Pub Date : 2024-11-29DOI: 10.1016/j.jmsy.2024.11.015
Mathias König , Herwig Winkler
Assistance systems play a crucial role in enhancing working conditions and efficiency in industrial assembly. In the context of Industry 4.0, it is important to determine the types of assistance systems that contribute to assembly goals as well as their economic benefits. First, the significance of the topic will be introduced, and the research questions will be presented. Second, the basic technical terms will be defined, and third, the research methodology of a structured literature review (SLR) will be delineated. The fourth section presents an overview of the ergonomic and information assistance systems used in operational practice and academic test set-ups. It further explains the reasons for using assistance systems in assembly and their economic benefits, particularly in terms of reducing assembly times and errors. In the fifth section, the research perspectives of the respective publications are evaluated and summarized in both a qualitative and quantitative way. The present mixed-methods study is not generalizable due to its limitations such as a small sample size, the geographical scope of the study, type of databanks, time of publication and language of the reviewed articles, and methods of data collection. It does, however, identify potential areas for future research and provide recommendations for further investigation.
{"title":"Investigation of assistance systems in assembly in the context of digitalization: A systematic literature review","authors":"Mathias König , Herwig Winkler","doi":"10.1016/j.jmsy.2024.11.015","DOIUrl":"10.1016/j.jmsy.2024.11.015","url":null,"abstract":"<div><div>Assistance systems play a crucial role in enhancing working conditions and efficiency in industrial assembly. In the context of Industry 4.0, it is important to determine the types of assistance systems that contribute to assembly goals as well as their economic benefits. First, the significance of the topic will be introduced, and the research questions will be presented. Second, the basic technical terms will be defined, and third, the research methodology of a structured literature review (SLR) will be delineated. The fourth section presents an overview of the ergonomic and information assistance systems used in operational practice and academic test set-ups. It further explains the reasons for using assistance systems in assembly and their economic benefits, particularly in terms of reducing assembly times and errors. In the fifth section, the research perspectives of the respective publications are evaluated and summarized in both a qualitative and quantitative way. The present mixed-methods study is not generalizable due to its limitations such as a small sample size, the geographical scope of the study, type of databanks, time of publication and language of the reviewed articles, and methods of data collection. It does, however, identify potential areas for future research and provide recommendations for further investigation.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 187-199"},"PeriodicalIF":12.2,"publicationDate":"2024-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747215","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-29DOI: 10.1016/j.jmsy.2024.11.014
Ruoxin Wang , Chi Fai Cheung , Yikai Zang , Chunjin Wang , Changlin Liu
Large aperture aspheric optical surfaces (LAAOS) have been applied in many industries, but their high requirements for precision and efficiency make manufacturing more challenging. Robotic polishing is a representative computer-controlled optical surfacing technique to manufacture LAAOS with low-cost and high-efficiency. However, how to achieve the highest material removal rate (MRR) involves many process parameters. It is difficult to determine the optimal parameter settings since the complex relationships among them. In this paper, a novel Bayesian optimized differential evolution based on deep learning method is proposed to optimize the MRR, in which the designed deep neural network is responsible for MRR modeling and Bayesian optimized differential evolution is used for MRR optimization. Bayesian optimization is used to find the best hyperparameter of differential evolution method so as to improve optimization performance. To evaluate the proposed method, a series of robotic polishing experiments are conducted to build the MRR model. The optimization performance comparison experiments show the superiority of our proposed method, which increases MRR by an average of 0.16.
{"title":"Material removal rate optimization with bayesian optimized differential evolution based on deep learning in robotic polishing","authors":"Ruoxin Wang , Chi Fai Cheung , Yikai Zang , Chunjin Wang , Changlin Liu","doi":"10.1016/j.jmsy.2024.11.014","DOIUrl":"10.1016/j.jmsy.2024.11.014","url":null,"abstract":"<div><div>Large aperture aspheric optical surfaces (LAAOS) have been applied in many industries, but their high requirements for precision and efficiency make manufacturing more challenging. Robotic polishing is a representative computer-controlled optical surfacing technique to manufacture LAAOS with low-cost and high-efficiency. However, how to achieve the highest material removal rate (MRR) involves many process parameters. It is difficult to determine the optimal parameter settings since the complex relationships among them. In this paper, a novel Bayesian optimized differential evolution based on deep learning method is proposed to optimize the MRR, in which the designed deep neural network is responsible for MRR modeling and Bayesian optimized differential evolution is used for MRR optimization. Bayesian optimization is used to find the best hyperparameter of differential evolution method so as to improve optimization performance. To evaluate the proposed method, a series of robotic polishing experiments are conducted to build the MRR model. The optimization performance comparison experiments show the superiority of our proposed method, which increases MRR by an average of 0.16.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 178-186"},"PeriodicalIF":12.2,"publicationDate":"2024-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747213","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-29DOI: 10.1016/j.jmsy.2024.11.017
Mohammad Mehdi Keramati Feyz Abadi, Chao Liu, Ming Zhang, Youxi Hu, Yuchun Xu
Energy poses a significant challenge in the industrial sector, and the abundance of data generated by Industry 4.0 technologies offers the opportunity to leverage Artificial Intelligence (AI) for enhancing energy efficiency (EE) in manufacturing processes, particularly within manufacturing systems. However, fully realizing AI's potential in addressing energy challenges requires a comprehensive review of AI methodologies aimed at overcoming obstacles in energy-efficient manufacturing systems. This article provides a systematic review that combines both quantitative and qualitative analyses of literature from the past ten years, focusing on mitigating prevalent energy efficiency challenges in manufacturing systems through AI-related methodologies. These challenges include Monitoring and Prediction, Real-Time Control, Scheduling, and Parameters Optimization. The AI-related solutions proposed in the reviewed research articles utilize Machine Learning (ML), Deep Learning (DL), and Reinforcement Learning (RL) techniques, either individually or in combination with other methods. A total of 67 journal papers on manufacturing systems, addressing the mentioned energy challenges through AI-related approaches, have been identified and thoroughly reviewed. As a result of this review, an Energy Efficient-Digital Twin (EE-DT) framework is proposed, demonstrating how a DT, equipped with AI techniques, can be applied to solve energy issues in manufacturing systems. This study provides scholars with a comprehensive guideline for selecting various types of AI methods to address common challenges in energy-efficient manufacturing systems, while also highlighting some promising future research directions.
{"title":"Leveraging AI for energy-efficient manufacturing systems: Review and future prospectives","authors":"Mohammad Mehdi Keramati Feyz Abadi, Chao Liu, Ming Zhang, Youxi Hu, Yuchun Xu","doi":"10.1016/j.jmsy.2024.11.017","DOIUrl":"10.1016/j.jmsy.2024.11.017","url":null,"abstract":"<div><div>Energy poses a significant challenge in the industrial sector, and the abundance of data generated by Industry 4.0 technologies offers the opportunity to leverage Artificial Intelligence (AI) for enhancing energy efficiency (EE) in manufacturing processes, particularly within manufacturing systems. However, fully realizing AI's potential in addressing energy challenges requires a comprehensive review of AI methodologies aimed at overcoming obstacles in energy-efficient manufacturing systems. This article provides a systematic review that combines both quantitative and qualitative analyses of literature from the past ten years, focusing on mitigating prevalent energy efficiency challenges in manufacturing systems through AI-related methodologies. These challenges include Monitoring and Prediction, Real-Time Control, Scheduling, and Parameters Optimization. The AI-related solutions proposed in the reviewed research articles utilize Machine Learning (ML), Deep Learning (DL), and Reinforcement Learning (RL) techniques, either individually or in combination with other methods. A total of 67 journal papers on manufacturing systems, addressing the mentioned energy challenges through AI-related approaches, have been identified and thoroughly reviewed. As a result of this review, an Energy Efficient-Digital Twin (EE-DT) framework is proposed, demonstrating how a DT, equipped with AI techniques, can be applied to solve energy issues in manufacturing systems. This study provides scholars with a comprehensive guideline for selecting various types of AI methods to address common challenges in energy-efficient manufacturing systems, while also highlighting some promising future research directions.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 153-177"},"PeriodicalIF":12.2,"publicationDate":"2024-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747214","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-28DOI: 10.1016/j.jmsy.2024.11.013
Pei Wang , Yixin Cui , Haizhen Tao , Xun Xu , Sheng Yang
The integrated multi-objective optimization of machining parameters for improved machining quality and efficiency is important in batch milling systems. Due to the change of the batch milling system state, the continuous use of the same machining parameters may lead to degradation in quality and efficiency for workpieces in batches. Machining parameter optimization is usually determined by manual experience or trial-and-error methods, making it difficult to achieve a synergistic consideration of both quality and efficiency. To address this issue, a novel multi-task deep reinforcement learning method for machining parameter optimization in a batch machining system is proposed. Firstly, a reliable parallel joint estimation model of multiple machining quality and efficiency indicators is established using a multi-task time series estimation method, which can learn the correlation of these indicators to improve estimation accuracy. Then, the parameter optimization problem is formalized as a Markov decision process supported by a reinforcement learning virtual environment and an agent. The reinforcement learning virtual environment with the joint estimation model is constructed to improve the accuracy of optimized machining parameters for the collaborative optimization of quality and efficiency indicators. Within the virtual environment, time series sequential state, sequential action, multi-objective reward function, and constraint conditions adapted to the joint estimation model are defined to repeatedly evaluate different machining parameters. The agent with a multi-head attention and a dynamic weight adjustment mechanism is designed to improve the stability of the optimization process. Finally, experiments on a real machining dataset of thin-walled parts show that compared with the traditional deep reinforcement learning algorithm, the optimization effect of the proposed framework is improved by 9 %−12 %, and the standard deviation is decreased by 9 % −18 %.
{"title":"Machining parameter optimization for a batch milling system using multi-task deep reinforcement learning","authors":"Pei Wang , Yixin Cui , Haizhen Tao , Xun Xu , Sheng Yang","doi":"10.1016/j.jmsy.2024.11.013","DOIUrl":"10.1016/j.jmsy.2024.11.013","url":null,"abstract":"<div><div>The integrated multi-objective optimization of machining parameters for improved machining quality and efficiency is important in batch milling systems. Due to the change of the batch milling system state, the continuous use of the same machining parameters may lead to degradation in quality and efficiency for workpieces in batches. Machining parameter optimization is usually determined by manual experience or trial-and-error methods, making it difficult to achieve a synergistic consideration of both quality and efficiency. To address this issue, a novel multi-task deep reinforcement learning method for machining parameter optimization in a batch machining system is proposed. Firstly, a reliable parallel joint estimation model of multiple machining quality and efficiency indicators is established using a multi-task time series estimation method, which can learn the correlation of these indicators to improve estimation accuracy. Then, the parameter optimization problem is formalized as a Markov decision process supported by a reinforcement learning virtual environment and an agent. The reinforcement learning virtual environment with the joint estimation model is constructed to improve the accuracy of optimized machining parameters for the collaborative optimization of quality and efficiency indicators. Within the virtual environment, time series sequential state, sequential action, multi-objective reward function, and constraint conditions adapted to the joint estimation model are defined to repeatedly evaluate different machining parameters. The agent with a multi-head attention and a dynamic weight adjustment mechanism is designed to improve the stability of the optimization process. Finally, experiments on a real machining dataset of thin-walled parts show that compared with the traditional deep reinforcement learning algorithm, the optimization effect of the proposed framework is improved by 9 %−12 %, and the standard deviation is decreased by 9 % −18 %.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 124-152"},"PeriodicalIF":12.2,"publicationDate":"2024-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747212","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-27DOI: 10.1016/j.jmsy.2024.11.008
Louis Schäfer, Stefan Tse, Marvin Carl May, Gisela Lanza
Today, manufacturers and suppliers are challenged to deliver customized products at the lowest possible cost and in increasingly shorter time frames, due to the increasing number of variants. Achieving this demands efficient production system planning. However, current planning in the manufacturing industry is heavily reliant on manual processes and individual expertise. Prior research tackles this issue by aiming to develop a comprehensive approach for assisted, model-based rough planning of production systems. This article focuses the optimization of variant-specific production systems. The basis for this is a process precedence graph that restricts the optimization of the assignment of process steps to stations. In the mathematical modeling of the Assembly Line Balancing Problem (ALBP), this work addresses complex constraints, including the selection of station equipment, the utilization of multiple robots per station and a non-discrete assignment of tasks. The approach developed is applied to the example of a Tier 1 automotive supplier, where the multi-criteria solution of the ALBP allows an evaluation of the planning result. To this end, this work compares the algorithmically generated solution both qualitatively and quantitatively with an example of manual expert planning. Thereby it demonstrates the broad, industrial applicability of the approach. Consequently, this research contributes to enhancing efficiency in production system planning, leading to sustainable reductions in both costs and time.
{"title":"Assisted production system planning by means of complex robotic assembly line balancing","authors":"Louis Schäfer, Stefan Tse, Marvin Carl May, Gisela Lanza","doi":"10.1016/j.jmsy.2024.11.008","DOIUrl":"10.1016/j.jmsy.2024.11.008","url":null,"abstract":"<div><div>Today, manufacturers and suppliers are challenged to deliver customized products at the lowest possible cost and in increasingly shorter time frames, due to the increasing number of variants. Achieving this demands efficient production system planning. However, current planning in the manufacturing industry is heavily reliant on manual processes and individual expertise. Prior research tackles this issue by aiming to develop a comprehensive approach for assisted, model-based rough planning of production systems. This article focuses the optimization of variant-specific production systems. The basis for this is a process precedence graph that restricts the optimization of the assignment of process steps to stations. In the mathematical modeling of the <em>Assembly Line Balancing Problem</em> (ALBP), this work addresses complex constraints, including the selection of station equipment, the utilization of multiple robots per station and a non-discrete assignment of tasks. The approach developed is applied to the example of a Tier 1 automotive supplier, where the multi-criteria solution of the ALBP allows an evaluation of the planning result. To this end, this work compares the algorithmically generated solution both qualitatively and quantitatively with an example of manual expert planning. Thereby it demonstrates the broad, industrial applicability of the approach. Consequently, this research contributes to enhancing efficiency in production system planning, leading to sustainable reductions in both costs and time.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 109-123"},"PeriodicalIF":12.2,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142721391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-27DOI: 10.1016/j.jmsy.2024.10.019
Jing Wang , Deming Lei , Debiao Li , Xixing Li , Hongtao Tang
Distributed energy-efficient hybrid flow shop scheduling problem (DEHFSP) with batch processing machines (BPMs) is rarely considered, let alone DEHFSP with BPMs and uncertainty. In this study, a fuzzy DEHFSP with BPMs at a middle stage and no precedence between some stages is presented, and a dynamic artificial bee colony (DABC) is proposed to simultaneously optimize the total agreement index, fuzzy makespan, and fuzzy total energy consumption. To produce high quality solutions, Metropolis criterion is used, dynamic employed bee phase based on neighborhood structure dynamic selection is implemented, and group-based onlooker bee phase with bidirectional communication is given. Migration operator is also adopted to replace scout bee phase. Extensive experiments are conducted, and the optimal combination of key parameters for DABC is decided by the Taguchi method. Comparative results and statistical analysis show that new strategies of DABC are effective, and DABC is highly competitive in solving the considered fuzzy DEHFSP.
{"title":"A dynamic artificial bee colony for fuzzy distributed energy-efficient hybrid flow shop scheduling with batch processing machines","authors":"Jing Wang , Deming Lei , Debiao Li , Xixing Li , Hongtao Tang","doi":"10.1016/j.jmsy.2024.10.019","DOIUrl":"10.1016/j.jmsy.2024.10.019","url":null,"abstract":"<div><div>Distributed energy-efficient hybrid flow shop scheduling problem (DEHFSP) with batch processing machines (BPMs) is rarely considered, let alone DEHFSP with BPMs and uncertainty. In this study, a fuzzy DEHFSP with BPMs at a middle stage and no precedence between some stages is presented, and a dynamic artificial bee colony (DABC) is proposed to simultaneously optimize the total agreement index, fuzzy makespan, and fuzzy total energy consumption. To produce high quality solutions, Metropolis criterion is used, dynamic employed bee phase based on neighborhood structure dynamic selection is implemented, and group-based onlooker bee phase with bidirectional communication is given. Migration operator is also adopted to replace scout bee phase. Extensive experiments are conducted, and the optimal combination of key parameters for DABC is decided by the Taguchi method. Comparative results and statistical analysis show that new strategies of DABC are effective, and DABC is highly competitive in solving the considered fuzzy DEHFSP.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 94-108"},"PeriodicalIF":12.2,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142721390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-26DOI: 10.1016/j.jmsy.2024.11.005
Hongcheng Li , Jian Peng , Yachao Jia , Rong Luo , Huajun Cao , Yunpeng Cao , Yu Zhang , Haihong Shi
Die-casting is an efficient and precise casting process, but it consumes significant energy and contributes to severe environmental pollution. The characteristic features of the die-casting process chain include high demand for energy and resources, dynamic synergy among multiple processing equipment, and mixed production of various products. These characteristics lead to challenges in carbon emission accounting, such as the problem of carbon emission data haze. To address this issue, this study analyzes the dynamic characteristics of carbon emissions in the die-casting process chain to identify the sources of carbon emissions. Subsequently, a multi-source carbon data collection scheme is developed based on these sources, and an information-physical fusion-based model for carbon source data collection and integration is established. Following this, the correlation between carbon sources in the die-casting process chain and the production process is elucidated, and a carbon emission accounting model for mixed production of multiple die-casting products is developed. For model parameterization, time-series power data are systematically integrated. Finally, using the dynamic characteristics of carbon emissions from typical die-casting production and the carbon source data model as a foundation, a case study is conducted on the carbon emissions from mixed production in the die-casting process chain. The results demonstrate the effectiveness, feasibility, and reliability of the proposed carbon emission accounting model. This study lays the foundation for optimizing carbon reduction in the die-casting process chain and supports the transition to a low-carbon die-casting workshop.
{"title":"Dynamic carbon emissions accounting in the mixed production process of multi-pressure die-castingproducts based on cyber physical production system","authors":"Hongcheng Li , Jian Peng , Yachao Jia , Rong Luo , Huajun Cao , Yunpeng Cao , Yu Zhang , Haihong Shi","doi":"10.1016/j.jmsy.2024.11.005","DOIUrl":"10.1016/j.jmsy.2024.11.005","url":null,"abstract":"<div><div>Die-casting is an efficient and precise casting process, but it consumes significant energy and contributes to severe environmental pollution. The characteristic features of the die-casting process chain include high demand for energy and resources, dynamic synergy among multiple processing equipment, and mixed production of various products. These characteristics lead to challenges in carbon emission accounting, such as the problem of carbon emission data haze. To address this issue, this study analyzes the dynamic characteristics of carbon emissions in the die-casting process chain to identify the sources of carbon emissions. Subsequently, a multi-source carbon data collection scheme is developed based on these sources, and an information-physical fusion-based model for carbon source data collection and integration is established. Following this, the correlation between carbon sources in the die-casting process chain and the production process is elucidated, and a carbon emission accounting model for mixed production of multiple die-casting products is developed. For model parameterization, time-series power data are systematically integrated. Finally, using the dynamic characteristics of carbon emissions from typical die-casting production and the carbon source data model as a foundation, a case study is conducted on the carbon emissions from mixed production in the die-casting process chain. The results demonstrate the effectiveness, feasibility, and reliability of the proposed carbon emission accounting model. This study lays the foundation for optimizing carbon reduction in the die-casting process chain and supports the transition to a low-carbon die-casting workshop.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 69-80"},"PeriodicalIF":12.2,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-26DOI: 10.1016/j.jmsy.2024.11.010
Donghai Wang , Shun Liu , Jing Zou , Wenjun Qiao , Sun Jin
Flexible robotic cells are pivotal in flexible and customized manufacturing. An effective scheduling policy for such cells can significantly reduce the makespan and improve the production efficiency. This study introduces an innovative end-to-end real-time scheduling method leveraging deep reinforcement learning (DRL) to minimize the makespan in a flexible robotic cell. We introduce a heterogeneous disjunctive graph model for a nuanced representation of the scheduling problem, which incorporates transportation through specific disjunctive arcs. The DRL utilizes Graph Neural Network (GNN) for model feature extraction and employs Proximal Policy Optimization (PPO) to train the scheduling agent. Our methodology can also better leverage the transport robot capacity to mitigate system blockage and deadlock. Numerical experiments are conducted to demonstrate the effectiveness of the proposed method.
{"title":"Flexible robotic cell scheduling with graph neural network based deep reinforcement learning","authors":"Donghai Wang , Shun Liu , Jing Zou , Wenjun Qiao , Sun Jin","doi":"10.1016/j.jmsy.2024.11.010","DOIUrl":"10.1016/j.jmsy.2024.11.010","url":null,"abstract":"<div><div>Flexible robotic cells are pivotal in flexible and customized manufacturing. An effective scheduling policy for such cells can significantly reduce the makespan and improve the production efficiency. This study introduces an innovative end-to-end real-time scheduling method leveraging deep reinforcement learning (DRL) to minimize the makespan in a flexible robotic cell. We introduce a heterogeneous disjunctive graph model for a nuanced representation of the scheduling problem, which incorporates transportation through specific disjunctive arcs. The DRL utilizes Graph Neural Network (GNN) for model feature extraction and employs Proximal Policy Optimization (PPO) to train the scheduling agent. Our methodology can also better leverage the transport robot capacity to mitigate system blockage and deadlock. Numerical experiments are conducted to demonstrate the effectiveness of the proposed method.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 81-93"},"PeriodicalIF":12.2,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the current injection molding (IM) industry, it remains challenging to monitor and estimate production quality promptly. It is costly and time-consuming to measure part quality manually after each production cycle ends, which results in quality defects difficult to be captured in time. In this case, a soft sensor is essential to model the IM process and predict the final quality in real time with multi-source industrial production data. However, traditional data-driven modeling methods fail to take advantage of the information in complex high-frequency data from in-mold sensors, resulting in an inaccurate IM model and unsatisfactory quality prediction performance. To solve this problem, this paper proposes a novel soft sensor framework based on a teacher-student structure. After specialized preprocessing of multiple sensor time series data, a GRU-based autoencoder with an attention mechanism (GRU-A-AE) is trained as a teacher model, extracting deep implicit features involving valuable time sequential information. Then, a cascaded relationship among shallow feature points from sensor signals, deep features, and final part weights is established using back propagation neural networks (BPNNs). To demonstrate its effectiveness and superiority, the proposed soft sensor is trained and tested with practical IM data under normal and fluctuating production conditions, respectively. Compared with conventional methods, our method has higher prediction accuracy with testing RMSE of 0.1049 and of 0.9950 under normal conditions, which proves more valuable information in high-frequency sensor signals are explored from the teacher model and IM production dynamics are captured precisely. In addition, its better prediction performance in the case of production condition fluctuation verifies its strong robustness.
在当前的注塑成型（IM）行业，及时监控和评估生产质量仍是一项挑战。在每个生产周期结束后手动测量零件质量既费钱又费时，导致难以及时捕捉质量缺陷。在这种情况下，必须使用软传感器对 IM 过程进行建模，并利用多源工业生产数据实时预测最终质量。然而，传统的数据驱动建模方法无法利用模内传感器复杂高频数据中的信息，导致 IM 模型不准确，质量预测性能不理想。为解决这一问题，本文提出了一种基于师生结构的新型软传感器框架。在对多个传感器时间序列数据进行专门的预处理后，基于 GRU 的自动编码器与注意力机制（GRU-A-AE）被训练为教师模型，提取涉及有价值的时间序列信息的深层隐含特征。然后，利用反向传播神经网络（BPNN）在传感器信号的浅层特征点、深层特征和最终部分权重之间建立级联关系。为了证明所提出的软传感器的有效性和优越性，分别在正常和波动的生产条件下用实际的 IM 数据对其进行了训练和测试。与传统方法相比，我们的方法具有更高的预测精度，正常条件下的测试均方根误差为 0.1049，R2 为 0.9950，这证明从教师模型中发掘了更多有价值的高频传感器信号信息，并精确捕捉了 IM 的生产动态。此外，它在生产条件波动情况下的预测性能也更佳，验证了其强大的鲁棒性。
{"title":"Novel deep learning based soft sensor feature extraction for part weight prediction in injection molding processes","authors":"Weilong Ding, Husnain Ali, Kaihua Gao, Zheng Zhang, Furong Gao","doi":"10.1016/j.jmsy.2024.11.011","DOIUrl":"10.1016/j.jmsy.2024.11.011","url":null,"abstract":"<div><div>In the current injection molding (IM) industry, it remains challenging to monitor and estimate production quality promptly. It is costly and time-consuming to measure part quality manually after each production cycle ends, which results in quality defects difficult to be captured in time. In this case, a soft sensor is essential to model the IM process and predict the final quality in real time with multi-source industrial production data. However, traditional data-driven modeling methods fail to take advantage of the information in complex high-frequency data from in-mold sensors, resulting in an inaccurate IM model and unsatisfactory quality prediction performance. To solve this problem, this paper proposes a novel soft sensor framework based on a teacher-student structure. After specialized preprocessing of multiple sensor time series data, a GRU-based autoencoder with an attention mechanism (GRU-A-AE) is trained as a teacher model, extracting deep implicit features involving valuable time sequential information. Then, a cascaded relationship among shallow feature points from sensor signals, deep features, and final part weights is established using back propagation neural networks (BPNNs). To demonstrate its effectiveness and superiority, the proposed soft sensor is trained and tested with practical IM data under normal and fluctuating production conditions, respectively. Compared with conventional methods, our method has higher prediction accuracy with testing RMSE of 0.1049 and <span><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span> of 0.9950 under normal conditions, which proves more valuable information in high-frequency sensor signals are explored from the teacher model and IM production dynamics are captured precisely. In addition, its better prediction performance in the case of production condition fluctuation verifies its strong robustness.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 58-68"},"PeriodicalIF":12.2,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699294","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-25DOI: 10.1016/j.jmsy.2024.11.009
Jieyang Peng , Dongkun Wang , Junkai Zhao , Yunfei Teng , Andreas Kimmig , Xiaoming Tao , Jivka Ovtcharova
The assembly of printed circuit boards (PCBs) is one of the standard processes in chip production, directly contributing to the quality and performance of the chips. In the automated PCB assembly process, machine vision and coordinate localization methods are commonly employed to guide the positioning of assembly units. However, occlusion or poor lighting conditions can affect the effectiveness of machine vision-based methods. Additionally, the assembly of odd-form components requires highly specialized fixtures for assembly unit positioning, leading to high costs and low flexibility, especially for multi-variety and small-batch production. Drawing on these considerations, a vision-free, model-agnostic meta-method for compensating robotic position errors is proposed, which maximizes the probability of accurate robotic positioning through interactive feedback, thereby reducing the dependency on visual feedback and mitigating the impact of occlusions or lighting variations. The proposed method endows the robot with the capability to learn and adapt to various position errors, inspired by the human instinct for grasping under uncertainties. Furthermore, it is a self-adaptive method that can accelerate the robotic positioning process as more examples are incorporated and learned. Empirical studies show that the proposed method can handle a variety of odd-form components without relying on specialized fixtures, while achieving similar assembly efficiency to highly dedicated automation equipment. As of the writing of this paper, the proposed meta-method has already been implemented in a robotic-based assembly line for odd-form electronic components. Since PCB assembly involves various electronic components with different sizes, shapes, and functions, subsequent studies can focus on assembly sequence and assembly route optimization to further enhance assembly efficiency.
{"title":"Meta-learning enhanced adaptive robot control strategy for automated PCB assembly","authors":"Jieyang Peng , Dongkun Wang , Junkai Zhao , Yunfei Teng , Andreas Kimmig , Xiaoming Tao , Jivka Ovtcharova","doi":"10.1016/j.jmsy.2024.11.009","DOIUrl":"10.1016/j.jmsy.2024.11.009","url":null,"abstract":"<div><div>The assembly of printed circuit boards (PCBs) is one of the standard processes in chip production, directly contributing to the quality and performance of the chips. In the automated PCB assembly process, machine vision and coordinate localization methods are commonly employed to guide the positioning of assembly units. However, occlusion or poor lighting conditions can affect the effectiveness of machine vision-based methods. Additionally, the assembly of odd-form components requires highly specialized fixtures for assembly unit positioning, leading to high costs and low flexibility, especially for multi-variety and small-batch production. Drawing on these considerations, a vision-free, model-agnostic meta-method for compensating robotic position errors is proposed, which maximizes the probability of accurate robotic positioning through interactive feedback, thereby reducing the dependency on visual feedback and mitigating the impact of occlusions or lighting variations. The proposed method endows the robot with the capability to learn and adapt to various position errors, inspired by the human instinct for grasping under uncertainties. Furthermore, it is a self-adaptive method that can accelerate the robotic positioning process as more examples are incorporated and learned. Empirical studies show that the proposed method can handle a variety of odd-form components without relying on specialized fixtures, while achieving similar assembly efficiency to highly dedicated automation equipment. As of the writing of this paper, the proposed meta-method has already been implemented in a robotic-based assembly line for odd-form electronic components. Since PCB assembly involves various electronic components with different sizes, shapes, and functions, subsequent studies can focus on assembly sequence and assembly route optimization to further enhance assembly efficiency.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"78 ","pages":"Pages 46-57"},"PeriodicalIF":12.2,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号