IET Collaborative Intelligent Manufacturing最新文献

Numerous files, such as records and logs, are generated in the process of equipment diagnosis and maintenance (D&M). These files contain lots of unstructured plain text. Knowledge in these files could be reused for similar equipment faults. In practice, knowledge presented in plain text is hard to acquire. Thus, automated named entity recognition (NER) and relation extraction (RE) methods based on pretrained encoders could be used to extract entities and relations and develop a structured knowledge graph (KG), thus facilitating intelligent manufacturing. However, equipment fault NER exhibits suboptimal performance with existing encoders pretrained on general-domain corpus. In this paper, domain-adaptation-based NER with information enrichment is proposed for developing an equipment fault KG. A domain-adapted encoder is tailored for equipment fault NER through domain-adaptive pretraining (DAPT). Update of word segmentation dictionary and adjustment of masking approach are implemented during DAPT for information enrichment, which helps make the most of the limited domain-specific pretraining corpus. Experimental results show that the F1 score of NER is improved by 1.22% using the domain-adapted encoder compared to its counterpart using the encoder pretrained on general-domain corpus. Furthermore, a reliable and robust question answering (QA) application of the developed equipment fault KG is also shown.

A novel algorithm combining bidirectional recurrent neural networks (BiRNNs) with temporal difference is proposed for multi-objective energy-efficient non-permutation flow-shop scheduling problem (NFSP). The objectives of this problem involve minimising both the makespan and total energy consumption. To begin, a mathematical model is formulated to represent the energy-efficient NFSP. Subsequently, the NFSP is transformed into a Markov decision process, where an action space comprising 28 scheduling rules is constructed. Considering the global and local features of NFSP, a set of 15 state features is extracted. Different reward functions are then defined to correspond to the specific objectives. Furthermore, the state features of NFSP are extracted using a multi-layer perceptron model based on BiRNNs. By utilising the TD(λ) algorithm to calculate the state value function, various policies are generated. In order to evaluate the proposed algorithm, a new test set for the energy-efficient NFSP is constructed, building upon classic benchmark problems. Finally, comparison experiments are conducted to demonstrate the effectiveness and efficiency of the proposed algorithm.

Robot perception is an important topic in artificial intelligence field, and tactile recognition in particular is indispensable for human–computer interaction. Efficiently classifying data obtained by touch sensors has long been an issue. In recent years, spiking neural networks (SNNs) have been widely used in tactile data categorisation due to their temporal information processing benefits, low power consumption, and high biological dependability. However, traditional SNN classification methods often encounter under-convergence when using membrane potential representation, decreasing their classification accuracy. Meanwhile, due to the time-discrete nature of SNN models, classification requires a significant time overhead, which restricts their real-time tactile sensing application potential. Considering these concerns, the authors propose a faster and more accurate SNN tactile classification approach using improved membrane potential representation. This method effectively overcomes model convergence problems by optimising the membrane potential expression and the relationship between the loss function and network parameters while significantly reducing the time overhead and enhancing the classification accuracy and robustness of the model. The experimental results show that the propose approach improves the classification accuracy by 4.16% and 2.71% and reduces the overall time by 8.00% and 8.14% on the EvTouch-Containers dataset and EvTouch-Objects dataset, respectively, when compared with existing models.

This work focuses on detecting defects in welding seams using the most advanced You Only Look Once (YOLO) algorithms and transfer learning. To this end, the authors prepared a small dataset of images using manual welding and compared the performance of the YOLO v5, v6, v7, and v8 methods after two-step training. Key findings reveal that YOLOv7 demonstrates superior performance, suggesting its potential as a valuable tool in automated welding quality control. The authors’ research underscores the importance of model selection. It lays the groundwork for future exploration in larger datasets and varied welding scenarios, potentially contributing to defect detection practices in manufacturing industries. The dataset and the code repository links are also provided to support our findings.

To adapt to the dynamic, diverse, and personalised needs of customers, manufacturing enterprises face the challenge of continuously adjusting their resource structure. This has led manufacturers to shift towards a smart cloud manufacturing mode in order to build highly flexible production logistics (PL) systems. In these systems, the optimal configuring of PL resources is fundamental for daily logistics planning and vehicle scheduling control, providing necessary resources for the entire PL segment. However, traditional resource configuration methods face limitations, such as incomplete information acquisition, slow response in resource configuration, and suboptimal configuration results, leading to high subsequent operational costs and inefficient logistics transportation. These issues limit the performance of the PL system. To address these challenges, the authors propose a digital twin-based optimisation model and method for smart cloud PL resources. The approach begins with constructing an optimisation model for the PL system considering the quality of service for a cloud resource is constructed, aiming to minimise the number of logistics vehicles and the total cost of the PL system. Additionally, a DT-based decision framework for optimising smart cloud PL resources is proposed. Alongside a DT-based dynamic configuration strategy for smart cloud PL resources is designed. By developing a multi-teacher grouping teaching strategy and a cross-learning strategy, the teaching and learning strategies of the standard teaching-learning-based optimisation algorithm are improved. Finally, numerical simulation experiments were conducted on the logistics transportation process of a cooperating enterprise, verifying the feasibility and effectiveness of the proposed algorithms and strategies. The findings of this study provide valuable references for the management of PL resources and algorithm design in advanced manufacturing modes.

Matheuristic is an optimisation methodology that integrates mathematical approaches and heuristics to address intractable combinatorial optimisation problems, where a common framework is to insert mixed integer linear programming (MILP) models as local search functions for evolutionary algorithms. However, since a mathematical programming formulation only tries to find the solution with the best objective value, matheuristics are rarely adopted to multi-objective scenarios asking for a set of Pareto optimal solutions, for example, vehicle routing problems and production scheduling problems. In this situation, the ɛ-constraint, which transforms multi-objective problems into single-objective formulations by considering selected objectives as constraints, seems to be a promising approach. First, an augmented ɛ-constraint-based matheuristic methodology (ɛ-MH) is proposed to apply the idea of ɛ-constraint to embedded MILP models, so that Pareto fronts obtained by meta-heuristics can be further improved by solving a set of MILP models. Afterwards, four speed-up strategies are developed to alleviate the computational burden resulting from repeatedly solving mathematical formulations, which also imply preferable scenarios for taking advantages of the ɛ-MH. Finally, several real-world bi-objective scheduling problems are discussed to present potential applications for the proposed methodology.

A flexible casting job shop scheduling problem (FCJSP) with batch processing machines is proposed based on the analysis of the flexible job shop scheduling problem (FJSP) and the study of the expendable casting process. Considering the makespan under the influence of the energy consumption, the authors apply the time execution window to the FCJSP model in conjunction with the characteristics of casting production. A hybrid particle swarm optimisation algorithm (HPSO) is developed to solve the FCJSP. The HPSO employs a block integration decoding rule to address scheduling integration. Particle swarm optimisation is used for global search, employing both discrete and continuous search strategies. Furthermore, the local search employs tabu search with neighbourhood operations based on knowledge-driven techniques. Simulation experiments demonstrate the feasibility of the proposed optimisation model. In the end, the HPSO algorithm has been successfully applied to the real expendable casting scheduling. The results demonstrate that it is more efficient and robust than previously reported algorithms.

The increasing need for safe and comfortable high-speed elevators due to the rise of super-tall buildings has led to a focus on vibration reduction modelling and optimisation. This article selects factors that have a significant impact on the vibration of high-speed elevator car systems through sensitivity evaluation to form a six-dimensional parameter space and establishes a multi-objective optimisation model for the car system. The Gibbis method and Radial Basis Function neural network are combined to sample and construct surrogate models, respectively. Meanwhile, a BA–EO algorithm that combines Bat algorithm and Extremal optimisation to adapt to a multidimensional parameter space is proposed here. In practical applications, the peak-to-peak value of vibration acceleration, which significantly affects human perception, is chosen as the objective function for vibration reduction optimisation. After optimisation, the vibrations of the car and car frame are decreased by 19% and 9%, respectively, which extend the service life of the high-speed elevator and enhance safety and comfort for passengers.

As the important resources and equipment of container terminals, berths and quay cranes (QCs) face various challenges in actual operations and their operation efficiency in turn affects the performance of the whole terminal. The authors investigate an integrated berth allocation and QC assignment and scheduling problem under the influence of various factors, including the two main factors of vessel arrival time uncertainty and tide, and the two secondary factors of berth deviation and interference between cranes. To formulate the problem, the authors develop a multi-factor robust scheduling model. A Genetic Algorithm (GA) with Brain Storm Optimisation based on the Contract Net Protocol (CNP) is designed to optimise the berth and QC scheduling scheme. Specifically, the authors use the GA for individual coding and population initialisation, use the brainstorming algorithm for clustering, and introduce the CNP for individual updating. The experimental results show that the designed algorithm can adapt the scheduling plan to complex environments and can improve the service level of terminals.

Prognostic and Health Management (PHM) is an emerging maintenance concept that is highly regarded by the scientific community and practitioners, as its adoption can bring economic, technical and environmental benefits to a company. PHM fully reflects the smart maintenance paradigm encompassing data collection, data manipulation, state detection, health assessment, prognostic assessment and advisory generation. Despite the undeniable benefits, there is still a large gap between the scientific and the real world. Several authors have investigated on the barriers to PHM implementation for companies, highlighting among them the lack of systematic approaches to its design and implementation. As a first contribution to this topic, the authors conducted a systematic literature review (SLR) to investigate the use of Decision Support Systems (DSSs) to support the PHM implementation. The SLR highlighted that few DSS had been developed and were limited to critical unit identification, maintenance strategy selection and data acquisition phase of PHM. Therefore, a conceptual framework for PHM implementation was provided as a second contribution. This framework summarises the decisions that should be addressed by a practitioner wishing to implement PHM services; moreover, it could lay the foundations for the development/improvement of the missing/existing DSSs for PHM implementation.