Pub Date : 2024-08-23DOI: 10.1016/j.conengprac.2024.106061
One of the main factors influencing machine tool feed system tracking performance is friction. By creating an accurate friction model and implementing feed-forward compensation based on the model, the negative impacts of friction can be efficiently reduced. The generalized Maxwell-slip (GMS) model is commonly used to model feed system friction; however, simple and effective parameter identification methods are lacking. In this paper, a parameter identification method based on a metaheuristic Gaussian swarm optimization (GSO) algorithm is proposed. The method divides the parameters into two parts via a theoretical derivation, and employs GSO to identify each part successively. The proposed GSO is a novel metaheuristic algorithm inspired by the Gaussian probability function. The excellent performance of the GSO ensures that the friction parameters can be accurately and quickly identified. The results of the simulation and physical identification experiments show that the proposed GSO-based identification method can accurately identify the parameters of the GMS model with average and maximum relative errors of 3.96% and 14.05%, respectively. The identified model can accurately predict the friction of the feed system. Additionally, after friction compensation, the tracking error was decreased by an average of 78.9%.
{"title":"Effective parameter identification of the GMS friction model for feed systems in CNC machines","authors":"","doi":"10.1016/j.conengprac.2024.106061","DOIUrl":"10.1016/j.conengprac.2024.106061","url":null,"abstract":"<div><p>One of the main factors influencing machine tool feed system tracking performance is friction. By creating an accurate friction model and implementing feed-forward compensation based on the model, the negative impacts of friction can be efficiently reduced. The generalized Maxwell-slip (GMS) model is commonly used to model feed system friction; however, simple and effective parameter identification methods are lacking. In this paper, a parameter identification method based on a metaheuristic Gaussian swarm optimization (GSO) algorithm is proposed. The method divides the parameters into two parts via a theoretical derivation, and employs GSO to identify each part successively. The proposed GSO is a novel metaheuristic algorithm inspired by the Gaussian probability function. The excellent performance of the GSO ensures that the friction parameters can be accurately and quickly identified. The results of the simulation and physical identification experiments show that the proposed GSO-based identification method can accurately identify the parameters of the GMS model with average and maximum relative errors of 3.96% and 14.05%, respectively. The identified model can accurately predict the friction of the feed system. Additionally, after friction compensation, the tracking error was decreased by an average of 78.9%.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142050300","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-22DOI: 10.1016/j.conengprac.2024.106047
In this paper, the formation control problem is considered for unicycle multi-agent systems whose kinematic models contain some external perturbations. The approach to addressing the problem involves the development of a homogeneity-based leader–follower formation control protocol, which takes into account bounded perturbations. It is shown that such a control protocol can be obtained if there is an external supervisor monitoring the group and broadcasting a limited amount of data to followers. Simulations as well as experimental results are performed to illustrate the effectiveness of the proposed control protocol using the QBot2 unicycle mobile robot.
{"title":"A generalized homogeneity-based formation control for perturbed unicycle multi-agent systems","authors":"","doi":"10.1016/j.conengprac.2024.106047","DOIUrl":"10.1016/j.conengprac.2024.106047","url":null,"abstract":"<div><p>In this paper, the formation control problem is considered for unicycle multi-agent systems whose kinematic models contain some external perturbations. The approach to addressing the problem involves the development of a homogeneity-based leader–follower formation control protocol, which takes into account bounded perturbations. It is shown that such a control protocol can be obtained if there is an external supervisor monitoring the group and broadcasting a limited amount of data to followers. Simulations as well as experimental results are performed to illustrate the effectiveness of the proposed control protocol using the QBot2 unicycle mobile robot.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142039792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-21DOI: 10.1016/j.conengprac.2024.106049
The P2 configuration plug-in hybrid electric vehicle (P2-PHEV) equipped with a multi-speed transmission has a high potential for recovering more regenerative energy, as the shifting strategy can be employed to adjust the working zone of the electric motor (EM). However, the existing shifting strategy designed for normal driving conditions cannot achieve optimal regenerative energy recovery. In this study, a shifting strategy used in the regenerative braking process is proposed. First, to make the EM provide more regenerative force during braking, a braking force distribution algorithm is devised while simultaneously considering braking stability and safety. Second, to realize maximum regenerative braking energy recovery, an optimal shifting strategy is designed for regenerative braking. Third, the classical braking process is analyzed and six thresholds are abstracted and optimized to establish a rule for restraining frequent gearshifts raised in the proposed optimal shifting strategy. Finally, the proposed strategy is verified under three standard cycles, results show that the proposed shifting strategy can recover considerable regenerative energy without frequent gearshifts.
配备多速变速器的 P2 配置插电式混合动力电动汽车(P2-PHEV)具有回收更多再生能量的巨大潜力,因为换挡策略可用于调整电动马达(EM)的工作区域。然而,现有的针对正常驾驶条件设计的换挡策略无法实现最佳的再生能量回收。本研究提出了一种用于再生制动过程的换挡策略。首先,为了使 EM 在制动过程中提供更多的再生力,设计了一种制动力分配算法,同时考虑了制动稳定性和安全性。其次,为实现最大的再生制动能量回收,设计了再生制动的最佳换挡策略。第三,对经典制动过程进行分析,并抽象和优化了六个阈值,从而建立了一个规则,用于抑制所提出的优化换挡策略中的频繁换挡。最后,在三个标准周期下对所提出的策略进行了验证,结果表明所提出的换挡策略可以在不频繁换挡的情况下回收大量再生能量。
{"title":"Downshifting strategy of plug-in hybrid vehicle during braking process for greater regenerative energy","authors":"","doi":"10.1016/j.conengprac.2024.106049","DOIUrl":"10.1016/j.conengprac.2024.106049","url":null,"abstract":"<div><p>The P2 configuration plug-in hybrid electric vehicle (P2-PHEV) equipped with a multi-speed transmission has a high potential for recovering more regenerative energy, as the shifting strategy can be employed to adjust the working zone of the electric motor (EM). However, the existing shifting strategy designed for normal driving conditions cannot achieve optimal regenerative energy recovery. In this study, a shifting strategy used in the regenerative braking process is proposed. First, to make the EM provide more regenerative force during braking, a braking force distribution algorithm is devised while simultaneously considering braking stability and safety. Second, to realize maximum regenerative braking energy recovery, an optimal shifting strategy is designed for regenerative braking. Third, the classical braking process is analyzed and six thresholds are abstracted and optimized to establish a rule for restraining frequent gearshifts raised in the proposed optimal shifting strategy. Finally, the proposed strategy is verified under three standard cycles, results show that the proposed shifting strategy can recover considerable regenerative energy without frequent gearshifts.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142021481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-20DOI: 10.1016/j.conengprac.2024.106048
The skid steering unmanned ground vehicle (SUGV) plays an important role in extremely harsh environments. Improving the autonomous control capability and energy efficiency of SUGV is urgently needed. This article presents a skid steering-based path tracking control strategy. In the upper controller, an improved model-free sliding mode controller (APMS) is used to calculate the yaw moment for tracking control. On the lower controller, the Snow Ablation Optimizer (SAO) is used to distribute the output torque of the drive motors, taking longitudinal force, yaw moment and energy consumption into account. Finally, the designed controller is validated through simulation under different operating conditions. The results show that the proposed coordination controller achieves good control performance, increases energy efficiency and at the same time ensures tracking accuracy.
{"title":"Path tracking and energy efficiency coordination control strategy for skid-steering unmanned ground vehicle","authors":"","doi":"10.1016/j.conengprac.2024.106048","DOIUrl":"10.1016/j.conengprac.2024.106048","url":null,"abstract":"<div><p>The skid steering unmanned ground vehicle (SUGV) plays an important role in extremely harsh environments. Improving the autonomous control capability and energy efficiency of SUGV is urgently needed. This article presents a skid steering-based path tracking control strategy. In the upper controller, an improved model-free sliding mode controller (APMS) is used to calculate the yaw moment for tracking control. On the lower controller, the Snow Ablation Optimizer (SAO) is used to distribute the output torque of the drive motors, taking longitudinal force, yaw moment and energy consumption into account. Finally, the designed controller is validated through simulation under different operating conditions. The results show that the proposed coordination controller achieves good control performance, increases energy efficiency and at the same time ensures tracking accuracy.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142012624","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-17DOI: 10.1016/j.conengprac.2024.106045
Due to the growing demand for robust autonomous systems, automating maintenance and fault mitigation activities has become essential. If an unexpected fault occurs during the travel, the system should be able to manage that fault autonomously and continue its mission. Thus, a robust fault mitigation system is needed that can quickly reconfigure itself in an optimal way. This paper presents a novel digital twin-based fault mitigation strategy that uses hierarchical control architecture. Here, a computationally efficient high-fidelity hybrid engine model is developed to simulate actual engine behavior. This hybrid engine model includes a neural network model representing the cylinder combustion process and well-studied physics-based analytical equations describing the remaining subsystems. This architecture uses a feedback controller on top of the control calibration map, generated offline using the hybrid model, to mitigate faults and modeling errors. The fault mitigation strategies are calibrated and validated through model-in-loop (MIL) and hardware-in-loop (HIL) simulations for various operating points using the Navistar 7.6 liters six-cylinder engine. The effectiveness of the proposed architecture in handling injector nozzle clogging, intake manifold leaks, and pressure shift faults is illustrated. The results demonstrate that the proposed architecture can completely overcome faults and maintain the desired torque in a few seconds. Moreover, the average accuracy of 96% is observed for the engine model compared to experimental data. It is anticipated that the proposed end-to-end architecture will be easily deployable on unmanned marine vessels and can be extended to accommodate other component faults.
{"title":"Digital twin-enabled autonomous fault mitigation in diesel engines: An experimental validation","authors":"","doi":"10.1016/j.conengprac.2024.106045","DOIUrl":"10.1016/j.conengprac.2024.106045","url":null,"abstract":"<div><p>Due to the growing demand for robust autonomous systems, automating maintenance and fault mitigation activities has become essential. If an unexpected fault occurs during the travel, the system should be able to manage that fault autonomously and continue its mission. Thus, a robust fault mitigation system is needed that can quickly reconfigure itself in an optimal way. This paper presents a novel digital twin-based fault mitigation strategy that uses hierarchical control architecture. Here, a computationally efficient high-fidelity hybrid engine model is developed to simulate actual engine behavior. This hybrid engine model includes a neural network model representing the cylinder combustion process and well-studied physics-based analytical equations describing the remaining subsystems. This architecture uses a feedback controller on top of the control calibration map, generated offline using the hybrid model, to mitigate faults and modeling errors. The fault mitigation strategies are calibrated and validated through model-in-loop (MIL) and hardware-in-loop (HIL) simulations for various operating points using the Navistar 7.6 liters six-cylinder engine. The effectiveness of the proposed architecture in handling injector nozzle clogging, intake manifold leaks, and pressure shift faults is illustrated. The results demonstrate that the proposed architecture can completely overcome faults and maintain the desired torque in a few seconds. Moreover, the average accuracy of 96% is observed for the engine model compared to experimental data. It is anticipated that the proposed end-to-end architecture will be easily deployable on unmanned marine vessels and can be extended to accommodate other component faults.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141998352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-16DOI: 10.1016/j.conengprac.2024.106040
This study focuses on the control strategy for a ground mobile robot (GMR) with independent three-axis six-wheel drive and four-wheel independent steering, performing double lane change trajectory tracking in complex scenarios. Initially, a dynamic model of the six-wheel independent drive and steering GMR was constructed. Utilizing Model Predictive Control (MPC) technology, the challenge of trajectory tracking at low speeds was effectively addressed. For high-speed conditions, by thoroughly analyzing the impact of the predictive time-domain, this study innovatively introduced an Adaptive Neuro-Fuzzy Inference System (ANFIS) to dynamically adjust the prediction horizon of the MPC. A novel trajectory tracking algorithm integrating MPC and ANFIS was developed, with the network structure being trained using backpropagation (BP) method and the least squares method. Compared to traditional MPC, this hybrid strategy significantly improves trajectory tracking accuracy and stability at high speeds, with computational efficiency increased by 48.65%. Additionally, the algorithm demonstrated excellent adaptability and control effectiveness in various rigorous tests, including different speed levels, complex steering paths, load changes, sudden obstacles, and variable terrain. A 70 km/h trajectory tracking experiment on a physical vehicle yielded a root mean square (RMS) error of 0.1904 m, verifying its superior tracking performance and practical reliability. This provides a pioneering solution for high-performance trajectory control of ground mobile robots.
{"title":"Research on ground mobile robot trajectory tracking control based on MPC and ANFIS","authors":"","doi":"10.1016/j.conengprac.2024.106040","DOIUrl":"10.1016/j.conengprac.2024.106040","url":null,"abstract":"<div><p>This study focuses on the control strategy for a ground mobile robot (GMR) with independent three-axis six-wheel drive and four-wheel independent steering, performing double lane change trajectory tracking in complex scenarios. Initially, a dynamic model of the six-wheel independent drive and steering GMR was constructed. Utilizing Model Predictive Control (MPC) technology, the challenge of trajectory tracking at low speeds was effectively addressed. For high-speed conditions, by thoroughly analyzing the impact of the predictive time-domain, this study innovatively introduced an Adaptive Neuro-Fuzzy Inference System (ANFIS) to dynamically adjust the prediction horizon of the MPC. A novel trajectory tracking algorithm integrating MPC and ANFIS was developed, with the network structure being trained using backpropagation (BP) method and the least squares method. Compared to traditional MPC, this hybrid strategy significantly improves trajectory tracking accuracy and stability at high speeds, with computational efficiency increased by 48.65%. Additionally, the algorithm demonstrated excellent adaptability and control effectiveness in various rigorous tests, including different speed levels, complex steering paths, load changes, sudden obstacles, and variable terrain. A 70 km/h trajectory tracking experiment on a physical vehicle yielded a root mean square (RMS) error of 0.1904 m, verifying its superior tracking performance and practical reliability. This provides a pioneering solution for high-performance trajectory control of ground mobile robots.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141998350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-16DOI: 10.1016/j.conengprac.2024.106042
Explicit reference governor (ERG) is an add-on unit that provides constraint handling capability to pre-stabilized systems. The main idea behind ERG is to manipulate the derivative of the applied reference in continuous time such that the satisfaction of state and input constraints is guaranteed at all times. However, ERG should be practically implemented in discrete-time. This paper studies the discrete-time implementation of ERG, and provides conditions under which the feasibility and convergence properties of the ERG framework are maintained when the updates of the applied reference are performed in discrete time. Specifically, using Zero-Order Hold (ZOH) discretization method, we develop an adaptive algorithm to adjust the gain of the discretized term based on actual measurements to maintain all properties of ERG when implemented in discrete-time. The proposed approach is validated via extensive simulation and experimental studies.
{"title":"Adaptive gain design for Zero-Order Hold discrete-time implementation of explicit reference governor","authors":"","doi":"10.1016/j.conengprac.2024.106042","DOIUrl":"10.1016/j.conengprac.2024.106042","url":null,"abstract":"<div><p>Explicit reference governor (ERG) is an <em>add-on</em> unit that provides constraint handling capability to pre-stabilized systems. The main idea behind ERG is to manipulate the derivative of the applied reference in continuous time such that the satisfaction of state and input constraints is guaranteed at all times. However, ERG should be practically implemented in discrete-time. This paper studies the discrete-time implementation of ERG, and provides conditions under which the feasibility and convergence properties of the ERG framework are maintained when the updates of the applied reference are performed in discrete time. Specifically, using Zero-Order Hold (ZOH) discretization method, we develop an adaptive algorithm to adjust the gain of the discretized term based on actual measurements to maintain all properties of ERG when implemented in discrete-time. The proposed approach is validated via extensive simulation and experimental studies.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141998351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-13DOI: 10.1016/j.conengprac.2024.106032
This paper proposes a novel hybrid offline–online neural identification-based robust adaptive control strategy for quadrotors subject to parameter uncertainties and external disturbances. A new method of using hybrid offline–online neural identification is developed to compensate for the residual force and moment caused by parameter uncertainties. Unlike previous methods that ignore the relevance of uncertainties in the time dimension, the proposed neural identification method mines temporal features of states from historical data by introducing long short-term memory (LSTM) networks, resulting in high identification accuracy. Furthermore, an online adaptation update law is designed to optimize the weights of the network estimates for strong robustness. Consequently, based on the identification of the network, a robust tracking controller on is constructed, which is capable of attenuating the bounded disturbances by introducing anti-disturbance components. Finally, numerical simulations and experiments in the real physical world are carried out to verify the performance. The experimental results demonstrate that the proposed strategy not only achieves more accurate uncertainty identification in comparison to the existing methods, but also realizes a 44.28% reduction in the root-mean-square error (RMSE) of the position under the lump uncertainties, which illustrates enhanced robustness and generalizability. Video: https://youtu.be/3kIG5fcQaVE.
{"title":"Hybrid offline–online neural identification-based robust adaptive tracking control for quadrotors","authors":"","doi":"10.1016/j.conengprac.2024.106032","DOIUrl":"10.1016/j.conengprac.2024.106032","url":null,"abstract":"<div><p>This paper proposes a novel hybrid offline–online neural identification-based robust adaptive control strategy for quadrotors subject to parameter uncertainties and external disturbances. A new method of using hybrid offline–online neural identification is developed to compensate for the residual force and moment caused by parameter uncertainties. Unlike previous methods that ignore the relevance of uncertainties in the time dimension, the proposed neural identification method mines temporal features of states from historical data by introducing long short-term memory (LSTM) networks, resulting in high identification accuracy. Furthermore, an online adaptation update law is designed to optimize the weights of the network estimates for strong robustness. Consequently, based on the identification of the network, a robust tracking controller on <span><math><mrow><mi>S</mi><mi>E</mi><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></math></span> is constructed, which is capable of attenuating the bounded disturbances by introducing anti-disturbance components. Finally, numerical simulations and experiments in the real physical world are carried out to verify the performance. The experimental results demonstrate that the proposed strategy not only achieves more accurate uncertainty identification in comparison to the existing methods, but also realizes a 44.28% reduction in the root-mean-square error (RMSE) of the position under the lump uncertainties, which illustrates enhanced robustness and generalizability. Video: <span><span>https://youtu.be/3kIG5fcQaVE</span><svg><path></path></svg></span>.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141979492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-12DOI: 10.1016/j.conengprac.2024.106041
Model Predictive Control (MPC) is a wide spread advanced process control methodology for optimization based control of multi-input and multi-output processes systems. Typically, a surrogate model of the process dynamics is utilized to predict the future states of a process as a function of input actions and an initial state. The predictive model is often a linear model, such as a state space model, due to the computational burden of the resulting optimization problem when utilizing nonlinear models. Recently, rectified linear unit (ReLU) based neural networks (NN) were shown to be mixed integer linear representable, thus allowing their incorporation into mixed integer programming (MIP) frameworks. However, the resulting MIP-based MPC problems are often computationally intractable to solve in real-time. The computational intractability of the reformulated NN-based optimization models is typically addressed in the literature by applying some form of bounds tightening approach. However, this in itself may have a large computational cost. In this work, a novel bound tightening procedure based on a multiparametric (MP) programming formulation of the corresponding MIP reformulated MPC optimization problems is proposed. Which tightening only needs to be computed and applied once-and-offline, thereby significantly improving the computational performance of the MPC in real-time. Some aspects of the effect of regularization during NN regression on the computational difficulty of these optimization problems are also investigated in conjunction with the proposed a priori bounds-tightening approach. The proposed method is compared to the base case without the parametric tightening procedure, as well as NN regularization through two optimal control case studies: (1) A ReLU NN-based MPC of an unstable nonlinear chemostat and, (2) a ReLU NN-based MPC of a nonlinear continuously stirred tank reactor (CSTR). Significant reductions in average time of 99.96% and 91.90% are observed, for the chemostat NN based MPC and CSTR NN based MPC, respectively.
模型预测控制(MPC)是一种广泛应用的先进过程控制方法,用于对多输入和多输出过程系统进行基于优化的控制。通常情况下,利用过程动态的代理模型来预测过程的未来状态,作为输入操作和初始状态的函数。预测模型通常是线性模型,如状态空间模型,这是因为在使用非线性模型时,由此产生的优化问题会造成计算负担。最近,基于整型线性单元(ReLU)的神经网络(NN)被证明是混合整型线性可表示的,因此可以将其纳入混合整型编程(MIP)框架。然而,由此产生的基于 MIP 的 MPC 问题往往在计算上难以实时解决。文献中通常通过应用某种形式的边界收紧方法来解决基于 NN 的重构优化模型的计算棘手性问题。然而,这种方法本身可能会产生很大的计算成本。在这项工作中,提出了一种基于相应 MIP 重构 MPC 优化问题的多参数(MP)编程表述的新型边界收紧程序。这种收紧只需计算和应用一次,而且是离线的,从而大大提高了 MPC 的实时计算性能。结合所提出的先验边界收紧方法,还研究了 NN 回归过程中正则化对这些优化问题计算难度的影响。通过两个优化控制案例研究,将所提出的方法与无参数紧缩程序的基本情况以及 NN 正则化进行了比较:(1) 基于 ReLU NN 的不稳定非线性化学恒温器的 MPC;(2) 基于 ReLU NN 的非线性连续搅拌罐反应器(CSTR)的 MPC。基于化学恒温器 NN 的 MPC 和基于 CSTR NN 的 MPC 的平均时间分别显著缩短了 99.96% 和 91.90%。
{"title":"A multiparametric approach to accelerating ReLU neural network based model predictive control","authors":"","doi":"10.1016/j.conengprac.2024.106041","DOIUrl":"10.1016/j.conengprac.2024.106041","url":null,"abstract":"<div><p>Model Predictive Control (MPC) is a wide spread advanced process control methodology for optimization based control of multi-input and multi-output processes systems. Typically, a surrogate model of the process dynamics is utilized to predict the future states of a process as a function of input actions and an initial state. The predictive model is often a linear model, such as a state space model, due to the computational burden of the resulting optimization problem when utilizing nonlinear models. Recently, rectified linear unit (ReLU) based neural networks (NN) were shown to be mixed integer linear representable, thus allowing their incorporation into mixed integer programming (MIP) frameworks. However, the resulting MIP-based MPC problems are often computationally intractable to solve in real-time. The computational intractability of the reformulated NN-based optimization models is typically addressed in the literature by applying some form of bounds tightening approach. However, this in itself may have a large computational cost. In this work, a novel bound tightening procedure based on a multiparametric (MP) programming formulation of the corresponding MIP reformulated MPC optimization problems is proposed. Which tightening only needs to be computed and applied once-and-offline, thereby significantly improving the computational performance of the MPC in real-time. Some aspects of the effect of regularization during NN regression on the computational difficulty of these optimization problems are also investigated in conjunction with the proposed a priori bounds-tightening approach. The proposed method is compared to the base case without the parametric tightening procedure, as well as NN regularization through two optimal control case studies: (1) A ReLU NN-based MPC of an unstable nonlinear chemostat and, (2) a ReLU NN-based MPC of a nonlinear continuously stirred tank reactor (CSTR). Significant reductions in average time of 99.96% and 91.90% are observed, for the chemostat NN based MPC and CSTR NN based MPC, respectively.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141979493","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-10DOI: 10.1016/j.conengprac.2024.106039
The accuracy of rolling force prediction is key to improving the precision of strip thickness control. The compressive load required for the strip in the rolling process is not only related to the size of the billet and process parameters such as deformation speed, temperature, and reduction, but also to the deformation boundary conditions of the billet between the rolls, such as the wear state of the rolls, lubrication conditions, etc. These influencing factors are interrelated and constantly changing, which is particularly prominent in small-batch and multi-specification intermittent production modes. The existing rolling force prediction models are constructed based on the rolling deformation mechanism through numerous simplifications. Due to challenges in fully and accurately characterizing various complex rolling deformation processes, their mapping relationships with process parameters, and constantly changing boundary conditions, the accuracy of the simplified rolling force prediction model is difficult to meet the control requirements of actual production. This paper proposes a physics-guided data-driven (PGDD) rolling force modeling method. It separates rolling condition features into static and dynamic parts using mechanistic and empirical knowledge and introduces a machine learning framework that integrates these parts for modeling. In this framework, the static feature fitting part can establish the influence of process parameters such as billet chemical composition, size, rolling speed, temperature, etc. on the rolling force. Meanwhile, the dynamic feature fitting part is responsible for the collaborative modeling of influencing factors reflecting the evolution rules of roll state, learning the cumulative effects of various complex processing states from a large amount of time-series data formed by different combinations of rolling conditions. Experiments with real production condition data show that the proposed physics-guided data-driven modeling method can accurately predict the rolling force under complex and variable conditions, and its adaptability and accuracy are superior to the online original model and traditional data-driven model.
{"title":"A physics guided data-driven prediction method for dynamic and static feature fusion modeling of rolling force in steel strip production","authors":"","doi":"10.1016/j.conengprac.2024.106039","DOIUrl":"10.1016/j.conengprac.2024.106039","url":null,"abstract":"<div><p>The accuracy of rolling force prediction is key to improving the precision of strip thickness control. The compressive load required for the strip in the rolling process is not only related to the size of the billet and process parameters such as deformation speed, temperature, and reduction, but also to the deformation boundary conditions of the billet between the rolls, such as the wear state of the rolls, lubrication conditions, etc. These influencing factors are interrelated and constantly changing, which is particularly prominent in small-batch and multi-specification intermittent production modes. The existing rolling force prediction models are constructed based on the rolling deformation mechanism through numerous simplifications. Due to challenges in fully and accurately characterizing various complex rolling deformation processes, their mapping relationships with process parameters, and constantly changing boundary conditions, the accuracy of the simplified rolling force prediction model is difficult to meet the control requirements of actual production. This paper proposes a physics-guided data-driven (PGDD) rolling force modeling method. It separates rolling condition features into static and dynamic parts using mechanistic and empirical knowledge and introduces a machine learning framework that integrates these parts for modeling. In this framework, the static feature fitting part can establish the influence of process parameters such as billet chemical composition, size, rolling speed, temperature, etc. on the rolling force. Meanwhile, the dynamic feature fitting part is responsible for the collaborative modeling of influencing factors reflecting the evolution rules of roll state, learning the cumulative effects of various complex processing states from a large amount of time-series data formed by different combinations of rolling conditions. Experiments with real production condition data show that the proposed physics-guided data-driven modeling method can accurately predict the rolling force under complex and variable conditions, and its adaptability and accuracy are superior to the online original model and traditional data-driven model.</p></div>","PeriodicalId":50615,"journal":{"name":"Control Engineering Practice","volume":null,"pages":null},"PeriodicalIF":5.4,"publicationDate":"2024-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141963996","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}