRectangular Improvement Method for Plan View Pattern of Plates During the Angular Rolling Process
The effect of the angular rolling process on the plan view pattern of a plate was studied, and the rectangular influencing factors and improvement methods for this process were proposed in this paper. DEFORM (v11.0) finite element software was used to simulate the processes of conventional rolling and angular rolling, and the degree of rectangularity of plates under different rolling process conditions was compared. A formula to characterize the degree of rectangularity of plates was established; the closer this value is to one, the better the degree of rectangularity. Considering the actual rolling process conditions, the range of theoretically calculated rectangular rotation angles was extended to obtain the optimum rectangular rotation angle using the finite element simulation method. In the two-pass angular rolling process, the optimal rectangular angle of the second pass was 14.275° when the first pass was 15°. The optimal rectangular angle of the plate was 19.008° when the first pass’ angle was 20°. Two-pass angular rolling is different to four-pass rolling, and the simulation results showed that (1.0012) was less than (1.0015) and (1.0034) was less than (1.0055). The rectangularity degree of the four-pass process was better than the two-pass process. Angular rolling experiments were carried out, and the actual data show that the characteristic rectangular value of the rolled piece was 1.003 during the four-pass process and 1.014 during the two-pass process. This verified that separating the one-group two-pass angular rolling process from the one-group four-pass angular rolling process can improve the rectangular degree of the rolled plate, thereby increasing the yield rate. This provides a theoretical basis for industrial applications. - Open Access
- 2 December 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (23) , 3816
Fixed-Time Command-Filtered Control for Nonlinear Systems with Mismatched Disturbances
This article concerns the issue of adaptive fuzzy command-filtered fixed-time control in the context of a category of nonlinear systems characterized by mismatched disturbances and unknown nonlinear functions. The backstepping-based disturbance observers are created to alleviate the effects of mismatched disturbances and Fuzzy logic systems are brought into play to model those terms that are unknown. To address the complexity explosion issue in traditional backstepping control, this paper utilizes fixed-time command filters (FTCFs) to design a novel control approach. Moreover, filtering error compensation mechanisms are developed to eliminate the errors introduced by the FTCFs. This paper derives a novel adaptive fixed-time control protocol that successfully conquers the difficulties posed by unknown nonlinear functions and mismatched disturbances. The protocol, implemented within a backstepping framework, guarantees the boundedness of all signals and tracking errors within fixed time. The efficacy of the derived control protocol is illustrated through simulation examples.
- 2 December 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (23) , 3816
Open AccessFixed-Time Command-Filtered Control for Nonlinear Systems with Mismatched Disturbances
This article concerns the issue of adaptive fuzzy command-filtered fixed-time control in the context of a category of nonlinear systems characterized by mismatched disturbances and unknown nonlinear functions. The backstepping-based disturbance observers are created to alleviate the effects of mismatched disturbances and Fuzzy logic systems are brought into play to model those terms that are unknown. To address the complexity explosion issue in traditional backstepping control, this paper utilizes fixed-time command filters (FTCFs) to design a novel control approach. Moreover, filtering error compensation mechanisms are developed to eliminate the errors introduced by the FTCFs. This paper derives a novel adaptive fixed-time control protocol that successfully conquers the difficulties posed by unknown nonlinear functions and mismatched disturbances. The protocol, implemented within a backstepping framework, guarantees the boundedness of all signals and tracking errors within fixed time. The efficacy of the derived control protocol is illustrated through simulation examples.
- Open Access
- 8 November 2024
- Published by MDPI AG in Preprints.org
Rectangular Improvement Method for Plan View Pattern of Plate During Angular Rolling Process
The effect of the angular rolling process on the plan view pattern of the plate is studied, the rectangular influence factor and improvement method for plate angular rolling process are proposed in this paper. Finite element models are established for the conventional rolling and angular rolling processes respectively, and with DEFORM software, the numerical simulation is carried out. A characterization formula reflecting the rectangular degree of the rolled piece is constructed and the rectangular degree of the plates under different rolling process conditions are compared. Considering the actual rolling process conditions, the range of theoretically calculated rectangular rotation angles is extended to obtain the optimum rectangular rotation angle with finite element simulation method. The one-group two-pass angular rolling is separated to one-group four-pass rolling and the simulation results show that the rectangular degree of the four-pass process is better than the two-pass process. Angular rolling experiments are carried out, the actual data shows that the rectangular characteristic value of the rolled piece is 1.003 with the four-pass process, and 1.014 with the two-pass process. It verifies that with separating one-group two-pass angular rolling process to one-group four-pass angular rolling process can improve the rectangular degree of the rolled plate.
- 8 November 2024
- Published by MDPI AG in Preprints.org
Open AccessRectangular Improvement Method for Plan View Pattern of Plate During Angular Rolling Process
The effect of the angular rolling process on the plan view pattern of the plate is studied, the rectangular influence factor and improvement method for plate angular rolling process are proposed in this paper. Finite element models are established for the conventional rolling and angular rolling processes respectively, and with DEFORM software, the numerical simulation is carried out. A characterization formula reflecting the rectangular degree of the rolled piece is constructed and the rectangular degree of the plates under different rolling process conditions are compared. Considering the actual rolling process conditions, the range of theoretically calculated rectangular rotation angles is extended to obtain the optimum rectangular rotation angle with finite element simulation method. The one-group two-pass angular rolling is separated to one-group four-pass rolling and the simulation results show that the rectangular degree of the four-pass process is better than the two-pass process. Angular rolling experiments are carried out, the actual data shows that the rectangular characteristic value of the rolled piece is 1.003 with the four-pass process, and 1.014 with the two-pass process. It verifies that with separating one-group two-pass angular rolling process to one-group four-pass angular rolling process can improve the rectangular degree of the rolled plate.
- Open Access
Digital Model of Automatic Plate Turning for Plate Mills Based on Machine Vision and Reinforcement Learning Algorithm
Plate turning is an essential step in the plate rolling process. The traditional control mode relies on the manual observation of billets and mainly manual operation. Manual plate turning becomes an external disturbance of the automatic control system of plate mills, which reduces the reproducibility and accuracy of the rolling process. The automatic plate turning function is urgently needed to improve the control level of the rolling line. In this paper, based on the improved image processing algorithm, the position and angle information of the billet conversion process are detected in real time, and the real-time processing of detection data in a complex production environment is realized. Based on the change in the billet rotation angle in the actual plate turning process, a mathematical model is constructed to simulate the plate turning process. On this basis, the digital model and optimization algorithm for automatic plate turning based on reinforcement learning are established, and the automatic optimization of plate turning speed and accuracy is completed. The field application of data-driven plate turning systems replaces manual plate turning control. The plate turning angle detection error of the system is ≤2°. The average plate turning time of each billet is greatly shortened compared with manual plate turning mode, and the fastest time can be shortened by more than 1 s, which greatly improves the production efficiency and is of great significance for improving the automatic control level and digital upgrade of plate mills.
Open AccessDigital Model of Automatic Plate Turning for Plate Mills Based on Machine Vision and Reinforcement Learning Algorithm
Plate turning is an essential step in the plate rolling process. The traditional control mode relies on the manual observation of billets and mainly manual operation. Manual plate turning becomes an external disturbance of the automatic control system of plate mills, which reduces the reproducibility and accuracy of the rolling process. The automatic plate turning function is urgently needed to improve the control level of the rolling line. In this paper, based on the improved image processing algorithm, the position and angle information of the billet conversion process are detected in real time, and the real-time processing of detection data in a complex production environment is realized. Based on the change in the billet rotation angle in the actual plate turning process, a mathematical model is constructed to simulate the plate turning process. On this basis, the digital model and optimization algorithm for automatic plate turning based on reinforcement learning are established, and the automatic optimization of plate turning speed and accuracy is completed. The field application of data-driven plate turning systems replaces manual plate turning control. The plate turning angle detection error of the system is ≤2°. The average plate turning time of each billet is greatly shortened compared with manual plate turning mode, and the fastest time can be shortened by more than 1 s, which greatly improves the production efficiency and is of great significance for improving the automatic control level and digital upgrade of plate mills.