628
Views
0
CrossRef citations to date
0
Altmetric
Editorial

Smart manufacturing enabled by intelligent technologies

ORCID Icon, ORCID Icon, &

Rapidly evolving global initiatives highlighted a manufacturing future that is connected, smart, resilient, human-centric, and sustainable for producing high-value-added products and services. Manufacturing systems, therefore, must change: (1) new manufacturing control strategies are required to enable flexible production of heterogeneous manufacturing jobs at dynamic batch sizes with mass efficiency; (2) factories need to be resilient and self-organizing to adapt to disruptions in customer demands, factory conditions and supply chain changes; and (3) machines need to be sympathetic with its operators, to name a few.

These anticipated changes are made possible by the extensive adoption of artificial intelligence technologies. This special issue compiles 13 original research articles highlighting theoretical and technological breakthroughs in quality checking, monitoring, and control for distinctive manufacturing processes and systems.

The first paper, ‘A Multi-level spatial feature fusion-based transformer for intelligent defect recognition with small samples toward smart manufacturing system’, by Yiping Gao, Xinyu Li and Liang Gao, developed a multi-scale spatial feature fusion-based transformer (ViT) architecture for small-sample product defect recognition. The proposed method simulates human vision to extract the multi-level features of defects and three improved ViTs are built to fuse the features.

The second paper, ‘Smart manufacturing under limited and heterogeneous data: a sim-to-real transfer learning with convolutional variational autoencoder in thermoforming’, by Milad Ramezankhani, Mehrtash Harandi, Rudolf Seethaler and Abbas S. Milani, deals with the challenge of predicting product quality in advanced manufacturing due to sparse and diverse data. It introduces a sim-to-real transfer-learning framework using wide-and-deep learning to process structured sensory data and thermal images separately. A Convolutional Variational Autoencoder (ConvVAE) learns compact representations of thermal images, demonstrating superior performance in an industrial case study despite limited data. The model’s predictions align with theoretical expectations and data statistics, confirmed by an explainability analysis using SHAP values.

The third paper, ‘In-process monitoring of the ultraprecision machining process with convolution neural networks’, by K Manjunath, Suman Tewary, Neha Khatri and Kai Cheng, focuses on in-process monitoring and quality control in ultra-precision machining (UPM), a critical aspect of the manufacturing industry. Traditional methods struggle with the dynamic and noisy UPM environment, requiring intelligent monitoring. The paper explores the feasibility of using convolutional neural network (CNN) to classify abnormal and normal machining based on vibrational signals transformed into time-frequency-based log-spectrogram images. The proposed CNN algorithm achieves an 85.92% accuracy in classifying these images, demonstrating its potential for effective in-process monitoring of UPM.

The fourth paper, ‘The deep learning-based equipment health monitoring model adopting subject matter expert’, by Jr-Fong Dang, introduces a deep learning-based framework for equipment health monitoring (EHM) in the context of Industry 4.0, leveraging advanced sensor data. It integrates subject matter expert knowledge and employs a sliding window strategy for real-time EHM. The optimal window size is determined using autocorrelation functions and an empirical study validates the effectiveness of this framework. By identifying optimal hyperparameters that minimize loss and validate window size, the proposed algorithm outperforms other machine learning models. Additionally, a general framework for maintaining equipment performance is outlined, showcasing the practicality and effectiveness of the approach.

The next paper, ‘A state-space model-based temperature control system for laser remanufacturing molten pool’, by Guo-Zhe Yang, Tong-Ming Liu, Bo-Xue Song, Xing-Yu Jiang, Zi-Sheng Wang, Ke-Qiang Chen and Wei-Jun Liu, introduces a real-time online control method for managing molten pool temperature during laser remanufacturing. It uses a state space model to address challenges related to establishing a control mechanism and dealing with multiple variable constraints. The system developed utilizes the state space model and an incremental integral separation PID control, employing a single temperature feedback point to overcome delays seen in traditional feedback cycles. The experiments conducted demonstrate the effectiveness of this system in eliminating sudden and gradual temperature changes during deposition, significantly enhancing thermal accumulation, and consequently improving the quality and stability of remanufactured parts.

Starting to focus on system-level scheduling and control, the sixth paper, ‘Digital-twin-based job shop multi-objective scheduling model and strategy’, by Zhuo Zhou, Liyun Xu, Xufeng Ling and Beikun Zhang, introduces a digital-twin-based job shop scheduling strategy to overcome issues in traditional scheduling methods. It focuses on minimizing completion time, tardiness, and energy consumption. Using cloud-edge computing, it devises an improved scheduling approach and enhances the genetic algorithm for better optimization. The effectiveness of these enhancements is validated using standard datasets and practical problems.

The seventh paper, ‘Dynamic modeling and analysis of multi-product flexible production line’, by Muhammad Waseem, Chen Li and Qing Chang, addresses mass customization challenges in manufacturing by proposing a novel approach – multi-product dynamic production systems. Unlike traditional Flexible Manufacturing Systems (FMS) that halt production due to single machine failures, this system accommodates extended failures while maintaining operation. It introduces a dynamic modeling method to study system behavior during disruptions like machine failures and demand changes, defining real-time metrics like permanent production loss (PPL) and demand dissatisfaction. Additionally, the paper presents a real-time analysis method to assess PPL and demand dissatisfaction. Numerical case studies validate the effectiveness of this approach, showcasing the fidelity of the model and the efficiency of the real-time analysis method in multi-product production lines.

The eighth paper, ‘Dynamic spatiotemporal scheduling of hull parts under complex constraints in shipbuilding workshop’, by Xiangdong Wang, Xiaofeng Hu and Chunhua Zhang, addresses the digital transformation in shipbuilding, focusing on the spatiotemporal scheduling problem within hull construction at the subassembly level. It introduces a multi-queue two-level optimization method and a dynamic scheduling model based on multi-directed acyclic graphs (multi-DAG) and queueing theory. The paper describes complex constraints in processing tasks and resources, proposing a model to minimize average part time. It suggests priority determination methods for queue events and allocation strategies for space and workers to enhance efficiency. The proposed method undergoes simulation experiments using workshop data and is compared to heuristic and combination rules using queueing indicators and space utilization versus time graphs to demonstrate its effectiveness.

The ninth paper, ‘Research on dynamic scheduling and perception method of assembly resources based on digital twin’, by Yunrui Wang, Yao Wang, Wengzhe Ren, Zhengli Wu and Juan Li, tackles challenges in assembly plants caused by uncertain resources, proposing a dynamic scheduling method using digital twin technology. It introduces a model and explains how assembly resources operate within this digital twin. It also explores a detailed perception method using the Petri network, creating models for key assembly resources. By simulating a frame factory, it provides real-time data, aiding in smoother implementation of assembly plans and resource scheduling.

The tenth paper, ‘Optimal selection of cloud manufacturing resources based on bacteria foraging optimization’, by Yanjuan Hu, Leiting Pan, Wenjun Lv and Zhanli Wang, presents cloud manufacturing (CMfg) to merge computing and manufacturing resources for efficient distribution. To streamline CMfg, it proposes a method using bacterial foraging optimization: it establishes evaluation criteria, builds a mathematical model, determines weights using Analytic Hierarchy Process (AHP) and entropy weight method, and applies BFO to optimize resource selection. The NGW51 reducer case study validates the effectiveness of BFO in improving CMfg resource allocation compared to existing methods.

The eleventh paper, ‘GNN-based deep reinforcement learning for MBD product model recommendation’, by Yuying Hu, Zewen Sheng, Min Ye, Meiyu Zhang and Chengfeng Jian, addresses the growing demand for better digital twin model recommendations amid increased usage in product delivery. It introduces a Graph Neural Network (GNN)-based deep reinforcement learning (DRL) approach to categorize models based on semantic features, enhancing the reuse of non-geometric aspects often overlooked in existing methods. The experiment demonstrates the method’s superior accuracy and its ability to effectively cater to user design needs compared to traditional recommendation algorithms.

The twelfth paper, ‘A systematic investigation on barriers to effective implementation of human-robot assembly line: a BWM approach’, by Bo Tian, Mukund Janardhanan and Marina Marinelli, investigates the Human-Robot Assembly Line (HRAL) as a versatile manufacturing system for modern personalized production. It examines the factors influencing HRAL implementation from drivers, benefits, and barriers, to guide practitioners in identifying suitable applications and overcoming challenges. Using heterogeneous resource theory, it breaks down HRAL into 11 features, showing how these contribute to its advantages and limitations. A systematic Triple-I framework analyzes implementation barriers, emphasizing productivity and accuracy limitations. The study proposes a barrier-elimination plan to support successful HRAL integration by addressing these challenges.

The final paper, ‘Readiness assessment for smart manufacturing system implementation: multiple case of Indian small and medium enterprises’, by Monica Shukla and Ravi Shankar, explores the challenge faced by small and medium enterprises (SMEs) in developing countries as they strive to adapt to smart manufacturing. It introduces a readiness assessment model with 5 stages, 6 building blocks, and 22 dimensions. Using ‘Stepwise Weight Assessment Ratio Analysis (SWARA)’, the model evaluates these dimensions, demonstrated through its application to five Indian SMEs, offering recommendations for improvement. This model serves as a self-assessment tool for SMEs to evaluate their readiness and plan for future transformation.

We would like to thank all the reviewers who gave their significant comments and suggestions for improving the published papers in this special issue. Thank you all the contributors to make the publication of this special issue. Special thanks should be given to Professor Stephen T. Newman, Editor-in-Chief of the International Journal of Computer Integrated Manufacturing, who gave his great support. We hope that this special issue will bridge the academic and practitioners so as to enhance smart manufacturing in the future.

Reprints and Corporate Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

To request a reprint or corporate permissions for this article, please click on the relevant link below:

Academic Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

Obtain permissions instantly via Rightslink by clicking on the button below:

If you are unable to obtain permissions via Rightslink, please complete and submit this Permissions form. For more information, please visit our Permissions help page.