Industrial Internet technologies have promoted the emergence and advancement of various smart service-oriented manufacturing (SOM) paradigms in the context of Industry 5.0. Existing research has considered energy consumption as an optimization objective to perform service allocation and balance the economic and environmental sustainability. Despite the significant progress, most approaches focused on reducing energy usage alone, without integrating the market mechanisms of carbon trading into the service allocation process. In contrast, carbon trading models not only account for the direct environmental impacts of energy consumption but also leverage market-based incentives to enable enterprises to dynamically balance economic performance and carbon emission during service allocation. To address the issue, this paper proposes a carbon emission-aware service allocation framework tailored for Industrial Internet-enabled SOM. By integrating dynamic carbon emission constraints into a multi-objective optimization process and employing multiple objective particle swarm optimization-non dominated sorting genetic algorithm-II (MOPSO-NSGA-II), the proposed approach strives to achieve an optimal balance among production efficiency, service quality, cost, and environmental impact. A case study is conducted to validate the superiority of the proposed approach in addressing carbon emission-aware service allocation problem.
Lightweight structural design with enhanced mechanical properties is considered a significant requirement in the field of advanced manufacturing industries such as aerospace engineering, biomedical engineering, and automotive engineering. Lattice structures have been considered as potential solutions to obtain structures with enhanced mechanical properties such as strength to weight ratio and energy absorption capacity. This research aims to carry out a comparative analysis of eleven different lattice structures inspired by nature using Fused Deposition Modelling with PLA as the chosen material. FEA analysis was conducted to analyse the mechanical behaviour of the lattice structures with respect to stress-strain behaviour, displacement, reaction force, structural efficiency, and energy absorption capacity. The results indicate the significance of lattice structures with different topologies to obtain enhanced mechanical properties. The lattice structure with the highest energy absorption capacity was the Truncated Cube lattice structure with 539.08 N.mm energy absorption capacity while maintaining the highest structural efficiency.
The trends of mass customization and personalized production lead to increased manufacturing complexity, requiring human involvement for flexibility, particularly in High-Mix, Low-Volume (HMLV) assembly. These production systems impose challenges on operators, highlighting the growing importance of providing cognitive support. Digital assembly instructions have emerged as a solution, offering step-by-step guidance to operators. However, creating and maintaining these instructions is labor-intensive, especially in HMLV contexts, due to numerous product variants and frequent design changes. This study introduces a framework for integrating product variability in digital assembly instructions and streamlining the instruction authoring process. A semantic model based on the industrial standard ISA-95 is proposed to integrate engineering information into assembly instructions. Additionally, a methodology is presented that incorporates a 150% workflow for managing instructions for a product family, from which specific variant configurations can be derived. The system suggests relevant instruction content during the authoring process of a new product variant to enhance the reuse of previously written instructions. A methodology to handle engineering changes within the assembly instructions has also been developed to warrant consistency with the product design. Initial testing has demonstrated promising results, including substantial time savings and improved consistency.
Industry 5.0 emphasises the integration of human intelligence with advanced automation, promoting collaboration between humans and machines in manufacturing environments. This paper presents an Extended Reality (XR)-driven Digital Twin (DT) framework for intelligent robotic laser cleaning, designed to enhance human-robot collaboration through immersive and intuitive interfaces. The system integrates a YOLOv11x-OBB-based defect-detection module, adaptive path-planning algorithms, and gesture-based robot control within Mixed Reality (MR) and Virtual Reality (VR) environments. The DT, implemented in Unity and synchronised with a UR10e collaborative robot via ROS2 (Robot Operating System 2)-RTDE (Real-Time Data Exchange) communication, enables real-time monitoring, trajectory validation, and bidirectional data exchange between physical and virtual systems. Experimental results demonstrate accurate defect localisation (mAP@0.5 = 0.995, mAP@0.5:0.95 = 0.977), responsive hand-gesture interaction (438 ms average latency), and stable DT synchronisation. Furthermore, the proposed defect-targeted path-planning strategy reduced laser-cleaning time by 86% compared with full-surface cleaning, highlighting its potential for industrial productivity and sustainability. The results confirm that combining XR, DT, and AI-based perception establishes an effective human-centric framework for intelligent robotic surface processing, supporting future deployment in Industry 5.0 manufacturing systems.
The increasing adoption of artificial intelligence (AI) and machine learning (ML) in manufacturing is driven by the need for more personalised, efficient and adaptive production systems. However, industrial AI systems must also comply with emerging governance frameworks, particularly ISO/IEC 42001:2023, the AI management system standard. Despite its relevance, manufacturers often struggle to translate the standard's high-level requirements into practical design and implementation decisions.This article proposes a reference architecture to support the design, deployment and governance of AI systems in manufacturing in accordance with ISO/IEC 42001:2023. The architecture defines the main system layers, functional components, and governance mechanisms required to ensure lifecycle control, traceability, operational integration, and alignment with compliance. Given the recent publication of the standard, this work provides an early contribution to operationalising AI management system requirements within industrial architectures. The architecture is applied to a real-world metal manufacturing use case focused on predictive maintenance. The implementation demonstrates its feasibility and improves system traceability and operational performance, including overall equipment effectiveness and maintenance-related downtime. Although validated in a specific context, the modular and layered design supports adaptation to other manufacturing environments, considering domain constraints, available infrastructure and legacy system integration.
In today's dynamic industrial environment, shaped by volatile markets and resource competition, manufacturers must adapt swiftly to remain viable. These circumstances often demand physical changes that may affect the underlying Control Program (CP). Adapting the CP entails reconfiguring systems between operational states, where the associated complexity and effort are crucial factor for evaluating the economic feasibility of alternative strategies. While effort estimation has been extensively explored in software engineering, existing approaches do not quantify the adaptation effort of CP, particularly in Cyber-Physical Production Systems (CPPSs). This paper introduces a complexity index-based graph model for CP adaptation effort estimation and operationalises it via scaling index-to-time-factor to produce quantitative time and cost estimates under scarce industrial data. Unlike architecture-led task listings and qualitative frameworks, the approach yields scenario-level quantitative KPIs that capture change propagation across physical, functional, and code layers. A network-graph model encapsulates these indices as scenario-specific state graphs, enabling change tracking and visualisation. The quantification method then computes adaptation effort from graph differences. The approach is validated through adaptation estimates for one lab-scale and three industrial automation cells. Results demonstrate the approach effectiveness in evaluating architecture, implementation, and runtime effort, offering a practical tool for strategic adaptation decisions.
The advances in robotics, coupled with the growing integration of cutting-edge technologies such as Artificial Intelligence and, specifically, Deep Learning within manufacturing, have enabled the development of flexible robotic systems capable of understanding their environment and executing tasks autonomously. This paper introduces a unified vision-based framework for industrial assembly applications that integrates Deep Learning - based object pose estimation and automated quality inspection within a single system architecture. The framework employs two complementary perception pipelines, combining semantic segmentation and direct 6D pose estimation models to accurately handle objects with diverse geometries and precision requirements. In parallel, a Deep Learning - based object detection module enables real-time quality inspection by identifying missing, misaligned, or incorrectly placed components. The overall framework is designed from a deployment-oriented perspective, supporting seamless integration of perception outputs within robotic control and monitoring systems while satisfying real-time industrial constraints. The framework is applied and validated on an industrially inspired case study from the consumer electronics sector, focusing on the assembly and inspection of a trimmer head and a monitor cover. The obtained results demonstrate accurate pose estimation, effective defect detection, and system-level integration in a realistic production context, highlighting the applicability of the proposed approach to flexible manufacturing environments.
Metal Additive Manufacturing (AM) processes are increasingly adopted because they combine high part quality with improved cost and time efficiency compared to conventional methods. However, Directed Energy Deposition (DED) processes still require extensive effort during setup to identify process parameters that ensure part quality. Given the complexity of the underlying physical phenomena, physics-based process modeling is essential for efficient process setup and planning. This work presents a fast macro-scale thermal simulation model for the wire laser-based DED (DED-LB) process, which, although less mature, is highly promising. The enthalpy method is employed to model heat diffusion, enabling reliable prediction of the temperature field and melt pool geometrical characteristics within computational times suitable for engineering applications. The results show consistently good performance, validated through correlation with experimental data for two feedstock materials, multiple geometries, and a range of process parameters.
Human-robot collaboration (HRC) technology can facilitate high-variety, low-volume (HVLV) assembly systems. However, simultaneously empowering both the operator and the cobot to adopt changes remains challenging. This study proposes a bidirectional HRC empowerment model to achieve high team fluency in the HVLV assembly scenario. Deep learning-based augmented reality (AR) assembly assistance is offered to empower operators for a wide range of assembly tasks, and a hidden semi-Markov model (HSMM)-based prediction is proposed to enhance the robot's cognitive capability in relation to the operator. An AR and HSMM empowered HRC assembly station, featuring two agents (the operator and the cobot), is implemented using the current model. The team fluency metrics and the NASA-TLX mental workload instrument are measured for four typical assembly scenarios: manual assembly, basic HRC, HSMM-assisted HRC, and HSMM and AR-assisted HRC assembly. The experiment results confirm that the proposed model increases assembly efficiency by 13% and decreases the overall workload by 9.6%. The proposed bidirectional HRC empowerment model, which integrates AR technology and the HSMM algorithm, contributes to both fluency and mental workload in diverse HRC assembly scenarios within the HVLV systems.
Detecting faults promptly and accurately in manufacturing is crucial for maintaining efficiency, productivity, and product quality. This research explores the potential of using autoencoders, including models such as Long Short-Term Memory (LSTM) networks, Bidirectional Long Short-Term Memory (BiLSTM), WaveNet, and hybrid models that combine 1D Convolutional Neural Networks (1D CNN) with BiLSTM layers, for real-time fault detection in drilling operations. The autoencoders were trained on data from normal operational conditions to learn standard patterns and then tested on datasets containing fault instances. Results show that autoencoders with BiLSTM and LSTM layers excelled at capturing temporal dependencies. Hybrid models, such as WaveNet-BiLSTM and CNN-BiLSTM, which integrate LSTM and convolutional layers, significantly enhanced fault detection performance, achieving perfect accuracy, precision, recall, and F1 scores with just 10% of normal training samples. The models' effectiveness was also evaluated using shorter time windows, resulting in an accuracy of 98%. This advancement facilitates the integration of such models into intelligent manufacturing systems for proactive maintenance and operational optimization, potentially reducing downtime and maintenance costs while maintaining high product quality.
Control plans are built around diverse monitoring systems that are crucial in implementing a manufacturer's quality strategy. However, maintaining applicable, relevant, and up-to-date control plans remains challenging due to increasing product variety, changing equipment and processes, and high expert turnover. In the Industry 4.0 context, where quality control is automated, there is a lack of integration between monitoring systems and their associated control plans. This hinders the ability to track the evolution of control plans across their design, configuration, and execution phases. Addressing these challenges, a Reference Data Model (RDM) was developed to map the relationships between quality and production components and monitoring systems. This model generates automated control plans in compliance with industry standards. To comprehensively evaluate the effectiveness of the control strategy, a Control Plan Performance Effectiveness (CPPE) indicator was proposed to facilitate the control plan review process. The CPPE comprises three metrics - completeness of control, realization of sampling, and quality of control equipment - and supports the Zero Defect Manufacturing (ZDM) strategy by ensuring a 'first-time-right' approach to quality control. The application of this indicator was illustrated through two case studies from the semiconductor industry, where the CPPE successfully identified areas of poor detection and redundant control methods.
Production - logistics synchronization systems face increasing challenges from volatile market demand and stringent resource constraints in the era of personalized manufacturing. Conventional forecasting and static resource configuration methods often fail to deliver responsiveness, cost efficiency, and adaptability under such uncertainty. To address these limitations, this study develops an integrated Predictive Opti-State Control (POsC) framework for dynamic resource configuration. Guided by the POsC strategy, a hybrid Temporal Convolutional Network - Long Short-Term Memory (TCN - LSTM) model is employed to capture nonlinear demand fluctuations and long-term dependencies across multiple product categories. Based on forecasting outputs, a dynamic resource configuration model is formulated to minimize total system costs, including production, transition, storage, and rental costs, under multi-resource constraints. To efficiently solve the resulting combinatorial optimization problem, an Improved Simulated Annealing (ISA) algorithm is designed, which enhances global exploration and local exploitation capabilities. The proposed approach is validated through an industrial case study of a leading paint manufacturer in the Guangdong - Hong Kong - Macao Greater Bay Area, where idle equipment capacity and excessive rental costs frequently arise due to demand uncertainty. Experimental analysis shows that the integrated framework significantly outperforms conventional forecasting and heuristic configuration methods, achieving superior performance in demand prediction accuracy, resource utilization, cost reduction, and system flexibility.
